Blackwell Science, LtdOxford, UKAJSPAsian Journal of Social Psychology1367-2223Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005August 2005 82173190Original ArticleGroup-based cooperationToshio Yamagishi et al.

Asian Journal of Social Psychology (2005) 8: 173–190

Comparisons of Australians and Japanese on group-based cooperation Toshio Yamagishi and Yosuke Makimura Hokkaido University, Japan

Margaret Foddy Carleton University, Canada

Masafumi Matsuda NTT, Japan

Toko Kiyonari Mc Master University, Canada

Michael J. Platow Australian National University, Australia

A cross-societal experiment with 49 Australian and 56 Japanese participants examined if the group heuristic account of ingroup-favoring behavior in a Prisoner’s Dilemma game can be extended beyond the minimal group situation to a situation involving an enduring social category (i.e. participant’s nationality). Participants played a Prisoner’s Dilemma game five times, each time with a different partner. Two of the five partners were ingroup members, two were outgroup members, and the nationality of one partner was not known. Furthermore, one of the two ingroup (or outgroup) partners knew that the participant was a member of the same (or the other) nationality, and the other did not know it. The results indicated that the knowledge that the partner had about the nationality of the participant exerted an effect only when the partner was an ingroup member. No major difference was found between Australian and Japanese participants. An outgroup-favoring cooperation pattern was observed, but that pattern was shown to be a result of fairness concerns among Australian participants and of positive stereotypes of Australians among Japanese participants. Key words: cross-cultural experiment, in-group relationship, prisoner’s dilemma, reciprocity, trust.

favoritism,

inter-group

Correspondence: Toshio Yamagishi, Graduate School of Letters, Hokkaido University, N10 W7 Kita-ku, Sapporo 060-0810, Japan. Email: [email protected] © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

174

Toshio Yamagishi et al.

Introduction It has long been known that shared group membership increases cooperation with ingroup members by generating positive evaluations of fellow ingroup members and/or expectations of generalized reciprocity (Kramer & Brewer, 1984; Brewer & Kramer, 1986; Kramer, 1991; Wit & Wilke, 1992; Kramer et al., 1993; Kramer & Goldman, 1995; Jin & Yamagishi, 1997; Kollock, 1997; de Cremer & van Vugt, 1999; van Vugt & de Cremer, 1999). The cooperationenhancing effect of shared group membership was originally explained with reference to social identity theory (Tajfel & Turner, 1986; Hogg & Abrams, 1988; Hogg, 1992), due mostly to the fact that the effect was observed even in the minimal group situation, an experimental setting originally developed by Tajfel et al. (1971). The minimal group experiment (Tajfel et al., 1971) produced a finding that participants treat a member of their own arbitrarily constituted group more favorably than a member of a different, but equally arbitrary, group. Trivial features that are shared by members and the lack of face-to-face interaction or interdependence of rewards define the minimal group. For example, participants may be divided into groups of ‘over-estimators’ and ‘under-estimators’ according to their estimation of the number of dots displayed on a screen, or by their supposed preference for paintings by Klee versus Kandinsky. Participants then allocate monetary rewards between a member of their own minimal group and of the other minimal group. There seems to be no reason to expect participants in this experiment to make group-based differential allocations (e.g. to give more to someone who shows a similar preference for an artist to whom they have probably never been exposed before) and, yet, participants in the minimal group experiment, whether in the original study by Tajfel et al. (1971) or in a large number of subsequent experiments, have shown such favoritism (Billig & Tajfel, 1973; Turner, 1975; Turner et al., 1979; Hartstone & Augoustinos, 1995). One explanation offered for the ingroup favoritism observed in the minimal group paradigm derives from social identity theory. Within the framework of this theory, people’s self-concepts are comprised of both personal identities (who they are as unique individuals, different from all others) and social identities (who they are as group members, sharing important characteristics with others). In the minimal group paradigm, categorization is imposed on people by an experimenter in a laboratory context. People adopt the categorization as self-defining in this particular context. Moreover, the categorization leads people to engage in social comparison (Festinger, 1954; Turner et al., 1987) with other relevant categories as a means of understanding the value of their own category membership, and, hence, social identity. Positively discrepant comparisons - comparisons that favor group members’ own category over another - are assumed to reflect positively on group members, leading to an enhancement of their social identities in this context (Oakes & Turner, 1990; Platow et al., 1997; Rubin & Hewstone, 1998). In the minimal group paradigm, only one value-laden dimension - the distribution of valued resources - is available for group members on which to make this comparison. As a result, group members actively strive to positively differentiate their ingroup from a relevant outgroup on this available dimension as a means of social identity enhancement (Diehl, 1989; Leyens et al., 1994; Rubin & Hewstone, 1998). More recently, however, this explanation of ingroup-favoring behavior in the minimal group experiment has been challenged by an alternative account that claims that the presence of a salient ingroup activates a ‘group heuristic’ or a naïve theory of groups (Karp et al., 1993; Jin et al., 1996; Jin & Yamagishi, 1997; Yamagishi et al., 1998, 1999; Yamagishi & Kiyonari, 2000; Kiyonari, 2002). According to Yamagishi and colleagues, people have a naïve © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

