Toward an understanding of the endogenous nature of group identification in games

The authors thank Jonathan Baron, Emily Corse, Rachel Croson, Sean Duffy, Astri Muren, Debbie Prentice, Mary Steffel, Jack Worrall, and participants of the ESA conference at Caltech and the SABE/IAREP conference in Rome for helpful comments. The authors offer special thanks to Johanna Hertel. This research was supported by Rutgers University Research Council Grants #202171 and #202344.

Abstract

It is commonly assumed that identification with a social group is constant throughout the play of a one-shot game in the absence of feedback. Our experiment provides evidence that challenges this assumption. Subjects play one of two versions of the prisoner's dilemma game. These versions are distinguished by the relative attractiveness of the uncooperative action. The version with a relatively attractive uncooperative action is referred to as the Easy Game and the other as the Difficult Game. Subjects who play the Difficult Game, exhibit a change in group identification which is related to their selected action. No such relationship exists within the Easy Game. Additionally, the change primarily occurs after the action is selected rather than upon inspection of the game.

Keywords:

• Group identification
• Experimental game theory
• Endogenous preferences
• Social identity
• Decision difficulty

Notes

The authors thank Jonathan Baron, Emily Corse, Rachel Croson, Sean Duffy, Astri Muren, Debbie Prentice, Mary Steffel, Jack Worrall, and participants of the ESA conference at Caltech and the SABE/IAREP conference in Rome for helpful comments. The authors offer special thanks to Johanna Hertel. This research was supported by Rutgers University Research Council Grants #202171 and #202344.

1 For our purposes, we define group identification to be the degree to which the subject feels that she belongs to the group and favorably regards membership in the group.

2 For instance, see Abrams and Hogg (1999), Brown, Condor, Mathews, Wade, and Williams (1986), Gaertner, Mann, Murrell, and Dovidio (1989), Garcia, Tor, Bazerman, and Miller (2005), Grieve and Hogg (1999), Hogg, Cooper-Shaw, and Holzworth (1993), Hogg and Grieve (1999), Hogg and Hardie (1991, 1992), Reid and Hogg (2005), and Swann, Kwan, Polzer, and Milton (2003).

3 For instance, see Aguiar, Brañas-Garza, Espinosa, and Miller (2010), Ahmed (2007), Brewer and Kramer (1986), Charness, Rigotti, and Rustichini (2007), Chen and Li (2009), Dawes, Van de Kragt, and Orbell (1988), Eckel and Grossman (2005), Kramer and Brewer (1984), McLeish and Oxoby (2007, 2011), and Wit and Wilke (1992).

4 Although Güth et al. (2008) use items adapted from Gaertner et al. (1989) rather than Grieve and Hogg (1999).

5 Also see Bowles (1998), Canegallo, Ortona, Ottone, Ponzaono, and Scacciati (2008), Schotter, Weiss, and Zapater (1996), Smith (2012), and Vlaev and Chater (2006).

6 Also see Izuma et al. (2010), Liberman and Forster (2006), Sharot, De Martino, and Dolan (2009), Shultz and Lepper (1996), Shultz, Leveille, and Lepper (1999), and Steffel (2010). Note that some of these papers are often considered to be in the cognitive dissonance rather than decision difficulty literature, despite that they are also cited by decision difficulty papers. The cognitive dissonance literature focuses on the mental discomfort associated with situations in which attitudes and actions do not coincide. Difficult decisions can cause cognitive dissonance, yet these two concepts are distinct.

7 Rapoport and Chammah (1965) for more on this index.

8 The written instructions and the power point slides are available from the corresponding author upon request.

9 Liberman, Sammuels, and Ross (2004) show that referring to a prisoner's dilemma game as the “Wall Street Game” induces less cooperative behavior than referring to it as the “Community Game.”

10 Note that Brañas-Garza (2007) examines behavior in the dictator game where some subjects are exposed to a similar statement, and some are not. The author finds that the subjects who were exposed to this sentence acted more generously in a dictator game. In contrast to Brañas-Garza (2007), each of our subjects are exposed to this sentence. It is possible that, upon observing this statement, the subjects acted more prosocially than they would have otherwise. However, due to the different designs, a comparison between our paper and Brañas-Garza (2007) is difficult.

11 A logistic regression with choice as the dependent variable and the categorical session variable as an independent variable suggests that the session variable is not significant (p = 0.43).

12 With p-values of 0.40 and 0.19, respectively.

13 The first regression in Table 4 is identical to the second regression in Table 3. Both are included in order to facilitate the comparisons to the other regressions in the tables.