Queensland University of Technology. Institute of Health and Biomedical Innovation School of Nursing and Midwifery Faculty of Health.

Transcription

1 Queensland University of Technology Institute of Health and Biomedical Innovation School of Nursing and Midwifery Faculty of Health Examination of Relationships and Mediating Effects of Self-Efficacy, Locus of Control, Coping and the Practice Environment on Caring Efficacy and Job Satisfaction in Australian Registered Nurses Carol Reid RN, BHSc (Nursing), Grad Cert Health Studies (Sexual Health), MAPPSc (Nursing) Professor Debra Anderson, Dr Cameron Hurst This dissertation is submitted to fulfil the requirement of the award of the Doctor of Philosophy at Queensland University of Technology March 2012

2 Contents Contents... ii Statement of Original Authorship... viii Acknowledgements... ix List of Tables... x List of Figures... xi List of Abbreviations... xii Abstract... xiii Related Presentations and Publications... xv Chapter 1: Introduction Introduction Background and Significance of this Study Statement of Problem Aim Objectives Study Study Research Questions and Hypotheses Research Questions for Study Hypotheses for Study Research Questions for Study Hypotheses for Study Conceptual Framework Caring in Nursing Caring Efficacy Social Cognitive Theory (Efficacy Theory) Efficacy Expectations Sources of Efficacy Expectations Caring, Self-Efficacy and Coping Self-Efficacy in Nursing and Job Satisfaction Self-Efficacy, Caring, Work Locus of Control, Socio-Demographics and Job Satisfaction Summary of Conceptual Framework Definition of Terms Summary Chapter 2: Literature Review Introduction Self-Efficacy in Nursing Caring Efficacy Caring in Nursing and Socio-Demographics Self-Efficacy and Socio-Demographics Work Locus of Control Locus of Control and Socio-Demographics Coping Styles ii

3 2.5.1 Coping and Socio-Demographics The Professional Nursing Practice Environment The Professional Nursing Practice Environment and Self-Efficacy The Nursing Practice Environment and Socio-Demographics Job Satisfaction Job Satisfaction and Socio-Demographics Summary Chapter 3: Research Plan: Methodology Introduction Study 1: Pilot Study Research Questions Aims Research Design and Methodology Research Population Recruitment Methods Sample Size Sampling Strategy Measures/Assessment Tools Data Management and Analysis Test-Retest Reliability Validity Testing Study 2: Cross-Sectional Survey Design Research Questions Aims Research Design and Methodology Research Population and Sampling Frame Recruitment Methods Sample Size Sampling Strategy Introduction to the Assessment Tools Demographic Questionnaire General Self-Efficacy Scale Work Locus of Control Scale Ways of Coping Questionnaire Revised Nursing Work Index (NWI-R) Brisbane Practice Environment Measure Caring Efficacy Scale Job Satisfaction Survey Data Management Normality Assumption Multicollinearity Missing Data Data Analysis Data Quality Psychometric Testing of the Instruments Study Test-Retest Reliability Study Construct Validity Reliability iii

4 Timeline for this Study Ethical Considerations Summary Chapter 4: The Research Findings Study 1 (Pilot Study) Introduction Objectives Results The Questionnaire Description of Sample Test-Retest Reliability of the Scales Summary of Findings Pilot Study Discussion and Conclusion Chapter 5: Research Findings Study Introduction Description of the Sample Personal Characteristics of the Sample Organisational Characteristics Research Question 1: What is the Level of Perceived Caring Efficacy and Job Satisfaction of an Australian Registered Nurse Population? Caring Efficacy and Job Satisfaction Research Question 2: Do Levels of Caring Efficacy and Job Satisfaction in an Australian Population of Registered Nurses Vary According to the Tested Socio-Demographic Variables? Associations between Caring Efficacy Levels and Categorical Organisational Characteristics and Job Satisfaction Levels and Categorical Organisational Characteristics Associations between Caring Efficacy Levels and Categorical Personal Characteristics and Job Satisfaction Levels and Categorical Personal Characteristics Summary of Findings Research Question 3: Do the Socio-Demographic Factors Examined Explain Levels of Caring Efficacy and Job Satisfaction in the Tested Australian Registered Nurse Population? Correlation Analysis of Continuous Variables and Caring Efficacy and Job Satisfaction Bivariate Analysis of Categorical Variables Multivariable Modelling of Caring Efficacy and Job Satisfaction using Socio-Demographic Variables Multivariable Modelling of Caring Efficacy Effect Modification Model Diagnostics Multivariable Modelling of Job Satisfaction Effect Modification Model Diagnostics Summary Is the Brisbane Practice Environment Measure a Valid and Reliable Instrument for Use in an Australian Population of Registered Nurses? Reliability Construct Validity iv

5 5.6.3 Confirmatory Factor Analysis Group Analysis Research Question 4: Does a High Level of Self-Efficacy in Nursing Enhance Relationships Between Work Locus of Control, General Self- Efficacy, Coping Styles, the Professional Nursing Practice Environment and Job Satisfaction? Internal Consistency Structural Equation Modelling Modelling Relationships between Variables Using Structural Equation Model Correlation Analysis Showing Associations Among Scales General Self-Efficacy, Practice Environment and Work Locus of Control Model Fit Potential Causal Pathways and Caring Efficacy Potential Causal Pathways and Job Satisfaction Caring Efficacy and Job Satisfaction Modelling Summary of Findings Research Question 5: Can Caring Efficacy Mediate the Relationship among Work Locus of Control, General Self-Efficacy, Coping Styles, Professional Nursing Practice Environment, Predetermined Socio- Demographic Variables and Job Satisfaction? Mediation Analysis Sobel s Test Mediation of Caring Efficacy and General Self-Efficacy and that of Work Locus of Control and Practice Environment Summary of Findings Chapter 6: Discussion Introduction Associations of Socio-Demographic Factors with Caring Efficacy and Job Satisfaction Organisational Characteristics, Caring Efficacy and Job Satisfaction Personal and Continuous Socio-Demographic Characteristics, Caring Efficacy and Job Satisfaction Age, Years of Experience and Years in Current Job Years in Current Position Specialty Area Gender Education Qualifications Marital Status Job Status Associations Between Caring Efficacy and Job Satisfaction The Validity and Reliability of the Brisbane Practice Environment Measure (BPEM) Model Testing Modelling Analyses and Caring Efficacy Modelling Analyses and Job Satisfaction Mediational Analysis Summary v

6 Chapter 7: Conclusion Introduction Summary of Study Key Findings of Study Description of the Sample Personal Characteristics of the Sample Organisational Characteristics Research Question 1: What Is the Level of Caring Efficacy and Job Satisfaction of the Registered Nurse Population in Australia? Research Question 2: Do Levels of Caring Efficacy and Job Satisfaction among the Registered Nurse Population in Australia Vary According to Age, Gender, Marital Status, Level of Education, Years of Experience, Years in Current Job, Employment Status, Geographical Location, Specialty Area, Health Sector and Australian State? Associations between Caring Efficacy Levels and Organisational Characteristics and Job Satisfaction Levels and Organisational Characteristics Associations between Caring Efficacy Levels and Personal Categorical Characteristics and Job Satisfaction Levels and Personal Categorical Characteristics Research Question 3: Do the Variables Age, Sex, Marital Status, Level of Education, Years of Experience, Years in Current Job, Employment Status, Geographical Location, Specialty Area, Health Sector and Australian State Predict Caring Efficacy and Job Satisfaction? Associations between Caring Efficacy Levels and the Continuous Socio-Demographic Variables Associations between Job Satisfaction Levels and Continuous Socio-Demographic Variables Associations between Caring Efficacy Levels and Job Satisfaction Levels Associations between Caring Efficacy and Job Status, Geographical Location, Education Qualifications, Marital Status, Specialty Area, Health Sector, Gender, State, Age, Years of Experience and Years in Current Job Following an ANOVA Multivariable Analysis (GLM) of Significantly Associated Socio- Demographic Variables and Caring Efficacy Associations between Job Satisfaction and Employment Status, Geographical Location, Education Level, Marital Status, Specialty Area, Health Sector, Gender, State, Age, Years of Experience and Years in Current Job Following an ANOVA Multivariable Analysis of Significantly Associated Socio- Demographic Variables and Job Satisfaction Is the Brisbane Practice Environment Measure a Valid and Reliable Instrument for Use in an Australian Population of Registered Nurses? Research Question 4: Does a High Level of Caring Efficacy among Registered Nurses Enhance Relationships between General Self- Efficacy, Work Locus of Control, Coping Styles, the Professional Nursing Practice Environment and Job Satisfaction Levels? vi

7 Structural Equation Modelling Correlation Analysis Showing Associations among the Scales of General Self-Efficacy, Practice Environment and Work Locus of Control Potential Causal Pathways and Caring Efficacy Potential Causal Pathways and Job Satisfaction Caring Efficacy and Job Satisfaction Modelling Research Question 5: Can Nursing Self-Efficacy Mediate a Relationship among the Factors of Work Locus of Control, General Self Efficacy, Professional Nursing Practice Environment, Predetermined Socio-Demographic Variables and Job Satisfaction? Strengths of Study Limitations of Study Implications Implications for Theorisation Implications for Practice Implications for Further Research Recommendations Summary References Appendices Appendix 1: Research Higher Degree Initiative Award Appendix 2: National Research and Scholarship Fund Award Appendix 3: Introduction to Structural Equation Modelling Using AMOS Certificate Appendix 4: Membership Honor s Society of Nursing Sigma Theta Tau International Appendix 5: Plain Language Statement of Information and Instructions Appendix 6: Human Ethics Approval Certificate Appendix 7: Demographic Questionnaire About You Appendix 8: General Self-Efficacy Scale and Permission for Use Appendix 9: Work Locus of Control Scale and Permission for Use Appendix 10: Ways of Coping Questionnaire and Permission for Use Appendix 11: Revised Nursing Work Index (NWI-R) and Permission for Use Appendix 12: Brisbane Practice Environment Measure and Permission for Use 193 Appendix 13: Caring Efficacy Scale and Permission for Use Appendix 14: Job Satisfaction Survey and Permission for Use Appendix 15: Budget for Project Appendix 16: Bland Altman Plots vii

8 Statement of Original Authorship The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made. Signed: On: / / viii

9 Acknowledgements I wish to express sincere thanks and appreciation to all of those people who made this dissertation possible. I would not have been able to complete this thesis without the assistance, support and guidance of the following people. In particular, I would like to express immense gratitude to my principal supervisor Professor Debra Anderson for her invaluable support, encouragement and guidance for this last and very important part of my PhD journey. I also especially appreciate my associate supervisor, Dr Cameron Hurst, for his invaluable contribution of statistical consultation in data analysis and for his instruction guidance and encouragement throughout this time. I would also like to express my sincere thanks to Professor Mary Courtney, for her invaluable support, encouragement and guidance throughout the majority of my PhD journey. She was indeed a great mentor. Much appreciation is also extended to the Royal College Nursing Australia and Queensland Health for the financial support to conduct this PhD project. Thanks also to my PhD colleagues from the Royal Brisbane and Women s Hospital for our shared experiences and support, which made this journey easier. Finally I would like to express my sincere thanks to my family for their support and encouragement over the time and a particular thank you must go to my husband, Greg for his support, patience, and encouragement during this PhD journey. ix

10 List of Tables Table 1: Explanatory Variables and Measurement Strategies Table 2: Demographics Comparison of Nursing Midwifery Labour Force, 2008 (AIHW, 2010) and Study Statistics, Table 3: Description of the Sample Table 4: Levels of Caring Efficacy and Job Satisfaction for Different Levels of the Categorical Socio-Demographic Variables Table 5: Modelling Approach (A Multivariable Analysis Using a GLM) (Bursac et al., 2008) Table 6: Correlation of Caring Efficacy, Job Satisfaction, Age in Years, Length of Time in Current Job, Years of Experience as a Registered Nurse Table 7: EFA of the 5 Factors of B-PEM (Pattern Matrix) Table 8: Factor Correlation Matrix of B-PEM Table 9: Betas and Standardised Betas of 28 items of the B-PEM Table 10: Cronbach s Alpha (ɑ) Analysis Showed High Internal Consistency (>0.06) for Each of the Scales and Subscales Table 11: Correlations Among General Self-Efficacy, Work Locus of Control and Practice Environment Table 12: Regression and Standardised Regression Weights: Outcome Variables Explained by Personal Control and Environmental Factors Table 13: Mediation Analysis (Three Steps) Table 14: Sobel s Test for Indirect Effect Table 15: Mediation Effect of Caring Efficacy to Job Satisfaction x

11 List of Figures Figure 1: Conceptual Framework Figure 2: Research Design Figure 3: General Linear Modelling Analyses of Socio-Demographic Variables and Caring Efficacy and Job Satisfaction Figure 4: Model for CFA of the Five Factors. Item numbering coincides with that given in Table Figure 5: Final Model Components Figure 6: Unmediated Model (Kenny, 2009) Figure 7: Mediation Model (MacKinnon, Fairchild & Fritz, 2007) Figure 8: Caring Efficacy and Job Satisfaction Model xi

12 List of Abbreviations BPE-M CE CES CFA CFI CI EFA GLM GSE GSES JS JSS ML NWIR PAF PCA PE RMSEA SEM UK US VIF WCQ WLC WLCS WOC Brisbane Practice Environment Measure Caring Efficacy Caring Efficacy Scale Comparative Factor Analysis Comparative Fit Index Confidence Interval Exploratory Factor Analysis General Linear Modelling General Self-Efficacy General Self-Efficacy Scale Job Satisfaction Job Satisfaction Survey Maximum Likelihood Nursing Work Index Revised Principal Axis Factoring Principal Component Analysis Practice Environment Root Mean Square Error of Approximation Structural Equation Modelling United Kingdom United States of America Variance Inflation Factor Ways of Coping Questionnaire Work Locus of Control Work Locus of Control Scale Ways of Coping xii

13 Abstract Background to the Problem: Improving nurses self-efficacy and job satisfaction may improve the quality of nursing care to patients. Moreover, to work effectively and consistently with professional nursing standards, nurses have to believe they are able to make decisions about their practice. In order to identify what strategies and professional development programmes should be developed and implemented for registered nurses in the Australian context, a comprehensive profile of registered nurses and factors that affect nursing care in Australia needs to be available. However, at present, there is limited information available on a) the perceived caring efficacy and job satisfaction of registered nurses in Australia, and b) the relationships between the demographic variables general self-efficacy, work locus of control, coping styles, the professional nursing practice environment and caring efficacy and job satisfaction of registered nurses in Australia. This is the first study to 1) investigate relationships between caring efficacy and job satisfaction with factors such as general self-efficacy, locus of control and coping, 2) the nursing practice environment in the Australian context and 3) conceptualise a model of caring efficacy and job satisfaction in the Australian context. Research Design and Methods: This study used a two-phase cross-sectional survey design. A pilot study was conducted in order to determine the validity and reliability of the survey instruments and to assess the effectiveness of the participant recruitment process. The second study of the research involved investigating the relationships between the socio-demographic, dependent and independent variables. Socio-demographic variables included age, gender, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector, state and marital status. Other independent variables in this study included general self-efficacy, work locus of control, coping styles and the professional nursing practice environment. The dependent variables were job satisfaction and caring efficacy. xiii

14 Results: A confirmatory factor analysis of the Brisbane Practice Environment Measure (B-PEM) was conducted. A five-factor structure of the B-PEM was confirmed. Relationships between socio-demographic variables, caring efficacy and job satisfaction, were identified at the bivariate and multivariable levels. Further, examination using structural equation modelling revealed general self-efficacy, work locus of control, coping style and the professional nursing practice environment contributed to caring efficacy and job satisfaction of registered nurses in Australia. Conclusion: This research contributes to the literature on how socio-demographic, personal and environmental variables (work locus of control, general self-efficacy and the nursing practice environment) influence caring efficacy and job satisfaction in registered nurses in Australia. Caring efficacy and job satisfaction may be improved if general self-efficacy is high in those that have an internal work locus of control. The study has also shown that practice environments that provide the necessary resources improve job satisfaction in nurses. The results have identified that the development and implementation of strategies for professional development and orientation programmes that enhance self-efficacy and work locus of control may contribute to better quality nursing practice and job satisfaction. This may further assist registered nurses towards focusing on improving their practice abilities. These strategies along with practice environments that provide the necessary resources for nurses to practice effectively may lead to better job satisfaction. This information is important for nursing leaders, healthcare organisations and policymakers, as the development and implementation of these strategies may lead to better recruitment and retention of nurses. The study results will contribute to the national and international literature on self-efficacy, job satisfaction and nursing practice. xiv

15 Related Presentations and Publications Publications in preparation Reid, C. L., Anderson, D., Hurst, C. & Courtney, M. Is the Brisbane practice environment measure (B-PEM) a valid and reliable instrument for use in an Australian population of registered nurses? Journal of Advanced Nursing. Reid, C. L., Anderson, D., Courtney, M. & Hurst, C. Socio-demographic factors affecting caring efficacy and job satisfaction in Australian registered nurses. Journal for Clinical Nursing. Reid, C. L., Anderson, D., Courtney, M. & Hurst, C. Examination of relationships and mediating effects of socio-demographics, general self-efficacy, locus of control, coping and the practice environment on caring efficacy and job satisfaction in Australian registered nurses. Nursing Research. Conference papers Reid, C. L., Anderson, D., Hurst, C. (Accepted for oral presentation 30 July-3 August, 2012) Relationships and mediating effects of self-efficacy, locus of control, coping and the practice environment on self efficacy in nursing and job satisfaction in Australian Registered Nurses" Honor Society of Sigma Theta Tau International 23rd International Nursing Research Congress, Brisbane, Queensland Australia Reid, C. L., Anderson, D., Hurst, C. & Courtney, M. (29 October 2 November 2011) Factors affecting caring efficacy and job satisfaction in Australian registered nurses (Abstract published). Honor Society of Nursing, Sigma Theta Tau International, 41st Biennial Convention, Grapevine, TX, USA. Reid, C. L., Anderson, D., Hurst, C. (22 24 September 2011) Examination of the relationships between general self-efficacy, work locus of control and the nursing practice environment on caring efficacy and job satisfaction in Australian registered xv

16 nurses. Fifth Pan-Pacific Nursing Conference and Seventh Nursing Symposium on Cancer Care, Hong Kong. Seminar presentations Reid, C. (2006). Examination of relationships and mediating effects of general selfefficacy upon work locus of control, ways of coping and the practice environment on caring efficacy and job satisfaction in Australian registered nurses. Paper presented at the Postgraduate Research Development Forum, School of Nursing and Midwifery, QUT. Reid, C. (2011). Examination of relationships and mediating effects of general selfefficacy upon work locus of control, ways of coping and the practice environment on caring efficacy and job satisfaction in Australian registered nurses. Paper presented at the Postgraduate Research Development Forum, School of Nursing and Midwifery, QUT. Awards/Achievements A scholarship from Queensland Health Research Higher Degree Support Initiative, (Appendix 1) A scholarship from Royal College of Nursing Australia Research scholarships (Appendix 2) Certificate of Completion for Introduction to Structural Equation Modelling using AMOS from Australian Consortium for Social and Political Research Incorporated in 2007 (Appendix 3) A member of Sigma Theta Tau International May 2011 (Appendix 4). xvi

17 Chapter 1: Introduction 1.1 Introduction Nursing, according to the International Council of Nurses (2010), encompasses autonomous and collaborative care of individuals of all ages, families, groups and communities, sick or well and in all settings. Nursing includes the promotion of health, prevention of illness, and the care of ill, disabled and dying people. Advocacy, promotion of a safe environments, research, and participation in shaping health policy and in patients and health systems management, and education are also key nursing roles. Reports suggest that improving nurses self-efficacy (i.e. nurses confidence and ability to do their job) and job satisfaction (JS) may improve the quality of nursing care to patients (Oulton, 2006; Aiken, Clarke & Sloane, 2000; Shorr, 2000). Moreover, to work effectively and consistently with professional nursing standards, nurses need to be able to make decisions about their practice (Manojlovich, 2005a). In order to identify what strategies and professional development programmes should be developed and implemented for registered nurses in the Australian context, a comprehensive profile of registered nurses and factors that affect nursing care in Australia needs to be available. However, presently there is limited information available on the perceived levels of self-efficacy and job satisfaction of registered nurses in Australia. There is also limited or no information on the relationships existing between the socio-demographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state) and job satisfaction and caring efficacy. Additionally a dearth of information is available on relationships that exist between the personal control variables (general self-efficacy [GSE], work locus of control [WLC], coping styles), the organisational variable, the professional nursing practice environment (PE) and caring efficacy (CE) and job satisfaction (JS) of registered nurses in Australia. 1

18 This chapter commences by providing an overview of the background and significance of this study. It will include the aim, objectives, research questions and research hypotheses tested in this study. The conceptual framework and definition of terms is provided along with the outline of the structure of this dissertation. Chapter 2 presents a review of the literature including general self-efficacy, work locus of control, coping styles, the professional nursing practice environment, caring efficacy, and job satisfaction. Chapter 3 provides a detailed description of the methodology used in both of the research studies. It includes, research design, research population, recruitment methods, sample size, assessment tools, data management and ethical considerations. Chapter 4 discusses the research findings and discussion of the pilot study. Chapter 5 discusses the research findings of study 2 including the psychometric evaluation of items selected from the Brisbane Practice Environment Measure for an Australian registered nurse population. Chapter 6 presents the research findings and discussion of socio-demographic factors affecting caring efficacy and job satisfaction in Australian registered nurses. The research findings of the Structural Equation Modelling are also presented. These are discussed in terms of relationships between general self-efficacy, work locus of control, coping styles and the professional nursing practice environment, socio-demographics (personal and organisational factors) and the outcome variables caring efficacy and job satisfaction in Australian registered nurses. Mediational analysis is also discussed in terms of caring efficacy as the mediator for the socio-demographic variables general self-efficacy, work locus of control, coping and the professional nursing practice environment to job satisfaction. The conclusions are presented in Chapter 7. This chapter incorporates the key findings and limitations of the two research studies. Following this, the implications of the research findings for clinical practice and further research are presented. Recommendations for future research are also included. 2

19 1.2 Background and Significance of this Study Approximately per cent of the global healthcare workforce is represented by nurses, while in the Australian healthcare workforce, nurses represent 55 per cent (Productivity Commission, ; WHO, 2006). However, there is currently a global nursing shortage, which is compromising the future goals of improving health and well-being (International Council of Nurses [ICN], 2010; Productivity Commission, 2007; WHO, 2006). Moreover, the effects of demographic factors in Australia such as the aging workforce (Schofield, 2007; AIHW, 2010) and other socio-demographic factors such as retirement; an increase in mobility of nurses between occupations; a slowing in the labour market; a decline in undergraduate commencements; increased average age of new nursing graduates and difficulties with retention will continue to have an effect upon this shortage in Australia until Meanwhile, a careful balance of both policy initiatives and recruitment of new nurses is required to improve nursing retention in order to support the aging population (Schofield, 2007). In addition, it has been reported that in Australia there were 30,000 qualified nurses no longer working in the healthcare system at the time of this current study (New South Wales Nurses Association, 2007), while Oulton (2006) reported that nursing shortages vary as a result of certain socio-demographic variables including type of nurse, geographic, location, sector, service and organisation. These variables need further investigation. Further currently in Australia it is reported that from 2004 to 2008, the number of employed nurses increased by 11.8%, from 243,916 to 272,741 (AIHW, 2010). However it is not reported whether this has made any difference to the preexisting nursing shortage. Additionally the ICN (2003) has reported that unsatisfactory work environments which support excessive workloads, lack of support for staff, workplace violence, stress, burnout, wage disparities, inequalities and little opportunities for decision making add to this shortage. Oulton (2006) reveals that there is not only a real shortage but also a pseudo-shortage (i.e. enough nurses are available, but not enough are willing to work under existing workplace conditions). The ICN ( ) reports that several countries continue to be 3

20 challenged by both underemployment and unemployed health professionals and also with extreme shortages. It is reported that job satisfaction affects the quality of patient care, (Aiken, Sloane, Clarke, Poghosyan, Cho and You et al., 2011; Takase, 2001). Nurses who are dissatisfied with their workplace distance themselves from their tasks and patients (Demerouti et al., 2000) and this affects the nurse patient relationship (Clarke & Aiken, 2006; Takase, 2001). Further, nurses who think that their efforts are not fully supported are inclined to leave the profession (Duffield, Roche, O Brien-Pallas and Catling-Paull, 2009). It is also reported that nurses may be highly dissatisfied with their work, with reports that there is a loss of a professional focus and nurses practicing isolated, routine tasks rather than higher level skills involving independent judgement and thought (Shorr, 2000). A United States study found that nurses might also lose confidence in their ability to perform professional nursing practice (selfefficacy) in the clinical setting according to the standards, regardless how much control they have to make decisions. This study indicated self-efficacy could be an important factor in contributing to professional nursing practice behaviours (Manojlovich, 2005a). Another more recent study showed collective efficacy (an expanded concept of self-efficacy) within a nursing unit had a positive effect upon improving nursing performance (Lee & Ko, 2010). At the present time there is limited or no information available on a) caring efficacy and job satisfaction of Australian registered nurses and b) the relationships between the socio-demographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), personal control variables (general self-efficacy, work locus of control, coping styles), and the organisational variable (professional nursing practice environment), caring efficacy and job satisfaction of Australian registered nurses. Therefore, study 2 will provide a comprehensive profile of registered nurses and the variables that affect nursing practice in Australia. This current study will improve upon previous international studies in the following ways. It was suggested by Coates (1997) that caring efficacy (nurses perceived selfefficacy) be measured with other variables such as coping styles, general self- 4

21 efficacy and work locus of control to further assess relationships between these personal control variables. This study will investigate these variables in relation to caring efficacy. It was also suggested that a random selection from a diverse population of nurses be used. A random sample from a diverse range of nurses in Australia will be examined in relation to caring efficacy. Further, social desirability response effect bias should be reduced as it is a limitation for studies where participants self-rate their own behaviours. In both study 1 and 2, the voluntary anonymous return of the questionnaires via the Australian postage service assured participant confidentiality. 1.3 Statement of Problem In order to provide care to patients, nursing as a profession relies on the establishment of caring relationships with people who are in need of health care. Watson, Jackson, and Borbasi (2005, p. 114) stated that caring is essential to the theory and practice of nursing. The current study developed from discussions the principal investigator had with nurses who were frustrated about their perceived lack of confidence and their inability to provide optimal care for their patients in an acute care setting within Queensland Health. These nurses believed they were not able to access the appropriate multidisciplinary specialist resources, which were available at the hospital, in order to ensure patients received adequate pain management. Furthermore they were not confident they would be able to initiate contact with this specialist service as they also believed it was a medical referral service only. The nurses indicated that their recommendations to the medical staff were not followed up and also, they believed that they were not supported by the nursing management when the problem was reported. It was suggested to these nurses that they could contact the specialist team and speak to nursing staff, however they reported they were still not confident they could do so and hence no action was taken. 5

22 1.3.1 Aim The aim of this study was to describe the factors that affect self-efficacy in nursing and job satisfaction in the Australian context and provide a basis to develop a model that identifies nurses self-efficacy and job satisfaction Objectives In order to achieve the aims of the pilot study (Study 1) and the main study (Study 2) the following objectives were identified Study 1 The objectives of study 1 were: 1. to test the validity and reliability of the Caring Efficacy Scale (CES) (Coates, 1997), Job Satisfaction Survey (JSS) (Spector, 1985), General Self-Efficacy Scale (GSES) (Schwarzer & Jerusalem, 1995), Work Locus of Control Scale (WLCS) (Spector, 1988), Ways of Coping Questionnaire (WCQ) (Folkman & Lazarus, 1988) and the Nursing Work Index Revised (NWIR) (Aiken & Patrician, 2000) ; 2. to assess the feasibility of all questions for this study; and 3. to assess the effectiveness of the participant recruitment process Study 2 The objectives for study 2 were: 1. to measure the levels of perceived caring efficacy and job satisfaction of registered nurses in Australia; 2. to measure the associations of caring efficacy and job satisfaction in the registered nurse population in Australia according to age, gender, marital status, level of education, years of experience, years in current job, 6

23 employment status, geographical location, specialty area, health sector and Australian state; 3. to identify if the socio-demographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state) predict caring efficacy and job satisfaction levels in an Australian registered nurse population; 4. to explore if a high level of caring efficacy enhances the bivariate relationships between work locus of control, general self-efficacy, coping styles, the professional nursing practice environment and job satisfaction; and 5. to explore whether caring efficacy mediates the relationship among the variables of socio-demographics, work locus of control, general self-efficacy, coping styles, professional nursing practice environment and job satisfaction Research Questions and Hypotheses Research Questions for Study 1 The research questions of study 1 (pilot study) are: 1. Is the Caring Efficacy Scale (Coates, 1997) a valid and reliable instrument for the Australian registered nurse population? 2. Is the Job Satisfaction Survey (Spector, 1985) a valid and reliable instrument for the Australian registered nurse population? 3. Is the General Self-Efficacy Scale (Schwarzer & Jerusalem, 1995) a valid and reliable instrument for the Australian registered nurse population? 4. Is the Work Locus of Control Scale (Spector, 1988) a valid and reliable instrument for the Australian registered nurse population? 5. Is the Ways of Coping Questionnaire (Folkman & Lazarus, 1988) a valid and reliable instrument for the Australian registered nurse population? 6. Is the Nursing Work Index Revised a valid and reliable instrument for the Australian registered nurse population (Aiken et al., 2000)? 7. Is the Caring Efficacy Scale feasible for use in a cross-sectional study for the Australian registered nurse population? 7

24 8. Is the Job Satisfaction Survey feasible for use in a cross-sectional study for the Australian registered nurse population? 9. Is the General Self-Efficacy Scale feasible for use in a cross-sectional study for the Australian registered nurse population? 10. Is the Work Locus of Control Scale feasible for use in a cross-sectional study for the Australian registered nurse population? 11. Is the Ways of Coping Questionnaire feasible for use in a cross-sectional study for the Australian registered nurse population? 12. Is the Nursing Work Index Revised feasible for use in a cross-sectional study for the Australian registered nurse population? Hypotheses for Study 1 1. The Caring Efficacy Scale is a valid and reliable instrument for the Australian registered nurse population. 2. The Job Satisfaction Survey is a valid and reliable instrument for the Australian registered nurse population. 3. The General Self-Efficacy Scale is a valid and reliable instrument for the Australian registered nurse population. 4. The Work Locus of Control Scale is a valid and reliable instrument for the Australian registered nurse population. 5. The Ways of Coping Questionnaire is a valid and reliable instrument for the Australian registered nurse population. 6. The Nursing Work Index Revised is a valid and reliable instrument for the Australian registered nurse population. 7. The Caring Efficacy Scale is feasible for use in a cross-sectional study for the Australian registered nurse population. 8. The Job Satisfaction Survey is feasible for use in a cross-sectional study for the Australian registered nurse population. 9. The General Self-Efficacy Scale is feasible for use in a cross-sectional study for the Australian registered nurse population. 10. The Work Locus of Control Scale is feasible for use in a cross-sectional study for the Australian registered nurse population. 8

25 11. The Ways of Coping Questionnaire is feasible for use in a cross-sectional study for the Australian registered nurse population. 12. The Nursing Work Index Revised is feasible for use in a cross-sectional study for the Australian registered nurse population? Research Questions for Study 2 The research questions of study 2 (main study) are: 1. What is the level of caring efficacy and job satisfaction of the registered nurse population in Australia? 2. Do levels of caring efficacy and job satisfaction in the registered nurse population in Australia vary according to age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state? 3. Do the variables age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state predict caring efficacy and job satisfaction? 4. Do levels of self-efficacy in registered nurses vary according to work locus of control, coping styles, general self-efficacy, the professional nursing practice environment and the (outcome variables) caring efficacy and job satisfaction levels? 5. Does a high level of caring efficacy enhance relationships between work locus of control, general self-efficacy, coping styles, professional nursing practice environment and job satisfaction? 6. Does caring efficacy mediate the bivariate relationship among the scales of work locus of control, general self-efficacy, coping styles, professional nursing practice environment and job satisfaction? 9

26 Hypotheses for Study 2 The following hypotheses were tested in study 2: 1. High levels of perceived caring efficacy and job satisfaction exist in registered nurses in Australia. 2. Levels of caring efficacy and job satisfaction in a registered nurse population in Australia vary according to age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state. 3. Age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state explain caring efficacy and job satisfaction levels in an Australian registered nurse population. 4. A high level of caring efficacy enhances relationships between work locus of control, general self-efficacy, coping styles, the professional nursing practice environment and job satisfaction. 5. Levels of self-efficacy in registered nurses vary according to work locus of control, coping styles, general self-efficacy, the professional nursing practice environment and the (outcome variables) caring efficacy and job satisfaction levels. 6. Caring efficacy mediates the bivariate relationship among the variables of socio-demographics, work locus of control, general self-efficacy, coping styles, the professional nursing practice environment and job satisfaction. 1.4 Conceptual Framework The conceptual framework was developed by connecting a group of concepts by propositions. This is also referred to as a theoretical model (Clifford, 1997). The conceptual framework for this study emanates from the theoretical framework of caring efficacy which is reflective of Watson s transpersonal caring theory (Watson 1985, 2005) and social learning theory (Bandura, 1977, 1986) which refers to the belief that one has the ability and confidence to get things done in given 10