175

theory of how groups operate, and one important feature of it is the belief that the group is the place where generalized exchanges occur, and where people can expect generalized reciprocity. Given the fact that humans have lived and cooperated in small bands through most of their evolution (Cosmides, 1989; Cosmides & Tooby, 1992), it is no surprise that they would have acquired a naïve theory of how people behave in a group situation. A consequence of having such a naïve theory of groups is the expectation that people give preferential treatment to the members of their own group. Ingroup-favoring behavior in the minimal group situation is thus a generalized reciprocation of the favorable treatment they expect to be given by members of their own group. This general heuristic applies also to one’s own behavior towards other group members. The idea of the group heuristic briefly summarized above is complementary to the notion of ‘schema of an outgroup’ - ‘learned beliefs or expectations that intergroup relations are competitive, unfriendly, deceitful, and aggressive’ - discussed by Insko and Schopler (1998, p. 80). Insko and colleagues (Insko & Schopler, 1998; Insko et al., 1998) argue that the ‘discontinuity effect’ observed in their studies is at least partly caused by the schema of outgroups. The group heuristic is a mirror image of the outgroup schema. It is a set of beliefs or expectations that ingroup relations will be cooperative, friendly, truthful, and peaceful. These two together form a system of beliefs that groups are characterized by a system of generalized exchange, and intergroup relations that lack the system of generalized exchange are where cooperation is wasted at best, and exploited at worst. The existence of the group heuristic (Yamagishi et al., 1999) cannot be directly tested, but a set of predictions concerning its operation has been successfully confirmed. First, in a replication study of the original minimal group experiment, Karp et al. (1993) found that participants allocated more money to a member of their own group than to a member of the other group. This ingroup-favoring behavior in reward allocation completely disappeared, however, when it was made clear to the participants that the reward they would receive in the experiment did not depend on the other participants’ behavior. Karp and colleagues achieved this manipulation by paying participants a predetermined amount, independent of the outcome of the assignment of outcomes to others in the minimal group. In the original minimal group experiment, each participant allocated money to one ingroup member and one outgroup member. As all participants faced the same situation, they were in the situation of being allocated money from ingroup and outgroup members as well, albeit not from the particular others to whom they allocated money. What Karp and colleagues eliminated is the latter aspect of ‘mutual fate control’ (Thibaut & Kelley, 1959), with the result that ingroup-favoring behavior disappeared. If ingroup favoritism relied only on a desire to positively distinguish the ingroup, it should not matter whether mutual fate control is present. This experiment and similar other conceptual replications of it (Jin et al., 1996; Jin & Yamagishi, 1997; Yamagishi et al., 1999; Yamagishi & Kiyonari, 2000; Kiyonari, 2002) have repeatedly shown that eliminating the expectation of ingroup-favoring treatment by other members of their own group eliminates ingroup-favoring behavior among participants faced with the minimal group situation. In one such experiment (Jin & Yamagishi, 1997; reported in English in Yamagishi et al., 1999), participants played a Prisoner’s Dilemma (PD) game with an ingroup member and an outgroup member. As expected, they cooperated more with an ingroup member than with an outgroup member. Thus, Jin and Yamagishi successfully replicated the ingroup-favoring behavior in the PD game observed in earlier research (Kramer & Brewer, 1984; Brewer & Kramer, 1986; Kramer, 1991; Wit & Wilke, 1992; Kramer et al., 1993; Kramer & Goldman, 1995; Jin & Yamagishi, 1997; Kollock, 1997; de Cremer & van Vugt, 1999; van Vugt & de Cremer, 1999). However, this ingroup-favoring behavior (i.e. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

176

Toshio Yamagishi et al.

