Single-Item Measure of Risk Orientation

Transcription

1 Single-Item Measure of Risk Orientation Cherie D. Maestas, Florida State University Scott Bell, Florida State University William Pollock, SUNY Stonybrook For related publications / working papers see: Ehrlich, Sean and Cherie Maestas Risk, Risk Orientation and Policy Opinions: The Case of Free Trade Political Psychology. 31(5): Maestas, Cherie and Pollock, William M., Measuring Generalized Risk Orientation with a Single Survey Item (May 3, 2010). Available at SSRN:

2 Risk Orientation (Affect) General affective response to facing risky or uncertain situations Trait-based Correlates with stable personality characteristics Related to genetic characteristics Independent of context or framing Distinct from risk perception and behavior Maestas, Bell & Pollock Florida State University, APSA 2013

3 Effects of Risk-Orientation Risk Aversion / Acceptance Risk Perception Political behavior & Preferences Maestas, Bell & Pollock Florida State University, APSA 2013

4 Prior Measurement Strategies Comparative Gambles Stock or market ownership Multi-item scales, usually 6 to 20 items Problems: Too much space Lack of generality Behavioral rather than affective Maestas, Bell & Pollock Florida State University, APSA 2013

5 EGSS 2 Summer 2011 Affect Towards Risk Question In general, people often face risks when making financial, career, or other life decisions. Overall, do you feel comfortable, uncomfortable, or neither comfortable nor uncomfortable taking risks? Maestas, Bell & Pollock Florida State University, APSA 2013

6 EGSS 2 Summer 2011 Affect Towards Risk Question In general, people often face risks when making financial, career, or other life decisions. Overall, do you feel comfortable, uncomfortable, or neither comfortable nor uncomfortable taking risks? Maestas, Bell & Pollock Florida State University, APSA 2013

7 Split-Half Design: List Version List (Risk-L) Extremely comfortable taking risks Moderately comfortable taking risks Slightly comfortable taking risks Neither comfortable nor uncomfortable taking risks Slightly uncomfortable taking risks Moderately uncomfortable taking risks Extremely uncomfortable taking risks Maestas, Bell & Pollock Florida State University, APSA 2013

8 Split-Half Design: Branched Version Branching (Risk-B) Comfortable taking risks Uncomfortable taking risks Neither comfortable nor uncomfortable taking risks How (uncomfortable / comfortable) do you feel taking risks? Extremely (comfortable / uncomfortable) Moderately (comfortable / uncomfortable) Slightly (comfortable / uncomfortable) Maestas, Bell & Pollock Florida State University, APSA 2013

9 0 Percent Pooled Distribution EGSS 7 Item Risk Scale, Pooled = Ext. Uncomfortable; 7=Ext. Comfortable Maestas, Bell & Pollock Florida State University, APSA 2013

10 0 5 0 Percent Percent Comparison: 2010 CCES Results 7 Item Risk Scale (Factor Scores) CCES 2010 Pre-election Wave (N=4309) Risk Orientation, Single-Item List CCES Pre-election 2010 (N=4482) = Ext. Risk Averse 1 = Ext. Risk Acceptant = Ext. Uncomfortable 7 = Ext. Comfortable Maestas, Bell & Pollock Florida State University, APSA 2013

11 0 Percent Distributions of Risk-B and Risk-L List Version (N=641) Risk Orientation Branched Version (N=590) = Ext. Uncomfortable 4 = Neutral 7 = Ext. Comfortable EGSS-2 unweighted Maestas, Bell & Pollock Florida State University, APSA 2013

12 -----Cutpoints omitted Maestas, Bell & Pollock Florida State University, APSA 2013

13 Substantive Examples Political Participation Prospective Sociotropic & Pocketbook Assessments Political Affect Policy Preference Note: All models run with weights and with full set of demographic and substantive controls Maestas, Bell & Pollock Florida State University, APSA 2013

14 Political Participation Attend Rally (+) Donate money (+) Button-wearing (?) Phoning/Writing an Official (?) Scaling for all Q s: -----Cutpoints and other controls omitted =Never; 1=Any at all Maestas, Bell & Pollock Florida State University, APSA 2013

15

16 Sociotropic and Pocketbook Prospective Econ. Eval (+): 1=Much Worse, 2=Much Better -----Cutpoints and other controls omitted Financial Worry (-): 1=Not at all 5=Extremely

17

18 Affect Fear (Afraid) (-) Hope (+) Scaling on both questions: 1=Not at all, 5=Extremely; -----Cutpoints and other controls omitted -----

19

20 Policy Support Unemployment (-): 1 = <26 weeks; 3 = >26 weeks; Govt Jobs (+): 1=Intervention 7=No Intervention Healthcare (-): -----Cutpoints and other controls omitted =Oppose 7=Support

21

22

23 List or Branched?

24 0 Percent Distributions of Risk-B and Risk-L List Version (N=641) Risk Orientation Branched Version (N=590) = Ext. Uncomfortable 4 = Neutral 7 = Ext. Comfortable EGSS-2 unweighted Maestas, Bell & Pollock Florida State University, APSA 2013

25

26 Rally and Donate List vs. Branch Attend Rally (+) Donate money (+) Scaling for all Q s: 0=Never; 1=Any at all -----Cutpoints and other controls omitted -----

27

28 Economy List vs. Branched Prospective Econ. Eval (+): 1=Much Worse, 2=Much Better Financial Worry (-): 1=Not at all 5=Extremely -----Cutpoints and other controls omitted -----

29

30 Policy: List vs. Branch Govt Jobs (+): 1=Intervention 7=No Intervention Healthcare (-): 1=Oppose 7=Support -----Cutpoints and other controls omitted -----

31

32 Conclusions Single item risk orientation measure useful & cheap in terms of space List version produces results similar to multi-item scales in many but not all cases Branched magnifies extremes of scale; List is more similar to multi-item scales Both predict wide variety of outcomes