27 circumstances. Efficacy beliefs are about judgements on one s performance in specific domains. For example, the domain considered for the development of the Caring Efficacy Scale (Coates, 1997) was nurses beliefs that they can express caring orientations, attitudes and behaviours as well as develop caring relationships with patients (Coates, 1997). The Caring Efficacy Scale was developed by Coates in 1997 and was used in this current study. Furthermore, to assist in understanding the nature of nursing in Australia, the study aimed to explore the relationships among sociodemographic variables, personal control variables (general self-efficacy, work locus of control and ways of coping) and the environmental variable (practice environment) and caring efficacy and job satisfaction. The development of this conceptual framework is necessary as there is a limited amount of literature available discussing these variables in the Australian context Caring in Nursing Many nurse theorists have identified that the concepts of care and caring are associated with nursing and these have defined nursing since the development of the profession (Alligood and Tomey, 2010). Early nurse theorists set the way for nursing in the United States of America and Australia. Leininger (1988, 2002) and Watson (1999) identified that caring in nursing is essential and suggested that human caring involves values, a will and a commitment to care, knowledge, caring actions, and consequences (Watson, 1999, p. 29). Furthermore Leininger (1991, p. 35) suggests that care is the essence of nursing and the central, dominant and unifying focus of nursing. Whereas Gaut (1992) describes caring as the presence and personal commitment to being involved in a meaningful process. Watson s early theory (Watson, 1985) is a form of humanism with its origin in metaphysics. Nursing according to Watson, puts the main focus on the process of human care for individuals, families and groups and aims to reach a higher degree of harmony within their mind, body and soul which is achieved through caring transactions (McCance, 2003, p. 103). In relation to nursing, the views of Boykin and Schoenhofer (1993) are that all people are caring and they maintain the idea that persons as caring involves a commitment to know self and others as caring (McCance, 2003, p. 103). 11

28 Coates (1997) described the nature of caring as a process, a presence, responding to patient needs, using transformational relationships with a holistic and creative approach in order to assist patients as their needs change (Amendolair, 2007, Sadler, 2000). Nursing, according to Leininger (1988, p. 154) suggests care should be supported by the concepts of health and environmental contexts. For Leininger (1988, p. 140) caring: in the generic sense refers to those assistive, supportive, or facilitative acts towards or for another individual or group with evident or anticipated needs to ameliorate or improve a human condition or lifeway. Leininger (1988, p. 140) separates the generic sense of caring from professional caring by defining professional caring as those cognitive and culturally learned behaviours, techniques, processes, or patterns that enable or help an individual, family, or community to improve or maintain a favourable healthy condition or lifeway. Leininger (cited in Ray, 1981, p. 27) believes caring is the intellectual, theoretical, heuristic, and central practice focus of nursing. It is reported, that for nurses to display caring, they need to be able to develop trust with the person receiving the care (Saddler 2000). Nurses are professionally knowledgeable and implement skilled and therapeutic interventions (Sadler, 2000). However to do so, nurses need to be able to be present (Amendolair, 2007). Presence is described by Tavernier (2006, p. 152) as being transformative to the patient care experience, having the potential to hinder or facilitate the healing process Covington (2003) describes caring presence as the connection that happens between the nurse and the patient and is how nurses express caring. Caring presence is further defined by Covington, (2003, p.304) as a way of being, a way of relating, a way of being-with and being there, and as a nursing intervention. In contrast to the caring presence of nurses, early work from (Blegen, 1993) identified that nurses were dissatisfied with their workplace by disengaging from patients, showing emotional and physical exhaustion (burnout) or either left or intended to leave the organization and profession. A study by Wade et al. (2008) conducted in the United States of America suggested it is essential for organisations to support quality of care in nursing in order to create an appropriate environment for nurses to be satisfied in their work. Traditionally, the nursing profession looks after the sick with care and compassion. Additionally, human caring is described by today s theorists including, Watson (2005) and Leininger (1988) as being central to the discipline and profession of 12

29 nursing. Leininger (1988) identified that nursing is in the unique position to provide quality care by means of in-depth knowledge of meaning and experiences in relation to healing and the provision of health care promotion to those who are well, unwell, disabled and dying. Coates (1997) reported that current healthcare environments however, are showing trends towards emphasizing the importance of accountability with healthcare programs and are concerned more with costs and outcomes whereas nursing tradition values process and quality of patient care. Therefore the importance of further exploring and assessing the nature of caring through research is essential as well as pursuing the implications of the research for nursing practice and education Coates (1997, p. 53) Caring Efficacy Caring efficacy is the belief or ability of a person to convey a caring orientation and build up caring relationships with patients. The emphasis on the caring relationship and the caring experience in Watson s transpersonal caring theory (1996, 1988, 1979) assisted in developing the Caring Efficacy Scale (Coates, 1997). Self-efficacy is people s judgments of their capabilities to organize and execute courses of action required to attain designated types of performances. It is concerned not with the skills one has but with judgments of what one can do with whatever skills one possesses (Bandura, 1986, p. 391). Watson describes human caring theory in practice as allowing the commitment and consciousness of the nurse to transcend (or at least attempt to transcend) the physical material surface and reach beyond, to touch the human center of the person (Watson, 1988, p. 176). Moreover, earlier nursing theorists such as Newman (1986), Parse (1981) and Rogers (1970) define human beings as being open system energy fields that continuously engage in an exchange of energy with the environment, energy centres and patterns of consciousness: Watson, (1988, p. 177) suggests: Nursing within a transpersonal caring perspective attends to the human center of both the one caring and the one being cared for; it embraces a spiritual, even metaphysical dimension of the caring process; it is concerned with preserving human dignity and 13

30 preserving humanity in the fragmented, technological, medical curedominated systems. Human caring theory is reported by Watson (1988) to be the highest form of commitment. The ethic of caring is described by Watson (1988) as nurses taking on a personal authorization to employ a caring awareness and ethic. It has also been reported that most nurses work from a caring model and also usually as employees of a system shaped and controlled by medicine (Watson, 1988). This human caring theory challenges the traditional models of health-illness and science, suggesting they are inadequate for both patients and the caring-healing processes experienced by nurses in transpersonal caring moments: The expanded dimensions of human caring theory are imbedded in an ethic and human value of caring as a moral ideal, with an ontological and epistemological commitment to preserve and restore the human center in theory and practice. In the human center, human caring and healing become transpersonal and intersubjective and open up a higher energy fieldconsciousness that has metaphysical, transcendent potentialities (Watson, 1988, p. 181). The contribution of nursing in the promotion of self-healing of the whole person uses nursing interventions such as decisional control, reminiscence, journaling, therapeutic touch, presence, active listening and humor (Marckx, 1995, p. 45). The goal of the transpersonal caring theory is to assist people toward a level of harmony or health within the mind, body and soul. This allows the experience of self-healing within the person. Coates (1997) reported that self-efficacy theory provides a connection between human beliefs and behaviours in environmental situations and therefore informs the definition and assessment of caring (Coates, 1977, p. 54) Social Cognitive Theory (Efficacy Theory) Self-efficacy includes the confidence in capability to regulate one s motivation, thought processes, emotional states and the social environment as well as levels of behavioural attainment (Maibach & Murphy, 1995, p. 38). Bandura (1986), in describing self-efficacy theory asserted that personal mastery expectations are the primary determinants of behavioural change and accordingly, 14

31 this theory suggests there are two types of expectations responsible for significantly influencing behaviour: outcome expectancies (i.e. the perception that particular behaviours will result in particular outcomes), and self-efficacy expectancy or the belief that one is able to accomplish certain behaviours. Self-efficacy beliefs are influenced by four sources of information: performance accomplishments (the feeling of success), vicarious experience (observing the successful achievement of others), verbal persuasion (verbally convincing people they can succeed) and physiological information (the physical and emotional capacity of someone s ability). The strongest predictor of self-efficacy is that of performance accomplishments or mastery. Behaviour change is influenced by a person s selfefficacy (Mayer et al., 2005; Lenz & Shortridge-Baggett, 2001; Bandura, 1977 ) and efficacy expectations predict future performances (Hickey, Owen & Froman, 1992) Efficacy Expectations Bandura (1977) further describes efficacy expectations as differing on certain dimensions that may also have important performance implications. These dimensions include magnitude, generality and strength. Magnitude refers to how individuals deal with tasks according to difficulty. Some people may be limited to simple tasks, while others will extend themselves to moderate or even the most difficult tasks. Expectations that differ in generality are described as those experiences that create limited mastery expectations, or those that inspire a general sense of effficacy expanding way beyond the specific treatment circumstances. The varying degree of the strength of expectancies illustrated by weak expectations are easily extinguishable by noncoroborating events, while those who have expectations of mastery over the experience will perservere in their efforts to cope despite the noncoroborating experiences. It has been established that mastery expectations influence performance (Bandura, 1977) Sources of Efficacy Expectations Self-efficacy is influenced by four indicators i.e., performance accomplishments, vicarious experience, verbal persuasion, and physiological arousal (Gist et al., 1992). 15

32 Peformance accomplishments: This source of efficacy is based on personal mastery experiences (i.e. success raises mastery expectations and repeated failures especially at the start of an experience, diminish them). The development of strong efficacy expectations by repeated successes and sustained effort reduces the negative effect of ocassional failure (Bandura, 1977). Vicarious experience (observing others): While observing others is regarded as a less effective source of self-efficacy compared to the actual experience of a task, it does help people judge their own self-efficacy. Other people become role models and display information such as the degree of difficulty to carry out a behaviour. Those observing are then able to judge their own abilities upon which they can base their estimation of success (Lenz & Shortridge-Baggett, 2002). Verbal persuasion: This source of self-efficacy is regularly and easily used often to support the other sources. Health professionals convince people that they can be successful with a task by means of instructions, suggestions and advice. This is, however, an even weaker source of self-efficacy when used by itself, because it does not concern the personal experiences of those being encouraged. However, if people can be convinced they are capable of doing the task and do believe they can, they will persevere provided that the task is realistic (Lenz & Shortridge-Baggett, 2002). Physiological information (self-evaluation of physiological and emotional states): People judge their capacity to be successful at a task based upon how they are feeling; for example, stress, anxiety, tension, depression are a guage of personal deficiency and when strength and perserverance is required, fatigue and pain are indicators of lower physical efficacy (Lenz & Shortridge-Baggett, 2002). In summary, a person s judgement of their self-efficacy is specific to both the job at hand and the situation. Furthermore people use these judgements in relation to achieving a goal. This means individuals judge the effects of their actions and then the interpretation of the effects of these actions assist to create people s efficacy beliefs e.g. results which are interpreted as successful increase self-efficacy whilst those interpreted as not successful lessen it (Bandura, 1977). 16

33 1.4.6 Caring, Self-Efficacy and Coping Studies from the United States of America (USA) have reported that behaviour such as nurses removing themselves from the patient s bedside except when required for assigned tasks, disengaging themselves from intense healthcare situations and locating themselves at the nurses station instead of being with the patients may be the result of a lack of self-efficacy (Manojlovich, 2005b; Coates, 1997). How employees manage the challenges of their jobs is influenced by their perceived selfefficacy (Murphy, 2005). Manojlovich (2005b) found strong relationships between self-efficacy and structural empowerment (environmental factors) and also strong relationships between self-efficacy and professional practice behaviours. Selfefficacy is the reason employees either manage a situation well using problem solving coping, or remove themselves from the situation and resort to dysfunctional coping styles (Bandura, 1997). Self-efficacy will be measured in this study using the General Self-Efficacy Scale (Schwarzer & Jerusalem, 1995) Self-Efficacy in Nursing and Job Satisfaction There is a paucity of literature on job satisfaction and the practice environment in nursing in the Australian context and therefore it was decided, in this current study, to replace professional practice behaviours with job satisfaction and further develop a model of nursing self-efficacy, and the nursing practice environment with personal control factors (general self-efficacy, work locus of control and coping) as suggested in Coates (1997) Self-Efficacy, Caring, Work Locus of Control, Socio-Demographics and Job Satisfaction Self-efficacy is reported to be correlated with work locus of control, and both of these may affect job satisfaction (Judge & Bono, 2001; Judge, Locke & Durham, 1998). Relationships between socio-demographic variables and job satisfaction have also been identified by (Shields and Ward, 2001). Further, job satisfaction has been found to be associated with job retention (Zangaro & Soeken, 2007). In addition to this, recent research has also found organisational factors can affect job satisfaction 17

34 in nurses (Duffield, et al., 2009). The Job Satisfaction Survey (Spector, 1985) was used to measure job satisfaction in this current study. The previous literature by Coates (1997) also informed the development of a conceptual model for this current study, which was evaluated using the structural equation modelling (SEM) technique. 1.5 Summary of Conceptual Framework In order to address the above research questions (see page 9), the following conceptual framework is proposed (see Figure 1). According to the conceptual framework of this research, this study investigates the relationships between sociodemographic, personal control and organisational variables, and caring efficacy and job satisfaction. Perceived self-efficacy is a powerful tool for people when coping with stress (Lazarus et al., 1987). Factors characteristic to nursing include stress, frequent job turnover and job dissatisfaction (Wong et al., 2001). Further a meta-analysis by Zangaro and Soeken (2007) reported autonomy, job stress, management leadership style, workplace communication and nurse physician collaboration as variables in the nursing literature that mediated or moderated nurses job satisfaction. Furthermore Zangaro and Soeken (2007) found nurses often used problem-solving strategies as a way of coping with the demands of the job and stressful situations. Additionally, it is reported that people with high levels of perceived self-efficacy use problem-solving strategies to manage workplace issues, while those with low levels of perceived self-efficacy resort to more dysfunctional coping styles. Bandura (1997) identified that if a person s perceived self-efficacy is high, better coping strategies may be used in difficult work situations. Information such as this would be valuable for nursing leaders and healthcare organisations to enhance self-efficacy with nursing staff (Manojlovich, 2005a). Coping as measured by the Ways of Coping Questionnaire (Folkman and Lazarus, 1988) was investigated in this study in relation to caring efficacy and job satisfaction. 18

35 A meta-analysis by Bono and Judge (2001) found that general self-efficacy and work locus of control were predictors of job satisfaction. The socio-demographic factor of age has also been found to be related to job satisfaction (Amendolair, 2007; Tyson & Pongruengphant, 2004; Spector, 1985). Further, internal work locus of control is reported to increase with the socio-demographic variables age, length of work tenure, gender (males), increased education levels and certain industry types. It has been found that people with an internal locus of control have higher levels of satisfaction compared to those with an external locus of control (Letvak, 2002; Spector, 1985, Rotter, 1966). Those with an external locus of control are more likely to feel alienated from the workplace and show sabotage, aggression and withdrawal behaviours as a way of dealing with frustration with a job. Moreover, people who have an internal locus of control are reported to cope better than externals (Duffy et al., 1977), while those with an external locus of control have been found to have fewer coping skills (Anderson, 1977). Externals also distance themselves from the circumstances of constant failure, which may involve leaving the job (Forte, 2005). More importantly, patient care may be compromised and negative outcomes for nurses occur if nurses are dissatisfied with their work (Aiken et al., 2011; Begat et al., 2005). Locus of control for this study was measured by the Work Locus of Control Scale (Spector, 1982) This information informed the development of the conceptual framework for this study. No studies were identified that have examined the levels of caring efficacy and job satisfaction in registered nurses in Australia. Also no studies have investigated the socio-demographic variables, personal control and environmental variables described above in relation to caring efficacy and job satisfaction in registered nurses in Australia. 19

36 Personal characteristics (e.g. age, gender, marital status, education levels, years of experience, years in current job, speciality and employment status) General Self-Efficacy Characteristics of organisation location (e.g. rural, remote, regional and metropolitan) & Australian State Work Locus of Control Caring Efficacy Behaviour Job Satisfaction Health sector (public or private) Ways of Coping Nursing Practice Environment Figure 1: Conceptual Framework 20

37 1.6 Definition of Terms For the purposes of classification, several key terms are defined here including selfefficacy, caring efficacy, locus of control, coping and job satisfaction. Self-efficacy is defined as people s judgments of their capabilities to organize and execute courses of action required to attain designated types of performances. It is concerned not with the skills one has but with judgments of what one can do with whatever skills one possesses (Bandura, 1986, p. 391). Caring Efficacy is the conviction or the belief in one s ability to express a caring orientation and to develop caring relationships with clients or patients (Coates, 1997, p. 53). Watson (1988, p. 176) describes human caring theory in practice as allowing commitment and consciousness of the nurse to transcend (or at least attempt to transcend) the physical material surface and reach beyond, to touch the human center of the person. Locus of control is the extent to which people believe that they have control over their own fate (Thomas, Sorensen and Eby, p. 1057, 2006). Coping may be defined as cognitive and behavioural efforts to manage specific external and/or internal demands appraised as taxing or exceeding the resources of the individual (Folkman & Lazarus, 1988, p. 6). This definition is defined by the authors of the Ways of Coping Questionnaire which was used in this current study. Job satisfaction is regarded as an emotional affective response to a job or particular aspects of a job (Locke, 1976). Socio-demographics is defined as pertaining to, or characterized by a combination of sociological (=related to sociology) and demographic (= relating to populations) characteristics (MacMillan, 2010). 21

38 1.7 Summary Several studies have investigated associations between different socio-demographic, psychosocial and environmental factors and job satisfaction, but there is very little evidence available that discusses these factors in registered nurses in the Australian context. Further, a small number of studies examining caring efficacy exist and have found that environmental and personal control factors such as self-efficacy are influential on the development of professional practice behaviours in nurses in America (Manojlovich, 2005a; Manojlovich, 2005b). This current study extends this work by examining caring efficacy in the context of personal control variables (coping styles, work locus of control and self-efficacy) such as that suggested by Coates (1997), along with the professional nursing practice environment and job satisfaction in the Australian context. This is the first time these personal control and organisational variables have been studied together in registered nurses in the Australian context in relation to caring efficacy and job satisfaction. There is sufficient evidence in the literature that supported the inclusion of these variables in this current study. In addition a conceptual model of caring efficacy and job satisfaction for registered nurses in Australia was developed as a useful construct for this research. For this purpose, a Structural Equation Modelling technique was used to investigate the relationships among the factors of coping styles, work locus of control, general self-efficacy and the professional nursing practice environment and the outcome variables caring efficacy and job satisfaction concurrently. Testing for construct validity of the Brisbane Practice Environment Measure was also conducted, as this measure had not at the time of this current study undergone such testing. To achieve the purpose of this research, two studies were developed. Study 1, a pilot study, was conducted in order to determine the face validity and testretest reliability of the survey instruments, to assess the effectiveness of the participant recruitment process and to determine the sample size required for the subsequent study. The comments and missing data (face validity) indicated that the 22

39 Nursing Work Index Revised was not valid for the Australian registered nurse population. Study 2 was conducted to examine relationships between the socio-demographic variables (potential confounders), personal control variables (self-efficacy, work locus of control, coping styles), and the organisational variable (the professional nursing practice environment) and caring efficacy and job satisfaction. The evaluation of the psychometric properties of the Brisbane Practice Environment Measure (B-PEM) (Webster et al., 2009), (which replaced the Nursing Work Index Revised) in an Australian registered nurse population was conducted. The following chapter will present a review of the literature including general selfefficacy, work locus of control, coping styles, the professional nursing practice environment, caring efficacy, and job satisfaction. 23

40 Chapter 2: Literature Review 2.1 Introduction In order to understand the background to this research study, this chapter provides an overview of the literature and the rationale for the study of socio-demographics, personal control and organisational variables, and caring efficacy and job satisfaction. This review begins with a review of previous caring efficacy and selfefficacy research and a discussion of self-efficacy in nursing. This will be followed by a discussion of the construct of caring efficacy, general self-efficacy, work locus of control, coping styles, the professional nursing practice environment and job satisfaction. Finally, the rationale for the inclusion of these variables in the proposed model and the expected pattern of relationships among them are discussed. In order to identify the relevant literature relating to this topic a comprehensive literature review was undertaken. Electronic databases, including the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, Ovid MEDLINE (all years), and Ovid CINAHL (to December 2011) and PubMed (1993 to December, 2011) were searched. Reference lists of potentially useful articles were also searched. Only published articles in the English language were reviewed. There were no restrictions by date. The titles and abstracts of references identified by the search strategy used above were assessed. Full reports of all potentially relevant literature were then retrieved. Reference lists of retrieved studies were screened to identify further studies, which were also retrieved. The current global nursing shortage is well reported. According to Oulton (2006) a global report by the World Health Organization (2006) on the nursing shortage found that by 2010 Africa would have required a further 1 million more health professionals, whilst in Australia in 2010 there was an expected shortage of 40,000 nurses. By 2012 it is suggested the United States will be in need of 1 million nurses, Canada is expected to be short 113,000 nurses by the year 2016 and by 2025, Denmark will be short 22,000 nurses. This nursing shortage is damaging the future of 24

41 international health systems and undermining the capacity of nursing to meet patients needs (Oulton, 2006). It has also been shown that nursing shortages are associated with increased mortality, staff violence, accidents/injuries, cross infection and adverse postoperative events, (Oulton, 2006, p.374) An international review examined the importance of work environments and hospital outcomes in nine countries (Aiken, Sloane, Clarke, Poghosyan, Cho, You, Finlayson, et al., 2011). The finding was that in the majority of studies, more than one third of nurses working in hospitals were dissatisfied with their work. Nurses dissatisfaction with work can lead to emotional exhaustion and burnout and further affect patient outcomes (Aiken et al., 2011; Aiken & Sloane, 1997). Further, it has been found in the United Kingdom that 14.3% of newly qualified nurses and midwives do not enter the profession and in Australia, 66 % of new graduates report they contemplated leaving within their first year of practice (International Council of Nurses, 2003). Globally, nurses have reported several reasons for their discontent including understaffing, lack of human resources, poor skill mix and an inability to give the care they want to give. In addition, the nurses who have left organisations reported issues such as a lack of access to professional development, stress, workplace violence, bullying and harassment, lack of support and feeling valued, unacceptable pay and lack of opportunity for autonomy and control over their work (Oulton, 2006). Nurses who believe they do not have the ability in difficult situations to initiate the motivation and resources required to do the work (self-efficacy), may have a negative affect on patient care (Manojlovich, 2005a). Furthermore, it is believed that people with high levels of perceived self-efficacy are more likely to persist in managing difficult situations and continue through difficult times (Gist & Mitchell, 1992). In a later study, Judge and Bono, (2001) found high levels of self-efficacy in individuals was positively related to higher levels of job satisfaction. 25

42 2.2 Self-Efficacy in Nursing An earlier study by Laschinger, Sabiston, and Kutszcher (1997), suggested that nurses who have control over the content and context of their work may show an increased contribution in decision making. This has been in the past regarded as essential for professional practice (Alexander, 1982). Despite this, Manojlovich (2005a) reports there remains a lack of control over the context and content of nurses work and further suggests that power for many nurses remains intangible. Moreover, important characteristics of nursing practice today include nurses having the ability to develop and continue therapeutic relationships with patients, nurses having autonomy and control over the practice environment (Scott, Sochalski & Aiken, 1999) and nurses having more involvement in decision making (Aiken et al., 2011). In addition, employee satisfaction is enhanced when organisations offer access to authority (Laschinger, Shamian & Thomson, 2001). Despite this, nurses continue to complain of feeling powerless in their ability to make decisions (Manojlovich, 2005a). Nurses should have the confidence and the belief that they have the necessary knowledge and control over all the activities required of them in their practice in order to make decisions (Australian Nursing and Midwifery Council, 2005;Manojlovich, 2005a), otherwise, effective and consistent practice within the professional standards may be compromised (Manojlovich, 2005a). Nurses need to trust that they have the authority necessary to provide skilled care and be comfortable as decision-makers and care providers (Manojlovich, 2005a). A strong relationship has been reported by Manojlovich (2005b) to exist between nurses self-efficacy and nurses professional practice behaviours, which ultimately may affect the quality of patient care that nurses provide. Behaviour such as nurses removing themselves from the patient s bedside except when required for assigned tasks, disengaging themselves from intense healthcare situations and locating themselves at the nurses station instead of being with the patients may assist nurses to cope with the work environment; however, these behaviours may be the result of a lack of self-efficacy (Manojlovich, 2005b). 26

43 In the past, the responsibilities of nurses have been extensive in the workplace, and have been accompanied by limited authority (Manjolovich, 2005b; Parker, Tuckett, Eley & Hegney, 1993). Bandura (1986) found that when perceived control is high, confidence is improved, making tasks less stressful and more worthwhile (Bandura, 1986). Therefore, self-efficacy should be encouraged in nursing to improve practice (Manojlovich, 2005a). Perceived self-efficacy also affects the way in which employees manage challenging job demands. Employees must believe they have the essential behavioural, cognitive and motivational resources required for the job or they will focus on the daunting parts of the task, use insufficient effort and fail (Stajkovic & Luthans, 1998). They either develop strategies to manage the circumstances better or separate from the situation and expend the least effort required to manage the task (Manojlovich, 2005a). These two types of adaptive strategies are discussed by Bandura (1977, p. 466): People who perceive they have a high level of self-efficacy use problemsolving coping to enhance their work situation. Whereas those who don t, believe there is not much they can do to alter the stressful aspects of their job and choose dysfunctional ways of coping to manage their stress. Therefore, if a person s perceived self-efficacy is high, better coping strategies may be used in difficult work situations. Information such as this is valuable for nursing leaders to enhance self-efficacy of nursing staff (Manojlovich, 2005a). For nurses to remain competent and increase self-efficacy they require access to the necessary resources in the workplace (Mayer, Andrusyszyn & Iwasiw, 2005). Further, different levels of perceived self-efficacy influence the amount of effort made when new behaviours are learned e.g. high levels will result in a person applying energy to a task, persisting longer, setting more challenging goals and continuing in the face of adversity until it is mastered and the new behaviour is learned (Bandura, 1986; 1994). High levels of perceived self-efficacy may also take place when people face less difficult tasks (self-efficacy magnitude) (Mayer et al., 2005). 27

44 Self-efficacy is enhanced when threatening situations are removed. This occurs more frequently in those individuals who have coped successfully in the past as opposed to those who have repeatedly failed (Bandura, 1977). Perceived self-efficacy at work influences the way employees manage the requirements and challenges of their jobs (Murphy, 2005). Self-efficacy is therefore an explanation of why employees with the same level of skill adjust to, manage the situation better, or remove themselves from a situation and only provide the minimum effort required to perform tasks. In summary there is some support for the proposition that high levels of self-efficacy (confidence and ability) are essential for nurses to work within the standards and effectively manage the demands of the workplace. 2.3 Caring Efficacy A US study by Manojlovich (2005a) established that a significant relationship exists between self-efficacy in nursing and professional practice behaviours, which in turn provides information about the determinants of practice behaviours and the sources of those determinants. This study (Manojlovich 2005a) also found that environmental, self-efficacy and personal factors had an influence over the development of professional practice behaviours. A lack of self-efficacy in nursing (i.e. nurses believing they do not have the ability to activate the required motivation, cognitive resources and courses of action required to have control over their work) may have an effect on nurses choosing to stay away from patients, gather at the nurses station and remove themselves from intense healthcare situations. While this is not functional professional practice, it is an option for reducing stress in extremely challenging work situations (Begat, Ellefsen & Severinsson, 2005; Manojlovich, 2005a). Professional nursing practice behaviours have been described as the ability to establish and maintain patient relationships, autonomy, decision making and control over practice and collaborative relationships particularly with the physicians at a unit level. Other self-efficacy research in nursing has also shown that high levels of self-efficacy may be an indicator for improved professional nursing practice behaviours (Le Blanc, Schaufeli, Salanova Llorens & Nap, 2010; Lee & Ko, 2010; Manojlovich, 2005a; Scott et al., 1999). 28

45 2.3.1 Caring in Nursing and Socio-Demographics Benner (2001) reported that years of experience may improve nurses knowledge, capability, and confidence to deliver care. A Taiwanese study by Lee and Ko (2010), found that older nurses with more years of experience showed higher levels of caring than younger nurses with less experience. This study also found higher education levels showed higher caring scores as measured by the CARE-Q. Lee also found differences between specialty areas and caring behaviours. Nurses in eemergency units showed lower levels of caring behaviours compared to paediatric units, surgical units, and psychiatric units (which showed the highest levels). A study by Schofield, Tolson, Arthur, Davies, and Nolan (2005) found age, qualifications and length of time working in the specialty area of geriatric nursing showed higher levels of caring attributes. These previous studies have assisted in informing some of the sociodemographic variables to be used in the current study Self-Efficacy and Socio-Demographics No studies were found that examined socio-demographics and self-efficacy in nurses. This study will add to the literature on self-efficacy in registered nurses in Australia. 2.4 Work Locus of Control Locus of control (internal versus external control of reinforcement) may be best described as where people through a lifetime of social learning, acquire a generalized expectation about the source of reinforcement for their actions (Ashkanasy, 1985, p. 1328). It is a person s beliefs about the sources of control over the reinforcement that has been received. Storms and Spector 1987, identified that work locus of control had become an important construct which was useful for comprehending human behavior in workplaces as well as identifying a person's responsibility for learning. 29

46 Rotter (1966) described locus of control as an internal-external control continuum. Those with an internal locus of control believe that reinforcement comes from their own personalities, and those with an external locus of control see reinforcement as a result of fate, chance or powerful others. Thus, it is believed that internals are better adjusted and cope better than externals. Further, it is reported those with an external locus of control may believe they cannot do anything personally to make things happen for them, whereas those with an internal locus of control believe they have the ability to make it happen. Those with an internal locus of control have a great need to value what they are putting their energy into, both for their own achievement or to assist others in achieving (Maylor, 2001). Forte (2005) noted that employees with an internal locus of control could make the decision of what appropriate behaviour was, while those with an external locus of control decided what appropriate behaviour was after watching others. Those with an internal locus of control also have been shown to be more creative and productive. They believe they are utilising their potential therefore achieving more. It has been suggested that people who experience constant failure distance themselves from the circumstances of the failure. This may involve obtaining a different job, moving or suggesting their failure was caused by outside factors (Forte, 2005). Externality has also been reported to be associated with intentions to quit in a job (Spector & Michaels, 1986). People who score high in externality are also reported to have less job satisifaction, feel more estranged from the work setting and are less involved in their job compared to internals (Duffy, Shiflett & Downey, 1977). Moreover, externals are more likely to show sabotage, aggression and withdrawal behaviour as a result of frustration within an organisation, whereas frustrated internals are more likely to work out ways to quickly resolve the problem (Spector, 1978; Storms & Spector, 1987). Work locus of control has been found to correlate with general locus of control and job satisfaction and it is specifically related to the work domain (Spector, 1988). Judge and Bono (2001) found general self efficacy was related to locus of control and both of these were found to be correlated with job satisfaction. People s belief 30

47 that they have limited control over circumstances and limited opportunities to partake in decision making in their work can have an effect upon job satification (Keller, 1984; Organ & Greene, 1974). People with these beliefs (i.e. have an external locus of control) have been shown to have fewer coping strategies (Anderson, 1977). One nursing study by Schmitz, Neumann and Oppermann (2000) found those nurses with an external locus of control were associated with feelings of stress and burnout. The results support the view that a perceived degree of control may be seen to be influential in supporting nurses to cope better with stress and burnout (Schmitz et al., 2000). This may result in improved nursing job satisfaction and retention (Zangaro & Soeken, 2007). Work locus of control was tested in relation to caring efficacy and job satisfaction in the current study. A meta-analysis of locus of control at work by Ng, Sorensen and Eby (2006) found that those with an internal locus of control were positively associated with better work outcomes such as: positive job and social experiences and more job motivation. The Work Locus of Control Scale (Spector, 1988) was used in the present study as it is reported to be a more precise measurement for work behaviour than the General Locus of Control Scale (Rotter, 1966) Locus of Control and Socio-Demographics Rotter s theory (1966) argues that internal locus of control increases with age, length of work tenure, higher management level, gender (males), increased education levels and certain industry types. It has been reported in other studies that a significant relationship exists between locus of control and increasing age (Letvak, 2002; Spector, 1985). In these studies, older people who demonstrated an internal locus of control reported higher levels of satisfaction with their work compared to those with an external locus of control. As relationships between personal characteristics including age have also been found to relate to job satisfaction in some studies and as our workforce is continuing to age (Letvak, 2002; Spector, 1985), relationships between these characteristics were investigated in the current study. 31

48 2.5 Coping Styles Coping may be defined as cognitive and behavioural efforts to manage specific external and/or internal demands appraised as taxing or exceeding the resources of the individual (Folkman & Lazarus, 1988, p. 6). Coping refers only to those adaptational activities that involve effort: It does not refer to all the things we do in relating to the environment (Folkman & Lazarus, 1988, p. 8). For example, there is a difference between the commencement stages of skill acquisition, which are stressful and involve effort, and later when the skills become automatic (Folkman & Lazarus, 1988). Coping behaviours are an example of factors that limit the effects of psychological distress and are described as either problem focused (changing the source of stress) or emotion focused (regulating stressful emotions) (Fleishman & Fogel, 1994). Coping may involve managing or altering the problem (problem-focused coping), or regulating the emotional response to the problem (emotion-focused coping) (Lazarus & Folkman, 1984, p. 179). Although people use both problem and emotionfocused coping when dealing with a problem, it has been found that people may use problem-focused coping when managing work-related problems (Lazarus & Folkman, 1984). It has also been shown in another study (Wong, Leung & So, 2001, p. 188) that nursing is characterised by stress, frequent job turnover, and job dissatisfaction. This Hong Kong study reported that nurses use problem-solving strategies such as discussing problems with other nursing colleagues and identification of different ways of problem management, as ways of coping with stressful situations (Wong et al., 2001). A study conducted on nursing students in London, investigating general coping, direct coping and suppression, by Parkes (1984) found that people with an internal locus of control used high levels of direct coping and low levels of suppression when they perceived circumstances as being changeable, whereas externals (external locus of control) used high levels of suppression and low levels of direct coping in the same situation. Further, an earlier study by Lazarus and Folkman (1987) identified that perceived self-efficacy assists people with coping with stress. 32