cooperating more with an ingroup member than with an outgroup member) disappeared in a new condition introduced by Jin and Yamagishi, in which the participant knew whether the partner was an ingroup or outgroup member but the partner did not know the group membership of the participant (and the participant knew that the partner did not know it). In this ‘unilateral-knowledge’ condition, only focal participants knew the social categories of the two players, and they could thus not expect ingroup-favoring treatment from another ingroup member who was unaware of their shared group membership. Eliminating the expectation for ingroup-favoring behavior from an ingroup member in this condition was sufficient to eliminate ingroup-favoring behavior by the participant. However, a more favorable evaluation of ingroup members’ traits (e.g. trustworthiness, generosity) was not eliminated by this manipulation; participants rated their ingroup members more favorably than outgroup members, even when the partner did not know that the participant was a member of the same group. Participants clearly identified themselves with the ingroup and had positive evaluations of their fellow group members, yet they did not behave in the PD in an ingroup-favoring manner. The same finding was reported in another, partial replication experiment by Kiyonari (2002). The results of these studies, highlighting the importance of expectations of ingroupfavoring behavior from other ingroup members to the production of biased reward distributions, provide strong support for the group heuristic interpretation of ingroup-favoring behavior in the minimal group situation. It is an open question, however, whether the findings, especially findings about the effects of expectations, are generalizable to non-minimal group situations, or to non-Japanese populations. There are several reasons to suspect that they may not. First, minimal categorization may serve more as a maximal than a minimal group (Leyens et al., 1994, p. 61) in the sense that group membership is the only relevant and salient information provided to the participants. It is a pure form of group, devoid of any ‘contaminating’ features often associated with enduring groups outside of the laboratory. It is thus easier for participants to expect that others will use group categories in making their decisions, leading to expectations of ingroup-favoring treatment from other ingroup members. When other features associated with the particular groups people face outside the laboratory (such as stereotypes associated with the groups) are more salient, people may not as easily activate the group heuristic and expect other ingroup members to treat them favorably. Particular characteristics of the ingroup stereotype relevant to fair treatment may then become more important determinants of ingroup bias. Although several studies confirmed that findings in laboratory minimal group research can be generalized to enduring groups (Amancio, 1989; Mullen et al., 1992), they were not conducted as an explicit test of the group heuristic theory, and thus did not incorporate the important conditions that block generalized expectations of reciprocity (e.g. provision of noncontingent reward to the participant making allocations, eliminating mutual knowledge of shared group membership).1 Only one study (Makimura & Yamagishi, 2002) has so far examined whether the effect of eliminating expectations of ingroup-favoring behavior can be generalized to non-minimal group situations. In this study, Makimura and Yamagishi replicated the Karp et al. (1993) experiment described above, using two groups of participants that had been engaged in ongoing interactions. The result of this experiment showed no effect of the knowledge manipulation. Whether or not the partner knew of the participant’s group membership, the participant allocated more money to a member of his or her own group than to a member of the other group. However, the ‘indiscriminate’ ingroup-favoritism observed among the participants in this study was shown not to be based on depersonalized social identity. Rather, Makimura and Yamagishi showed that it was based on interpersonal bonds © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

177

with people in their own group (cf. Hewstone et al., 1982). These results of Makimura and Yamagishi (2002) provide grounds to suspect that the effect of eliminating expectations of ingroup-favoring behavior, and thus the operation of the group heuristic, may be limited to the minimal group situation. Another reason to suspect that the operation of the group heuristic is limited to the minimal group situation is that its operation may be overridden by the effect of group identity when the group provides a strong basis for identity. One of the purposes of the current study is to examine whether the effect of eliminating expectations of ingroup-favoring behavior by other ingroup members observed in the earlier studies by Yamagishi and colleagues also occurs in other types of non-minimal groups, those defined by enduring social categories. Such categories, like minimal groups, are different from existing groups of friends or workmates in that there will never be any face-to-face contact among the majority of members of the group. Nevertheless, people do show strong identification with large categories such as nationality, university, ethnic groups, and so on, and display a positive ingroup bias in their evaluations of group members (Mullen et al., 1992; Foddy & Hogg, 1999; Brown, 2000), and will place more trust in ingroup members (Brewer, 1981, 1999). It has frequently been assumed by many social identity theorists that the ‘real’ social categories and the ‘mere categories’ of the minimal group will follow the same principles; that is, both will invoke ingroup favoritism (Deaux, 2000). Using this logic, and allowing our argument about the importance of expectations of reciprocity among group members, one would therefore expect that the effects of mutual knowledge of shared category membership found for the minimal groups in the studies reported above would also occur for enduring social categories. If there are no such effects of mutual knowledge of shared category membership, the wider importance of expectations of generalized reciprocity in generating ingroup trust and cooperation will be compromised. To study the impact of shared membership in existing groups, we used the category of nationality, which is a large, anonymous, but meaningful group, with which people identify to varying degrees, and for which there may be generalized expectations of ingroup favoring behavior, despite a lack of personal ongoing contact among group members. A different line of reasoning also leads one to suspect that the findings concerning the importance of expectations for ingroup-favoring behavior from other ingroup members as a determinant of ingroup favoritism may be limited to Japanese participants. Japanese are supposed to have a culture of collectivism (Hofstede, 1980; Triandis, 1995; Smith & Bond, 1999), and it is widely believed that membership in ingroups is more important in collectivist cultures than in individualist cultures for a wide range of social processes, including self-definition (Markus & Kitayama, 1991), cooperation (Karau & Williams, 1993), leadership (Meade, 1985), and so on. In contrast, members of a more individualistic culture may be less affected by their membership ingroups and social categories, with greater emphasis given to de-contextualized logic (Nisbett et al., 2001). The naïve theory of how groups operate may be different (e.g. expecting ingroup-favoring behavior less) for members of more individualistic cultures. Thus, the group heuristic may be less easily activated, or even when it is activated, the expectation of ingroup-favoring behavior by others may be weaker among members of an individualistic culture. Thus, members of an individualistic culture may have weaker expectations for ingroup-favoring behavior from other ingroup members. As a consequence, eliminating such expectations may not produce as pronounced an effect among individualists as among collectivists. The second purpose of our study was to examine, by comparing Japanese participants’ behavior with that of Australian participants,2 whether the effect observed in Yamagishi and colleagues’ © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

178

Toshio Yamagishi et al.

experiments conducted in Japan with Japanese participants could be generalized to participants from a more individualistic culture. In the experiment presented below, we replicated the Jin and Yamagishi (1997) study using nationality groups rather than minimal groups. For Australian participants, the ingroup was Australian and the outgroup were Japanese participants. For Japanese participants, the ingroup was Japanese and the outgroup were Australian participants. We test the following hypothesis derived from the group heuristic argument concerning the effect of the knowledge manipulation: whether the partner knows the group membership of the player, that is, whether there is mutual recognition of shared group membership. We had no strong theoretical basis to expect that the results obtained in the original experiment (Jin & Yamagishi, 1997) with minimal groups would not be replicated with existing nationality categories, apart from the possibility that such effects might be weaker among Australian participants. It is predicted that expectations of generalized reciprocity exist when the player knows that her partner is aware of the fact that they both belong to the same group. Otherwise, the player cannot expect her partner to provide favorable treatment to her. When a player interacts with an outgroup member, knowledge of group membership will not matter, as there is no expectation of a system of generalized exchanges across groups. Hypotheses Hypothesis 1. The knowledge manipulation will have an effect such that PD players who interact with an ingroup member will cooperate more when their partner knows that they both belong to the same group (mutual knowledge condition) than when their partner does not know if they belong to the same group (unilateral knowledge condition). Hypothesis 2. The knowledge manipulation will not have an effect when PD players interact with outgroup members. Furthermore, we propose the following hypothesis based on the argument that groups will play a more important role in collectivist culture than in individualist culture. Hypothesis 3. The effect of the knowledge manipulation predicted in Hypothesis 1 will be stronger among Japanese participants than among Australian participants.

Method Overview of the experiment Participants played a Prisoner’s Dilemma game five times, once in each of the within-subject conditions to be explained below. Two of the five experimental conditions were the standard ingroup-outgroup conditions. For Australian participants, Australians were the ingroup and Japanese participants were the outgroup. For Japanese participants, Japanese participants were the ingroup and Australian participants were the outgroup. In the ingroup condition, participants played the Prisoner’s Dilemma game with a member of their own nationality group; in the outgroup condition, they played the same game with a member of the other nationality group. They were informed of the nationality of the partner, but not of the partner’s personal identity. The third condition was a control condition in which they played © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

179

the Prisoner’s Dilemma game with a partner whose nationality was not specified participants were told that the partner was either a Japanese or an Australian. In the fourth condition, participants played the Prisoner’s Dilemma game with an ingroup member, but they were told that the partner (who was an ingroup member) did not know the nationality of the participant. That is, participants knew that the partner was a member of their own nationality, but the partner did not know whether the participant was a member of his or her nationality group (and the participant knew that the partner did not know it). This operationalized the conditions of mutual or non-mutual knowledge referred to in the hypotheses. Similarly, in the fifth condition they played with an outgroup member who did not know the participant’s nationality. Thus, except for the control condition in which participants did not know the group identity of the partner, group identity of the partner was crossed with the knowledge the partner had about the participant’s nationality. In the mutualknowledge conditions (the first and the second conditions) both players knew about the nationalities of the two, and in the unilateral-knowledge conditions (the fourth and the fifth conditions), the participant knew the nationality of the partner and the partner did not know the nationality of the participant. Summarizing the five conditions, in the IN-MK (ingroup/ mutual-knowledge) condition, both the participant and the partner knew that they shared the same nationality, in the OUT-MK (outgroup/mutual-knowledge) condition, both of them knew that their nationalities were different, in the control condition neither knew the other’s nationality, in the IN-UK (ingroup/unilateral-knowledge) condition, the participant knew that the partner was a member of the same nationality group and was informed that the partner did not know the participant’s nationality, and finally in the OUT-UK (outgroup/unilateralknowledge) condition, the participant knew that the partner was an outgroup member and was informed that the partner did not know the participant’s nationality. In each of the five Prisoner’s Dilemma games, each participant was given 100 yen (or AUD 1.00) as an endowment and was asked how much of it to give to the partner (in increments of 10 yen or 10 cents). The amount provided by the participant was doubled in value and was given to the partner. For the purpose of the experiment, in cross-national exchanges, 100 yen were exchanged for AUD 1.00, though the actual exchange rate at the time of experiment was about 65 yen for AUD 1.00. The purchasing power of AUD 1.00 was about the same of that of 100 yen. Design and participants We used a 2 ¥ 5 (game conditions) mixed design in which the participant nationality (Australia vs Japan) was crossed with the five within-subjects conditions described above. The first within-subjects condition presented to the participant was the control condition, in which neither the participant nor the partner knew each other’s nationality. The order of presentation among the other four within-subjects conditions was then randomized. Fortynine Australian students from La Trobe University in Melbourne, Australia, and 57 Japanese students from Hokkaido University in Sapporo, Japan, participated in the study. The Prisoner’s Dilemma game The PD game used in the current experiment took the form of resource exchanges rather than the standard form involving a payoff matrix.3 In each of the 5 within-subjects conditions, the participant was provided with an initial endowment of AUD 1.00 or 100 yen and then asked how much of the endowment to provide to the partner in increments of 10 cents or 10 yen. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