49 A more recent study by Golbasi (2008) of Turkish nurses found greater levels of job satisfaction were associated with positive coping strategies. This study contributes to a growing body of evidence demonstrating the importance of coping strategies to nurses' job satisfaction. The relationships between coping, caring efficacy and job satisfaction were tested in this current study Coping and Socio-Demographics Folkman and Lazarus, (1980) found very little difference in relation to coping and gender differences when type of stressful encounter was controlled for. They reported that women s sources of stress were more health focussed resulting in emotion-focussed coping and men focussed more on work issues as stressful encounters, which resulted in more problem-focussed coping. Another more recent study of Iranian nurses (Laal & Aliramaie, 2010) found significant associations between age, job experience, tenure, state and a positive application of coping methods such as listening to music, reading, talking with a partner or colleague and exercise. There were also significant differences found between gender and job experience and negative responses to stress in this study. Males who had been working in the job less than 5 years were found to be more difficult and impatient. 2.6 The Professional Nursing Practice Environment Kanter s Theory of Structural Power in Organisations (1979) affirmed that informal and formal power in workplaces makes organisational structures accessible and hence empowers workers. Formal power occurs when jobs provide opportunities for flexibility, visibility, and creativity. Furthermore this early work by Kanter showed formal power also comes as a result of jobs that are regarded as relevant and central to the organization. Meanwhile informal power is determined by relationships and associations with peers, subordinates, and superiors either internal or external to the organization. Power is defined by Kanter as the ability to mobilize resources and achieve goals in (Armstrong & Laschinger, 2006, p. 125). Further to this, Armstrong and Laschinger, (2006), in their recent exploratory study, found organisations that provide their staff with empowering work conditions will increase 33

50 recruitment and retention of nurses. This will also create a more positive climate of safety for patients, which is supportive of an improvement in the quality of care provided. The structure of the work environment provides an important relationship to staff attitudes and behaviours in the workplace. In addition, perceived access to power and opportunity demonstrate a relationship with staff behaviours and attitudes in workplaces. Kanter (1977) suggested that individuals display different behaviours depending on whether certain structural supports (power and opportunity) were in place. This framework provides an explanation to negative workplace behaviours for example, staff turnover. Power according to Kanter, (1977) refers to the ability to access and marshal the necessary resources, information, and support to get the job done (Kanter, 1977). Moreover (Kanter, 1977) identified workplaces that provide accessibility to the four social structure factors of the environment i.e., information (data, technical knowledge, and necessary expertise required for the job) support (guidance and feedback received from other staff to improve efficiency), access to resources (the ability to acquire essential materials, supplies, money, and personnel needed to meet organizational outcomes) and opportunities (growth, mobility, and the option to increase knowledge and skills) promote effectiveness and satisfaction in employees. (Kanter, 1977; Laschinger, 1996). Additionally jobs that offer discretion and are central to the organisational purpose, provide access to the structures of empowerment. Strong connections with peers, superiors and other people within the organisation provide the opportunity for increased access to these resources (Armstrong& Laschinger, 2006). Kanters s theory has been widely tested in nursing studies and has linked empowerment to job satisfaction, nurses perceptions of autonomy and control over their practice environments and also to leader behaviour. The nursing practice environment is further defined as organizational characteristics of a work setting that facilitate or constrain professional nursing practice (Lake & Friese, p. 2, 2006). There is limited literature identifying differences in nursing practice environments (i.e. organisational factors that have an influence upon nursing practice) and yet it is recognised that these environments are essential for addressing 34

51 the nursing shortage (International Council of Nurses, 2006). It is reported that human resources (staffing) and the social organisation of the work (practice environment) have an effect upon patient outcomes and may be direct or indirect. It was noted by Aiken et al. (2000) that practice environment research reported in the 1980s and 1990s compared characteristic differences and trends between magnet hospitals and nonmagnet hospitals. It was noted in magnet hospitals that attracting staff was high and staff turnover low (Aiken et al., 2000). From this earlier research and later research from the United States of America (USA) (Sengin, 2003), certain organisational traits that support nurses in their practice have been reported, which enable them to conduct complex functions and provide optimal patient care. These traits also positively affect job satisfaction and incorporate autonomy, interpersonal communication-collaboration, professional practice, recognition, administrationmanagement practices, job-task requirements, advancement, work conditions and the physical environment, pay and fairness or distributive justice (International Council of Nurses, 2006; Sengin, 2003; Blegen, 1993). In addition, recent Australian research conducted in one state in Australia by Duffield et al. (2009) has also found work environment factors have a positive effect on job satisfaction. These factors include autonomy in nursing, control over practice and leadership at the ward or unit level. This study also found that older nurses with dependents were more satisfied. Moreover, it has also been identified that nursing leaders are well positioned to have an influence upon and further improve the environment for nursing staff (Duffield et al., 2009; Webster, Flint & Courtney, 2009; Paliadelis, 2008; Paliadelis & Cruickshank, 2008). Improving the nursing environment by promoting respectful relationships among health care providers, empowering nurses to make decisions, investing in nursing leadership to extend career trajectories and developing standards of practice that foster quality of care potentially produce benefits to the health system (Kazanjian, Green, Wong & Reid, 2005, p. 115). These benefits include, better job satisfaction and prevention of burnout for nursing staff, and improved workforce retention (Kazanjian et al., 2005). Further, it has been found that in areas of higher registered nurse staffing levels and positive work environments, associations have been made with lower morbidity, patient satisfaction and health-related quality of life (Aiken et al., 2008; Hayes, O Brien-Pallas, Duffield, Shamian & Buchan et al., 2006; 35

52 Laschinger & Leiter, 2006; Kazanjian et al., 2005) as well as improved health status of nurses (Gershon et al., 2007) The Professional Nursing Practice Environment and Self-Efficacy Bandura (1989) found that people who perceive they have the ability to mobilise the necessary resources are able to get the job done. Furthermore self-efficacy may determine the levels of motivation nurses have to work through obstacles to get the job done. In the workplace, people with high levels of personal self-efficacy may find their perceived self-efficacy is diminished due to the lack of empowering structures in the hospital or work environment. Alternatively, self-efficacy may be a requirement for nurses in the workplace in order to utilise what structures are available (Manojlovich, 2005a). Nurses s perceptions of the work environment rather than objectivety influences the variablility in response to the workplace (Sprietzer, 1996) The Nursing Practice Environment and Socio-Demographics Nedd (2006) investigated empowerment and intent to stay in nurses in Florida in the United States of America and found individual nurse characteristics were not significantly related to intent to stay in this sample. This finding was consistent with Kanter's theoretical expectation that work behaviour and attitudes, such as intent to stay, are not so much related to personal characteristics as they are related to perceived access to workplace empowerment structures within the organization. This previous research by Nedd (2006) focussed on factors that could be changed, such as organisational factors that are within the realm of management and not on individual characteristics of the nurses (age, gender, education, years of nursing experience and number of years in current job). 2.7 Job Satisfaction An early study conducted in the USA found relationships exist between work satisfaction, employee performance and client outcomes in human services 36

53 (Spector,1985). More recent research found dissatisfaction among nurses is a major factor contributing to the nursing shortage (Aiken et al., 2011). This has resulted in inadequate staffing levels in hospitals required for the provision of safe and effective care (Aiken et al., 2011; Aiken et al., 2002). Further, nurses relationships with patients can be compromised if nurses are dissatisfied with their work and may result in them distancing themselves from patients thus affecting patient care (Begat et al., 2005; Aiken, Smith & Lake, 1994). The results of a meta-analysis that was conducted by Blegen (1993) found in order of the strongest relationships to nurses job satisfaction were the variables: stress, commitment, communication with supervisor, autonomy, recognition, communication with peers, fairness, locus of control, age, years of experience, education, and professionalism. The results of a more recent meta-analysis by Zangaro and Soeken (2007) found that having autonomy, good collaboration between nurses and physicians in an organization, and the reduction of job stress, were the main variables that were constant with improved job satisfaction. It was also found in this review that over the last 12 years the increased correlation between job stress and job satisfaction was mainly caused by new technology, staff shortages, unpredictable workloads and poor workflow. Stressful hospital environments have been long stated as a major factor contributing to job satisfaction and intention to leave a job. Zangaro and Soeken (2007) concluded that nurses are leaving hospitals in search of less stressful work environments, consistent work hours, and jobs that provide more satisfaction. From the literature it can be seen that there is a need for nursing leaders to develop strategies to improve the work environment to make it more appealing to nurses in order to improve patient outcomes. The Job Satisfaction Survey used in this current study was founded on the theoretical position that job satisfaction characterizes an affective or attitudinal reaction to a job. Job satisfacton is often referred to as an emotional affective response to a job or aspects of a job (Locke 1976; Smith, Kendall, & Hulin, 1969). Smith et al. (1969) suggests that being satisfied with the different aspects of a job comes from the cognitive process of comparing the current job aspects with the individuals belief system. Three approaches are discussed by Locke (1976) in relation to the reasons for job attitudes. First, there may be discrepancies about what the job offers and what 37

54 the person expects; second the degree to which jobs provide for individual needs and third the degree in which individual ideals are satisfied (Locke, 1976). Early research by Spector (1985) implied that a person would stay in a satisfying job and avoid or quit a dissatisfying one. It has been shown that withdrawal behaviour, turnover, absenteeism and withdrawal intentions all correlate with job satisfaction (Spector 1985). Pay, promotional opportunities, fringe benefits, contingent awards (appreciation and recognition) supervision, co-workers, nature of work itself, communication and operating procedures (rules, procedures and red tape) all measured either positive or negative attitudes to a job. The personal variable age was also found to be related to job satisfaction in this study. The job satisfaction survey was developed from this early research for use to measure job satisfaction in human services including nursing (Spector, 1985). A systematic review conducted by Van Saane, Sluiter, Verbeek and Frings-Dresen (2003) investigating the quality of job satisfaction instruments for use as evaluative tools in hospital environments found the Job Satisfaction Survey met the criteria for use in this context for this current study. Further early research by Spector (1982) found relationships exist in the workplace between locus of control, job satisfaction and retention. Individuals with an internal locus of control were found to be more likely to have better job satisfaction, as they were less likely to stay where they were dissatisfied (Spector, 1982). Those with an internal locus of control were also shown to have longer job tenure (Spector & Michaels, 1986), and they were more likely to be successful in organisations. It was reported by Spector and Michaels that when those with an internal locus of control are challenged by being shown differences between acceptable performance and their actual performance, they will improve their actual performance to meet the standards. For those with an external locus of control a strong relationship existed between job satisfaction and intentions to leave (Spector & Michaels, 1986). A meta analysis of four personality traits (which included locus of control) in relation to job satisfaction and job performance conducted by Judge and Bono (2001) found locus of control was among the best dispositional predictors of job satisfaction and job performance. 38

55 Judge et al. (1998) suggested that individuals with high levels of generalised selfefficacy would have increased job satisfaction as they would be able to manage difficult situations and also continue in the face of adversity to achieve the outcomes they value. A meta-analysis conducted by Bono and Judge (2003) looked at the relationship of core self-evaluation traits including generalised self-efficacy and locus of control in relation to job satisfaction. The results indicated these traits were significant predictors of job satisfaction (Bono & Judge, 2003) Job Satisfaction and Socio-Demographics The results of a meta-analysis that was conducted by Blegen (1993) found sociodemographics with the strongest relationships to nurse s job satisfaction were the variables: age, years of experience and education. More recently a meta-analysis of 31 studies was performed (Zangaro & Soeken, 2007). Their results found nurses working in specialty areas tended to have clearer role expectations and sought more challenging jobs It was reported nurses with more years of experience have higher education levels and became authorities in specialty areas. They also tended to have more autonomy and opportunities in the work areas. Age was defined by generation in this current review of the literature and the following three generations have been used. These include, Baby Boomers (those nearing retirement), Generation Xers, and Nexters (the youngest group). The findings showed Generation Xers generally don t stay in jobs that don t meet their expectations. A lack of communication with Generation Xers also resulted in increased stress and decreased retention. Job stress was found to precede job satisfaction, and also found to predict young nurses intention to leave an organization (Zangaro & Soeken, 2007). No studies have been found that have explored relationships between the sociodemographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), general self-efficacy, work locus of control, coping styles, the professional nursing practice environment, caring efficacy and job satisfaction in the Australian context, yet previous literature discussed in this current study indicate relationships do exist between most of these variables and may also influence patient outcomes and job satisfaction. Coates (1997) suggested further 39

56 studies relating to caring efficacy should also include instruments that assess personal control variables such as general self-efficacy, work locus of control and coping styles. 2.8 Summary This chapter has reviewed the theoretical and empirical literature in relation to caring efficacy. The review has given a synthesis of existing caring, self-efficacy and job satisfaction research demonstrating the nature of this literature. The review also discussed the important theoretical contributions that have been facilitated by the research, while highlighting the gaps evident in the research. It was important to determine whether these variables could be constructed to illustrate the relevance between the socio-demographic, personal control variables, the environmental variable and outcomes of caring efficacy and job satisfaction. Early seminal work conducted on caring efficacy (Coates, 1997), job satisfaction (Spector, 1982), coping (Folkman & Lazarus, 1988), locus of control (Spector, 1988) self-efficacy (Bandura, 1977) and the work place environment (Kanter, 1977) was central to the development of the framework for this current research study. The next chapter will describe the research methodology for this study and will describe the instruments used. The ethical issues associated with this study will also be discussed. 40

57 Chapter 3: Research Plan: Methodology 3.1 Introduction This chapter will describe the research design of two research studies. The population, sample, research setting, research procedure, data collection methods, measures, data management and analysis and ethical considerations for each study will be described. To explore the proposed research questions, the two studies will be undertaken using quantitative research methods. The conceptual framework, as previously discussed in Chapter 1 of this study, was based on the literature reviewed. The research design for this study is shown in Figure 2. Study 1 A pilot test of ethical considerations, population, sample, research setting, measures, data collection methods, data management, data analysis. Study 2 Cross-sectional survey of Australian Registered nurses. Measures: sociodemographics, general self-efficacy, locus of control, practice environment, coping, caring efficacy and job satisfaction. Figure 2: Research Design Quantitative research uses deductive processes which are based on the conventions of a positive philosophy of science. Quantitative approaches enable the researcher to numerically measure a number of dependent and independent variables. This can 41

58 then be analysed using descriptive and inferential statistical methods. Within the quantitative paradigm, a research question (theory) is developed at the beginning of the study by the quantitative researcher in order to test or verify the theory (Neuman, 1994). 3.2 Study 1: Pilot Study Research Questions A pilot study (Study 1) was conducted prior to Study 2 in order to identify any flaws in the design of the study. The measures used in this study were identified from the Introduction (Chapter 1) and Literature review (Chapter 2). The research questions of study 1 were: 1. Is the Caring Efficacy Scale a valid and reliable instrument for the Australian registered nurse population? 2. Is the Job Satisfaction Survey a valid and reliable instrument for the Australian registered nurse population? 3. Is the Work Locus of Control Scale, General Self-efficacy Scale, Ways of Coping Questionnaire and Nursing Work Index Revised valid and reliable instruments for use in the Australian registered nurse population? 4. Is the Caring Efficacy Scale feasible for use in a cross-sectional study for the Australian registered nurse population? 5. Is the Job Satisfaction Survey feasible for use in a cross-sectional study for the Australian registered nurse population? 6. Is the Work Locus of Control Scale, General Self-efficacy Scale, Ways of Coping Questionnaire and Nursing Work Index Revised feasible for use in the Australian registered nurse population? 42

59 3.2.2 Aims The aims of study 1 were: 1. To pilot test and undertake preliminary testing of content and face validity and reliability of the Caring Efficacy Scale, Job Satisfaction Survey and the Nursing Work Index Revised. 2. To assess feasibility of all questions for this study. 3. To assess the effectiveness of the participant recruitment process. 3.3 Research Design and Methodology To achieve these aims, a cross-sectional survey research design was used in study Research Population The study population and criteria for selection included registered nurses in Australia. A total of 100 participants were recruited from the database of the Australian professional and industrial organisation Recruitment Methods Nurses were randomly selected from the register by a member of the staff of the participating Australian professional and industrial organisation and stratified according to gender. Each individual participant was mailed two survey questionnaires (the same) containing socio-demographic questions and the measures previously identified; a Plain Language Statement of Information (see Appendix 5) about the study and 2 reply paid envelopes with the return address. The participants were asked to complete one of the questionnaires and send it back to the chief investigator in one of the reply paid envelopes via Australia Post. They were also asked to complete the second questionnaire 1 week later and send this back in the second reply paid envelope via Australia Post in order to test reliability of the questionnaires in this population. The voluntary anonymous return of the 43

60 questionnaires via Australia post indicated consent to participate. All of the nurses who entered this research project did so voluntarily and were free to withdraw from the study at any time with no penalty. The phone numbers of both the researchers and the Queensland University of Technology, Human Research Ethics Committee were made available to the participants if they required further information about the project or were concerned with the ethics of this research. The participants received evidence in the Plain Language Statement of Information that ethical approval had been sought Sample Size As a pilot study, the emphasis was not on establishing statistical significance on effect, but to provide information for analytical planning purposes of the main study. A sample size of 100 registered nurses was chosen for the pilot study in order to gauge the sources and magnitudes of variation within the Australian registered nursing population and also to pretest the instruments. Ordinarily, at least pretests are required (Polit & Beck, 2004) Sampling Strategy In order to achieve the aims of study 1, the sample was chosen from the same population as that of study Measures/Assessment Tools In total, 7 measures were used in study 1 and all participants received the questionnaire containing all of these measures. The first measure consisted of a socio-demographic questionnaire (i.e. pertaining to, or characterised by a combination of sociological and demographic characteristics) (MacMillan, 2010), to identify personal and organisational characteristics of the participants. These included age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state. The demographic questionnaire was developed by the chief investigator (see Appendix 7). It also consisted of the General Self-Efficacy Scale 44

61 (Schwarzer & Jerusalem, 1995), Work Locus of Control (Spector, 1988), Ways of Coping Questionnaire (Folkman & Lazarus, 1988), Nursing Work Index Revised (Aiken & Patrician, 2000), Caring Efficacy Scale (Coates, 1997), and the Job Satisfaction Survey (Spector, 1985).The details of the research instruments used in study 1 will be discussed when addressing study Data Management and Analysis Data entry was performed by the chief investigator. All statistical data was analysed using the Statistical Package for the Social Sciences (SPSS) (version 17.0, SPSS, Inc., Chicago, IL). Data collected was coded, verified and was visually checked. The use of a verifying program eliminated data inaccuracies. Data cleaning was implemented to check for outliers that may not be legitimate entries and for codes that were incorrect. Consistency checks were undertaken to ensure answers were consistent with each other where possible (Polit & Hungler, 1999) Test-Retest Reliability The reliability test is defined as the ratio of the true variance to the total variance of the measurement (Vehkalahti, Puntanen & Tarkkonen, 2006). Preliminary testing of the reliability of the questionnaires for this study was undertaken using Bland Altman analysis (Bland & Altman, 1986). Test-retest reliability was used to examine the reliability of all the instruments used in study 1. Test-retest reliability is the assessment of stability over time for a set of scores on an instrument (De Von et al., 2007). Repeated administration was conducted over a 1-week period for the instruments to be measured in study 1 including Caring Efficacy Scale, Job Satisfaction Survey, General Self-Efficacy Scale, Work Locus of Control Scale, Nursing Work Index Revised and Ways of Coping Questionnaire. A time interval of at least 1 week between test and retest is required (Walsh & Betz, 2001). The need for adequate time between the tests is to prevent a memory effect (De Von et al., 2007). 45

62 3.3.8 Validity Testing Face validity provided the opportunity to see that participants understood and responded to the items on the scales used in this current study (De Von et al., 2007). Data were analysed as follows: 1. Descriptive statistics were calculated as percentages in order to demonstrate how variables were distributed in this study sample. 2. Bland Altman plots were generated to show test-retest reliability over a 1- week interval of the General Self-Efficacy Scale, Ways of Coping Questionnaire, Work Locus of Control Scale, Caring Efficacy Scale, Job Satisfaction Survey and the Nursing Work Index Revised. 3. Face validity was used to test the validity of the instruments in an Australian registered nurse population. 3.4 Study 2: Cross-Sectional Survey Design Research Questions The following research questions were developed for study What is the level of caring efficacy and job satisfaction of the registered nurse population in Australia? 2. Do levels of self-efficacy in registered nurses vary according to work locus of control, coping styles, general self-efficacy, the professional nursing practice environment and (outcome variables) caring efficacy and job satisfaction levels? 3. Do the variables age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state predict caring efficacy and job satisfaction levels? 46

63 4. Does a high level of caring efficacy enhance relationships between work locus of control, general self-efficacy, coping styles, the professional nursing practice environment and job satisfaction? 5. Can caring efficacy mediate the relationship among the variables of work locus of control, general self-efficacy, coping styles, professional nursing practice environment and job satisfaction? Aims The aims of study 2 are: 1. to describe the factors that affect caring efficacy and job satisfaction of registered nurses in the Australian context and 2. to provide a basis to develop a model that improves nurses perceived caring efficacy and job satisfaction in the Australian context Research Design and Methodology A cross-sectional research design was undertaken to examine the relationships between socio-demographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), personal control variables (general self-efficacy, work locus of control and coping styles) and the organisational variable, (the nursing practice environment) and the outcome variables caring efficacy and job satisfaction Research Population and Sampling Frame The study population and criteria for selection included registered nurses in Australia who were at the time members of a specific Australian industrial and professional organisation. The sampling frame must capture, statistically, the target population (Upton & Cook, 2008). Sampling frames for population-based cohort studies include any well-defined 47

64 population (Szklo, 1998). At the time of this current study there was no national regulatory authority for nurses in Australia. In the first instance the nursing and midwifery regulatory organisations for each state were approached prior to conducting the study. The nursing and midwifery regulatory bodies from two major states of Australia were not able to participate in this study. Regulatory bodies of the two Australian territories were also not able to provide the services required to recruit participants. Following this, Australia s largest professional and industrial nursing, midwifery and assistants in nursing organisation (200,000 members) with branches in each state and territory, agreed to participate in the recruitment of the participants for this study at a national level. It was reported in 2008, that the number of registered and enrolled nurses employed as nurses in Australia was 272,741 (AIHW, 2010). It was therefore expected that a representative population of registered nurses could be obtained from this organisation. Four major Australian states from a total of six agreed to participate. Numbers were too small for the processes of randomisation and stratification according to gender from the members of the two territories and the smaller state (Tasmania) and were therefore not included in this study. Western Australia (one larger state) was unable to participate at the time of this study. Further interpretation of representation of the registered nurse population for this study was made difficult as there was limited information available. The Australian Nursing Federation (ANF) (2011) recognised there was deterioration in the quality of the way the data from the AIHW (2008) was reported therefore making the information difficult to interpret. This is a limitation for this current study, as it was difficult to obtain the necessary information from the population based sample, in order to generalise this to the study population Recruitment Methods Two thousand registered nurses were randomly selected and stratified according to gender to ensure the avoidance of sampling bias and sampling error. Sampling bias is the overrepresentation or the underrepresentation of part of the study population in terms of an attribute pertinent to the research question. Sampling error occurs when there are differences between population values and the values of the sample (Polit & Hungler, 1999). Sample size will be discussed in the next section. The eligible participants in this study were selected from the register by staff of the Australian 48

65 industrial and professional nursing organisation. The states participating in this study included Queensland, New South Wales, Victoria and South Australia. Other states and territories were not included as the membership numbers were too small or the states/territories declined the offer to participate. Surveys were sent to registered nurses as follows: Victoria, 450 females and 50 males; New South Wales, 750 females and 85 males; Queensland, 500 females and 50 males; and South Australia, 100 females and 15 males. Each individual participant was mailed a Plain Language Statement of Information about the study, the survey questionnaire as used in study one and a reply paid return addressed envelope. The measure The Nursing Work Index Revised was excluded from the questionnaire and replaced by The Brisbane Practice Environment Measure. The voluntary anonymous return of the questionnaires via Australia post indicated consent to participate. Follow-up reminders were mailed to all randomly selected participants after four weeks to maximize the response rate required i.e., 32% of the 2000 potential participants (see Sample Size). All nurses who entered this research project did so voluntarily and were free to withdraw from the study at any time with no penalty. Phone numbers of both the researchers and the QUT Research Ethics Committee were made available to the participants in the event they required further information about the project or were concerned with the ethics of this research. The participants were informed of ethical approval from the Queensland University of Technology, Human Research Ethics Committee in the Plain Language Statement of Information Sample Size Sample size calculation for the Structural Equation Modelling (SEM) (including confirmatory factor analysis (CFA) method is a very complex matter. The sample size was calculated for the current study based on the assumption that there would be a 32 % response rate (Kaplowitz, Hadlock & Levine, 2004) and that approximately ten respondents per item of our largest survey instruments would be received (Ways of Coping Questionnaire with 66 items) as suggested by Nunnally (1978). 49

66 There were 31.9% respondents who returned the survey (n = 639) giving 9.68 respondents per item for this largest instrument, but 26 participants for the next largest instrument (Job Satisfaction Survey with 36 items). It was assumed that this sample size adequately powered the general linear modelling (GLM) technique (analysis of variance [ANOVA] and regression) Sampling Strategy A stratified random sampling strategy was used in this study. Probability or random sampling involves some form of random selection in choosing the elements. It has gained recognition as a good approach for assessment because greater confidence can be placed in the representativeness of probability samples. Random sampling involves a process in which each element in the population has an equal and independent chance of being selected (Polit & Hungler, 1999). Stratified random sampling is a variant of simple random sampling in which the population is first divided into two or more strata or subgroups. The aim of stratified sampling is to enhance representativeness. Stratified sampling designs subdivide the population into homogenous subsets from which an appropriate number of elements can be selected at random. In stratified sampling, a decision about a person s status in a stratum is made before a sample is randomly selected. The most common method of drawing a stratified sample is to group together those elements that belong to a stratum and to select randomly the desired number of elements (Polit & Hungler, 1999). In this study, gender was used as the stratified variable from the accessible population of the database of the specified Australian Industrial and Professional Nursing Organisation. This is because in Australia, registered nurses are predominately female. Demonstrating this, 91% of nurses in 2006 were female (Australian Bureau of Statistics, 2008). From these participants the response rate for this current study was 639 (31.9 %). 50

67 3.4.8 Introduction to the Assessment Tools The confounding variables, independent variables, dependent variables and measures used to collect the data in this study are shown in Table 1. Confounding variables include the socio-demographic variables such as age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state. The independent variables include general self-efficacy, work locus of control, coping and the professional nursing practice environment and the dependent variables include caring efficacy and job satisfaction. Table 1: Explanatory Variables and Measurement Strategies Variable Domain Instrument Job Satisfaction Dependent Variable Job Satisfaction Survey Caring Efficacy Dependent Variable Caring Efficacy Scale Locus of Control Independent Variable Work Locus of Control Coping Independent Variable Ways of Coping Questionnaire General Self-Efficacy Independent Variable The General Self-Efficacy Scale The nursing practice environment Independent Variable Revised Nursing Work Index (Study 1) The Brisbane Practice Environment Measure (Study 2) Socio-demographic Information Confounding Variables Age Gender Marital status Education level Years of experience Years in current job Job status Geographical location Speciality area Health sector State In order to examine the variables of interest, three areas were examined in detail: socio-demographic variables, independent variables and dependent variables. To examine the demographic variables a descriptive survey was used to collect related information of personal and work-related factors (designed by the researcher). The independent variables were examined using the instruments General Self-Efficacy Scale (Schwarzer & Jerusalem, 1995), Work Locus of Control Scale (Spector, 1988), the Ways of Coping Questionnaire (Folkman & Lazarus, 1988); and the Brisbane Practice Environment Measure (Webster, Flint & Courtney, 2009). Outcomes of general self-efficacy, work locus of control, coping styles and the professional 51

68 nursing practice environment were examined using the instruments Job Satisfaction Survey (Spector, 1985) and Caring Efficacy Scale (Coates, 1997). A more detailed description of the measures is outlined in the following section Demographic Questionnaire A personal data sheet was utilised to collect socio-demographic information related to personal characteristics (age, gender, marital status, level of education, years of experience, years in current job, employment status and specialty area) and organisational information (geographical location, health sector and Australian state) as discussed in the literature. Personal categorical variables were explained as: marital status which comprised of single, married, widowed, separated/divorced; education levels was represented by certificate, bachelor, graduate certificate/diploma, masters and PhD; employment status was described as full time, part time and casual employment and specialty area was explained as midwifery, medical/surgery, intensive care, aged care, paediatrics, psychiatry, operating room/recovery, community, support services (non-clinical), and emergency nursing. Organisational categorical variables were described as: geographical location which was represented by metropolitan, provincial/regional, rural and remote; health sector which was explained by private hospital, public hospital, residential/aged care and community domiciliary and Australian state which was represented by Queensland, New South Wales, Victoria and South Australia (AIHW, 2008). The demographic questionnaire is shown in Appendix General Self-Efficacy Scale Self-efficacy was measured by the General Self-Efficacy Scale developed by Schwarzer and Jerusalem (1995). This is a 10-item self-report, Likert-type scale reflecting an optimistic self-belief that one can manage difficult tasks or cope with adversity. This scale is based on Bandura s approach, which measures situationspecific beliefs or the ability to carry out a specific action. The ten items are made to measure perceived self-efficacy such as facilitating goal-setting, effort investment, 52

69 persistence in the face of barriers and recovery from setbacks. Each item is related to successful coping and infers an internal-stable attribution of success. In the operative sense perceived self-efficacy is related to subsequent behaviour and therefore is useable in clinical practice and behaviour change. The scale is uni-dimensional, and Cronbach s alphas ranged from 0.76 to 0.90 with most in the high 0.80s. It is a 4- item scale from 1 = Not at all true to 4 = exactly true with a final composite score ranging from 10 to 40. The General Self-Efficacy Scale and permission of use is shown in Appendix Work Locus of Control Scale The Work Locus of Control Scale describes the causes to which individuals attribute their successes and failures (Forte, 2005, p. 65). The original I-E Locus of Control Scale was developed by Rotter in 1966 and measured the generalised expectancies for internal versus external control of reinforcement. The Work Locus of Control Scale (Spector, 1988) is a 16-item instrument that assesses control beliefs in the workplace. It is a domain-specific scale. The format is a summated rating and has 6 responses to choose from: disagree very much, disagree moderately, disagree slightly, agree slightly, agree moderately, agree very much and is scored from 1 to 6, respectively. The total score range is from 16 to 96. The coefficient alpha ranges from 0.80 to The Work Locus of Control Scale and permission of use is shown in Appendix Ways of Coping Questionnaire The Ways of Coping Questionnaire was developed by Folkman and Lazarus (1988) and is an assessment of how people think and act following a certain stressful encounter. It is a 66-item; self-report questionnaire measuring cognitive and behavioural practices used to manage a certain stressful occurrence. There are eight subscales that include confrontive coping, distancing, self-controlling, seeking social support, accepting responsibility, escape avoidance, planful problem solving and positive reappraisal. It has a 4-point Likert scale and requires responses from 0 (Does not apply) to 3 (Used a great deal). Internal consistency coefficients ranged from 0.61 to 0.79 (Folkman & Lazarus, 1988). There are two ways of scoring this 53

70 questionnaire: raw and relative scoring. Raw scoring provides the information regarding the extent each method of coping was used in a specific encounter and relative scoring describes how much each score relates to all the combined scores. Permission to use the Ways of Coping Questionnaire was obtained for study 1 (2,100 copies) (Appendix 10). Raw scoring was used in this current study Revised Nursing Work Index (NWI-R) The Revised Nursing Work Index (NWI-R) (Aiken & Patrician, 2000) was developed from the Nursing Work Index (Kramer & Hafner, 1989) which in turn was developed from the findings from research on magnet hospitals. This instrument was designed to be an inclusive list of factors that have bearing on environments conductive to quality care provisions and job satisfaction. (Aiken & Patrician p. 148, 2000). The NWI-R contains 55 of the original 65 NWI items but has a total of 57 questions. Those items which did not specifically relate to a professional practice environment were discarded. The current items represent important traits of the practice environment which are very predictive of important nurse and inpatient outcomes (Aiken, Clarke & Sloan, 2002). The NWI-R is scored using a summative rating scale ranging from 1 (strongly disagree) to 4 (strongly agree). Higher scores indicate higher numbers of attributes depicted by each item present in the work environment. (Flynn, Carryer & Budge, 2005). The Chronbach s Alpha for the NWI- R was 0.96 (Aiken et al, 2000). The Revised Nursing Work Index and permission of use for study 1 is shown in Appendix Brisbane Practice Environment Measure The Brisbane Practice Environment Measure (Webster et al., 2009) was developed in a large sample of registered nurses in Brisbane, Australia. It is based on the reality and experiences of nurses working lives. The original 33-item model consisted of 5 thematic constructs developed according to a constant comparative method of analysis and which was validated for use in a metropolitan acute public hospital population of nurses (Webster et al., 2009). The Brisbane Practice Environment Measure is a work environment specific, self-report questionnaire that measures the nursing practice environment. Content validity of the Brisbane Practice Environment 54