180

Toshio Yamagishi et al.

The amount of money contributed by the participant to the partner was doubled by the experimenter before it was given to the partner. Similarly, the participant received twice the amount of contribution made by the partner. This exchange of monetary resources represented a Prisoner’s Dilemma. The joint profit was maximized when both contributed the most, each receiving twice the maximum contribution of AUD 1.00 or 100 yen (i.e. AUD 2.00 or 200 yen). However, the participant could keep the endowment to him- or herself by not contributing to the partner; the less he or she contributed to the partner, the more he or she saved for him- or herself. Thus, not contributing any was the ‘rational’ strategy. However, if the two participants adopted this ‘rational’ strategy of not contributing any, they each ended up with the original endowment of AUD 1.00 or 100 yen, which is half of what they could have earned if they both contributed the maximum. No feedback of the partner’s decisions was provided after each ‘exchange’. Procedure The experiment was conducted in social psychology laboratories at Hokkaido University, Japan, and La Trobe University, Australia. Each laboratory consisted of several rooms for the participants and a control room. Each participant room was equipped with a computer, which was networked to a host computer located in the control room. Upon arrival, each participant was immediately led to his or her room and asked to wait until all the participants had arrived and the computer screen told him or her when to start. Each participant was alone throughout the experiment in a room, aside from times when the experimenter checked on progress. The experiment started when all the scheduled participants arrived, or after about 10 min of waiting if not all the scheduled participants had arrived. Instructions appeared on participants’ screens telling them that they were in a study of ‘interpersonal transactions within and across two societies: Australia and Japan’. After reading the instructions, participants completed five ‘trading periods’, each representing a within-subjects condition. Each trading period started with the display of nationality information (whether the partner in this trading period was Japanese or Australian, and the knowledge information (whether the partner had been told the nationality of the participant). Then, the participant was asked to decide how much of the endowment (AUD 1.00 or 100 yen) to provide to the partner. When the amount to provide to the partner was entered (by clicking a button displayed on the computer screen) participants completed a set of brief postdecisional questions on the computer, which assessed comprehension of the relevant experimental manipulations. After the participants had completed all five trading periods, they completed a postexperimental questionnaire on the computer. In the postexperimental questionnaire, participants’ identification with their own nationality group was measured with seven items.4 They rated how they felt about their own nationality group on each of these seven items on a nine-point scale. In addition to identification with their own nationality group, we also measured in the postexperimental questionnaire the valence of nationality stereotypes on attributes that are considered to be related to group-based trust, by asking which of the two nationality groups better fitted each of four adjectives: trustworthy, generous, likable, friendly (1 = characterizes outgroup more; 9 = characterizes ingroup more; 5 = describes both equally). Finally, the amount the participant made through the ‘trading’ was displayed on the computer screen, followed by a message thanking the participant. Participants were individually paid in their own rooms and separately discharged. The computer program used in the experiment was originally developed in Japan using the Japanese language, and was later translated into English by the first and the second authors. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