71 Measure has been previously tested. Further validation of the psychometric properties of the model was initially conducted using exploratory factor analysis (EFA) and subsequently using Confirmatory Factor Analysis (CFA) (Flint et al., 2010). Weston (2009) concluded that instruments developed to measure concepts such as autonomy and control, which reflect work environments, must be valid and clear in what they are measuring. The Brisbane Practice Environment Measure, which was developed for a specific registered nursing population employed in a metropolitan acute care hospital, was further validated in this current study using a larger and more diverse Australian registered nurse population. Cronbach s alpha coefficient for the Brisbane Practice Environment Measure was The permission for use of the Brisbane Practice Environment Measure for study 2 is shown in Appendix Caring Efficacy Scale The Caring Efficacy Scale alumni self-report form developed by Coates (1997) is a 30-item, 6-point, Likert-type scale (strongly disagree 3 to strongly agree +3), which assesses nurses caring efficacy (i.e. confidence relating to ability to express a caring orientation and develop caring relationships with patients). This 30-item scale consists of 23 positively-worded and seven negatively-worded items and indicates a high level of internal consistency; Cronbach s Alpha for this scale was (Coates, 1997). The scale was developed based on Watson s theory of transpersonal caring (Watson, 1979, 1988, 1996) and Bandura s (1977, 1986) social learning theory. The Caring Efficacy Scale and permission of use is shown in Appendix Job Satisfaction Survey The Job Satisfaction Survey (Spector, 1985) has 36 items and is a nine-facet scale with four items each measuring employee attitudes in relation to their job and aspects about their job. It has a summated rating scale design with six choices per item (strongly disagree to strongly agree). The items are written in both directions (positive and negative) so some items are reverse scored. The 9 facets include pay, promotion, supervision, fringe benefits, contingent rewards (performance-based awards), operating procedures (required rules and procedures), co-workers, nature of 55

72 work and communication. Scores for total job satisfaction range from 0.36 to 0.216, with each item scored from 1 to 6. The coefficient alpha for the total scale was The Job Satisfaction Survey and permission for use are shown in Appendix Data Management Data management and entry was performed by the chief investigator. All statistical data was analysed using SPSS and the AMOS (Analysis of Moment Structures) version 17.0 program (Arbuckle, 2006). A p-value of less than 0.05 (p < 0.05) was used to signify statistical significance. Data collected was coded, verified and was visually checked. The use of a verifying programme eliminated data inaccuracies. Data cleaning was implemented to check for outliers that were not legitimate entries and for codes that were incorrect. Outliers are described by Hair, Black, Babin, Anderson & Tatham, (2006) as observations whose characteristics are identified as being different from most of the other observations. They may be a result of erroneous data entry or a coding mistake. Similarly, an unusual event or observation may also be responsible for an outlier. Consistency checks ensured answers were consistent with each other where possible. An appropriate analysis file was created using a database management package. All information was documented, and a code book was developed for listing each variable by the chief investigator Normality Assumption An assumption of normality is essential for SEM with maximum likelihood (ML) estimation as used in this study. Violation of the assumption of normality will result in imprecise statistical tests in the SEM analysis (Tabachnick & Fidell, 2007). Univariate and multivariable normality were checked for all of the variables, and univariate normality was measured using descriptive statistics and graphical analyses. The mean is within ten per cent of the median and the standard deviation less than half of the mean. The nature of the distribution of all of the variables was described using values of skewness and kurtosis. Univariate normality was considered when skewness and kurtosis were inside 3 to +3. Graphical analyses 56

73 using histograms demonstrated normality (Hair et al., 2006). Multivariable normality was evaluated using the AMOS programme Multicollinearity Multicollinearity is where two or more variables are very closely linearly related (Field, 2009, p. 790); it shows that separate variables are in fact measuring the same thing. If collinearity increases between the variables, the unique variance explained by each independent variable is reduced and the shared prediction percentage increases, thus decreasing the predictive power of each independent variable (Hair et al., 2006). Multicollinearity is often determined using the Variance Inflation Factor (VIF). This is the ratio of the total standardised variance to the unique variance ratio. A VIF greater than 10.0 means there are severe problems with multicollinearity (Hair et al., 2006). Multicollinearity is also calculated using a squared multiple correlation between each measured variable and all the others and by viewing the correlation matrix. Multicollinearity is indicated when there is a squared multiple correlation of greater than 0.90 (R 2 smc > 0.90) (Hair et al., 2006) Missing Data Missing data can be the result of errors in data collection, data entry or omission of the answers by respondents (Hair et al., 2006). Missing data results in the reduction of sample size for data analysis leading to flawed results such as convergence failures, biased parameter estimates and inflated fit indices (Shah & Goldstein, 2006). Examination of missing data was undertaken using the SPSS programme. Descriptive statistics were used to confirm whether the missing data occurred randomly or in a systematic pattern. Missing information was evaluated by level of importance and treated appropriately either by deletion of the variable, substitution using the mean or estimation of the missing value (Schlomer, Bauman & Card, 2010). 57

74 Data Analysis Data analysis was undertaken after data preparation and management. Data were analysed using SPSS and AMOS. A p-value of less than or equal to 0.05 ( 0.05) was set to specify statistical significance. Biases were determined in preliminary analysis. Pilot testing of the database and data analysis was undertaken at the end of study 1, prior to sending out the final questionnaire. The data was described and expressed as follows: 1. Nominal data was described by medians and ranges and expressed in percentages. 2. Chi-squared analysis was used to analyse comparisons between groups. 3. Two-dimensional contingency tables were constructed for each of the characteristics of respondents to show cross tabulation of the frequencies of two variables. 4. Ordinal and numeric data was described by means and standard deviations. 5. ANOVA and t-tests were used to analyse the relationships between two or more groups. Descriptive statistics were calculated for the socio-demographic variables, the personal control variables general self-efficacy, work locus of control, coping styles, the professional nursing practice environment and the outcome variables caring efficacy and job satisfaction. Percentages and frequencies were used for gender, marital status, geographical location, health sector, employment status, specialty area and state, while means and standard deviations were used for age, level of education, years of experience and years in current job. 58

75 Inferential statistics such as Pearson s correlational analysis were used to: 1. detect relationships between age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector, Australian state and caring efficacy and job satisfaction 2. detect relationships between general self-efficacy, caring efficacy, work locus of control, coping styles and the practice environment and job satisfaction. Cronbach s alpha and reliability coefficients were calculated to test for internal consistency of tools. Construct validity of the Brisbane Practice Environment Measure (Webster et al., 2009) was examined using Confirmatory Factor Analysis (CFA). CFA is mostly used to appraise the latent structure of an instrument. It can also be used to confirm the number of dimensions of the instrument as well as the pattern of any relationship between an item and a factor (Browne, 2006). It is possible to identify if a theoretical measurement model is valid using CFA, and EFA examines data in order to identify possible constructs (Hair et al., 2006). CFA can also be a useful method to evaluate whether a predetermined factor model is a good fit for the data. EFA identifies the factor structure of a set of variables, which is based on data rather than theory (Floyd & Widaman, 1995). In this study the Brisbane Practice Environment Measure, which was developed using a population from one metropolitan public hospital in Brisbane Australia, was used to explore whether the predetermined factor models provide a good fit for the data of a more diverse group of Australian registered nurses. It is important to note that the Brisbane Practice Environment Measure had not undergone CFA at the time of this current analysis. This instrument was developed from a sample of registered nurses in a metropolitan acute care hospital in Queensland. It was therefore necessary to identify the validity of this new instrument on this more diverse Australian population. 59

76 Multiple regression analysis was used to: 1. explore whether age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state could predict caring efficacy levels and job satisfaction 2. explain whether self-efficacy, work locus of control, coping styles and the professional nursing practice environment could predict caring efficacy and job satisfaction. Structural Equation Modelling analysis was used in this study for a number of reasons. Firstly all of the variables including the multiple independent and dependent variables could be examined at the same time (Buhi, Goodson & Neilson, 2007). SEM provides a rigorous and refined approach to analysis that enables relationships between the socio-demographic, personal control and organisational variables and the outcome variables to be examined all together. The SEM can also control for measurement errors as it gives a specific estimate for assessing and correcting error variance parameters (Byrne, 2001). The multiple regression has no direct way of correcting for identified levels of measurement errors for the dependent and the independent variables (Hair et al., 2006) Data Quality The staff of the Australian industrial and professional nursing organisation was appointed to identify, randomly allocate and stratify the participants who were suitable for the study. The chief investigator was responsible for the data collection, checking and inputting all data into SPSS for data analysis. This was further checked by a database management package. Reliability of the measurements was checked by taking the measurement twice and recording it. All data was entered into a database and all written records were then checked again by the chief investigator. All the information was stored in a database and was password protected and stored in a locked room. 60

77 Psychometric Testing of the Instruments Study Validity Testing Face validity and test-retest reliability were used to examine the reliability and validity of all the instruments used in study 1 as discussed in Introduction to the Assessment Tools. Face validity or translational validity, the easiest and weakest form of validity testing, provides a subjective assessment of validity. While it does not indicate that the instrument measures the construct of interest, it provides the opportunity to see how potential participants may understand and respond to the items (De Von et al., 2007) Test-Retest Reliability Test-retest reliability is the assessment of stability over time for a set of scores on an instrument (De Von et al., 2007). Repeated administration was conducted over a oneweek period for the instruments to be measured in study 1 including the Caring Efficacy Scale, Job Satisfaction Survey, General Self-Efficacy Scale, Work Locus of Control Scale, Nursing Work Index Revised and Ways of Coping Questionnaire. A time interval of at least one week between test and retest was required (Walsh & Betz, 2001). The time between tests is used to help prevent a memory effect (De Von et al., 2007). Bland Altman plots were generated to examine the test-retest agreement of the Caring Efficacy Scale, Job Satisfaction Survey, General Self- Efficacy Scale, Work Locus of Control Scale, Nursing Work Index Revised and Ways of Coping Questionnaire Study 2 The reliability testing of the instruments was undertaken in study 2 on the Caring Efficacy Scale, Job Satisfaction Survey, General Self-Efficacy Scale, Work Locus of Control Scale, Brisbane Practice Environment Measure and Ways of Coping Questionnaire. Construct validity testing and reliability testing was also undertaken for the Brisbane Practice Environment Measure in this study. 61

78 Construct Validity The construct validity of the Brisbane Practice Environment Measure was examined using CFA. CFA is performed to confirm the number of underlying dimensions of an instrument and the patterns of the relationships between an item and a factor (Browne, 2006). CFA provides the opportunity for researchers to determine whether a theoretical measurement model is valid. EFA, however, explores data to detect potential constructs (Hair et al, 2006). CFA is a useful method to evaluate whether a pre-specified factor model provides a good fit to the data, and EFA identifies the factor structure of a set of variables based on data rather than theory (Floyd & Widaman, 1995). In this current study, CFA was used to determine the structure of the subscales of the original 33-item Brisbane Practice Environment Measure Reliability Internal consistency reliability is the similarity of items within a scale or how well the items on a tool fit together conceptually (De Von et al., 2007). The measure of internal consistency is used to examine the consistency of responses across items. It is based on the inter-correlation between items. Cronbach s alpha coefficient is the most commonly used statistic for the measurement of internal consistency reliability (De Von et al., 2007). Cronbach s alpha coefficients were used to assess each subscale and the overall score of the measures in this current study. The value of a reliability coefficient below 0.60 is not acceptable and when above 0.90 the scale should be shortened, since a higher Cronbach s alpha may suggest redundancy of items (DeVellis, 2003) Timeline for this Study The study, a cross sectional survey, consisted of two research studies. Study 1 tested the face validity and reliability of the measurement instruments; assessed the feasibility of all questions for this study; and tested the effectiveness of the participant recruitment process. Study 2 (the main study) was conducted in 2008 and 62

79 described the factors that affect self-efficacy in nursing and job satisfaction in the Australian context as well as provided a basis to develop a model that identified nurses self-efficacy and job satisfaction Research Costs/Budgets Financial support for this study was received from the Queensland University of Technology, Queensland Health and the Royal College of Nursing Australia. The research budget detail for this study is given in Appendix Ethical Considerations Ethical approval to conduct this research project was obtained from the Queensland University of Technology Human Research Ethics Committee in The Australian Nursing Federation accepted ethical approval from this university committee for the research study to be conducted at a national level. This approval is shown in Appendix 6. In relation to the survey data collection, it was acknowledged that there could be potential risks to participants. Potential risks to participants were identified as a remote risk of psychological distress due to the completion of the questionnaire. However, this risk was reduced by instructing respondents that participation was voluntary. If they agreed to participate, they returned the completed questionnaires. If not, they could simply choose not to complete the questionnaires without further comment or penalty. Further, they could contact the researcher or ethics committee any time if they had any concerns related to this study. There were no personal benefits to the participants for completing the questionnaires; however, for nurses to effectively function in the workplace they should have an understanding and be in control of work-related activities along with the confidence that they have the ability to do so. Questionnaires contained a participant information sheet on the front page to introduce the objectives of this study. Confidentiality was assured through the use of staff from the Australian industrial and professional 63

80 nursing organisation and who also identified relevant participants from the registers and sent the anonymous questionnaires. Data was analysed and stored on a secure network drive, and back-up disks were kept in a locked filing cabinet. No information about the project was published in any form that would allow any individual to be recognised. 3.6 Summary This chapter presented the research methods used in the study. It included the research design, the procedure for data collection, the instruments used, as well as evidence of their validity and reliability, and the statistical analyses employed. This study is a cross-sectional survey. A random sample of registered nurses in Australia was recruited from the register of the Australian industrial and professional nursing organisation. All participants were asked to complete a questionnaire comprising of 6 validated instruments and a socio-demographic questionnaire. These 6 instruments are the General Self-Efficacy Scale, Work Locus of Control Scale, Ways of Coping Questionnaire, Brisbane Practice Environment Measure, Caring Efficacy Scale and Job Satisfaction Survey. These instruments have provided evidence of acceptable reliability and validity. Data was collected from July to September After the completion of data collection, all statistical data was analysed using SPSS. Descriptive analysis (such as percentages, frequencies and means) and inferential statistics (such as correlations, multiple regression, ANOVA, mediaton and SEM) were used to analyse the data and test the proposed research hypotheses. The results of study 1, the pilot study, are presented in the next chapter. 64

81 Chapter 4: The Research Findings Study 1 (Pilot Study) 4.1 Introduction The pilot study was conducted prior to the main study in order to determine the feasibility of the main study and to identify any problems that may have been encountered or anticipated as it continued. A pilot study is an essential part of the overall preparation for the main study. It must also have the provision for the procurement of basic data from the pilot population. It is a systematic and planned opportunity for the researcher to obtain the opinions of the future respondents. This allows an exploration to find out whether the pilot population differed from the main study population and in what way the main study population along with their distinctive characteristics could affect the study. This chapter presents the results of this pilot study. It presents a description of the sample and reliability and validity testing of the instruments Caring Efficacy Scale (CES), Job Satisfaction Survey (JSS), General Self-Efficacy Scale (GSES), Work Locus of Control (WLC), Nursing Work Index Revised (NWIR) and Ways of Coping Questionnaire (WCQ) Objectives The pilot study provided the opportunity to pretest the questionnaires in relation to the instructions, layout, wording and length. Modifications were then made for the main study. 4.2 Results The Questionnaire The format, wording and length of the questionnaire were considered. A small number of grammatical errors were identified, but the respondents were able to follow instructions correctly. There was an initial response rate of 24 (time 1 = T1) 65

82 and 11 one week later (time 2 = T2). Due to the low response rate, the cover page was redesigned for ease of reading. Face validity of the all instruments was undertaken. Readability and acceptability of the questionnaire was indicated through the respondent s responses on the questionnaire. The feasibility of all the questions to this study was assessed and it was decided, following comments, responses and missing data from the respondents indicating issues with relevance to some of the questions on the Nursing Work Index Revised, to replace the instrument with the newly developed Brisbane Practice Environment Measure (Webster et al., 2009) for use in study 2. Test-retest reliability analysis was conducted on the Caring Efficacy Scale, General Self-Efficacy Scale, Work Locus of Control Scale, Ways of Coping Questionnaire, Job Satisfaction Survey and Nursing Work Index Revised over a one-week interval. The instruments showed good test-retest reliability. Bland Altman plots were generated to examine the test-retest agreement of the scales Description of Sample A total of 24 registered nurses participated in study 1 following the first mail out of the questionnaire. The sample comprised 87.5 % (n = 21) female registered nurses. The majority of nurses in this sample were over 40 years of age, 75 % (n = 18) and 62.5 % (n = 15) were married. Fifty per cent of this sample (n = 12) had spent 0 5 years in their current job, however the majority of nurses in this sample (70.8 %) reported having greater than 16 years of experience as a registered nurse. Certificate level education qualification was reported at 33.3 % (n = 8), and those with a bachelor s degree and postgraduate qualifications were reported as 58.3 % of the sample. The majority of this sample of nurses, 83.3 % (n = 20), worked in permanent employment (full time or part time). Forty-five per cent (n = 11) of nurses in this sample worked in a public hospital. The majority of nurses reported working in the metropolitan area. The largest specialty area that was represented was medical and surgical, 33.3 % (n = 8). 66

83 4.2.3 Test-Retest Reliability of the Scales The test-retest reliability of each measure was examined by means of a Bland Altman plot (Bland & Altman, 1986). The results of the reliability testing of each of the scales showed good test-retest reliability (see Appendix 16) Face Validity of the Scales The face validity or translational validity of the instruments Caring Efficacy Scale, Job Satisfaction Survey, General Self-Efficacy Scale, Work Locus of Control, Nursing Work Index Revised and Ways of Coping Questionnaire indicated that all of the instruments except Nursing Work Index Revised measured the construct of interest in this population of Australian registered nurses. 4.3 Summary of Findings The outcomes of a Bland Altman analysis showed good test-retest reliability of the CES, GSES and WLCS. The outcomes of the Bland Altman analyses showed good test-retest reliability of the subscales Confrontive Coping, Distancing, Self- Controlling, Seeking Social Support, Accepting Responsibility, Escape Avoidance, Planful Problem Solving and Positive Reappraisal of the WCQ. The outcomes of a Bland Altman analysis also showed good test-retest reliability of the subscales Pay, Promotion, Supervision, Fringe Benefits, Contingent Awards, Operating Conditions, Co-workers, Nature of Work, Communication and the Total Score of the JSS. The outcomes of a Bland Altman analysis showed good test-retest reliability of the subscales Nurse Participation in Hospital Affairs, Nursing Foundations for Quality of Care, Nurse Manager Ability, Leadership and Support of Nurses, Staffing and Resource Adequacy, Collegial Nurse-Physician Relations and the Total Scores of the NWIR. 67

84 4.4 Pilot Study Discussion and Conclusion The preliminary analysis showed the wording and length of the questionnaire was acceptable and understandable by participants, but as a result of the low response rate, the cover page was redesigned to ensure ease of reading. Readability and acceptability of the questionnaire was indicated through the respondents answers. As a result of the comments, responses and missing data from the respondents, indicating issues with relevance to some of the questions of the NWIR, it was subsequently replaced with the Brisbane Practice Environment Measure (B-PEM) for study 2. The randomisation and the mailing process were efficient. Reliability of the questionnaires was shown using Bland Altman analysis, but due to the small sample size, only face validity testing of the scales was possible. The following chapter discusses the results of the national survey from registered nurses who were at the time members of the Australian industrial and professional organisation. The aim of study 2 was to gain an understanding of the relationships between the socio-demographic variables, WLC, GSE, coping styles, the professional nursing PE, and CE and JS in this population. 68

85 Chapter 5: Research Findings Study Introduction This chapter presents the results of study two, the cross sectional survey of registered nurses in Australia who were members of an Australian industrial and professional organisation. This chapter initially shows a description of the sample. The results of bivariate analysis between socio-demographic variables and caring efficacy and job satisfaction are shown (e.g means, standard deviations range and level of significance) followed by multivariable modelling to determine significant associations between the variables. Reliability testing (Chronbach s Alpha) is usually presented prior to the validation testing; however as the reliability of the other instruments which include, (General Self-Efficacy Scale (GSES), Work Locus of Control Scale (WLCS), Ways of Coping Questionnaire (WCQ), Caring Efficacy Scale (CES) and Job Satisfaction Survey (JSS) have been validated elsewhere, the assessment of the reliability of these instruments will be presented after the validation of the Brisbane Practice Environment Measure (B-PEM). The proposed theoretical (conceptual) model presented in Chapter 1, was examined using bivariate statistics and Structural Equation Modelling (SEM) analysis. Finally mediation analyses were undertaken to see if CE confounded the relationship between the various covariates and JS. These covariates included age, years of experience, years in current job, education levels, GSE and WLC. 5.2 Description of the Sample All participants at the time of this study were members of a specified Australian professional and industrial nursing organisation. A sample of 2000 registered nurses from the register were stratified and randomised according to gender from the 69

86 selected organisation s database. There were 639 respondents. (31.9% of the individuals surveyed) who participated in this survey (Table 2) Personal Characteristics of the Sample In this sample there were 91.9 % female respondents. Most of the nurses were married [65.5 %] and the majority of nurses in this sample were employed in either permanent part time or casual employment [88.6 %]. The majority of the nurses in this sample were more than 40 years of age [72.4 %]. Eleven [1.7 %] reported being over the age of 65 years whilst the age range was 20 to 76 years of age. The following specialty areas were identified: Medical/Surgical; Intensive Care/Operating Room/Department Emergency Medicine; Paediatrics; Aged and Community care; Mental Health; Support Services and Midwifery. The largest single group of respondents in this study was represented by nurses working in medical /surgical nursing [31.6 %]. Interestingly 50% of this sample had been working in their current job for five years and less (see Table 2). Table 2: Demographics Comparison of Nursing & Midwifery Labour Force, 2008 (AIHW, 2010) and Study Statistics, AIHW statistics (2008) Majority of population (Subgroup) % Study statistics (2008) Majority of population (Subgroup) Gender Female 90.6 Gender Female 91.9 Sector Public 67.3 Sector Public 61.5 Employment Part time 47.9 Employment Part time 43.2 status status Specialty area Medical and Surgical 33.1 Specialty area Medical and Surgical 31.9 % Average Age in years 44.1 years Average age in years 46years 70

87 Socio-demographic variables of the nursing and midwifery labour force are shown in Table 3, page 73. The variables reported in AIHW (2010) do not entirely represent all of the variables used in the study. This 2008 report examined the variables, age, gender, health sector, employment status and specialty area, whereas the current study also included the variables, marital status, level of education, years of experience, years in current job, geographical location and Australian state. The data presented in Table 3 identified that the nursing and midwifery population in 2008 was comparable to the study respondents from the industrial and professional organisation in the same year. The nursing population working in aged care was also shown to be comparable in the data reported from the AIHW, 2010 (11.4%) to the current study data (12.5%). 71

88 Table 3: Description of the Sample Variable Number % Variable Number % Gender Current Employment Status Male Full time Female Part time Casual Age in years Part time & Casual Sector Private Hospital Public Hospital Residential Aged Care > Community/Domiciliary Marital status Geographical location Single Metropolitan Married Provincial/regional Widowed Rural Separated/Divorced Remote 3.5 Educational qualifications Specialty area Certificate Midwifery Diploma Medical/Surgical Bachelor Intensive care Graduate Certificate/Graduate Diploma Aged care Masters Paediatrics PhD 2.3 Psychiatry Operating room/recovery

89 Table 3 (continued) Variable Number % Variable Number % Years of experience Community Support services Emergency State > Queensland New South Wales Years in current job Victoria South Australia >

90 5.2.2 Organisational Characteristics Four states of Australia were included in this study. The participants were from Queensland [27.4 %], New South Wales [39.7 %], Victoria [23.2 %] and South Australia [6.3 %]. Four sectors (private, public, residential and aged care and community/ domiciliary) were identified with the majority of registered nurses in this sample working in the public sector [60.7 %], and the least [8.1 %] in residential and aged care. The geographical locations (metropolitan, provincial/regional and rural remote) were identified. More than half of the respondents worked in the metropolitan geographical location [59.2 %] with the least number from rural and remote areas [16.2 %] (see Table 3). 5.3 Research Question 1: What is the Level of Perceived Caring Efficacy and Job Satisfaction of an Australian Registered Nurse Population? Caring Efficacy and Job Satisfaction In this study, mean CE scores were found to be with a standard deviation Mean JS scores were with a standard deviation Values for CE ranged from 3.47 to 6.00 and JS ranged from 1.56 to One hundred per cent (100 %) of nurses sampled showed high CE scores on average (> 3.0) and 80.8 % showed high JS scores on average (> 3.0) in this sample. 5.4 Research Question 2: Do Levels of Caring Efficacy and Job Satisfaction in an Australian Population of Registered Nurses Vary According to the Tested Socio-Demographic Variables? Associations between Caring Efficacy Levels and Categorical Organisational Characteristics and Job Satisfaction Levels and Categorical Organisational Characteristics The means, standard deviations, medians, ranges and p values for the categorical personal and organisational characteristics in association with CE and JS of the sample were examined (see Table 4). 74

91 When examining CE and JS levels by Australian states (Queensland, New South Wales, Victoria and South Australia) in this current study, results of the descriptive statistics showed nurses from South Australia had the highest levels of CE and JS. Nurses from New South Wales in this sample, showed the lowest levels for CE. JS levels were found to be significantly lowest in nurses from NSW (see Table 4). On examination of CE and JS levels for geographical location in this population, it appeared registered nurses who worked in the rural and remote area have the highest levels of CE while nurses working in the metropolitan area had the lowest. However registered nurses in this sample working in the metropolitan area, have the highest JS levels in contrast to that of nurses working in the rural remote area. The results indicate that nurses in this sample working in rural and remote areas are more confident in their abilities to provide care for patients, whereas those in the metropolitan areas have higher levels of JS. These results were not found to be significant (see Table 4). Registered nurses in this study, who worked in the community /domiciliary sector, had the highest levels of CE and nurses who worked in the private sector, had the highest levels of JS. These results were not found to be significant. Nurses working in the public sector, although not found to be significant, had the lowest levels of CE and JS in this sample Associations between Caring Efficacy Levels and Categorical Personal Characteristics and Job Satisfaction Levels and Categorical Personal Characteristics When we examined CE and JS by sex in this sample, the levels for both of these outcomes were highest in the female population in this current study. These results were not found to be significant. Marital status results in this sample showed CE levels were highest in nurses who were widowed or divorced. The lowest levels were reported by those who were single and these results were found to be significant. In contrast, the highest levels of JS were found in those nurses who were married and 75

92 lowest in those who were widowed or divorced. These results were not found to be significant (see Table 4). When we examined CE for education qualifications in this sample, those with the qualification of certificate had the highest levels and these results were shown to be significant. Those with a diploma qualification had the lowest levels and these were not shown to be significant. Nurses, who had the highest levels of JS, although not significant, were those with master s or PhD level qualifications. Nurses with the lowest levels of JS were those with a certificate level qualification and this result was shown to be significant in this sample (see Table 4). In relation to job status, levels of CE in registered nurses in this sample who worked part time were higher than those in full time employment. In contrast full time workers showed higher levels of JS than part time employees. Casual employees in this sample showed the lowest levels of CE and JS. These results were not found to be significant in this sample (see Table 4). With regard to specialty area, it was found nurses in this sample who worked in the midwifery area had the highest levels of CE, whilst nurses working in mental health had the lowest levels. JS levels were highest in those nurses working in medical or surgical nursing areas. Nurses working in paediatrics showed the lowest levels of JS. These results were not shown to be significant (see Table 4) Summary of Findings Levels of CE and JS in this sample of nurses varied according to the different sociodemographics. It was shown nurses in this sample from the state of South Australia had the highest levels of CE and JS; however these results were not found to be significant. In this sample those who worked in the geographical location of rural and remote showed the highest levels of CE whereas those working in the metropolitan area showed higher levels of JS. These results were not found to be significant. Registered nurses in this sample working in the community and domiciliary health sector were found to have the highest levels of CE, whilst those in the private sector had the highest levels of JS. Nurses in this sample working in the public sector 76

93 showed the lowest levels of CE and JS. These results were not found to be significant. Female nurses in this sample were found to have the highest levels of CE and JS. These results were not found to be significant. The nurses in this sample who were widowed or divorced showed the highest levels of CE and those who were married were found to have the highest levels of JS. Those who were single showed the lowest levels of CE and this result was found to be significant. In relation to education qualifications, CE levels were found to be highest in the nurses in this sample who had a certificate level of qualification. These nurses also showed the lowest levels of JS. Both of these results were found to be significant. Nurses in this sample with masters and PhD levels of education had the highest JS levels. This result was not found to be significant. Registered Nurses in this sample who worked part time showed the highest levels CE and those working fulltime showed the highest levels of JS. These results were not found to be significant. Nurses in this sample working in the specialty area of midwifery were found to have the highest levels of CE and those working in the medical and surgical areas reported higher levels of JS. These results were not found to be significant. 77

94 Table 4: Levels of Caring Efficacy and Job Satisfaction for Different Levels of the Categorical Socio-Demographic Variables Caring Efficacy Job status Geographical location Education qualifications Marital status Number Mean(SD) Median Range P Job Satisfaction Number Mean(SD) Median Range P Full time, N= (0.512) Full time, N = (0.789) Part time, N= (0.470) Part time, N = (0.689) Casual, N = (0.511) Casual, N = (0.766) Metropolitan, N= 373 Provincial/Regional, N= 156 Rural/Remote, N = (0.481) Metropolitan, N = (0.557) Provincial/Regional, N = (0.438) Rural/Remote, N = (0.738) (0.744) (0.712) Certificate, N = (0.452) ** Certificate, N = (0.784) ** Diploma, N = (0.514) Diploma, N = (0.762) Bachelor, N = (0.522) Bachelor, N = (0.701) Graduate Certificate/Diploma, N = 181 Masters/PhD, N = (0.484) Graduate Certificate/Diploma, N = (0.518) Masters/PhD, N = (0.727) (0.714) Single, N= (0.529) ** Single, N = (0.699) ** Married, N = (0.478) Married, N = (0.730) Widowed/Divorced, N = (0.495) Widowed/Divorced, N = (0.797)

95 Caring Efficacy Specialty area Table 4 (continued) Number Mean(SD) Median Range P Job Satisfaction Number Mean(SD) Median Range P Medical/Surgical, N 5.027(0.501) Medical/Surgical, 3.823(0.734) = 202 N = 202 Intensive 5.028(0.511) Intensive 3.520(0.756) Care/Operating Care/Operating Room/Department Room/Department Emergency Emergency Medicine, N = 137 Medicine, N = 138 Paediatrics, N = (0.480) Paediatrics, N = (0.560) Aged/Community 5.154(0.478) Aged/Community, 3.714(0.749) care, N = 138 N = 138 Mental Health, N = 5.026(0.579) Mental Health, N = 3.591(0.775) Support Services, N 5.114(0.432) Support Services, N 3.550(0.626) = 68 = 64 Midwifery, N = (0.417) Midwifery, N = (0.706) Sector Private, N = (0.530) Private, N = (0.792) Public, N = (0.491) Public, N = (0.698) Residential/Aged 5.180(0.437) Residential/Aged 3.625(0.842) Care, N = 52 Care, N = 52 Community Domiciliary, N = (0.472) Community Domiciliary, N = (0.695) ** Sex Female, N= (0.488) Female, N = (0.729) Male, N = (0.531) Male, N = (0.811) State Queensland, N = (0.517) Queensland, N = (0.722) New South Wales N 5.044(0.485) New South Wales, 3.554(0.755) ** = 253 N =253 Victoria, N = (0.486) Victoria, N = (0.690) South Australia, N = (0.431) South Australia, N = (0.595)

96 5.5 Research Question 3: Do the Socio-Demographic Factors Examined Explain Levels of Caring Efficacy and Job Satisfaction in the Tested Australian Registered Nurse Population? The data for research question three were analysed using SPSS. Descriptive analyses were used to examine all variables and were represented as percentages. Correlation analysis was undertaken to examine associations among continuous data (age, years of experience as a registered nurse and years in current job, CE and JS). One-way ANOVAs were conducted to examine bivariate relationships between the categorical, socio-demographic variables and the CE and JS outcomes. When CE or JS were considered in terms of a factor (categorical predictor), then a one-way ANOVA was used. When CE or JS was considered in terms of a continuous predictor, a simple linear regression was used. A multivariable analysis was then conducted following the purposeful selection of covariates approach (Bursac et al., 2008) (Table 5) to build a GLM. Table 5: Modelling Approach (A Multivariable Analysis Using a GLM) (Bursac et al., 2008) Approach 1. Bivariate tests were used to check for associations among outcome variables (caring efficacy and job satisfaction) and individual predictors. 2. A multivariable model (GLM) was run with all covariates (whose P values in the step 1 bivariate analysis were <0.25). 3a. Using the model from step 2, a model including parameter estimates (including only those variables identified as significant in step 2 analysis) was run (covariates remaining significant were noted). 3b. The model was then run with covariates with P values > 0.05 from the model in step 3a (i.e. those excluded from the preliminary multivariable model). These variables were added to the model (from step 3a) one at a time to see if they: a) had become significant in the interim. b) altered existing slopes (i.e. were confounders) from the model in step 2 (i.e. the present confounders altered β by > 20%). Note: A confounder for a categorical or continuous variable will be one where all coefficients associated with a factor alter by more than 20% when the confounder is included. 4. Using the model from step 4, a multivariable model was run again, now including those variables excluded in step 1 (i.e. bivariate analysis). Note: As in step 3, a confounder for a categorical or continuous variable will be one where all coefficients associated with a factor alter by more than 20% when the confounder is included. 5. Model diagnostics. Residual plots were used to check if the model obtained was a good fit to the data (i.e. to check if the residual is normally distributed), with a mean of zero and a common variance. 6. Check for affect modification of factors (i.e. introduce 2-way interactions into the model). 80