181

Results We first examine whether the Japanese and Australian participants identified with their respective nationality more strongly than with the other nationality by analyzing their responses to postexperimental questions. Identification with own nationality group The participants’ levels of identification with their own nationality group was measured with seven items (see end note 4). As a preliminary factor analysis of the seven items on identification with own nationality group in each country revealed that the rivalry item does not fit with the rest, we decided to delete this item from the analysis. We constructed an ingroup-identity scale with the remaining six items (a = 0.82 for Australians; 0.74 for Japanese), of which the mean was not statistically different between the two nationality groups (5.30 for Australians and 4.87 for Japanese; t(103) = 1.63, ns.) Both Japanese participants and Australian participants showed a similar level of identification with the relevant national group. In addition to the above items, we asked participants how close they felt toward their own nationality group compared to the other nationality group and how similar they thought they were to their own nationality group compared to the other nationality group using a ninepoint scale. The mean of these two items, similarity/closeness measure of ingroup identification, was significantly greater than the scale midpoint of 5 among both Australian participants (6.50), t(48) = 9.51, p < 0.0001, and Japanese participants (6.41), t(55) = 7.71, p < 0.0001. The difference between the two nationality groups was not significant. This similarity/closeness measure of ingroup identification positively correlated with ingroupidentification scale both among Australian participants (r = 0.35, p < 0.05) and Japanese participants (r = 0.62, p < 0.0001). Nationality stereotypes We then constructed an ingroup boasting stereotype scale by taking the mean of the four evaluation items (a = 0.73 for Australians; 0.70 for Japanese). A clear nationality difference emerged on this scale; Australian participants thought that Australians were more trustworthy, generous, likable, and friendly than Japanese (5.45, SD = 0.80), whereas Japanese participants thought they were less trustworthy, etc. (4.17, SD = 1.09). The nationality difference was significant, t(103) = 6.77, p < 0.0001. Furthermore, the Australian mean was significantly greater than the scale midpoint of 5.0, t(48) = 3.92, p < 0.01, whereas the Japanese mean was significantly less than the midpoint, t(55) = 5.71, p < 0.0001. These results indicate that Australians have more strongly positive, category-based stereotypes of themselves than do Japanese. The ingroup-boasting stereotype scale was not correlated either with the ingroup identity scale (r = -0.07, ns.) or the similarity scale (r = 0.09, ns) among Japanese participants. Among Australian participants, it was moderately correlated with the similarity scale (r = 0.30, p < 0.05), but not with the ingroup identity scale (r = 0.18, ns). Cooperation levels Level of cooperation in each trial was calculated as the total amount of money transferred to the other person out of the initial endowment (range 0–100), and is displayed in Table 1. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

182

Toshio Yamagishi et al. Table 1 Mean cooperation levels for Australian and Japanese participants Knowledge manipulation

Australian participants (N = 49)

Japanese participants (N = 56)

Total (N = 104)

Ingroup

Mutual Knowledge Unilateral Knowledge

47.76 (26.95) 41.63 (26.88)

27.50 (23.61) 24.11 (22.79)

36.95 (27.07) 32.29 (26.17)

Outgroup

Mutual Knowledge Unilateral Knowledge Unknown (Control)

45.71 (27.39) 44.08 (27.98) 41.22 (25.47)

31.61 (22.22) 29.11 (22.26) 25.18 (17.27)

38.19 (25.64) 36.10 (26.07) 32.67 (22.84)

Partner’s group

Unknown

Standard deviations within parentheses.

Before testing the following set of contrasts, we conducted a nationality by gender of the participant by within-subject conditions (i.e. the five within-subject measures of cooperation), mixed design ANOVA. Only the main effect of nationality, F1,101 = 18.14, p < 0.0001, and the main effect of the within-subject conditions, F4,404 = 2.86, p < 0.05, were found to be significant. Australians were shown to be more cooperative overall than Japanese, as shown in Table 1. A similar pattern was obtained for the cooperation levels in the control condition in which the participant did not know whether the partner was Australian or Japanese, F1,101 = 17.24, p < 0.0001. As gender of the participant did not have any significant effect in this analysis, we dropped gender from the test of our hypotheses. Test of hypotheses We then tested the three hypotheses using a set of two planned orthogonal comparisons. Hypothesis 1 predicts that the cooperation level in IN-MK is higher than that in IN-UK, and Hypothesis 2 predicts that the cooperation levels in OUT-MK is not different from OUT-UK. Two orthogonal planned comparisons were used to test these hypotheses. The two withinsubjects comparisons were crossed with the nationality groups in this analysis. The result of the first comparison shows that the predicted effect of the first contrast is significant, F1,412 = 6.14, p < 0.05. The cooperation rate was higher in the IN-MK condition (M = 36.95 SD = 27.07) than in the IN-UK condition (M = 32.29, SD = 26.17). The interaction effect between the participant’s nationality and the first contrast was not significant, F1,412 = 0.51, ns. These results confirm the first hypothesis about the effect of knowledge manipulation for the ingroup partner, but do not support the third hypothesis concerning cultural differences. The second contrast tests for the effect of the knowledge manipulation for the two outgroup conditions, OUT-MK versus OUT-IK. While the mean cooperation rate in the OUT-MK condition (M = 38.19, SD = 25.64) is slightly higher than that in the OUT-UK condition (M = 36.10, SD = 26.07), the difference is not statistically significant, F1,412 = 1.16, ns. Nor does the contrast have a significant interaction with nationality, F1,412 = 0.05, ns. These results are consistent with the second hypothesis. The results of the above analyses involving contrasts 1 and 2 together support Hypothesis 1 and 2. On the other hand, we did not find support for the third hypothesis concerning the predicted cultural differences. In addition to the above two contrasts, the following two contrasts listed in Table 2 have been tested. Contrast 3 tests the ‘main’ effect of the partner’s group, either ingroup or outgroup, in the usual ANOVA, although this contrast includes data © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