97 5.5.1 Correlation Analysis of Continuous Variables and Caring Efficacy and Job Satisfaction Pearson s correlations coefficients (r) were calculated to gauge the association among the covariates (continuous variables) (age, years of experience, years in current job, CE and JS). In this study, the covariates age, years of experience and years in current job were all highly positively correlated with each other (all r > 0.40, p <0.001). Age and years of experience were also correlated with CE (r > 0.1, 0.001) and should therefore be considered further in any future model development of CE (Table 6). There were no significant interactions between these covariates (age, years of experience and years in current job) and JS. It was also found that CE and JS were positively correlated with each other (r > 0.1: p < 0.001), indicating that high levels of CE might predict better JS in this sample of nurses. Table 6: Correlation of Caring Efficacy, Job Satisfaction, Age in Years, Length of Time in Current Job, Years of Experience as a Registered Nurse Age 1 Age Length of time in current job Years in job *** 1 Years of experience as a registered nurse 0.758*** 0.463*** 1 Years of experience as a registered nurse Caring efficacy Caring efficacy *** *** 1 Job satisfaction *** 1 Job satisfaction Correlations were significant at the p *** Level of significance (2-tailed) Bivariate Analysis of Categorical Variables A One Way ANOVA was undertaken to determine if CE levels were explained by the categorical variables, job status, geographical location, education qualifications, marital status, specialty area, health sector, sex and state (with the level of significance α 0.05). Marital status was found to explain CE (p <0.001) and CE was 81

98 found to be explained by health sector (p=0.009). The results showed the variable widowed or divorced explained CE levels. The community health sector was found to explain CE levels in the nurses in this sample. A One Way ANOVA was conducted to determine if the categorical variables, job status, geographical location, education qualifications, marital status, specialty area, health sector, gender and state explained JS levels (with the level of significance α = 0.05). The categorical variables state, health sector and specialty area were found to explain JS levels (p < 0.005) in this study. This means nurses who work in the state of South Australia had significantly higher JS levels. In this study, the variable private sector was found to explain JS levels (p < 0.005). Nurses working in the specialty area of midwifery showed significantly higher levels of JS. To determine if variation in CE and JS can be explained by age, years of experience and years in current job a simple linear regression was done. The results revealed that age explained CE levels on average (p = 0.023). The results also revealed that age explained JS levels on average (p = 0.014). This means that older nurses in this study had higher levels of CE and JS Multivariable Modelling of Caring Efficacy and Job Satisfaction using Socio-Demographic Variables Multivariable Modelling of Caring Efficacy A purposeful selection of covariates approach (Bursac, et al., 2008) was used to build a multivariable model. This approach used the following steps. A multivariable model (GLM) with CE and all the covariates whose P values (in bivariate analysis) were < 0.25 was run. These covariates included marital status, health sector, age and years of experience. The variable years of experience was excluded as it was too highly correlated with age to include in the same model. This high P value threshold was used to not exclude partially important factors and covariates that may become significant with the addition of other variables in the model. 82

99 We then ran a model with all the covariates that were significant in this preliminary bivariate analysis with CE. These covariates included marital status, health sector and age. Following this, only age remained significant with CE. We then ran a model with all the covariates that were not significant in the preliminary multivariable model. These included marital status and health sector. Another model was run with all the covariates that were not significant (P values were > 0.05) in the initial bivariate analyses and these included state, job status, specialty area, education qualifications, geographical location, sex and years in current job. These factors did not become significant nor were they identified as confounders as they didn t alter the existing βs by more than 20%. Therefore none of these covariates were found to be confounders. The remaining covariate age remained significant with CE Effect Modification A full factorial analysis was not run to test the significance of an interaction effect as only the variable age remained in the model following the previous steps Model Diagnostics We also generated the residuals from our model to see if they were normally distributed around zero with constant variances. Perusal of the residuals did confirm this was the case Multivariable Modelling of Job Satisfaction We then ran a multivariable model (GLM) with JS as the outcome variable again using a purposeful selection of covariates approach for model building, and all the covariates whose P values (in bivariate analysis) were < These covariates included specialty area, health sector and Australian states. A model was run with all the covariates that were significant in this preliminary bivariate analysis with JS. Following this, these three covariates were found to remain significant with JS. A model with all the covariates that were not significant in the bivariate analyses was then run, which included age, years in current job, job 83

100 status, geographical location, education qualifications, marital status and sex. None of these covariates became significant, however the covariates years in current job and education qualifications altered the βs by more than 20% i.e. both were found to be confounders with the covariate, specialty area. The covariate state became nonsignificant when job status, geographical location and education qualifications were added to the model. The two remaining covariates, specialty area and health sector remained significant with JS Effect Modification A full factorial analysis was run with those variables remaining in the model (specialty area and health sector) after the previous steps to test the significance of an interaction effect with these two variables. The result showed that there were no significant interactions with specialty area and health sector Model Diagnostics Residuals from our model were produced to see if they were normally distributed around zero and had constant variances. Perusal of the residuals did confirm this was the case Summary In summary, the results have identified that some of the socio-demographic variables of Australian registered nurses, do explain variation in caring efficacy and job satisfaction at the bivariate level of statistical analysis (see Figure 3). The covariates age, years of experience, years in current job were all highly positively correlated with each other. Age and years of experience were also highly correlated with caring efficacy and should therefore be considered further in any future model development of caring efficacy. No significant interactions were identified between the covariates age, years of experience and years in current job and job satisfaction. However, it was found that caring efficacy and job satisfaction were positively correlated with each other. This demonstrated that high levels of caring efficacy are associated with better job satisfaction in this sample of nurses. Further, multivariable analyses found 84

101 relationships existed between age and caring efficacy. Multivariable analyses found relationships also existed between specialty area, health sector, and job satisfaction. 1. Socio-Demographic Factors (Personal Characteristics) Caring efficacy expectation Age 2. Socio-Demographic Factors (Organizational factors) Job satisfaction Health sector and Specialty Pay Promotion Supervision Fringe Benefits Contingent Awards Operating conditions Co-Workers Nature of Work Figure 3: General Linear Modelling Analyses of Socio-Demographic Variables and Caring Efficacy and Job Satisfaction 85

102 5.6 Is the Brisbane Practice Environment Measure a Valid and Reliable Instrument for Use in an Australian Population of Registered Nurses? The B-PEM is a work environment-specific instrument. It contains 33 items and is a self-administered questionnaire that measures the nursing PE (Webster et al., 2009). Individual items are measured on a 5-point Likert scale (strongly disagree to strongly agree) with a view of measuring specific characteristics present in the current nurses workplace. The items are scored from 1 to 5, with six items reversed scored Reliability Reliability analysis identified 5 subscales, and this analysis suggested that all subscales were reliable, with a Cronbach s alpha coefficient of these subscales of for professional development, for management support, for rostering, for out of depth and for workloads (see Table 10, p. 86). Normally reliability is presented before the validation of the instrument, however as we are assessing reliability of other instruments (GSES, WLCS, WCQ, CES and JSS), we have decided to present the reliability of these instruments after the validation of the B-PEM. 86

103 Table 7: EFA of the 5 Factors of B-PEM (Pattern Matrix) Factor Professional Development 27 There is support for professional development in my area In this area, clinical resources are adequate Off line time is offered for professional development There is time for staff development I have access to the information I need to do my job There is a high level of clinical expertise I can access Continuity of care is considered in this area There is equity in staff development opportunities Opportunities for advancement are available in this.405 organisation Management Support 14 My line manager is approachable I feel supported by my line manager My skills are acknowledged My line manager is responsive to emergent leave.570 requirements 31 My line manager is ready to help out in the clinical area I am acknowledged when I put in extra effort I feel like just a number I am treated as an individual.434 Rostering 30 Our roster complies with roster regulations There is equity in rostering in this area I am able to change my roster if necessary I participate in roster development.509 Out of Depth 19 I feel intimidated when working in this area In this area staff get away with bad behaviour I am asked to operate outside my scope of practice I am thrown in at the deep end Workloads 17 The workload is overwhelming in this area Staff workloads are equal The skill mix is about right in this area

104 5.6.2 Construct Validity First a principal component analysis (PCA) was performed on the 33 items originally included in the B-PEM to determine the number of factors. This yielded a five factor solution with eigenvalues greater than 1. We concluded that the number of constructs underlying the data coincided with the number components whose eigenvalues were >1. Principal axis factoring (PAF) with oblique rotation (promax) was then performed to elucidate the composition of the 5 factors. Pattern coefficients whose absolute value was less than 0.40 were excluded from further consideration. The PAF suggested that only 28 out of the original 33 items included in the B-PEM were instrumental in describing the PE subscales for this Australian registered nurse population. Items (2) Performance and appraisal is completed in this area, (4) I feel respected in the way people speak to me, (7) It is difficult to influence change in this area, (8) There is a great team spirit in my work area and (29) I enjoy coming to work of the B-PEM were excluded from further consideration. Perusal of the factor correlation matrix in the EFA indicated moderate to strong correlations between factors ranging from to (Table 8). All the 28 items were statistically significantly loaded (P 0.05) on the 5 subscales of the B-PEM. This suggests that any subsequent measurement model should allow for correlation among subscales (an oblique measurement model). On examination of the five factors, it was found that all of the Professional Development items loaded moderately to strongly positive on their factor and all of the items of the Management Support factor loaded moderately to strongly. One item, I feel just like a number loaded negatively on the latter factor as it was reversed scored. All the items of the Rostering factor loaded moderately to strongly and all the items on the Out of Depth factor loaded moderately to strongly also. All of the items on the Out of Depth factor loaded negatively as they were reversed scored. All three items of the Workloads factor loaded moderately to strongly on their factor with one item (the workload is overwhelming in this area), loading negatively as it was reversed scored. 88

105 Table 8: Factor Correlation Matrix of B-PEM Factor * All correlations were significant at the level of significance The five subscales identified were named as: 1. Professional Development: items relating to opportunities for advancement, skills development and equity of staff development opportunities (nine questions) 2. Management Support: items about line management support and feeling valued by the line manager (eight questions) 3. Rostering: items establishing the equity of rosters and flexibility of roster development (four questions) 4. Out of depth: items exploring whether nurses felt safe in their practice and whether bad behaviour of staff was tolerated (four questions) 5. Workloads: items about workloads being equal, overwhelming and whether the skill mix was satisfactory in the work area (three questions) Confirmatory Factor Analysis The five factors of the selected items were tested to evaluate data from the 28 items retained using CFA. All 28 items were significantly loaded on the respective factors (p< 0.05), professional development, management support, out of depth, workloads and rostering. The proposed measurement model is given in Figure 4. Item numbers in this figure coincide with the numbering in Table 7. Model fit was assessed using the Comparative Fit Index (CFI) and the Root Mean Square Error of Approximations (RMSEA). The raw and scaled χ 2 statistics were also inspected. Although the χ 2 = (p < 0.01) suggested a lack of fit, this fit index has been widely established to be upwardly biased with sample size and provides a poor measure of measurement model fit. It is included here only for reasons of convention (Smallman & Hurst, 89

106 2011). A model was assumed to adequately fit the data if it had a CFI > 0.09 (Brown, 2006) and a RMSEA < 0.05 (Browne & Cudeck 2005). The other fit indices suggested the measurement model fit the data well, χ 2 /df ratio = 2.845, RMSEA = and CFI = 0.906, (see Figure 4). The betas and standardized betas for the 28 items of the B-PEM are shown in Table 9. Table 9: Betas and Standardised Betas of 28 items of the B-PEM Factor Item # Β* βz Professional Development 27 There is support for professional development in my area In this area, clinical resources are adequate Off line time is offered for professional development There is time for staff development I have access to the information I need to do my job There is a high level of clinical expertise I can access Continuity of care is considered in this area There is equity in staff development opportunities Opportunities for advancement are available in this organisation Management Support 14 My line manager is approachable I feel supported by my line manager My skills are acknowledged My line manager is responsive to emergent leave requirements 31 My line manager is ready to help out in the clinical area I am acknowledged when I put in extra effort I feel just like a number I am treated as an individual Out of Depth 19 I feel intimidated when working in this area In this area staff get away with bad behaviour I am asked to operate outside my scope of practice I am thrown in at the deep end Workloads 17 The workload is overwhelming in this area Staff workloads are equal The skill mix is about in this area Rostering 20 There is equity in rostering in this area I am able to change my roster if necessary I participate in roster development Our roster complies with our roster regulations

107 Figure 4: Model for CFA of the Five Factors. Item numbering coincides with that given in Table 9. 91

108 Group Analysis To determine whether this is a single study population, multiple group analysis was performed to determine whether metropolitan and non-metropolitan (provincial/regional, rural and remote) nurses exhibit the same loadings and interfactor correlation of factors. It was revealed that there was no substantial difference in the structure for each group so it was decided that the CFA on the combined group should be retained. This further validates the use of these selected 28 items chosen from the B-PEM across the diverse Australian registered nurse population. In summary, the selected 28 items from the originally proposed 33 items of the B- PEM were valid for determining the effects of the PE upon registered nurses in Australia and may assist managers to evaluate the PE in order to improve JS and reduce burnout and job strain among nurses. 5.7 Research Question 4: Does a High Level of Self-Efficacy in Nursing Enhance Relationships Between Work Locus of Control, General Self-Efficacy, Coping Styles, the Professional Nursing Practice Environment and Job Satisfaction? Internal Consistency Prior to performing SEM, all scales used need to be established as reliable. A Cronbach s alpha coefficient was calculated for each of the subscales of the JSS, WCQ and B-PEM as well as the overall scores of the JSS, WCQ, B-PEM, GSES, WLCS and CES to determine internal consistency of the measure (Table 10). 92

109 Table 10: Cronbach s Alpha (ɑ) Analysis Showed High Internal Consistency (>0.06) for Each of the Scales and Subscales Variable Subscales ɑ General Self-Efficacy Scale Overall reliability Work Locus of Control Scale Overall reliability Brisbane Practice Environment Measure Overall reliability Professional Development (PE 1) Management Support (PE 2) Out of Depth (PE 3) Workloads (PE 4) Rostering (PE 5) Ways of Coping Questionnaire Confronting Coping (WOC 1) Distancing (WOC 2) Self-Controlling (WOC 3) Seeking Social Support (WOC 4) Accepting Responsibility (WOC 5) Escape Avoidance (WOC 6) Planful Problem Solving (WOC7) Positive Reappraisal (WOC 8) Caring Efficacy Scale (overall reliability) Job Satisfaction Scale (overall reliability) Pay (JSS 1) Promotion (JSS 2) Supervision (JSS 3) Fringe Benefits (JSS 4) Contingent Awards (JSS 5) Operating Conditions (JSS 6) Co-workers (JSS 7) Nature of Work (JSS 8) Communication (JSS 9) Structural Equation Modelling SEM was conducted to model possible causal connections among the identified variables GSE, WLC, coping styles and the professional nursing PE with the outcome variables CE and JS in this study. However, due to this being a crosssectional study design it was not the intent of the modelling to show causality. Maximum Likelihood was used to test the significance of the parameters (Anderson & Gerbing., 1988). 93

110 5.7.3 Modelling Model fit for the structural equation models was evaluated using the Comparative Fit index (CFI) and the Root Mean Square Error of Approximations (RMSEA). The raw and scaled χ² fit statistics were also examined and as previously discussed, these measures of model fit have been found to be upwardly biased in relation to sample size in studies of measurement model (Stallman & Hurst, 2011). In this study a model was presumed to represent an adequate fit if it had a CFI > 0.9 (Brown, 2006), and a RMSEA < 0.05 (Browne & Cudeck, 1993). Multiple-group analysis was used to determine whether inter-scale correlations and/or factor loading differed between the different education levels of Registered Nurses. The multiple-group analysis used an unconstrained model with all parameters allowed to vary without constraint Relationships between Variables Using Structural Equation Model Previous examination in study 2 of the multivariable analysis of socio-demographic variables and their effect upon CE and JS, found relationships existed between age and CE. Relationships were also found in the multivariable analysis to exist between years in current job, specialty area, health sectors, job status, education qualifications, geographical location and JS. The continuous, socio-demographic variables age, years of experience, years in current job and education levels were therefore included in the initial model. The standardised betas show that none of the continuous socio-demographic variables (age, years in current job, years of experience and education qualifications) were associated with any of the scale variables (GSE, WLC, coping styles and PE) in the structural model. Despite not being associated on the initial structural model, analyses of these variables were initially included to establish that they were indeed not significant. Three of the socio-demographic variables (i.e. everything but age) could not be shown to be associated with the scales (the 4 variables GSE, WLC, WOC and PE). Age was shown to be associated with all 4; however, it was noteworthy that the removal of all four socio-demographic variables led to a substantial improvement in the fit of the model. Even with age in the model, the 94

111 model could not be shown to fit. These variables were therefore excluded from subsequent modelling. Further, testing of the model was conducted between the exogenous variables (GSE, WLC, WOC and PE) and the endogenous variables (CE and JS). Including all WCQ domains in the single model resulted in an over-specified model. It was then decided to build the best model excluding these variables and then test each one at a time to observe whether they significantly contributed to explaining the adaptive outcome variables CE and JS. None of the WCQ could be shown to significantly contribute to the structural model and was therefore subsequently removed from the model. Strengths in associations between the variables were shown by using standardised betas. As only associations between the variables in this model were shown and not causal effects, the β z could be interpreted similarly to that of Pearson s coefficient (i.e. values nearing 1 or ( 1) are representative of strong positive or negative associations) Correlation Analysis Showing Associations Among Scales General Self-Efficacy, Practice Environment and Work Locus of Control The results of further correlation components of the model showed that the variable GSE was positively associated with PE (r = 0.28, p = < 0.001) and negatively associated with WLC (r = 0.26, p = <0.001). WLC was found to be negatively associated with PE (r = 0.49, p = < 0.001) (Table 11). Table 11: Correlations Among General Self-Efficacy, Work Locus of Control and Practice Environment Variable GSE WLC PE GSE 1 WLC * *** 1 1 PE *** *** *** Denotes P 0.05 level of significance. 95

112 5.8 Model Fit This model fit the data well overall, with χ² = , χ²/df = and CFI = All pathways were found to be significant except PE to CE (see Figure 5) Potential Causal Pathways and Caring Efficacy GSE positively explained CE (β = 0.38, p = 0.001). WLC negatively related to CE (i.e. as CE scores increased WLC scores decreased (β = 0.23, 0.001). Further, testing of the model found CE was explained by both GSE, which was positively related (β Z = 0.38, p < 0.001) and WLC, which was negatively related (β Z = 0.23, p = 0.001). PE was not significantly related to CE (β Z = 0.01, p = 0.85). Table 12 demonstrates these results Potential Causal Pathways and Job Satisfaction PE was shown to explain variation in JS (i.e. as PE scores increased, so did JS scores (β Z = 0.69, p = < 0.001). GSE was negatively related to JS (β Z = 0.09, p < 0.001) and WLC was found to be negatively related to JS (β Z = 0.20, p < 0.001). Table 12 demonstrates these results. Table 12: Regression and Standardised Regression Weights: Outcome Variables Explained by Personal Control and Environmental Factors Outcome Variable Risk/protective factor Β β Z P CE GSE <0.001 WLC <0.001 PE JS GSE <0.001 WLC <0.001 PE < Caring Efficacy and Job Satisfaction Modelling Further analysis conducted using the multiple squared correlation coefficient (R 2 ) showed the amount of variation in CE and JS accounted for by the model. It was 96

113 shown that 24 per cent of variation in CE can be accounted for by GSE, WLC and PE and 62 per cent of the variation in JS can be accounted for by these variables (see Figure 5). Figure 5: Final Model Components NB GSE (General Self-Efficacy); WLC (Work Locus of Control); PE (Practice Environment); CET (Caring Efficacy); JST (Job Satisfaction) 97

114 5.8.4 Summary of Findings The scale items GSE, WLC and PE were correlated with each other. All of these scale items were related to JS and GSE and WLC were related to CE. However, PE was not found to be related to CE. 5.9 Research Question 5: Can Caring Efficacy Mediate the Relationship among Work Locus of Control, General Self-Efficacy, Coping Styles, Professional Nursing Practice Environment, Predetermined Socio-Demographic Variables and Job Satisfaction? Mediation Analysis A mediation analysis was undertaken to see if CE confounded the relationship between the various covariates and JS. These covariates included GSE, WLC, PE, age, years of experience, years in current job and education qualifications. A mediational model is a causal model. This means the mediator in tandem with a covariate is the presumed cause of the outcome (mediation model) as opposed to hypothesizing a direct causal relationship existing between the independent variable and the dependent variable. Figure 6 and Figure 7 both demonstrate the explanation of the mediated and unmediated models. Multiple regression was used in this current study to estimate the paths c (total effect) and a, b and c 1 (direct effect) (see Figure 6 and Figure 7) for an explanation of the mediated model). Three steps (Baron & Kenny, 1986) were used to establish mediation in this study (see Table 13). These steps identified the existence of total zero-order relationships between the variables and identified if significant relationships occurred between the variables. No significant mediation occurs when the relationship between X and JS does not differ when a potential mediator is added. At first it seems c = c 1. In other words, CE does not mediate the relationship between X, a work-related trait (potential JS deterrent), and JS. Classical mediation is when c (the total relationship between X and JS) is substantially weakened (i.e. c > c 1 ) when we account for the mediator. Negative 98

115 confounding (masking) is when the relationship between PJS and JS is enhanced when we account for a mediator (e.g. c = 0.1, c 1 = 0.9). Table 13: Mediation Analysis (Three Steps) Model 1 Mediation Analysis Simple regression analysis was conducted with X predicting Y (Job Satisfaction). This estimated and tested path c and identified that there was an effect to be mediated. Y = β0 1 + cx + ϵ 1 Model 2 A simple regression was conducted with X predicting M to test path a. Y = β0 2 + c 1 X + bm + ϵ 2 Model 3 A simple regression analysis was conducted with M predicting Y to test the significance of path b. M = β0 3 + ax + ϵ 3 c (Total effect) X Y Figure 6: Unmediated Model (Kenny, 2009) 99

116 e 2 M (Mediating Variable) a b e 3 C 1 X (Independent Variable) Y (Dependent Variable) Figure 7: Mediation Model (MacKinnon, Fairchild & Fritz, 2007) In order to compare the two models in Figures 7 and 8 (i.e. to calculate the indirect effect), the steps presented in Table 13, page 106 (three regression models) need to be considered, because this is a generic procedure and was conducted for a number of mediation analyses in this present study Sobel s Test Following the generation of the three models presented in Table 13, the significance of the indirect pathway was then tested. A significant indirect effect establishes that a potential confounder does indeed mediate the relationship. The test is conducted to establish whether X effects Y through the pathway of (a) and (b) (MacKinnon, Fairchild & Fritz, 2007). Sobel s test (Sobel, 1982) was used for this analysis (Table 14) because it is a well-established method for testing significance of mediators. Table 14: Sobel s Test for Indirect Effect Analysis Outcome Step 1 c-c 1 = ab Indirect effect Step 2 SE (c c 1 2 ) = SE(ab) = E a b 2 + E 2 b a 2 Step 3 Confidence interval (CI) 95% for ab If zero not in CI indirect effect = mediation SE(ab) 100

117 Mackinnon, Lockwood, Hoffman, West & Sheets (2002) reported cross-sectional, correlational studies that observe the aetiology of a variable, provide information of the conceptual theory relating to mediating variables of a dependent variable such as JS in this study. However, in cross-sectional studies such as this study, it is difficult to establish the causal direction (i.e. whether CE mediates the relationship between GSE, WLC and PE to JS or vice-versa) Mediation of Caring Efficacy and General Self-Efficacy and that of Work Locus of Control and Practice Environment On observation of Table 15, it was noticed there was a significant indirect effect between JS and age. However, although at first CE seems to be a mediator for the relationship between age and JS, it should be noted that this relationship was not significant. To determine whether CE mediated the relationship between GSE, WLC and PE, mediational analyses were conducted. It was found that CE mediated the relationship between GSE and JS. CE did not mediate the relationship between WLC and JS or PE and JS (Table 15). Potential JS determinant Table 15: Mediation Effect of Caring Efficacy to Job Satisfaction Total effect (c) C z Direct Effect (c 1 ) C 1 z Indirect effect (c c 1 = ab) General Self ** , 0.579* Efficacy Education , Age , 0.1* Years of , experience Years in , Current Job Practice Environment ** , 0.01 Work Locus of Control ** , 0.08 * Denotes mediation effect; ** denotes significant correlation CI 101

118 a) Without a mediation effect (i.e. c = c 1) it is X (independent of any mediator) that explains JS (i.e. the relationship between X and JS is in no way mediated [partially explained] by CE). b) A significant effect (c 1 c equals a significant indirect effect) means that a significant part of the relationship of X and JS is mediated through CE (i.e. the nature of the direct relationship between X and JS is either less important or of a different level of and/or direction importance than it at first appears). In summary, caring efficacy mediated the relationship between general self-efficacy and job satisfaction. Caring efficacy did not mediate the relationship between any of the socio-demographic variables, work locus of control or practice environment and job satisfaction in this study Summary of Findings In this sample, levels of caring efficacy varied according to the different sociodemographics. High levels of caring efficacy were found in registered nurses who were from the state of South Australia and in those who worked in the rural and remote geographical location. Registered nurses who worked in the community and domiciliary sector were found to have high levels of caring efficacy, while those working in the public sector had the lowest levels in this sample. Female nurses and those nurses whose marital status was separated or divorced were found to have the highest levels of caring efficacy. In relation to education qualifications, caring efficacy levels were found to be highest in nurses with a certificate level of qualification. Registered nurses who worked part time and nurses who worked in the specialty area of midwifery showed the highest levels of caring efficacy. In this sample, job satisfaction was found to vary also according to the different socio-demographics. Job satisfaction was found to be highest nurses for working in South Australia and in those who worked in a metropolitan geographical location. Nurses who worked in the private sector in this sample had the highest levels of job satisfaction, while nurses working in the public sector in this sample showed the 102

119 lowest levels of job satisfaction. In this sample, female nurses and those nurses who were married showed the highest levels of job satisfaction. In relation to education qualifications in this sample, those nurses with a masters or PhD level were found to have the highest levels of job satisfaction. In this sample full-time nurses were found to have the highest levels of job satisfaction and nurses working in the medical and surgical specialty areas reported higher levels of job satisfaction. This study aimed to examine the relationships between socio-demographic variables, scale items and caring efficacy and job satisfaction in a population of Australian registered nurses. Results of bivariate analyses showed age, years of experience and years in current job were all highly positively correlated with each other. Age and years of experience were also correlated with caring efficacy and should be considered further in the future model development of caring efficacy. However, years of experience was found to be highly correlated with age and was therefore excluded from any multivariable analysis. In this sample, the nurses who were older and more experienced, showed higher levels of caring efficacy. Bivariate analysis showed no significant interactions between age, years of experience and years in current job and job satisfaction. Caring efficacy and job satisfaction were highly correlated with each other suggesting that high levels of caring efficacy might predict better job satisfaction in this sample of nurses. Results of a one-way ANOVA showed positive significant associations existed between those nurses who were widowed or divorced and caring efficacy. Positive significant associations were also found between nurses working in the community health sector in this sample and caring efficacy. Positive significant associations were found between those nurses who worked in the state of South Australia and job satisfaction. Positive significant associations were also found between those nurses who were working in the private sector and job satisfaction. Positive significant associations were found between those nurses working in the specialty area of midwifery and job satisfaction. The results of a simple linear regression revealed that age explained caring efficacy levels on average, and age also explained job satisfaction levels on average 103

120 indicating that older nurses in this study had higher levels of caring efficacy and job satisfaction. It was found after running a multivariable model (General Linear Model [GLM]) with caring efficacy and a GLM with job satisfaction that age remained significant with caring efficacy. Specialty area and health sector remained significant with job satisfaction. The psychometric properties of the selected 28 items from the original Brisbane Practice Environment Measure (B-PEM) were examined. It was found that the psychometric properties of the B-PEM demonstrated satisfactory validity to determine the factors professional development, management support, out of depth, workloads and rostering in a diverse Australian population of registered nurses. Structural Equation Modelling (SEM) showed the data fit well overall. All pathways were found to be significant except practice environment to caring efficacy. General self-efficacy positively explained caring efficacy. Work locus of control was negatively related to caring efficacy (i.e. as caring efficacy scores increased, work locus of control scores decreased). Further testing of the model found caring efficacy was explained by both general self-efficacy, which was positively related and work locus of control, which was negatively related. Practice environment was not significantly associated with caring efficacy. The practice environment (consisting of the subscales professional development, management support, out of depth, workloads and rostering) explained job satisfaction (i.e. as practice environment scores increased, so did job satisfaction scores). It was found that job satisfaction was explained by practice environment, which was positively related, general self-efficacy, which was negatively related and work locus of control, which was negatively related. Analysis using the multiple squared correlation coefficient (R 2 ) showed that 24 per cent of variation in caring efficacy can be accounted for by general self-efficacy, work locus of control and practice environment. 104

121 Self-Efficacy Expectation Caring Efficacy Expectation Work Locus of Control Job Satisfaction Nursing Environment Practice 2. Professional Development 3. Management Support 4. Out of Depth 5. Workloads 6. Rostering 1. Pay 2. Promotion 3. Supervision 4. Fringe Benefits 5. Contingent Awards 6. Operating Conditions 7. Co-workers 8. Nature of Work 9. Communication Figure 8: Caring Efficacy and Job Satisfaction Model Additionally, the results of a mediation analysis revealed caring efficacy mediated the relationship between general self-efficacy and job satisfaction. Caring efficacy did not mediate the relationship between any of the socio-demographic variables and job satisfaction, work locus of control and job satisfaction or practice environment and job satisfaction. The following chapter will discuss the findings of study 2 and a comparison of these findings to the existing literature will be presented. 105

122 Chapter 6: Discussion 6.1 Introduction This chapter will respond to the research questions and will present a discussion of the evidence found in study 2. A comparison of these findings to the existing literature will be made. This current study examined a number of relationships among socio-demographic (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), personal control (general self-efficacy, work locus of control and ways of coping) and organisational (practice environment) variables and caring efficacy and job satisfaction in an Australian population of Australian registered nurses. Firstly, a discussion of the socio-demographic variables in relation to caring efficacy and job satisfaction is presented. Secondly, the psychometric properties of selected items of the 33-item Brisbane Practice Environment Measure in an Australian sample of registered nurses are discussed. Analysis of the data examined relationships between the socio-demographic variables and the outcome variables, caring efficacy and job satisfaction. First of all, the means, standard deviations, medians and ranges for the categorical variables (gender, marital status, education qualifications, geographical location, job status, health sector and state) were calculated and reported. These variables pertained to the personal and organisational categorical variables. Pearson s correlation coefficient (r) analysis was undertaken for the continuous variables, age, years of experience as a registered nurse and length of time in current job (personal characteristics), caring efficacy and job satisfaction. One-way ANOVAs were conducted to examine if different levels of caring efficacy and job satisfaction existed in relation to the variables. Simple linear regression was undertaken to examine if relationships existed between the continuous variables and caring efficacy and the continuous variables and job satisfaction. A multivariable analysis was then conducted following the purposeful selection of covariates approach using a General Linear Model. 106

123 The results of the multivariable analysis of personal characteristics (age, years of experience, years in current job, gender, marital status, job status, education level and specialty area) and organisational (state, health sector and geographical location) were identified and described. Next, the relationships among socio-demographic variables, personal control factors and the outcome variables caring efficacy and job satisfaction are discussed. This part of the discussion will present the research framework model using the SEM analyses. Finally, the discussion will present a mediation analysis of caring efficacy in relation to the continuous socio-demographic factors, personal control and organisational factors and job satisfaction Associations of Socio-Demographic Factors with Caring Efficacy and Job Satisfaction In study 2, associations between socio-demographic factors and caring efficacy and socio-demographic variables and job satisfaction were determined in a sample of registered nurses from an Australian industrial and professional organisation. There are very few published empirical studies which have examined the caring efficacy scale and therefore the comparison of other studies are minimal. However a study by Amendolair (2007) found older nurses showed higher levels of caring efficacy; whereas a study by Lawrence (2002) reported that although not significant, older nurses scored lower on the Caring Efficacy Scale for age and years of experience. than nurses who were less than 30 years of age. As age and years of experience in nurses increased in the current study, confidence in their ability to conduct caring behaviours to patients was also shown to increase. These findings are similar to those of Bandura (1997) who reported, self-efficacy results from social support, experience, and feedback, all of which increase over time. Previous studies have identified that relationships exist between the sociodemographic variables age, years of experience, years in current job, marital status, education qualifications, geographical location and job satisfaction (Wade et al., 2008; Al-Hussami, 2008; ; Li, Lambert & Lambert, 2008; Hegney, McCarthy, Rogers-Clark, & Gorman., 2002; Shader, Broome, Broome, West, & Nash, 2001; 107