183

Table 2 Orthogonal planned contrasts used in the analysis Partner condition Knowledge condition

Ingroup Mutual

Contrast 1: Hypothesis 1 Contrast 2: Hypothesis 2 Contrast 3: Main effect of the partner Contrast 4: Control condition vs all others

Outgroup

Unilateral

Mutual

Unilateral

Unknown

1

-1

0

0

0

0

0

1

-1

0

/2

0

1

1

–1/2

–1/2

1

1

1

1

/2 /4

/4

/4

/4

-1

in the control condition for calculating the error term. Neither the contrast itself, F1,412 = 1.54, ns, nor the interaction with nationality, F1,412 = 0.72, ns, is significant. The last contrast, contrast 4, represents the difference between the control condition and the other conditions. Again, neither this contrast, F1,412 = 1.82, ns, nor the interaction with nationality, F1,412 = 0.02, ns, is significant. Outgroup-favoring pattern As shown in Table 1, both Australian and Japanese participants cooperated at about the same level in the ingroup/unilateral knowledge condition as in the control condition. This pattern is consistent with the pattern observed in the earlier studies that used the minimal groups rather than enduring categories (Jin & Yamagishi, 1997; Kiyonari, 2002). This pattern indicates that participants were indifferent to the information that their partner was from the same nationality when they had no grounds to expect a favorable treatment from their partner. In contrast, the results shown in Table 1 are unique in that the level of cooperation with outgroup, not ingroup, members was higher than in the control condition. The overall average cooperation level with outgroup partners (37.14) was significantly higher than that with an unknown partner (32.67), t(104) = 2.74, p < 0.01, two-tailed. This difference was observed both among Japanese (30.36 vs 25.18, t(55) = 2.31, p < 0.05, two-tailed) and Australian participants (44.90 vs 41.22, t(48) = 1.70, p < 0.10, two-tailed). In short, both Australian and Japanese participants cooperated with outgroup partners at a higher level than with unknown partners. Furthermore, the difference was more pronounced in the mutual knowledge condition in which the partner was aware of the participant’s nationality (5.52% point greater than in the control condition; t(104) = 3.46, p < 0.01) than in the unilateral knowledge condition (3.43% point greater than in the control condition; t(104) = 1.82, p < 0.08). We conducted additional data analysis to examine whether the unexpected outgroupfavoring pattern could be explained by fairness concerns of the participants (Singh et al., 1999). That is, participants may have been motivated to show that they were fair to the outgroup by behaving in a favorable manner (i.e. by being more cooperative) to the outgroup member. It is likely that such fairness concern is more prominent when the groups are national groups than when they are minimal groups. To examine whether this can explain the unexpected outgroup-favoring behavior pattern, we used the participants’ responses (nine© Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

184

Toshio Yamagishi et al.