124 Shields & Ward, 2001). Furthermore an Australian study by Duffield et al. (2009) found older nurses to have better job satisfaction and also were more likely to stay in their current job Organisational Characteristics, Caring Efficacy and Job Satisfaction From the four participating Australian states in this study, the majority (41.2%) of the respondents were from New South Wales. The results from these four Australian states in relation to caring efficacy showed nurses from South Australia had the highest levels of caring efficacy and nurses from New South Wales showed the lowest levels. No studies reporting on relationships between nurses working in the Australian states and caring efficacy were found. The results from these four Australian states in relation to job satisfaction levels showed nurses from South Australia had the highest levels of job satisfaction and nurses from New South Wales showed the lowest levels. No studies were found that investigated relationships between nurses working in the Australian states and job satisfaction. Further, the results of a one-way ANOVA found that a significant difference between the four Australian states and job satisfaction existed. However, after conducting a multivariable analysis using a GLM, the results showed the covariate state did not remain significantly associated with job satisfaction. Nurses in Australia work in different geographical locations that are identified as metropolitan, regional and rural and remote areas (Australian Institute of Health and Welfare, 2004). The results of this study found that levels of caring efficacy were lowest in nurses working in metropolitan areas. In contrast, nurses working in the rural remote areas had the highest levels of caring efficacy. No studies have investigated relationships between nurses working in the different geographical locations in Australia and caring efficacy. The results of this study found nurses working in the metropolitan areas had high levels of job satisfaction and nurses working in the rural remote areas had the lowest 108

125 levels of job satisfaction compared to other areas. Interestingly earlier Australian studies of nurses working in rural and remote areas reported high levels of job satisfaction (Hegney et al., 2000, 2002). However, job satisfaction levels on average in all geographical locations in this study were reported at a high level and therefore support these current results. To date, no Australian studies that have investigated the other geographical locations of work and job satisfaction in registered nurses in Australia were found. The four health sectors in Australia were identified and include private, public, residential and aged care, and community. Results of the analysis undertaken showed nurses who worked in the community health sector had the highest levels of selfefficacy in nursing (caring efficacy). Further, no studies reporting on relationships between the different health sectors in Australia and caring efficacy in registered nurses were found. The results of this study showed high levels of job satisfaction for all of the sectors. Further to this, the results of the multivariable analysis conducted in this current study showed significant relationships existed between the variable, health sector and job satisfaction. It was found that nurses who worked in the private health sector had the highest levels of job satisfaction. No studies reporting on job satisfaction levels and registered nurses working in the private health sector were found. In contrast though, it has been reported that recent changes in policies in western democracies including Australia have had a negative effect upon the perceptions of nurses job satisfaction for those working in public hospitals (Brunetto and Farr-Wharton, 2004). 109

126 6.1.3 Personal and Continuous Socio-Demographic Characteristics, Caring Efficacy and Job Satisfaction Age, Years of Experience and Years in Current Job The variables age and years of experience as a registered nurse were shown to be moderately correlated with caring efficacy. This means that in this sample, the nurses who were older and more experienced showed higher levels of caring efficacy. These results are supported by the study conducted by Amendolair (2007) where it was also identified that older experienced nurses showed higher caring efficacy scores. Although the number of years nurses worked in their job was correlated with age and years of experience, it was not found to be correlated with caring efficacy in this sample. A multivariable analysis was conducted and found significant relationships existed between age and caring efficacy. There is a role for the older and more experienced nurse who may be able to facilitate the orientation of new nurses in the clinical setting through role modelling, constructive feedback or verbal persuasion to increase their levels of self-efficacy. These self-efficacy strategies may improve patient outcomes as self-efficacy in nursing is believed to increase the time and effort nurses spend on patient problems. High levels of self-efficacy in nursing are also related to nurses job satisfaction which in turn has been identified as a key indicator for retention (Aiken et al., 2002). This information is important for nursing leaders at this time of global nursing shortages. Further, the covariates age, years of experience as a registered nurse and length of time in current job did not have any significant effect upon job satisfaction. Mixed results have been reported from other studies examining age and years of experience and nurses job satisfaction. Years of experience was found to predict nurses job satisfaction in one study (Li & Lambert., 2008), but the results of a meta-analysis (Blegen, 1993) showed correlations between age and job satisfaction and years of experience and job satisfaction were not significantly correlated. The results of the current study add to the literature based on these other historical studies. However, in contrast to these results, other studies (Al-Hussami, 2008; Li & Lambert., 2008; 110

127 Shields & Ward, 2001; Wade et al., 2008) have found that increasing age was highly positively related with job satisfaction in nurses Years in Current Position No relationships were found to exist between years in current position and caring efficacy. There are currently no studies reporting on relationships between the number of years that nurses worked in their current job and caring efficacy. No relationships were found to exist between years in current position and job satisfaction. However, in contrast to the results of job satisfaction in this current study, a recent US study found that the number of years nurses worked in their current position correlated positively with job satisfaction (Ulrich et al., 2007) Specialty Area It was found that nurses who worked in the specialty area of midwifery had the highest levels of caring efficacy in nurses. No research was found that discussed caring efficacy in relation to this specialty area. However, research conducted in an intensive care specialty area showed the importance of self-efficacy in nursing. The study by LeBlanc et al. (2009) identified that professional efficacy beliefs in nurses were related to collaborative practice and team commitment if adequate resources were available for them to do their job well. This in turn led to improved practice behaviours. The results of this current study found nurses who worked in the specialty area of mental health showed the lowest average levels of caring efficacy in this sample. A descriptive review of the literature (Dunn, Elsom and Cross, 2007) identified that mental health nurses may encounter patients who are aggressive, confused and have other cognitive deficits on a regular basis, which create stressful situations. A review by Dunn et al. (2007) supports the findings in this current study, as it has identified the critical need for the introduction of strategies to improve nurses self-efficacy for those working in the area of mental health. These strategies include the safe and effective management of patients and for their own safety. The outcomes identified 111

128 in the current study are vital. The development of future professional development programmes for nurses working in areas such as mental health, need to incorporate opportunities and support for nurses using strategies of self-efficacy such as repeated performances. This may further develop nurses confidence and ability to introduce new knowledge and skills with a caring orientation. This is also supported by Calabro, Mackey and Williams (2002) who indicated that increased self-efficacy in mental health nurses obtained from educational programmes can develop confidence and improve personal mastery in the management of aggression. Job satisfaction was highest in those nurses working in medical and surgical nursing. Further, the results of a multivariable analysis in this current study found that specialty area was significantly related to job satisfaction. Studies by Doiron and Jones (2006) and Shields and Ward (2001) support these results as they found registered nurses working in surgical wards at the time had higher levels of job satisfaction and were more likely not to leave their current job. There is currently no research available reporting on job satisfaction levels in nurses working in the area of medical nursing. Job satisfaction was found to be lowest in those nurses working in the specialty area of paediatrics in the current study. Interestingly and in contrast to these results, other studies examining job satisfaction and paediatric nursing found nurses working in this area were generally enthusiastic about their work and satisfied with their patient care (Ernst et al., 2004). An Australian study by Doiron and Jones (2006) revealed that job satisfaction levels were found to be highest in paediatric nurses. In a United Kingdom (UK) study by Shields and Ward (2001) found conformity existed in the results of nurses across specialty areas. However they also reported that nurses in that study who were working in paediatrics showed higher levels of job satisfaction in contrast to those working in mental health Gender The current study found female nurses in this sample had higher levels of caring efficacy than male nurses. Again, no studies reporting on relationships between gender and caring efficacy in registered nurses were found. Female nurses were also 112

129 found to have higher levels of job satisfaction as compared to male nurses. This is consistent with the results of the study by Shields and Ward (2001), where it was found lower levels of job satisfaction existed in male nurses Education Qualifications Levels of caring efficacy were examined with education qualifications and it was found that those with a certificate level qualification (the minimum requirement of registered nurse education in Australia) had the highest levels of caring efficacy. Coates (1997) also found no significant differences in caring efficacy between those with baccalaureate (lowest education level in that study) and higher education qualifications of masters and PhD. Notably, there were no nurses with a certificate level in Coates study. In 1994, the last of the changeover from an apprenticeship training model of registered nurse to an academic model of nursing education occurred (Commonwealth of Australia, 2002). Hence, it is likely the findings of the current study are a result of nurses in this group being older and more experienced. The current study found age and years of experience as a registered nurse were highly correlated with caring efficacy. When levels of job satisfaction with education qualifications were examined, it was found that those with a masters or PhD level of qualification had the highest level of job satisfaction, and in contrast, those with a certificate level qualification had the lowest job satisfaction levels. As reported by Al-Hassumi (2008), there is currently limited research available that discusses education levels and job satisfaction. Further, conflicting results have been reported from those that do exist. Education qualifications were found to be positively related to job satisfaction in nurses in the study by Al-Hussami (2008). In contrast, Adams et al. (2000) found no relationships existed between job satisfaction scores and nurses level of education. Further, a meta-analysis (Blegen, 1993) and a UK study (Shields & Ward, 2001) showed that nurses with higher levels of education had lower levels of job satisfaction. In contrast, Pfeffer and O Reilly (1987) have suggested that variations in education within work groups may have an effect on job attitudes and behaviours rather than the level of an individual s education. 113

130 Marital Status Caring efficacy was highest in nurses who were widowed or divorced. No studies that reported on relationships between caring efficacy and marital status were found. This current study adds to the literature on caring efficacy and the sociodemographic variable marital status. Job satisfaction levels were highest in those nurses who were married and lowest in those who were widowed or divorced. A previous study also found being married related positively and significantly with job satisfaction in nurses (Shields & Ward, 2001) Job Status On examination of job status in this current study, it was found that registered nurses who were employed on a permanent basis (full time or part time) had higher levels of caring efficacy compared to nurses who work in casual positions. Moreover, this current study adds to the literature on caring efficacy. The issue of non-permanent nursing staff having a reduced opportunity to build a relationship with patients was found by Jamieson, Williams, Lauder and Dwyer (2008). On further examination of job status in this current study, it was found that registered nurses who were employed on a permanent full-time basis had higher levels of job satisfaction compared to nurses who were working in casual positions. Studies reporting on relationships between job status and job satisfaction found a poor understanding by management existed of the needs of nurses of non-standard (nonpermanent) positions. The areas of poor understanding of these nurses include, lack of continuity of care, reduced matching of skills to the workplace, nurses working in unfamiliar contexts, a diminished opportunity for nurses to build a relationship with patients and reduced access for nurses to professional development (Jamieson et al., 2008; Lumley, Stanton & Barram, 2004). This poor understanding by management can lead to poorer patient outcomes on the occasions that casual nurses are working (Creegan, Duffield & Forrester 2003; Grinspun, 2003). These issues of staff 114

131 interactions and patient outcomes may result in reduced job satisfaction and nursing retention (Creegan et al., 2003) Associations Between Caring Efficacy and Job Satisfaction The results of correlation analysis conducted on caring efficacy and job satisfaction were strong in this current study. This means when levels of caring efficacy were high so were the levels of job satisfaction in this sample of Australian registered nurses. The results of this current study are supported by a meta-analysis, which reported relationships were found between self-efficacy and job satisfaction (Judge & Bono, 2001). General self-efficacy was highly correlated with job satisfaction in this meta-analysis. General self-efficacy and caring effciacy were highly correlated in this current study. Other studies have identified that self-efficacy in nursing influences nursing performance, which can affect the outcomes of organisations (Le Blanc et al., 2010: Lee & Ko, 2010; Manojlovich, 2005a). A high level of selfefficacy in nursing enables nurses to achieve the desired results and carry on with difficult challenges in order to achieve the outcomes they value (Manojlovich, 2005b). This information is important for nursing leaders and policymakers to ensure strategies that enhance self-efficacy in registered nurses are implemented in professional development programmes. In turn, this may improve job satisfaction and may have an effect upon recruitment and retention in nurses. In summary, results from this study have identified that relationships exist between the socio-demographic variables of Australian registered nurses, caring efficacy and job satisfaction at the bivariate level of statistical analysis. The findings showed that when levels of caring efficacy were high, the levels of job satisfaction were also high in this sample of Australian registered nurses. Further, multivariable analyses found relationships existed between age and caring efficacy. Multivariable analyses found relationships existed between the variables specialty area, health sector and job satisfaction. Furthermore, results of the correlation analysis conducted on caring efficacy and job satisfaction were strong in this sample of Australian registered nurses. 115

132 6.2 The Validity and Reliability of the Brisbane Practice Environment Measure (BPEM) This current study evaluated the psychometric properties of selected items chosen from the BPEM, based on the original 33 items of this measure in a diverse Australian population of registered nurses. Exploratory Factor Analysis with Principle Component Analysis using the original 33-item Brisbane Practice Environment Measure was undertaken to determine the structure of the subscales. Following this, Confirmatory Factor Analysis was conducted using the identified 28 items chosen from the BPEM following the pilot study. Confirmatory Factor Analysis revealed a five-factor solution of the 28 selected items that showed satisfactory fit, suggesting satisfactory construct validity of the factors. A 4-factor structure of the BPEM (Flint et al., 2010) has been recently developed using a similar approach in a population of registered nurses from a Brisbane metropolitan hospital. Confirmatory Factor Analysis was used to compare factor structures across the studies using the original BPEM. The psychometric properties of the selected 28 items from the original 33 item measure demonstrated satisfactory validity to determine the factors professional development, management support, out of depth, workloads and rostering in a diverse Australian population of registered nurses. There is growing evidence that nursing shortages and turnover rates are associated with job satisfaction (Aiken et al., 2011) and the nursing practice environment (American Association of Colleges of Nursing, 2002). Improving the nursing practice environment can potentially produce benefits to the health system such as better job satisfaction, prevention of burnout and improved workforce retention (Duffield et al., 2009; Kazanjian et al., 2005; Manojlovich, 2005b). Evidence shows that in areas of higher registered nurse staffing levels, associations have been made with lower patient morbidity and mortality, improved patient s satisfaction and health-related quality of life. Research has also found that certain organisational traits that support nurses in their practice enable them to conduct complex functions and provide optimal patient care (Aiken et al., 2008, Aiken et al., 2002). If 116

133 organisations have an understanding of the practice environment this asserts an opportunity for the examination of areas that require change. Consequently, this may retain or attract nurses to the organisation (Parker et al., 2010). In summary, the selected items from the original Brisbane Practice Environment Measure were valid for determining the effects of the practice environment upon registered nurses in Australia. This may assist managers to evaluate the practice environment in order to improve job satisfaction among nurses. These findings may provide vital information for nursing leaders concerned with the retention of experienced and knowledgeable staff. 6.3 Model Testing Bivariate analyses were conducted to examine the direction and strength of this study. Three of the continuous socio-demographic variables (years in current job, years of experience and education levels) could not be shown to be associated with the scales measuring general self-efficacy, work locus of control and practice environment. The socio-demographic variable age was shown to be associated with all three of the scales (general self-efficacy, work locus of control and practice environment), but after the removal of all four continuous socio-demographic variables (including age), it was found that there was substantial improvement in the fit of the model. All of these variables were then excluded from further model testing. Further testing of the model was conducted between the personal control variables (general self-efficacy, work locus of control and coping styles), the organisational variable (the nursing practice environment) and the outcome variables caring efficacy and job satisfaction. The ways of coping measure was not found to contribute to the structural model following data analysis using structural equation modelling in study 2 and was therefore removed. Job satisfaction is an emotional affective response to a job or aspects of a job (Locke, 1976). This current study showed that associations exist between the confidence of nurses to develop caring relationships and exhibit caring behaviours towards patients (caring efficacy) and job satisfaction. In addition, the nurses who had an internal 117

134 locus of control and high levels of general self-efficacy showed increased confidence in their ability to deliver caring behaviours and develop caring relationships with patients. Finally, older and more experienced nurses showed increased ability to express care and showed higher levels of job satisfaction. Therefore, these results suggest that nurses who have high levels of self-efficacy in nursing may have better job satisfaction. Higher levels of job satisfaction may reduce nurses intentions to quit (Shields & Ward, 2001) at a time where there is a shortage globally. Modelling showed significant relationships existed between the personal control variables (general self-efficacy and work locus of control) and the outcome variable caring efficacy. Significant relationships were also found to exist between the personal control variables (general self-efficacy and work locus of control), the organisational variable (practice environment) and the outcome variable (job satisfaction) Modelling Analyses and Caring Efficacy Amendolair (2007) examined the relationship between the nurses ability to express caring behaviours and to develop caring relationships with patients and job satisfaction. This study and the current study both showed high means on the Caring Efficacy Scale indicating the samples of nurses from both studies had high levels of confidence in developing caring relationships with patients. Both of these studies have also shown that the nursing participants feel positive in relation to their jobs. The current study therefore supports the conceptual model, which proposed that nurses who believed in their ability to develop caring relationships and show caring behaviours towards patients will show higher levels of job satisfaction. As previously identified, self-efficacy is an individual s belief that they have the confidence and ability to organise and conduct the required actions to get the job done (Bandura, 1997). In relation to the caring efficacy and job satisfaction in nursing, this relates to performing caring behaviours and developing caring relationships (caring efficacy) with patients. The results of modelling analysis showed that there was a positive relationship between general self-efficacy and caring efficacy in this current study (i.e. as levels 118

135 of general self-efficacy increased, so did the levels of caring efficacy). This means that caring efficacy levels were found to be high if general self-efficacy levels were high i.e., those with high levels of self-efficacy may believe they have the ability to convey a caring orientation and build up caring relationships with patients as reported by Coates (1997). This suggests that personal mastery experiences, vicarious experience (observing others), verbal persuasion and physiological information (self-evaluation of physiological and emotional states) as identified by Bandura (1986) are important influences on caring efficacy in registered nurses. Low levels of self-efficacy in nursing (i.e. nurses believing they do not have the ability to activate the required motivation, cognitive resources and courses of action required to have control over their work) may have an effect on nurses choosing to remove themselves from patients and from intense healthcare situations (Begat et al., 2005; Manojlovich, 2005a). Further, high levels of self-efficacy in nurses may be an indicator for enhanced professional nursing practice behaviours (Le Blanc et al., 2010; Lee & Ko, 2009; Manojlovich, 2005a). There is a dearth of studies reporting on relationships between caring efficacy and general self-efficacy in registered nurses. This current study is important because it adds to the current literature on self-efficacy in nursing. The results of this current study showed there were no significant associations found between the practice environment (features of hospitals able to attract and retain committed nursing staff) and caring efficacy. In contrast to the results of this current study, it was reported by Manojlovich (2005a) that the environment, as measured by structural empowerment (opportunity, resources, information and support) directly and significantly contributed to self-efficacy (as measured by the Caring Efficacy Scale). Other literature has suggested practice environments that support nurses in their practice have been found to result in better nurse and patient outcomes (Aiken et al., 2011; Lake, 2002). Further, individuals who see themselves as being able to perform across many situations (general self-efficacy) should be able to identify as being in control of their environment (Bono & Judge, 2003). Alternatively, perhaps some environments may not provide the necessary resources or necessary opportunities for nurses to engender self-efficacy in their nursing practice in order to take control of their work (Manojlovich, 2005a), hence the necessary confidence and 119

136 ability to develop a caring relationship and a caring orientation towards their patients is affected. Modelling also identified that those nurses with an internal work locus of control showed high levels of caring efficacy. There have been no previous studies that have reported on the relationship between work locus of control and caring efficacy in registered nurses. Nonetheless, general self-efficacy was found to be negatively associated with work locus of control in the current study. These results mean that nurses with high levels of self-efficacy were those with an internal work locus of control (internals). The personality construct of locus of control refers to people s beliefs or confidence they have control over the outcomes of events that affect them (internals) (Storms & Spector, 1987) and self-efficacy refers to people s confidence in relation to getting things done (Bandura, 1977). This current study has added further evidence to the literature indicating that general self-efficacy and locus of control are theoretically linked (Judge et al., 1998); and as found in earlier studies (Bandura 1977; Scherer et al., 1982), general self-efficacy may be even stronger when in association with locus of control. While it seems that high levels of perceived general self-efficacy and an internal locus of control should produce positive results, the literature also indicates that while people may believe that actions control the outcomes of future events, they may or may not believe they can actually undertake these actions to bring about the desired result (low self-efficacy) (Bandura, 1997). This information is important as it has been found that nurses with high levels of self-efficacy and an internal locus of control may work better in some specialty areas, for example, mental health (Dunn et al., 2007). There are no other current studies investigating general self-efficacy, work locus of control and caring efficacy in the nursing context. Nevertheless, this information is vital for nursing leaders when developing strategies for professional development programmes that enhance nurses perceived self-efficacy. It further provides valuable information to these leaders of the benefits of improved general self-efficacy in relation to work locus of control, and the potential for improved patient and nursing staff outcomes (Dunn et al., 2007) What is more, the development and 120

137 implementation of strategies which enhance nurses self-efficacy and work locus of control in specialty work units may also assist behaviour change in relation to the workplace requirements (Dunn et al., 2007) Modelling Analyses and Job Satisfaction It is surprising, that in contrast to what other studies have shown, there was a negative association between job satisfaction and general self-efficacy in this current study (Beta Z = 0.09, p<0.001). However, an association of is hardly substantial and its statistical significance may be an artefact of our large sample size rather than suggesting that there is indeed a negative association between job satisfaction and general self-efficacy. Path coefficient values 0.10 could signify a small effect, values of approximately 0.30 a medium effect and values of 0.50 a large effect (Kline, 1998). Moreover, other studies have found that high levels of selfefficacy in individuals positively relate to higher levels of job satisfation and job performance (Judge & Bono, 2001; Judge et al., 1998). It is believed that people with high levels of self-efficacy are more likely to persist in managing difficult situations and continue through hardships to achieve the outcomes they value. This may result in better job satisfaction (Gist & Mitchell, 1992). The results from this current study found job satisfaction was negatively explained by work locus of control. This means that nurses found to have an internal work locus of control showed higher levels of job satisfaction. The literature has identified that those with an internal locus of control believe that reinforcement comes from their own personalities and those with an external locus of control see reinforcement as a result of fate, chance or powerful others (Rotter, 1966). Thus, it is believed that those who show internality are better adjusted and cope better than those who show externality (Rotter, 1966). Further, the literature supports the results of this study, as work locus of control (which specifically relates to the work domain) has been found to correlate with general locus of control and job satisfaction (Muhonen & Torkelson, 2004; Spector, 1988). If people believe they have limited control over circumstances and limited opportunities to partake in decision making in their work, it can have an effect upon job satisfication (Maylor, 2001). Employees with an internal locus of control are able to make decisions of what constitiutes appropriate 121

138 behaviour, while those with an external locus of control decide what appropriate behaviour is after watching others. Internals have also been shown to be more creative and productive;they believe they are utilising their potential therefore achieving more (Bandura, 1977; Forte, 2005). Additionally, it has also been found in the workplace that supervisors who are internal have a more person-orientated manner (Duffy et al., 1977). In contrast, people who score high in externality are reported to have less job satisifaction, feel more estranged from the work setting and are less involved in their job compared to internals (Duffy et al., 1977). Hence, externals are more likely to show dysfunctional behaviour as a result of exasperation within an organisation, while frustrated internals are more likely to work out ways to quickly resolve the problem (Spector, 1978; Storms & Spector, 1987). Frustrated externals may leave their job, or refer to their failure as being caused by outside factors (Forte, 2005). Externality has been previously identified as being associated with intentions to quit a job (Spector & Michaels, 1986). While there are currently no studies currently reporting on relationships between job satisfaction and work locus of control in registered nurses, this is important information for nursing leaders and healthcare organisations when developing appropriate recruitment and retention strategies for nurses. Understanding the nature of the nursing profession as shown in this current study may be an important approach to improving job satisfaction in nurses (Coates, 1997; Manojlovich, 2005b). The results of this current study found that the nursing practice environment was positively related to job satisfaction. This means nurses who perceived the environment to be satisfactory had higher levels of job satisfaction. A study by Aiken et al., (2011) gives support to this outcome and other studies (Zangaro & Soeken, 2007; Brady-Schwartz, 2005) reported nurses leave unsatisfactory work environments. Additionally, it is reported that improved patient outcomes are related to satisfactory work environments (Aiken et al., 2011; Duffield et al., 2009; Brady- Schwartz, 2005). This current study has shown that nursing and organisational leaders need to incorporate such requisite features of the work environment of professional development, management support, out of depth, workloads and rostering in order to improve nurses job satisfaction and recruitment and retention of nurses in Australia. The characteristics of job satisfaction for this current study included: Pay, Promotion, Supervision, Fringe Benefits, Contingent Rewards, 122

139 Operating Conditions, Co-Workers, Nature of Work and Communication. Recent Australian research supports the result that the nursing practice environment is positively related to job satisfaction and also discusses that nursing leaders can have an influence upon satisfactory work environments for nursing staff (Duffield et al., 2009; Paliadelis, 2008; Paliadelis & Cruickshank, 2008). Nursing shortages continue to be an ongoing Australian and global issue (Oulton, 2006). It has also been identified that an examination of personal and organisational issues in nursing are needed to explore and develop strategies aimed at recruitment as well as retention of nurses in the Australian context (Middleton et al., 2010; Duffield et al., 2009). The practice environment is considered an integral part of nursing culture as it has a significant effect upon those employed within the organisation (Lake, 2007). By improving the nursing environment and promoting respectful relationships among healthcare providers, empowering nurses to make decisions, investing in nursing leadership to extend career trajectories, developing standards of practice that foster quality of care (Kazanjian et al., 2005, p. 115) healthcare organisations can potentially produce benefits to the health system such as better job satisfaction. Other potential benefits to the system include prevention of burnout and improved workforce retention (Kazanjian et al., 2005; Oulton, 2006). In summary, nursing leaders and healthcare organisations may benefit from identifying service gaps and issues related to the requisite features of the nursing practice environment as found in this current study. The information from this current study would greatly assist with informing the development of strategies to enhance healthy environments in the Australian context and may improve job satisfaction, a key feature that influences retention of nurses. 6.4 Mediational Analysis It was hypothesised that caring efficacy would mediate the relationship between the socio-demographic variables and job satisfaction and between the personal control variables (general self-efficacy, work locus of control and ways of coping) and job satisfaction. As none of the subscales of the ways of coping measure could be shown 123

140 to significantly contribute to the structural model it was subsequently removed from the model It was also hypothesised that caring efficacy would mediate the relationship between the nursing practice environment and job satisfaction. A mediation model was examined based on the structural regression model to consider the role of other variables in the relationships to job satisfaction. Further, when it was assumed caring efficacy was not a mediator, for every unit increase (i.e. increase of 1) we expected job satisfaction to increase by When caring efficacy was accounted for in the model, for every unit increase in general self-efficacy it was expected job satisfaction to increase by Baron and Kenny (1986) affirmed that as a mediator effect becomes larger, the direct relationship between the independent and outcome variables is reduced. Full mediation occurs when the relationship between the independent and the outcome variables is zero. The results of study 2 showed that general self-efficacy contributed to job satisfaction through caring efficacy. Although no previously published studies have examined the relationship between general self-efficacy and caring efficacy in relation to nursing practice, general self-efficacy may precede and even contribute to caring efficacy as this present study has shown. Organisations that support staff by providing the motivation, cognitive resources and courses of action necessary to be in command of one s work may have an effect upon nurses ability to show a caring orientation towards their patients (Coates, 1997). It is believed that people with high levels of self-efficacy are more likely to persist in difficult circumstances such as work environments, in order to achieve the outcomes they identify as important. This may result in better job satisfaction (Gist & Mitchell, 1992). Efforts to increase levels of self-efficacy in nursing practice are needed and should be supported by the theoretical determinants of self-efficacy (Manojlovich, 2005a); performance accomplishments (the feeling of success), vicarious experience (observing the successful achievement of others), verbal persuasion (verbally convincing people they can succeed) and physiological information (the physical and emotional capacity of someone s ability) (Bandura, 1977). 124

141 The results also showed that no mediation effect existed between caring efficacy and any of the continuous socio-demographic variables (age, years of experience, years in current job and education levels) to job satisfation. There was also no mediation effect found between caring efficacy and work locus of control to job satisfaction. Addtionally, no mediation effect was found to exist between caring efficacy and the organisational variable practice environment. No studies were found that have discussed any mediation effect between caring efficacy and socio-demographic variables to job satisfaction, caring efficacy, work locus of control and job satisfaction and caring efficacy and general self-efficacy and job satisfaction. Again, no studies reporting the mediation effect between caring efficacy, practice environment and job satisfaction were found. This is the first study to examine the mediation effects of caring efficacy upon the variables general self-efficacy, work locus of control and the practice environment, and job satisfaction. In summary, the impact that the personal control variables (general self-efficacy and work locus of control) have on job satisfaction have been well reported in the literature (Judge & Bono, 2001); however, the specific investigation of the interaction effect of socio-demographics, the personal control variables (general selfefficacy and work locus of control) and caring efficacy and job satisfaction in an Australian population of registered nurses is a very important contribution to clinical practice. The identification of the relationships that these socio-demographic and personal control variables have upon caring efficacy and job satisfaction contributes to the recognition of who is susceptible to having reduced levels of job satisfaction and caring efficacy and an understanding of what factors may contribute to improving these outcomes. Caring efficacy mediated the relationship between general self-efficacy and job satisfaction which suggests general self-efficacy may precede and even contribute to caring efficacy. Additionally, knowledge of these interactions indicate the urgent need for the development of strategies and professional development programmes that enhance self-efficacy and job satisfaction in registered nurses in Australia. 125

142 6.5 Summary This chapter has sought to discuss and emphasise the results and the findings in answer to the research questions. Identification of the personal control variables (general self-efficacy and work locus of control) and the organisational variable (practice environment) that may positively affect caring efficacy and job satisfaction have proven to be very important. The theoretical model developed provides a better understanding of how personal control variables (general self-efficacy and work locus of control) and the organisational variable (practice environment) have an influence upon caring efficacy and job satisfaction in a population of Australian registered nurses. This model also serves as useful pointers to nursing leaders and healthcare organisations to identify service gaps and issues related to the Australian nursing population, the practice environment and job satisfaction. Further, this research has shown the development of strategies that may enhance nurses selfefficacy, work locus of control and the professional nursing practice environment are imperative, as these may influence nurses caring orientation to their patients and job satisfaction. This model would support the development of strategies for professional development programmes that improve self-efficacy and work locus of control for nurses. These strategies would be developed according to the nurses age and years of experience. This model also provides valuable information which aids in determining the focus of interventions that may enhance the elements of the nursing practice environment which were further shown to have a major effect on job satisfaction. The development of these strategies may improve patient outcomes, along with the recruitment and retention of nurses at a time when there is a shortage and a pseudo shortage of registered nurses in Australia. The personal values and attitudes of nurses working in organisations provide support for the nurse to undertake the complex task of providing patient care. By having a belief in self (selfefficacy) the nurse has the capability to develop caring relations with their patients (Coates, 1997), find sense in the work they do (Fagermeon, 1997), and accomplish job satisfaction. 126

143 Chapter 7: Conclusion 7.1 Introduction The purpose of this study was to examine the relationships between sociodemographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), personal control variables (general selfefficacy, work locus of control and coping styles), the organisational outcome (the nursing practice environment) and the outcome variables (caring efficacy and job satisfaction) in Australian registered nurses. This study used a two-phase cross-sectional survey design. First, a pilot study was conducted to determine the validity and reliability of the survey instruments and to assess the effectiveness of the participant recruitment process. The second part of the research involved investigating the relationships between the above-mentioned socio-demographic variables, personal control variables, the organisational outcome and the outcome variables in a cohort of Australian registered nurses. To achieve this purpose, a number of research questions were developed. These were clearly identified and presented in chapter 1. This chapter summarises the pilot study (study 1) and discusses the key findings of the main study (study 2). A discussion of the strengths and limitations to the main study are then presented, followed by the implications for theorisation, clinical practice and further research. Recommendations for further work are then proposed. 7.2 Summary of Study 1 The format, wording and length of the questionnaire were found to be acceptable by the study population, and the respondents were able to follow the instructions correctly. However, owing to the low response rate (n = 24), the cover page was 127

144 redesigned to ensure ease of reading. Readability and acceptability of the questionnaire were indicated by the participants responses to the questionnaire. Validity testing was undertaken for all the instruments in this study. All of the questions in this study were assessed for feasibility. Comments, responses and missing data from the respondents highlighted the problems with some of the questions of the Nursing Work Index Revised. Thus, on the basis of a literature review of practice environment measures developed in Australia, the Nursing Work Index Revised was replaced with the Brisbane Practice Environment Measure (Webster et al., 2009) for study 2. The descriptive responses of this study were considered to be representative of those of the population of the main study. However, while the majority of the respondents had a bachelor s degree in the main study, the majority of registered nurses reported having a certificate alone in study 1. The randomisation and the mailing process were efficient. The test-retest reliability of each measure was examined by means of a Bland-Altman plot (Bland & Altman, 1986) and showed good test-retest reliability of the scales. Owing to the small sample size, only face validity testing was conducted. 7.3 Key Findings of Study 2 The study also determined relationships that exist among socio-demographic variables, work locus of control, coping, the professional nursing practice environment and caring efficacy and job satisfaction levels. Associations between socio-demographic variables and caring efficacy and socio-demographic variables and job satisfaction were also examined. A Confirmatory Factor Analysis was also undertaken to determine the reliability and validity of Brisbane Practice Environment Measure in a diverse group of Australian registered nurses. 128

145 7.4 Description of the Sample From a sample of 2000 registered nurses, 639 nurses (31.9% of the individuals surveyed) participated in this survey. This was deemed an adequate sample size for this study Personal Characteristics of the Sample In this study, the majority of respondents were women (91.9%). This was consistent with the report of the Nursing and Midwifery Labour Force Statistics in 2008 (AIHW, 2010). Most of the nurses were married (65.5%) and the majority were employed on permanent, part-time or casual basis. The part time findings in this current study are consistent with the above report. Most of the respondents were more than 40 years of age, and the age range was 20 to 76 years which is also consistent with the 2008 report. The following specialty areas were identified: medical/surgical; intensive care/operating room/department emergency medicine; paediatrics; aged and community care; mental health; support services and midwifery. Nurses working in medical/surgical nursing represented the single largest group of respondents (31.9%). This result was also consistent with the report from the Nursing and Midwifery Labour Force Statistics in 2008 (AIHW, 2008) Organisational Characteristics Four states of Australia were included in this study. The participants were from Queensland (28.4%), New South Wales (41.2%), Victoria (24.0%) and South Australia (6.5%). From the four sectors (private, public, residential and aged care, and community/domiciliary) the majority of registered nurses in this sample worked in the public sector (61.5%), whereas residential and aged care had the least number of nurses (12.5%). These results are consistent with those reported in the Nursing and Midwifery Labour Force Statistics in 2008 (AIHW, 2008). Of the three geographical locations (metropolitan, provincial/regional and rural remote) more than half of the respondents worked in metropolitan locations (59.2%) and rural and remote areas accounted for the least number of nurses (15.7%). The AIHW report (2008), 129

146 discussed in the previous paragraph, did not represent all the characteristics presented in this current study Research Question 1: What Is the Level of Caring Efficacy and Job Satisfaction of the Registered Nurse Population in Australia? High levels of caring efficacy and job satisfaction were found in this Australian sample of registered nurses. One hundred per cent (100 %) of nurses had high caring efficacy scores and 80.8 % showed high job satisfaction scores Research Question 2: Do Levels of Caring Efficacy and Job Satisfaction among the Registered Nurse Population in Australia Vary According to Age, Gender, Marital Status, Level of Education, Years of Experience, Years in Current Job, Employment Status, Geographical Location, Specialty Area, Health Sector and Australian State? Associations between Caring Efficacy Levels and Organisational Characteristics and Job Satisfaction Levels and Organisational Characteristics Levels of caring efficacy and job satisfaction were highest in nurses from South Australia and lowest for nurses from New South Wales. Nurses working in rural and remote areas had higher levels of caring efficacy than those in other locations, but this group showed the lowest levels of job satisfaction. In contrast, nurses working in the metropolitan areas reported higher levels of job satisfaction and the lowest levels of perceived caring efficacy. Registered nurses who worked in the community/domiciliary sector had the highest levels of perceived caring efficacy, and nurses who worked in the private sector had the highest levels of job satisfaction. Those working in the public sector had the lowest levels of perceived caring efficacy and job satisfaction. 130

147 Associations between Caring Efficacy Levels and Personal Categorical Characteristics and Job Satisfaction Levels and Personal Categorical Characteristics Caring efficacy and job satisfaction levels by gender showed that both of these outcomes were highest among the female population in this study. Caring efficacy levels were highest in nurses who were widowed or divorced and lowest in those who reported they were single. The highest levels of job satisfaction were reported by nurses who were married and lowest by those who were widowed or divorced. Associations between caring efficacy and education qualifications showed those with the certificate qualification had the highest levels of caring efficacy, while those with a diploma qualification had the lowest levels. Nurses who had the highest levels of job satisfaction were those with Master s degree or PhD level qualifications. Nurses with the lowest levels of job satisfaction were those with a certificate level qualification. In relation to job status, nurses who have permanent work arrangements such as fulltime or part-time work, have the highest levels of caring efficacy. They also showed higher levels of job satisfaction in contrast to casual employees in this study. With regard to specialty area, it was found nurses who worked in midwifery had the highest levels of perceived caring efficacy, while nurses working in mental health had the lowest levels. The levels of job satisfaction were highest in those nurses working in medical or surgical nursing areas, while nurses working in paediatrics showed the lowest levels. 131

148 7.4.5 Research Question 3: Do the Variables Age, Sex, Marital Status, Level of Education, Years of Experience, Years in Current Job, Employment Status, Geographical Location, Specialty Area, Health Sector and Australian State Predict Caring Efficacy and Job Satisfaction? Associations between Caring Efficacy Levels and the Continuous Socio- Demographic Variables Correlation analysis showed that age, years of experience and years in current job were all strongly positive. Age and years of experience were also highly correlated with caring efficacy, and the latter variable was therefore excluded from subsequent attempts to develop the caring efficacy model. Those nurses who were older and more experienced showed higher levels of caring efficacy Associations between Job Satisfaction Levels and Continuous Socio- Demographic Variables No significant associations were found between age, years of experience and years in current job and job satisfaction Associations between Caring Efficacy Levels and Job Satisfaction Levels. Caring efficacy and job satisfaction were found to be positively correlated with each other, which suggested that high levels of perceived caring efficacy may predict better job satisfaction in this sample of nurses. 132

149 Associations between Caring Efficacy and Job Status, Geographical Location, Education Qualifications, Marital Status, Specialty Area, Health Sector, Gender, State, Age, Years of Experience and Years in Current Job Following an ANOVA Caring efficacy had a significant positive association with the widowed or divorced status of nurses. They also showed that perceived caring efficacy had a significant positive association with working in the community health sector in this sample. Age explained the caring efficacy levels and job satisfaction levels in Australian registered nurses. This implied that the older nurses in this study had higher levels of caring efficacy and job satisfaction. The age range in this study was years Multivariable Analysis (GLM) of Significantly Associated Socio- Demographic Variables and Caring Efficacy The multivariable analysis with caring efficacy and marital status, health sector and age revealed that age was the only variable that remained significantly related to caring efficacy Associations between Job Satisfaction and Employment Status, Geographical Location, Education Level, Marital Status, Specialty Area, Health Sector, Gender, State, Age, Years of Experience and Years in Current Job Following an ANOVA Significant positive associations were found between working in the specialty area of midwifery and job satisfaction. Significant positive associations were also found between working in the private sector and job satisfaction and between working in the state of South Australia and job satisfaction Multivariable Analysis of Significantly Associated Socio-Demographic Variables and Job Satisfaction. The results of a multivariable analysis of job satisfaction, specialty area and health sector showed they were significantly associated with job satisfaction. 133

150 7.4.6 Is the Brisbane Practice Environment Measure a Valid and Reliable Instrument for Use in an Australian Population of Registered Nurses? The results showed that only 28 of the original 33 items included in the Brisbane Practice Environment Measure were instrumental in describing the practice environment subscales for this Australian registered nurse population. Items (2) Performance and appraisal is completed in this area, (4) I feel respected in the way people speak to me, (7) It is difficult to influence change in this area, (8) There is a great team spirit in my work area and (29) I enjoy coming to work, of the Brisbane Practice Environment Measure were excluded from further consideration. The results showed that the 28 items selected from the original 33 items of the Brisbane Practice Environment Measure were valid for determining the effects of the practice environment on this population of Australian registered nurses. Organisations and nursing leaders could regularly evaluate the practice environment in order to improve job satisfaction and reduce burnout and job strain among nurses (Manojlovich, 2005a) and promote the retention of nursing staff Research Question 4: Does a High Level of Caring Efficacy among Registered Nurses Enhance Relationships between General Self-Efficacy, Work Locus of Control, Coping Styles, the Professional Nursing Practice Environment and Job Satisfaction Levels? Structural Equation Modelling Three of the continuous socio-demographic variables (i.e. all except age) were not associated with the personal control variables (general self-efficacy, work locus of control and coping) and the organisational variable (practice environment). Age was shown to be associated with all four of the personal control and organisational variables; however, when all four socio-demographic variables (age, years of experience, years in current job and education qualifications) were excluded, the fit of the model improved substantially. These four variables were therefore excluded from further model testing. The ways of coping variable was not found to be relevant to the structural model and was therefore removed from the model. 134

151 Correlation Analysis Showing Associations among the Scales of General Self- Efficacy, Practice Environment and Work Locus of Control. The results for the correlation components of the model showed that general selfefficacy was positively associated with practice environment and negatively associated with work locus of control. Likewise, work locus of control was found to be negatively associated with practice environment. Internal consistency of the measures was determined for each of the subscales of the Job Satisfaction Survey, Ways of Coping Questionnaire, and the Brisbane Practice Environment Measure as well as the overall scores of the Job Satisfaction Survey, the Brisbane Practice Environment Measure, General Self-Efficacy Scale, Work Locus of Control Scale and the Caring Efficacy Scale by calculating the Chronbach s alpha coefficient Potential Causal Pathways and Caring Efficacy Further testing of the model showed that caring efficacy was explained by both general self-efficacy, which was positively related, and work locus of control, which was negatively related. Practice environment was not found to be significantly related to caring efficacy Potential Causal Pathways and Job Satisfaction Practice environment explained job satisfaction (i.e. as the practice environment scores increased, so did the job satisfaction scores). General self-efficacy was negatively related to job satisfaction, and work locus of control was negatively related to job satisfaction Caring Efficacy and Job Satisfaction Modelling The results of Structural Equation Modelling found that the model fit the data well. All pathways were found to be significant except the one from practice environment to caring efficacy. The results showed that 24 per cent of the variation in caring 135

152 efficacy could be attributed to general self-efficacy, work locus of control and practice environment. Further, 62 per cent of the variation in job satisfaction can be accounted for by general self-efficacy, work locus of control and practice environment. When all Ways of Coping domains were included in a single model, it resulted in an over-specified model, so the Ways of Coping measure was not found to contribute to the structural model and was therefore subsequently removed from the model Research Question 5: Can Nursing Self-Efficacy Mediate a Relationship among the Factors of Work Locus of Control, General Self Efficacy, Professional Nursing Practice Environment, Predetermined Socio-Demographic Variables and Job Satisfaction? Caring efficacy mediated the relationship between general self-efficacy and job satisfaction. Further, caring efficacy did not mediate the relationship between any of the socio-demographic variables and job satisfaction. Caring efficacy was not found to mediate the relationship between work locus of control and job satisfaction or practice environment and job satisfaction. In summary, the relationship between job satisfaction and caring efficacy may depend on a number of socio-demographic, personal control and organisational issues as shown in this study. This study provides important information for nursing leaders, as it contributes to a better understanding of the relationships that exist between the socio-demographic variables, personal control and organisational variables described and used in this study, which may contribute to more confident and skilled nursing practices and improved job satisfaction. 7.5 Strengths of Study 2 This study has a number of strengths that add to the literature on the nature of nursing and will benefit future studies. This is the first study that has extensively investigated the relationships among several socio-demographic, personal control (general self-efficacy, work locus of control an coping styles) and organisational 136

153 predictors (practice environment) and caring efficacy and job satisfaction as outcomes in a sample of Australian registered nurses by using psychometrically validated instruments. The literature on self-efficacy has demonstrated that people who perceive that they have high levels of efficacy believe they are able to get the job done. In contrast a lack of self-efficacy in nursing, that is, nurses believing they do not have the ability to motivate and decide on courses of action nor the cognitive resources required to have control over their work, may have an effect on nurses professional practice behaviours (Begat et al., 2005; Manojlovich, 2005a). Further, high levels of selfefficacy in nurses improve collaborative practice and also have a positive effect on team commitment (Le Blanc et al., 2010; Lee & Ko, 2010). However, no previous models have tested the associations between socio-demographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), personal control variables (general self-efficacy, work locus of control and coping styles) and the organisational variable practice environment and caring efficacy and job satisfaction as outcomes among registered nurses in the Australian context. Therefore, the findings of this study provide valuable information for the development of strategies for professional development programmes, which will enhance the levels of efficacy (mastery experiences, vicarious experience [observing others], verbal persuasion and physiological information [self-evaluation of physiological and emotional states]) (Bandura, 1977) among registered nurses. Improving self-efficacy in nurses may also improve nurses confidence and ability to develop caring relationships with their patients and may also improve their work locus of control. Further, the findings of this study showed 62 % of the variation in job satisfaction could be accounted for by general self-efficacy, work locus of control and practice environment. Strategies which enhance levels of self-efficacy and work locus of control may also improve job satisfaction in registered nurses in Australia. The development of strategies that incorporate the improvement of the practice environment as shown in this study (i.e., professional development, management support, rostering, out of depth and workloads) may improve job satisfaction levels 137

154 in Australian registered nurses and effect retention of nurses. The characteristics of job satisfaction that may be affected by the further development of these strategies include, Pay, Promotion, Supervision, Fringe Benefits, Contingent Rewards, Operating Conditions, Co-Workers, Nature of Work and Communication. Second, this is the first study to conceptualise a nursing self-efficacy model among registered nurses in Australia by using the Structural Equation Modelling technique. This study revealed that the Structural Equation Modelling approach was useful to gain a better understanding of the relationships that exist between socio-demographic variables, personal control and organisational variables and caring efficacy and job satisfaction among Australian registered nurses. Structural Equation Modelling has many advantages over correlational and regression analyses. It is a comprehensive approach for hypotheses testing with regard to the relationships between the variables in this study (Hoyle, 1995). It also tested the theoretical framework of the relationships between the variables (Rigdon, 1998) and the directional and nondirectional relationships between the variables (MacCallum & Austin, 2000) in this study. Finally, a rigorous research procedure was used to examine the reliability and validity of the Brisbane Practice Environment Measure. Psychometric testing of this instrument involved evaluation of construct, content, criterion and convergent validity along with internal consistency and reliability. This process will allow reproduction in other studies and correlation with previous findings. 7.6 Limitations of Study 2 Study 2 provided some important findings in the areas of socio-demographic variables, personal control and organisational variables and the outcome variables caring efficacy and job satisfaction; however, some limitations should be considered. This study had a cross-sectional survey design, which is a limitation in that no assumptions about causality can be made, and no assessment related to the impact of general self-efficacy or work locus of control and practice environment on caring 138

155 efficacy and job satisfaction can be made. Further, a longitudinal study would be recommended to better determine if the relationships between the variables in the present study persist over time. Therefore, future studies should employ a longitudinal design to evaluate the theoretical model suggested in these findings. It is acknowledged that the study sample is limited as it only included registered nurses who are members of an Australian industrial and professional organisation. It may be argued that registered nurses who choose to belong to this type of industrial and professional organisation may be different to those who do not. This sample was therefore representative of these nurses only, and future studies should investigate whether relationships exist between the socio-demographic factors, caring efficacy and job satisfaction in other samples of Australian registered nurses. Another potential limitation is the use of self-reported measures. Study participants were requested to self-report information on the measures assessed in this study; this may result in a social desirability response bias. Social desirability is described as the tendency of participants to rate their own behaviours on the basis of what they assume is consistent with social norms and expectations when completing selfreported measures (Jago, Baranowski, Cullen & Thompson, 2007). Social desirability can considerably affect self-reported health behaviours such as selfefficacy (Manojlovich, 2005a). This could be an issue in relation to general selfefficacy, work locus of control and caring efficacy in the present study. In order to minimise the social desirability bias, participants on recruitment were assured confidentiality and anonymity with regard to their answers with the provision of clear instructions on the questionnaire, as suggested by Klesges et al. (2004). Despite these limitations, the study findings offer a significant starting point for formulating research questions that will help future studies better explain the relevance of socio-demographic variables, personal control variables and organisational variables and caring efficacy and job satisfaction in registered nurses in Australia. Additionally the findings will inform the design of organisational interventions and educational approaches to improve caring efficacy and job satisfaction among registered nurses in Australia. 139

156 7.7 Implications This study provides an initial framework for exploring the relationships among socio-demographic variables, personal control and organisational variables and the outcomes caring efficacy and job satisfaction. It also adds to the nursing literature related to caring efficacy and job satisfaction in Australian registered nurses concerning socio-demographic variables, general self-efficacy, work locus of control and the nursing practice environment. The research findings provide a number of implications that can be used for theorisation, future practice and research Implications for Theorisation In this study, an initial theoretical model was developed using the Structural Equation Modelling technique to explore the relationships among sociodemographic, personal control and organisational variables and the outcomes caring efficacy and job satisfaction in Australian registered nurses. Self-efficacy in nursing is the ability to express a caring orientation and to develop caring relationships with patients (Coates, 1997). Previous studies have shown that self-efficacy in nursing may signify improved professional nursing practice behaviours (Pascale et al., 2010; Lee & Ko, 2010; Manojlovich, 2005a) and job satisfaction (Judge & Bono, 2001). Further, a lack of self-efficacy in nursing, with nurses perceiving that they lack the personal resources to take control of their work, may have an effect on patient care, and hence, nurses may remove themselves from challenging work situations to reduce stress (Begat et al., 2005; Manojlovich, 2005a). The results indicated that the personal control variable, general self-efficacy, had a significant negative association with the personal control variable work locus of control. This is supported by the literature (Judge et al., 1998), that those who have higher levels of perceived self-efficacy have an internal locus of control. General self-efficacy also had a significant positive association with caring efficacy in the current study. This means that the caring efficacy levels were high in nurses who had high levels of general self-efficacy. The results of this study showed nurses with high 140

157 levels of self-efficacy may also have an internal locus of control. This may lead to beliefs that nurses in this study are confident they can convey a caring orientation and build caring relationships with patients. Further, nurses who have high levels of perceived general self-efficacy showed positive attitudes towards the practice environment. This is important information for organisational and nursing leaders, in that high levels of perceived general self-efficacy among nurses may influence their caring efficacy and result in them having positive views of the practice environment. Healthcare organisations that create and provide the necessary resources that influence self-efficacy in nursing could ultimately have an effect upon the development of caring relationships and caring behaviours to the patients in these organisations and result in higher levels of job satisfaction. Nurses who are more satisfied with their work environments are more likely to stay in a job. Organisations that are able to retain knowledgeable and experienced nurses may result in improved patient and organisational outcomes. Further, work locus of control was found to have a significant negative association with practice environment and caring efficacy. The literature describes locus of control as a person s beliefs that he or she has the ability to control his or her environment. Thus, individuals with high levels of self-efficacy (confidence with respect to actions or behaviours) also have the perception that they have control over their environment (locus of control) (Bono & Judge, 2003). Therefore, nurses with an internal locus of control also believe they can be more involved in the organisation and are more likely to identify ways of dealing with obstacles in the workplace (Bandura, 1997). These internals may therefore be more likely to show better caring relationships with their patients. In contrast, externals are more likely to experience constant failure and detach themselves from the circumstances of failure which in turn may lead to them leaving the organisation (Bandura, 1997). This, too, is important information for those organisations and nursing leaders to develop strategies for the development of professional development and orientation programmes that support the improvement of selfefficacy in nurses, as this may lead to the retention of nurses. The relationship between caring efficacy and the practice environment was not confirmed in this study. However, it may be that self-efficacy in nursing is an important factor which 141

158 may enable nurses have a positive view of the work environment. This is because of the positive relationship found between general self-efficacy and caring efficacy and the positive association between general self-efficacy and the practice environment. Alternatively, some environments do not provide the necessary resources or necessary opportunities for nurses with high levels of general self-efficacy to take control of their work. Therefore the necessary confidence and ability to develop a caring relationship and a caring orientation towards their patients is hence affected (Manojlovich, 2005a). The results were unable to confirm the relationship between general self-efficacy and job satisfaction. However, this was possibly a result of the effect of the large sample size rather than a negative association between job satisfaction and general selfefficacy. Other studies have found that high levels of self-efficacy in individuals are positively related to higher levels of job satisfaction and job performance (Bono & Judge, 2003; Judge & Bono, 2001; Judge et al., 1998) The results of this current study showed that the practice environment explained job satisfaction. Higher levels of job satisfaction were noted among nurses who perceived that the practice environment provided the necessary resources of professional development, management support, equity of rosters and flexibility of roster development; safety for nurses in their practice; and equal and reasonable workloads with a satisfactory skill in the work area. This is supported by Aiken et al. (2008), who confirmed that better care environments were associated with better patient outcomes (i.e. lowered mortality rates), while poorer care environments led to greater job dissatisfaction among nurses and increased intention to leave. The results indicated that work locus of control was negatively related to job satisfaction. This implies that those with an internal locus of control (internals) had higher levels of job satisfaction, which is supported conceptually by the literature showing that positive individuals will make the effort to work in jobs that are more challenging (Judge et al., 1998). The perception is that more complex jobs are essentially more fulfilling. In this study by Judge et al. locus of control was shown to be highly correlated with self-efficacy. Hence people who judge themselves as important are able to cope with difficult situations in the workplace and are more 142

159 positive when they are challenged by the situations they encounter. Those who do not see themselves as important (externals) usually are negative towards the same challenging situations (Judge et al., 1998). The Structural Equation Modelling results showed that 24% of the variation in caring efficacy and 62 % of the variation in job satisfaction was accounted for by the model. This is very important information for nursing and organisational leaders to identify, develop and implement strategies that focus on the personal control variables (general self-efficacy and work locus of control) and the organisational variable (practice environment) to improve nursing outcomes such as the ability to express caring relationships and caring behaviours with their patients. Further, improved selfefficacy in nursing and job satisfaction may result in improved organisational (retention) and patient outcomes (Manojlovich, 2005 b). The results showed that caring efficacy mediated the relationship between general self-efficacy and job satisfaction (i.e. general self-efficacy contributed to job satisfaction through caring efficacy). The relationship between general self-efficacy and caring efficacy has not been previously examined, but general self-efficacy may precede and even contribute to caring efficacy, as this present study has shown. This theoretical model therefore provides evidence for nursing leaders and healthcare organisations for the development of professional development programmes that aim to increase levels of self-efficacy in nursing practice. These programmes need to be supported by the theoretical elements of self-efficacy (Manojlovich, 2005b). In summary the identification and development of strategies that further develop and improve personal control and organisational variables are integral components of improving caring efficacy and job satisfaction in registered nurses in Australia. This theoretical model provides an understanding of how personal control and organisational variables collaborate to influence the elements of self-efficacy and job satisfaction in nursing. This model also provides useful and important information for assisting nursing leaders and healthcare organisations determine the focus of professional development and orientation programmes that may increase the level of self-efficacy and job satisfaction in nurses. 143

160 7.7.2 Implications for Practice The study has some implications for clinical practice. First, age, years of experience as a registered nurse and length of time in current job were all strongly positively correlated with each other. Age and years of experience were also highly correlated with caring efficacy. This suggests that the nurses in this sample who were older and more experienced showed higher caring efficacy levels. The results of the multivariate analysis found significant relationships between age and caring efficacy. This has implications for nursing leaders who wish to develop appropriate orientation and professional development programmes for registered nurses. New nurses in a specific healthcare setting and those nurses who may not have the required years of experience may not have the confidence or ability to develop a caring orientation towards patients or for developing a caring relationship with patients. This suggests that orientation and professional development programmes should incorporate experiences of mastery, role modelling, positive self-appraisal and verbal support, which may improve caring efficacy in inexperienced and young registered nurses. The literature supports the encouragement of self-efficacy in nursing staff as it may provide nursing leaders with a valuable means for enhancing nursing practice at all levels. Second, the findings showed that job satisfaction and caring efficacy were highly correlated in this current study. This implies that high levels of caring efficacy led to high levels of job satisfaction in this sample of Australian registered nurses. Selfefficacy in nursing influences nursing performance, which is able to further influence organisational outcomes. High levels of self-efficacy in nursing enable nurses to achieve the desired results and overcome difficult challenges to achieve the outcomes they value (Manojlovich, 2005b). Nursing leaders and policymakers need to ensure that the strategies used for the development of professional development programmes, which are aimed at improving self-efficacy in registered nurses at all levels, are implemented. In turn, this may improve job satisfaction and may affect recruitment and retention of nurses. Finally, the findings of the model showed that the personal control variables (general self-efficacy and work locus of control) and the organisational factor (practice 144

161 environment) were related to caring efficacy and job satisfaction. It was found that 24% of the variation in caring efficacy could be accounted for by general selfefficacy, work locus of control and practice environment. Furthermore, 62% of the variation in job satisfaction could be accounted for by general self-efficacy, work locus of control and practice environment. The findings also showed that caring efficacy mediated the relationship between general self-efficacy and job satisfaction. Thus, there may be potential benefits in emphasising support for improving general self-efficacy among registered nurses to promote better caring efficacy and job satisfaction. These aspects should be taken into account when planning individualised and group training programmes. Organisations could survey their staff using measurements such as the Brisbane Practice Environment Measure to improve levels of job satisfaction and subsequent retention of registered nurses. Findings also showed a strong relationship between the nursing practice environment and levels of job satisfaction. Nursing, healthcare and organisational leaders and policymakers should recognise the importance of supporting the development and provision of resources associated with the environmental factors (professional development, management support, out of depth, workloads and rostering) identified in this study, which may improve job satisfaction in Australian registered nurses. This may in turn affect retention of nurses to stay in organisations that provide these necessary resources and support. Recruitment may also improve as nurses may also be attracted to these organisations. To further support this, healthcare organisations and nursing leaders in Australia should regularly use instruments that measure aspects of the nursing practice environment as shown in this study such as: professional development, management support, out of depth, workloads and rostering, in order to improve job satisfaction and possibly maintain experienced staffing levels. These environments may also improve the recruitment of nurses. To assist nurses in dealing with the ever-changing healthcare environment, nursing leaders and healthcare organisations should be aware of the different factors that affect self-efficacy and job satisfaction in nursing, as shown in this study. According to Manojlovich (2005b), nursing leaders should endeavour to foster and implement the theoretical elements of self-efficacy as described by Bandura (1977) to support nurses in their practice. An understanding of the characteristics of self-efficacy, 145

162 namely, performance accomplishments (personal mastery experiences), vicarious experience (role modelling), verbal persuasion (convincing people that they can be successful) and physiological information (self-evaluation of physiological and emotional states) (Bandura, 1977) can enable nursing leaders to develop professional development programmes aimed at enhancing self-efficacy for all levels of nurses. Hence, incorporation of a more eclectic approach such as a routine individual assessment of self-efficacy into professional development and orientation programmes rather than using a one-size-fits-all approach (Manojlovich, 2005a) to nursing education may provide useful information on the different types of resources required to enhance these programmes for nurses. Additionally, in this theoretical model, the high levels of general self-efficacy may improve work locus of control in nurses and their interaction with the practice environment. Thus, interventions that promote the development of confidence and abilities of registered nurses to have control over their work outcomes may enhance caring efficacy and job satisfaction in the workplace Implications for Further Research The study has several important implications for further studies. First, the covariates of age, years of experience as a registered nurse and length of time in current job were all highly positively correlated with each other. Age and years of experience were also highly correlated with caring efficacy, and they should therefore be considered in any future caring efficacy development model. Caring efficacy and job satisfaction were also positively correlated with each other, which suggest that high levels of caring efficacy may predict better job satisfaction in this sample of nurses and should be investigated further. The study found a number of significant associations between some of the sociodemographic variables, for example, marital status (widowed or divorced) and health sector (community health sector), and caring efficacy. Further, a number of significant associations were found between job satisfaction levels in nurses and the Australian states, the health sector and the specialty areas in which the nurses worked. The findings also showed that age explained the variation in caring efficacy 146

163 and job satisfaction: Older nurses in this study had higher levels of caring efficacy and job satisfaction. This current research offers the opportunity to further study relationships between the socio-demographics variables age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state of registered nurses, which have been identified as relevant to this research, and caring efficacy and job satisfaction in the Australian context. This study also provides the opportunity to further examine relationships between caring efficacy and job satisfaction in registered nurses in the Australian context. Future studies investigating job satisfaction in different populations of Australian nurses should consider further use of these variables in order to obtain a better understanding of the relationships that may exist. Finally, as this is a cross-sectional study, it was not the intent of the modelling to show causality. Although the findings and implications of the present study point to the need for the promotion of self-efficacy in new nurses in specific work areas and younger, less experienced nurses in order to enhance caring efficacy and job satisfaction, concurrent associations between the measures of general self-efficacy, work locus of control and the practice environment do not provide causal evidence for the effect of these variables on caring efficacy and job satisfaction. Prospective, longitudinal research is now required to determine the changes in the effect of these factors over time. Future studies could further investigate and explain how the factors in this model may vary in their inter-relationships in other nursing populations and cultures. The results of this study have also shown that there exists a significant need to further understand nurses confidence and ability to express caring behaviours and develop caring relationships in relation to age, personal control variables (general self-efficacy and work locus of control) and job satisfaction in nursing populations. In addition, further studies in the area of nursing management and nursing leaders are required in order to develop appropriate strategies for professional development programmes related to the enhancement of caring efficacy in nurses. 147

164 7.8 Recommendations To enhance clinical practice and extend further research, several recommendations are proposed as follows: 1. The Caring Efficacy Scale, Job Satisfaction Survey, General Self-Efficacy Scale, Work Locus of Control Scale and the Brisbane Practice Environment Measure need to be confirmed for their construct validity and internal consistency in future studies on Australian registered nurses. 2. The construct of socio-demographic factors needs to be re-evaluated and more variables need to be added or removed as appropriate. 3. There is also a vital need to investigate further the development of strategies to deal with the barriers that exist with the components of job satisfaction (pay, promotion, supervision, fringe benefits, contingent awards, operating conditions, co-workers, nature of work and communication) and the effects this has on nursing, nursing recruitment and retention, and on patient outcomes. 4. The subjects in this study were recruited from the database of a professional and industrial nursing organisation in Australia. Further studies that examine other nursing populations in Australia are required. 5. The development of future professional development and orientation programmes should include the theoretical traits of self-efficacy at all levels of nurses such as personal mastery, role modelling, verbal persuasion and physical and emotional ability. 6. Longitudinal studies are needed to investigate the effectiveness of strategies which enhance self-efficacy, work locus of control and the practice environment in relation to job satisfaction. 7. Age is an important predictor of self-efficacy. Nurses who are older and more experienced as well as having spent some time in their work areas could assist professional development programmes designed to improve self-efficacy in nurses by role modelling behaviour and verbal persuasion. 8. The concept of self-efficacy in nursing may develop further over time. Thus, a longitudinal study is needed to definitively determine the 148

165 relationship between the socio-demographic factors age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state, personal control factors, general self-efficacy and work locus of control, and the organisational factor, practice environment consisting of, professional development, management support, out of depth, workloads and rostering used in this study and the outcomes of caring efficacy and job satisfaction comprising of pay, promotion, supervision, fringe benefits, contingent awards, operating conditions, co-workers, nature of work and communication. 7.9 Summary In conclusion, this is the first study to examine the relationships between sociodemographic variables age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state, personal control factors, general selfefficacy and work locus of control, and organisational factor, the practice environment and the outcomes caring efficacy and job satisfaction by using the Structural Equation Modelling technique. Study 1 was conducted in order to determine the validity and reliability of the survey instruments and to assess the effectiveness of the participant recruitment process. Study 2 involved investigating the relationships of the socio-demographic variables (age, gender, marital status, level of education, years of experience, years in current job, employment status, geographical location, specialty area, health sector and Australian state), personal control variables (general self-efficacy, work locus of control and coping styles) and the organisational outcome (the practice environment) with the outcome variables (caring efficacy and job satisfaction) in Australian registered nurses. The issues surrounding the global nursing shortage are very complex in relation to the health care system. However nurses are also reporting that they are increasingly dissatisfied with their jobs and are leaving the profession at a time when there is an increased demand for care and an also a growing aging population. Nurses value 149

166 caring for patients and it is essential that younger and less experienced nurses are given every opportunity to develop personally and professionally using the elements of self-efficacy which may enhance their ability to develop caring relationships and express caring behaviours to their patients in order to achieve the outcomes they value. Nursing and organisational leaders therefore have a responsibility to ensure the development of strategies for professional development and orientation programmes which may enhance self-efficacy and as a result improve organisational and patient outcomes. Furthermore job satisfaction is an important criterion that enhances nursing performance, influences the intention of nurses to continue in the organization, achieve the outcomes they value in their profession, and supports the commitment of nurses to patient care. Healthcare organisations that provide the essential resources in the practice environment described in this study, may also improve job satisfaction in nurses. In addition to this, the development of selfefficacy in nurses may result in more favourable perceptions of the practice environment by nurses. This could essentially influence the recruitment and retention of nurses especially during this time of extensive reporting of a global nursing shortage. Nursing leaders need to identify factors that effect nurses job satisfaction, also those that impact upon a person s intent to stay the job in order to develop appropriate strategies that may enhance nursing retention. 150

167 References Aiken, L., Sloane, D., Clarke, S., Poghosyan, L., Cho, E., You, l., Finlayson, M., Kanai-pak, M. & Ungsuroch, A. (2011). Importance of work environments on hospital outcomes in nine countries. International Journal for Quality in Health Care, 23(4), Aiken, L., Clarke, S., Sloane, D., Lake, E. & Cheney, T. (2008). Effects of Hospital Care Environment on Patient Mortality and Nurse Outcomes. Journal of Nursing Administration. 38(5), Aiken, L., Clarke, S., & Sloane, D. (2002). Hospital staffing, organization, and quality care. International Journal for Quality Health Care, 14, Aiken L, Clarke S, Sloane D, Sochalski J, Silber J. (2002). Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. Journal of the American Medical Association 288: Aiken, L., Clarke, S., Sloane, D. and Sochalski, J. (2001). Nurses' reports on hospital care in five countries. Health Affairs 20(3), Aiken, L. H., Clarke, S. P. & Sloane, D. M. (2000). Hospital restructuring: Does it adversely affect care and outcomes? Journal of Nursing Administration, 30(10), Aiken, L., & Patrician, P. (2000). Measuring organizational traits of hospitals: The revised nursing work index. Nursing Research, 49(3), Aiken L. and. Sloane, D. (1997) Effects of Organizational Innovations in AIDS Care on Burnout among Hospital Nurses. Work and Occupations 24(4), Aiken, L., Smith, H., Lake E (1994). Lower Medicare mortality among a set of hospitals known for good nursing care. Medical Care 32,

168 American Association of Colleges of Nursing. Hallmarks of the professional nursing practice environment. American Association of Colleges of Nursing Web site. Published January Armstrong, K., & Laschinger, H. (2006). Structural empowerment, Magnet hospital characteristics and patient safety culture: Making the link. Journal of Nursing Care Quality, 21(2), Australian Bureau of Statistics Minor Group 254 Midwifery and Nursing Professionals ANZSCO- Australian and New Zealand Standard Classification of Occupations, First Edition, Last updated June F?opendocument. Australian Institute of Health and Welfare (2010). Nursing and midwifery labour force National health labour force series no. 43. Cat. no. HWL 44. Canberra: AIHW. Australian Nursing and Midwifery Council (2005). y_standards_rn.pdf Alexander J. Nursing Unit Organization: Its Effects on Staff Professionalism. Vol 4. Ann Arbor, Michigan: UMI Research Press; AL-Hussami, M. (2008). A Study of Nurses' Job Satisfaction: The Relationship to Organizational Commitment, Perceived Organizational Support, Transactional Leadership, Transformational Leadership, and Level of Education. European. Journal of Scientific Research, 22(2), Alligood, M. & Tomey, A. (2010). Nursing Theorists and Their Work, Seventh Edition. Mosby Incorporated. United States of America. Amendolair, D. (2007). Caring behaviours and jobs satisfaction. Proquest Dissertation and Theses. UMI Number

169 Anderson, C. (1977). Locus of control, coping behaviors, and performance in a stress setting, a longitudinal study. Journal of Applied Psychology, 62, Anderson, J., & Gerbing, D. (1988). Structural equation modelling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), Arbuckle, J. (2006). Amos (Version 7.0) (Computer Program). Chicago: SPSS. Armstrong-Stassen, M., & Cameron, S. (2003). Nurses' job satisfaction and turnover intentions over a six-year period of hospital downsizing and amalgamation. International Journal of Public Administration, 26(14), Armstrong, K. & Laschinger, H. (2006). Structural empowerment, magnet hospital characteristics, and patient safety culture. Journal of Nursing Care Quality, 21(2), Ashkanasy, N. (1985). Rotter s internal-external scale: Confirmatory factor analysis and correlation with social desirability for alternative scale formats. Journal of Personality and Social Psychology, 48(5), Australian Bureau of Statistics. (2008) Selected Health Occupations: Australia, Retrieved from: (Viewed, 25/07/11). Australian Bureau of Statistics. (2007). Commonwealth of Australia. Retrieved from Australian Institute of Health and Welfare (AIHW) (2010). Nursing and midwifery labour force Bulletin no. 81. Cat. no. AUS 130. Canberra: AIHW. Australian Institute of Health and Welfare (AIHW). (2009). Nursing and midwifery labour force National health labour force series no. 43. (Cat. No. HWL 44). Canberra: AIHW. Australian Institute of Health and Welfare (AIHW). (2004). Rural, regional and remote health: A guide to remoteness classifications. (AIHW Cat. No. PHE 53). Canberra: AIHW. 153

170 Australian Nursing Federation (2011). Submission to the productivity commission issues paper early childhood development workforce. Retrieved from Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W.H. Freeman. Bandura, A. (1994). Self-efficacy. In V. S. Ramachandran (Ed.), Encyclopedia of human behavior (Vol. 4, pp ). New York: Academic Press (Reprinted in H. Friedman [Ed.], Encyclopedia of mental health. San Diego: Academic Press, 1998). Bandura, A. (1989). Human agency in social cognitive theory. American Psychological Association, 44(9), Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), Baron, R. and Kenny, D. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology. 51(6), Begat, I., Ellefsen, B., and Severinsson, E. (2005). Nurses satisfaction with their work environment and the outcomes of clinical nursing supervision on nurses experiences of well-being a Norwegian study. Journal of Nursing Management, 13, Benner, P. (2001). From novice to expert: Excellence and power in clinical nursing practice (Commemorative edition). Upper Saddle River, NJ, Prentice Hall Health. Bland, J., & Altman, D. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. Lancet, i,

171 Blegen, M. J. (1993). Nurses job satisfaction: A meta-analysis of related variables. Nursing Research, 42(1), Bono, J.& Judge, T. (2003). Core Self-Evaluations: A Review of the Trait and its Role in Job Satisfaction and Job Performance European Journal of Personality. 17: S5 S18. Boykin, A. & Schoenhofer, S. (1993). Nursing as caring: A model for transforming practice. New York: National League of Nursing Publications. Brady-Schwartz, D. (2005). Further Evidence on the Magnet Recognition Program: Implications for Nursing Leaders. Journal of Nursing Administration. 35 (9) Brief, A. (1998). Cited in Weiss, H. M. (2002). Deconstructing job satisfaction: Separating evaluations, beliefs and affective experiences. Human Resource Management Review, 12, Browne, T. (2006). Confirmatory factor analysis for applied research (Methodology in the social sciences). New York: The Guilford Press. Browne, M. and Cudeck, R. (1993). Alternative ways of assessing model fit. In K.A. Bollen y J.S. Long (Eds.), Testing structural equation models. Beverly Hills, CA: Sage. Brunetto, Y. and Farr-Wharton, F. (2004). A case study examining the impact of public sector nurses perception of workplace autonomy on their job satisfaction: Lessons for management. International Journal of Management & Organisational Behaviour, 8 (5), Bucknall, T., and Thomas, S. (1997). Nurses reflections on problems associated with decision making in critical care settings. Journal of Advanced Nursing, 25, Buhi, E., Goodson, P. & Neilands, T. (2007). Structural Equation Modelling: A primer for health behaviour researchers. American Journal of Health Behaviour, 31(1),

172 Bursac, Z., Gauss, C., Williams, D. & Hosmer, D. (2008). Purposeful selection of variables in logistic regression. BioMed Central, 3(17). Byrne, B. (2001). Structural equation modelling with AMOS: Basic concepts, applications and programming. Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers. Calabro, K., Mackey, T., & Williams, S. (2002). Evaluation of training designed to prevent and manage patient violence. Issues in Mental Health Nursing, 23, Cheng, E. (2001). SEM being more effective than multiple regression in parsimonious model testing for management development research. Journal of Management Development, 20(7), Clarke, S. & Aiken, L. (2006). More nursing, fewer deaths. Quality and safety in Health Care, 15(1), 2-3. Clifford, C (1997).Nursing and health care research: a skills-based introduction. 2 nd Edition. Prentice Hall Europe, Wiltshire. Coates, C. (1997). The caring efficacy scale: Nurses self-reports of caring in practice settings. Advanced Practice Nursing, 3(1), Cohen-Mansfield, J., Lipson, S., & Horton, D. (2006). Medical decision making in the nursing home: A comparison of physician and nurse perspectives. Journal of Gerontological Nursing, 32(12), Collins, B.. (1974). Four components of the Rotter internal-external scale: Belief in a difficult world, a just world, a predictable world, and a politically responsive world. Journal of Personality and Social Psychology, 29(3), Commonwealth of Australia, Canberra (2002). National review of nursing education. Department of Education Science and Training. Covington, H. (2003). Caring presence: Delineation of a concept for holistic nursing. Journal of Holistic Nursing, 21(3),

173 Cowin, L., & Jacobson, D. (2003). The nursing shortage: Part way down the slippery slope. Collegian, 10(3), Craig, A., Franklin, J., & Andrews, G. (1984). A scale to measure locus of control of behaviour. British Journal of Medical Psychology, 57, Creegan, R., Duffield, C., & Forrester, K. (2003). Casualisation of the nursing workforce in Australia: Driving forces and implications. Australian Health Review, 26(1), Demerouti E., Bakker, A., Nachreiner, F. & Schaufeli, W. (2000). A model of burnout and life satisfaction amongst nurses. Journal of Advanced Nursing. 32(2), De Milt, D., Fitzpatrick, J., Ford, E. & McNulty, R. (2010). found NP s with intent to leave their current position cited reasons such as: little control over their practice, and limited opportunities for career advancement. Journal of the American Academy of Nurse Practitioners De Von, H., Block, M., Moyle-Wright, P., Ernst, D., Hayden, S., Lazzara, D., Kostas-Polston, E. (2007). A psychometric toolbox for testing validity and reliability. Journal of Nursing Scholarship, 39(2), DeVellis, R. (2003). Scale development: Theory and application (2nd ed.). Thousand Oaks, CA: Sage. Doiron, D., & Jones, G. (2006). Nurses retention and hospital characteristics in New South Wales. The Economic Record, 82(256), Duffield, C., Roache, M., O Brien-Pallas, L., Catling-Paull, C., & King, M. (2009). Staff satisfaction and retention and the role of the nursing unit manager. Collegian, 16, Duffy, P., Shiflett, S., & Downey, R. (1977). Locus of control: Dimensionality and predictability using Likert scales. Journal of Applied Psychology, 62(2),

174 Dunn, K., Elsom, S. & Cross, W. (2007). Self-efficacy and locus of control affect management of aggression by mental health nurses. Issues in Mental Health Nursing, 28, , Ernst, M., Franco, M., Messmer, P. & Gonzalez, J. (2004). Nurses job satisfaction, stress, and recognition in a paediatric setting. Paediatric Nursing. May-June. Fagermoen, M. (1997). Professional identity: Values embedded in meaningful nursing practice. Journal of Advanced Nursing, 25(3), Faragher, E., Cass, M., & Cooper, C. (2005). The relationship between job satisfaction and health: A meta-analysis. Occupational and Environmental Medicine, 62, Field, A. (2009). Discovering statistics using SPSS (3rd ed.). California: Sage Publications Inc. Fleishman, J. A., & Fogel, B. (1994). Coping and depressive symptoms among people with AIDS. Health Psychology, 13(2), Flint, A., Farrugia, C., Courtney, M. & Webster, J (2010). Psychometric Analysis of the Brisbane Practice Environment Measure (B-PEM) Journal of Nursing Scholarship, 42(1), Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7(3), Flynn, L., Carryer, J. & Budge, C. (2005). Organisational attributes valued by hospital, home care, and district nurses in the United States and New Zealand. Journal of Nursing Scholarship, 37, Folkman, S., & Lazarus, R. S. (1988). Ways of coping questionnaire permissions set manual, test booklet, scoring key. California: Mind Garden. Forte, A. (2005). Locus of control and the moral reasoning of managers. Journal of Business Ethics, 58,

175 Gaut, D. (Ed.). (1992). The presence of caring in nursing. New York: National League of Nursing Press. Gaynor, L., Gallasch, T., Yorkston, E., Stewart, S. & Turner, C. (2007). The future nursing workforce in Australia: Baseline data for a prospective study of the profile, attrition rates and graduate outcomes in a contemporary cohort of undergraduates. Australian Journal of Advanced Nursing, 25(2), Gershon, R., Stone, P., Zeltser, M., Faucett, J., Macdavitt, K., & Chou, S. (2007). Organizational climate and nurse health outcomes in the United States: A systematic review. Industrial Health, 45, Gist, M., & Mitchell, T. (1992). Self-efficacy: A theoretical analysis of its determinants and malleability. Academy of Management Review, 17, Golbasi, Z., Kelleci, M. & Dogan, S. (2008). Relationships between coping strategies, individual characteristics and job satisfaction in a sample of hospital nurses: Cross-sectional questionnaire survey. International Journal of Nursing Studies, 45, Gray, G. & Pratt, R. (1991). Towards a discipline of nursing. Churchill Livingstone, Melbourne. Gregson, S., Zhuwau, T., Ndlovu, J., & Nyamukapa, A. (2002). Methods to reduce social desirability bias in sex surveys in low-development settings. Experience in Zimbabwe. Sexually Transmitted Diseases, 29(10), Grinspun, D. (2003). Part-time and casual nursing work: The perils of healthcare restructuring. International Journal of Sociology and Social Policy, 23(8,9), Hair, J.., Jr., Black, W., Babin, B., Anderson, R.., & Tatham, R. (2006). Multivariate data analysis. Upper Saddle River, NJ: Pearson Prentice Hall. 159

176 Hayes, L., O Brien-Pallas, L., Duffield, C., Shamian, J., Buchan, J., Hughes, F., Stone, P. (2006). Nurse turnover: A literature review. International Journal of Nursing Studies, 43, Healy, C., & McKay, M. (2000). Nursing stress: The effects of coping strategies and job satisfaction in a sample of Australian nurses. Journal of Advanced Nursing, 31(3), Hegney, D. McCarthy, A. (2000). Job satisfaction and nurses in rural Australia. Journal of Nursing Administration, 30(7/8), Hegney, D., McCarthy, A., Rogers-Clark, C. & Gorman, D. (2002). Why Nurses are attracted to rural and remote practice. The Australian Journal of Rural Health. 10(3). Hegney, D., Plank, A. & Parker, V. (2003). Workplace violence in nursing in Queensland, Australia: a self-reported study. International Journal of Nursing Practice, 9(4), Heitlinger, A. (2003). The paradoxical impact of health care restructuring in Canada on nursing as a profession. International Journal of Health Services, 33(1), Hickey, M., Owen, S., & Froman, R. (1992). Instrument development: Cardiac diet and exercise self-efficacy. Nursing Research, 41(6), Hoffart, N., & Woods, C. (1996). Elements of a nursing professional practice model. Journal of Professional Nursing, 12, Hoyle, R. (1995). The structural equation modeling approach: Basic concepts and fundamental issues. In R. H. Hoyle (Ed.). Structural equation modeling: Concepts, issues, and applications (pp. 1 15). Thousand Oaks, CA: Sage Publications, Inc. International Council of Nurses. (2010). Definition of Nursing. Retreived from [Accessed 12 December, 2011]. 160

177 International Council of Nurses. (2007). Positive Practice Environments retrieved from LW_edit_Letterhead-eng.pdf (Accessed 23/03/12). International Council of Nurses. (2003). Global issues in the supply and demand of nurses. SEW News January March. Retrieved from [Accessed on July 17, 2006]. International Council of Nurses (2006). [ / images/ stories/ documents/ publications/ GNRI/ The_Global_Nursing_Shortage-Priority Areas_for_Intervention.pdf] webcite The Global Nursing Shortage: Priority Areas for Intervention. Geneva, Switzerland. Ironson, G., Smith, P., Brannick, M., Gibson, W., & Paul, K. (1989). Construction of a job in general scale: A comparison of global, composite and specific measures. Journal of Applied Psychology, 74(2), Jago, R., Baranowski, T., Baranowski, J., Cullen, K., & Thompson, D. (2007). Social desirability is associated with some physical activity, psychosocial variables and sedentary behaviour but not self-reported physical activity among adolescent males. Health Education Research, 22(3), Jamieson, L., Williams, L., Lauder, W., & Dwyer, T. (2008). The realities of parttime nursing: A grounded theory study. Journal of Nursing Management, 16, Judge, T., & Bono, J. (2001). Relationship of core self-evaluations traits selfesteem, generalized self-efficacy, locus of control, and emotional stability with job satisfaction and job performance: A meta-analysis. Journal of Applied Psychology, 86(1), Judge, T., Locke, E., & Durham, C. (1998). The dispositional causes of job satisfaction: A core evaluations approach. Research in Organizational Behaviour, 19, Kanter, R. (1977). Men and women of the corporation. New York, Basic Books. 161

178 Kanter, R. (1979). Power failure in management circuits. Harvard Business Review, 57(4), Kaplowitz, M., Hadlock, T., & Levine, R. (2004). A comparison of web and mail survey response rates. Public Opinion Quarterly, 68(1), Kazanjian, A., Green, C., Wong, J., & Reid, R. (2005). Effect of the hospital nursing environmenton patient mortality: A systematic review. Journal of Health Services Research and Policy, 10(2), A(7). Keller, R. (1984). The role of performance and absenteeism in the predicting of turnover. Academy of Management Journal, 27, Kenny, D. (2011). Klesges, L., Baranowski, T., Beech, B., Cullen, K., Murray, D., Rochon, J., Pratt, C. (2004). Social Desirability bias in self-reported dietary, physical activity and weight concerns measures in 8 10 year old African-American girls: Results from the girls health enrichment multisite studies (GEMS). Preventive Medicine, 38, S78 S87. Kline, R. (1998). Principles and Practice of Structural Equation Modeling. Guilford Press, New York. Kramer, M. & Hafner, L. (1989). Shared values: Impact on staff nurse job satisfaction and perceived productivity. Nursing Research, 38(3), Laal, M & Aliramaie., N. (2010). Nursing and Coping With Stress. International Journal of Collaborative Research on Internal Medicine & Public Health, 2(5), Lake, E. (2002). Development of the practice environment scale of the Nursing Work Index. Research in Nursing & Health, 25, Lake, E., & Friese, C. (2006). Variations in nursing practice environments: Relation to staffing and hospital characteristics. Nursing Research, 55(1),

179 Lake, E. (2007). The nursing practice environment. Medical Care Research & Review, 64 (Suppl 2), Laschinger, H., & Leiter, M. (2006). The impact of nurse work environments on patient safety outcomes: The mediating role of burnout engagement. The Journal of Nursing Administration, 36(5), Laschinger, H., Shamian, J., & Thomson, D. (2001). Impact of magnet hospital characteristics on nurses perceptions of trust, burnout, quality of care, and work satisfaction. Nursing Economics, 19(5), Lazarus, R. & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Lazarus, R. & Folkman, S. (1987). Transactional theory and research on emotions and coping. European Journal of Personality. 1(3), Le Blanc, P., Schaufeli, W., Salanova, M., Llorens, S. & Nap, R. (2010). Efficacy befiefs predict collaborative practice among intensive care nurses. Journal of Advanced Nursing. 66(3), ). Lee, T. & Ko, Y. (2010). Effects of self-efficacy, affectivity and Collective efficacy on nursing performance of hospital nurses. Journal of Advanced Nursing. 66(4), Leininger, M. (1988). The phenomenon of caring importance, research questions, and theoretical considerations. In Qualitative Research Methods in Nursing (Leimnger M. ed.), Grune & Stratton, Orlando, Florida. Leininger, M. (1991). Culture care diversity and universality: a theory of nursing. New York: National League of Nursing Press. Lenz, E. R., & Shortridge-Baggett, L. M. (2002). Self-efficacy in Nursing. Scholarly Inquiry for Nursing Practice, 15(3), Letvak, S. (2002). Retaining the older nurse. Journal of Nursing Administration, 32(7/8),

180 Lewis, M. & Urmston, J. (2000). Flogging the dead horse: the myth of nursing empowerment? Journal of Nursing Management. 8(4): Li, J. Lambert, M. & Lambert, V. (2008), Job satisfaction among intensive care nurses from the People s Republic of China. International Nursing Review. 55, Locke, E. (1976). Cited in Brief, A. & Weiss, H. (2001). Organizational behavior: Affect in the workplace. Annual Review of Psychology, 53, Lorig, K., Stewart, A., Ritter, P., Gonzales, V., Laurent, D., & Lynch, L. (1996). Outcome measures for health education and other health care interventions. Thousand Oaks, CA: Sage Publications. Lumley, C., Stanton, P., & Bartram, T. (2004). Casualisation friend or foe? A case study investigation of two Australian hospitals. New Zealand Journal of Employment Relations, 29(2), Lu, H., While, A., & Barriball, K. (2005). Job satisfaction among nurses: A literature review. International Journal of Nursing Studies, 42(2), Macmillan English Dictionary, (2010) Macmillan Publishers. MacKinnon, D. (2008). Introduction to Statistical Mediation Analysis. New York: Erlbaum. MacKinnon, D., Fairchild, A. & Fritz, M. (2007). Mediation Analysis. Annual Review of Psychology. 58, MacKinnon, D., Lockwood, C., Hoffman, J., West, S., & Sheets, V. (2002). A comparison of methods to test the significance of the mediated effect. Psychological Methods, 7, Maibach, E. & Murphy, D. (1995). Self-efficacy in health promotion research and practice: conceptualization and measurement. Health Education Research. 10(1),

181 Manojlovich, M. (2005a). Promoting nurses self-efficacy A leadership strategy to improve practice. Journal of Nursing Administration, 35(5), Manojlovich, M. (2005b). The effects of nursing leadership on hospital nurses professional practice behaviours. Journal of Nursing Administration, 35(7/8), Mayer, C., Andrusyszyn, M., & Iwasiw, C. (2005). Self-efficacy of staff nurses for health promotion counseling of patients at risk of stroke. Journal of Neuroscience Nursing, 24(4), Maylor, M. (2001). Senior nurses control expectations and the development of pressure ulcers. Nursing Standard, 15(45), Marckx, B. (1995). Watson s theory of caring: A model for implementation in Practice. Journal of Nursing Care Quality, 9(4), MacCallum, R., & Austin, J. (2000). Applications of structural equation modeling in psychological research. Annual Review of Psychology, 51, McCance, T. (2003). Caring in nursing practice: The development of a conceptual framework. Research and Theory for Nursing Practice: An International Journal. 17(2). Middleton, S., Griffiths R., Fernandez, R. & Smith, B. (2008). Nursing practice environment: How does one Australian hospital compare with magnet hospitals? International Journal of Nursing Practice. 14, Muhonen, T., & Torkelson, E. (2004). Work locus of control and its relationship to health and job satisfaction from a gender perspective. Stress and Health, 20, Murphy, S. (2005). A theory-based intervention to improve nurse s knowledge, selfefficacy, and skills to reduce PICC occlusion. Journal of Infusion Nursing, 28(3), Nedd, N. (2006). Perceptions of empowerment and intent to stay. Nursing Economics, 24(1),

182 Neuman, W. L. (1994). Social research methods (2nd ed.). Boston: Allyn & Bacon. New South Wales Nurses Association. (2007). Hospital based training plan under fire. Waterloo, NSW: NSW Nurses Association. Ng, T., Sorensen, K. & Eby, L. (2006). Locus of control at work: a meta-analysis. Journal of Organizational Behavior. 27, Nightingale F. (1859). Notes on nursing: what it is, and what it is not. London, UK. Harrison, Philadelphia, PA, Lippincott, Nunnally, J. (1978). Psychometric theory. New York: McGraw-Hill. Organ, D. & Greene, C. (1974). Role ambiguity, locus of control, and work satisfaction. Journal of Applied Psychology, 59, Oulton, J. (2006). The global nursing shortage: An overview of issues and actions. Policy, Politics and Nursing Practice, 7(3), 34S 39S. Page, A. (Ed.). (2004). Keeping patients safe: Transforming the work environment of nurses. Washington, DC: National Academies Press. Paliadelis, P.. (2008). The working world of nurse unit managers: Responsibility without power. Australian Health Review, 32(2), Paliadelis, P. & Cruickshank, M. (2008). Using a voice-centred relational method of data analysis in a feminist study exploring the working world of nursing unit managers. Qualitative Health Research, 18(10), Parkes, K. (1984). Locus of control, cognitive appraisal and coping in stressful episodes. Journal of Personality and Social Psychology, 46(3), Parker, D., Tuckett, A., Eley, R., & Hegney, D. (2010). Construct validity and reliability of the Practice Environment Scale of the Nursing Work Index (PES-NWI) for Queensland Nurses. International of Journal Nursing Practice, 16(4),

183 Parker, L. (1993). When to fix it and when to leave: Relationships among perceived control, self-efficacy, dissent and exit. Journal of Applied Psychology, 78(6), Parse, R. (1981). Man-living-health: A theory of nursing. New York, John Wiley. Pfeffer, J. and O Reilly, C. (1987). Hospital demography and turnover among nurses. Industrial Relations. 26(2). Polit, D., & Beck, C. (2004). Nursing research: Principle and methods (7th ed.). Philadelphia: Lippincott Williams & Wilkins. Polit, D., & Hungler, B. (1999). Nursing research: Principles and methods (5th ed.). London: Lippincott. Productivity Commission (2007), Annual Report , Annual Report Series, Productivity Commission, Canberra. Ray, M. (1981). A philosophical analysis of caring within nursing. In M. Leininger (Ed.), Caring: An essential human need (pp ). Detroit, MI: Wayne State University Press. Rigdon, E. (1998). Structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp ). Mahwah, NJ: Lawrence Erlbaum Associates, Publishers. Rogers, M.E. (1970). An introduction to the theoretical basis of nursing. Philadelphia: F.A. Davis. Rogers, A., Hwang, W., Scott, L., Aiken, L., & Dinges, D. (2004). The working hours of hospital staff nurses and patient safety. Health Affairs, 23(4), 202. Rotter, J. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs: General and Applied, 80(1), Whole No Saddler, J. (2000). A multiphase approach to concept analysis and development. In B. Rodgers & K. Knafl (Eds.), Concept Development in Nursing: 167

184 Foundations, Techniques, and Applications (2nd ed., p ). Philadelphia, PA: W.B. Saunders Company. Saeki, S., Okazaki, T., & Hachisuka, K. (2006). Concurrent validity of the community integration questionnaire in patients with traumatic brain injury in Japan. Journal of Rehabilitation Medicine, 38, Scherer, M., Maddux, J., Mercandante, B., Prentice-Dunn, S., Jacons, B., & Pogers, R. (1982). The self-efficacy scale: Construction and validation. Psychological Reports, 51, Schisler, C. (2006). Evaluating your nursing collection: a quick way to preserve nursing history in a working collection. Journal of the Medical Library Association, 95(3). Schlomer, G., Bauman, S. & Card, N. (2010). Best practices for missing data management in counseling psychology. Journal of Counseling Psychology, 57(1), Schmitz, N., Neumann, Opperman. R. (2000). Stress, burnout and locus of control in German nurses. International Journal of Nursing Studies, 37(2) Schofield, I., Tolson, D., Arthur, D., Davies, S., & Nolan. M. (2005). An exploration of the caring attributes and perceptions of work place change among gerontological nursing staff in England, Scotland and China (Hong Kong). International Journal of Nursing Studies, 42(2), Schofield, D. (2007). Replacing the projected retiring baby boomer nursing cohort BMC Health Services Research. 7, 87. Schultz, K., & Grimes, D. (2002). Generation of allocation sequences in randomised trials: Chance, not choice. The Lancet, 359(9), Scott, J., Sochalski, J., & Aiken, L. (1999). Review of magnet hospital research: Findings and implications for professional nursing practice. Journal of Nursing Administration, 29,

185 Sengin, K. (2003). Work-related attributes of RN job satisfaction in acute care hospitals. Journal of Nursing Administration, 33(6), Shader, K., Broome, M., Broome, C., West, M. & Nash, M (2001). Factors influencing satisfaction and anticipated turnover for nurses in an academic medical center. Journal of Nursing Administration 31(4), Shah, R., & Goldstein, S. (2006). Use of structural equation modelling in operations management research: Looking back and forward. Journal of Operations Management, 24(2), Shields, M. & Ward, M. (2001). Improving nurse retention in the National Health Service in England: The impact of job satisfaction and intentions to quit. Journal of Health Economics, 20, Shorr, A. (2000). Has nursing lost its professional focus? Nursing Administration Quarterly 25(1), Smith, P., Kendall, L. & Hulin, C. (1969). Measurement of satisfaction in work and retirement. Chicago, IL: Rand McNally. Sobel, M. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed), Sociological Methodology (pp ). Washington DC. American Sociological Association Quarterly, 25(1), Spector, P. (1978). Organizational frustration: A model and review of the literature. Personnel Psychology, 31, Spector, P. (1982). Behaviour in organisations as a function of employee s locus of control. Psychological Bulletin, 91, Spector, P. (1985). Measurement of human service staff satisfaction: Development of the Job Satisfaction Survey. American Journal of Community Psychology, 13(6), Spector, P. (1988). Development of the work locus of control scale. The British Psychology Society, 61,

186 Spector, P., & Michaels, C. (1986). Personality and employee withdrawal: Effects of locus of control on turnover. Psychological Reports, 59, SPSS for Windows, Rel (2001). Chicago: SPSS Inc. Stajkovic, A., & Luthans, F. (1998). Self-Efficacy and work-related performance: A meta-analysis. Psychological Bulletin, 124(2), Storms, P., & Spector, P. (1987). Relationships of organisational frustration with reported behavioural reactions: The moderating effect of locus of control. Journal of Occupational Psychology, 60, Schlomer, G., Bauman, S., & Card, A. (2010). Best practices for missing data management in counseling psychology. Journal of Counseling Psychology, 57(1), Schwarzer, R., & Jerusalem, M. (1995). Generalized self-efficacy scale. In J. Weinman, S. Wright, & M. Johnston (Eds.), Measures in health psychology: A user s portfolio. Causal and control beliefs (pp ). Windsor, UK: NFER-NELSON. Spreitzer, G. (1996). Social structural characteristics of psychological empowerment. Academy of Management Journal, 39(2), Stallman, H., & Hurst, C. (2011). Factor structure of the Frost Multidimensional Perfectionism Scale in university students. Australian Psychologist. 46(4) Tabachnick, B. & Fidell, L. (2007). Using Multivariate Statistics, Fifth Edition. Boston, Pearson Education, Inc. Takase, M. (2001). Nurse-environment mis-fit & nursing practice. Journal of Advanced Nursing. 35(6), Tavernier, S. (2006). An Evidence-based Conceptual Analysis of Presence, Holistic Nursing Practice. 20(3),

187 Tyson, P., & Pongruengphant, R. (2004). Five-year follow-up study of stress among nurses in public and private hospitals in Thailand. International Journal of Nursing Studies, 41, Ulrich, C., O'Donnell, P., Taylor, C., Farrar, A., Danis, M., & Grady, C. (2007). Ethical Climate, Ethics Stress, and the Job Satisfaction of Nurses and Social Workers in the United States. Social Science and Medicine. 65(8), Upton, G. & Cook, I. (2008). "A Dictionary of Statistics, 2nd edition revised." Oxford University Press. ISBN Van Saane, J. Sluiter, J., Verbeek, J. & Frings-Dresen, M. (2003). Reliability and validity of instruments measuring job satisfaction-a systematic review. Occupational Medicine, 53(3). Vehkalahti, K., Puntanen, S. &Tarkkonen. L. (2007). Effects of measurement errors in predictor selection of linear regression model. Journal of Computational Statistics & Data Analysis. 52(2). Wade, G., Osgood, B., Avino, K., Bucher, G., Bucher, L., Foraker, T., French, D. & Sirkowski, C. (2008). Influence of organizational characteristics and caring attributes of managers on nurses job enjoyment. Journal of Advanced Nursing 64(4), Walsh, W., & Betz, N. (2001). Tests and assessment (4th ed.). Upper Saddle River, NJ: Prentice Hall. Warr, P., Cook, J., & Wall, T. (1979). Scales for measurement of some work attitudes and aspects of psychological well-being. Journal of Occupational Psychology, 52, Watson, J. (1979). Nursing: The philosophy and science of caring. Boston, MA: Little, Brown and Company. Watson, J. (1985). Nursing human science and human care. NJ: Appleton-Century. 171

188 Watson, J. (1988). New dimensions of human caring theory. Nursing Science Quarterly, 1, Watson, J. (1996). Watson s theory of transpersonal caring. In P. H. Walker & B. Neuman (Eds.), Blueprint for use of nursing models: Education, research, practice, and administration (pp ). New York: National League for Nursing. Weston, J. (2009). Validity of instruments for measuring autonomy and control over nursing practice. Journal of Nursing Scholarship, 41(1), Watson, J., Jackson, D., & Borbasi, S. (2005). Tracing nurse caring: Issues, concerns, debates. In J. Daly, S. Speedy, D. Jackson, V. Lambert, & C. Lambert (Eds.), Professional nursing: Concepts, issues, and challenges (p ). New York: Springer Publishing. Webster, J., Flint, A., & Courtney, M. (2008). A new practice environment measure based on the reality and experiences of nurses working lives. Journal of Nursing Management, 17(1), White, A., & Spector, P. (1987). An investigation of age-related factors in the age job-satisfaction relationship. Psychology and Aging, 2(3), Wong, D., Leung, S., So, C. (2001). Differential impacts of coping strategies on the mental health of Chinese nurses in hospitals in Hong Kong. International Journal of Nursing Practice, 7, World Health Organization, (2006). The world health report 2006: working together for health Geneva 27, Switzerland. Zangaro, G. & Soeken, K. (2007). A meta-analysis of studies of nurses job satisfaction. Research in Nursing & Health, 30,

189 Appendices Appendix 1: Research Higher Degree Initiative Award 173

190 Appendix 2: National Research and Scholarship Fund Award 174

191 Appendix 3: Introduction to Structural Equation Modelling Using AMOS Certificate 175

192 Appendix 4: Membership Honor s Society of Nursing Sigma Theta Tau International 176

193 Appendix 5: Plain Language Statement of Information and Instructions 177

194 178

195 Appendix 6: Human Ethics Approval Certificate 179