point scale) to a postexperimental questionnaire item asking to what extent they were concerned ‘not to discriminate against the other person based on her/his nationality’. The analysis of this postexperimental questionnaire revealed an interesting difference between Australian and Japanese participants. First, Australian participants were much more concerned with fairness (M = 6.02, SD = 2.41) than were Japanese participants (M = 3.88, SD = 1.86; t(103) = 5.14, p < 0.0001). Furthermore, fairness concerns were positively related with the level of outgroup-favoring behavior among Australian participants but not among Japanese participants. Specifically, Australian participants whose level of fairness concerns was 4 or less (where 5 is the scale midpoint) cooperated 6.5 cents less with outgroup partners than with unknown partners, although the difference was not statistically different from zero, t(9) = 1.30, ns). Those whose level of fairness concern was 5 or more (i.e. those who were concerned ‘not to discriminate against the other person based on her/his nationality’ at least mildly) cooperated 6.3 cents more with outgroup partners than with unknown partners, and the difference was significantly different from zero, t(38) = 2.80, p < 0.01. The difference between the two groups was also significant, t(47) = 2.51, p < 0.05. These results indicate that the unexpected outgroup-favoring cooperation behavior was a phenomenon observed only among those who were concerned to be fair to other nationality groups. In contrast, the level of fairness concerns did not significantly affect their outgroup-favoring behavior among Japanese participants. Japanese participants whose level of fairness concern was 4 or less (i.e. those who did not care much about discriminating based on nationality) gave outgroup partners 6.0 yen more than they gave to unknown partners (which is significantly greater than zero, t(33) = 2.13, p < 0.05), whereas those who were concerned not to discriminate based on nationality (i.e. those whose level of fairness concern was 5 or more) cooperated with outgroup partners 3.9 yen more than with unknown partners, t(21) = 1.04, ns, and the difference between the two groups of Japanese participants was not statistically significant, t(54) = 0.47, ns. These results suggest that the outgroup-favoring cooperation among Japanese participants is not unique to those who were highly concerned to be fair to the outgroup. Thus, fairness concerns cannot explain outgroup-favoring behavior pattern among Japanese participants. A possible explanation of the outgroup-favoring behavior among Japanese participants is the high level of outgroup-favoring stereotypes found among them. As presented earlier, Japanese participants held positively valenced stereotypes of Australians. They expected Australians to be more trustworthy and generous than fellow Japanese. As expectations of the partner’s behavior are known to strongly affect players’ own behavior in the Prisoner’s Dilemma, the positive stereotypes of Australians among Japanese participants may explain the residual outgroup-favoring cooperation among Japanese participants. That is, Japanese participants may have expected a higher level of cooperation from Australians than from Japanese. By analyzing the postdecisional questionnaire items asking about the partner’s predicted cooperation level, it was found that they actually expected Australians to cooperate at a higher level than Japanese; they expected that an Australian partner would give them 9.38 cents more (t(55) = 3.84, p < 0.001), on average, than a partner whose nationality is not known, and 9.23 cents more (t(55) = 3.31, p < 0.01) than from another Japanese participant.5 Once expectations are controlled for, the residual outgroup-favoring pattern disappears among Japanese participants; they provided only 0.73 yen more to an Australian partner in the mutual knowledge condition than in the control condition, and the difference was not significantly different from zero, t(55) = 0.28, ns. In the unilateral knowledge condition, they, in fact, provided 3.20 yen less to an Australian partner than in the control condition, although it was not significantly different from zero, t(55) = 0.94, ns. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

185

Identity, stereotypes, and behavior Finally, we assessed whether the level of identification with own nationality group and valence of self-stereotypes, which we analyzed earlier, were related to behavior. Table 3 indicates correlations of the three identity scales (the ingroup identification scale, the similarity scale, and the ingroup boasting stereotype scale) with the level of cooperation in each of the five conditions. As shown in Table 3, only the similarity measure of ingroup identity among Australian participants had a systematic relation with behavior. The similarity measure of ingroup identity is negatively correlated with the cooperation level in all five conditions (which is only not significant in the ingroup/unilateral knowledge condition) among Australian participants. Australian participants who felt similar and close to their own nationality group were generally less cooperative, regardless of the nationality of their partner and the knowledge of the partner’s nationality, than those who felt less similar and less close to their own nationality group. Among Japanese participants, none of the three identity measures was correlated with their behavior. These results clearly indicate that strong identification with the ingroup does not make people behave in a more favorable manner to ingroup members than to outgroup members. The finding that the similarity measure of ingroup identification is negatively related to the general level of cooperation is puzzling. The only possible explanation of this negative effect of similarity on cooperation is that the similarity measure of ingroup identification reflects the level of the authoritarian personality (Adorno et al., 1950) - more authoritarian people feel closer to their ingroup, expect less cooperation from others, and cooperate less. This is an intriguing possibility, and a good topic for future studies.

Table 3 Correlations of the three identity scales (the ingroup identification scale, the similarity scale, and the ingroup boasting stereotype scale) with the level of cooperation in each of the five conditions Partner’s membership and knowledge conditions Scales

Nationality

IN-MK

IN-UK

OUT-MK

OUT-UK

Control

Ingroup identification

Australian Japanese

-0.27* -0.03

-0.05 -0.05

-0.15 -0.01

-0.13 -0.16

-0.23 -0.09

Similarity

Australian Japanese

-0.32** -0.19

-0.20 -0.08

-0.34** -0.07

-0.31** -0.18

-0.30** -0.06

Ingroup stereotype

Australian Japanese

-0.05 -0.25*

-0.09 -0.18

-0.05 -0.00

-0.05 -0.16

-0.00 -0.11

*p < 0.10, **p < 0.05. IN-MK, ingroup/mutual-knowledge condition; IN-UK, ingroup/unilateral-knowledge condition; OUTMK, outgroup/mutual-knowledge condition; OUT-UK, outgroup/unilateral-knowledge condition.

Discussion We examined whether the findings by Yamagishi and colleagues supporting the group heuristic account of ingroup favoritism observed in the minimal group situation with Japanese participants could be generalized to situations involving enduring social categories such as © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

186

Toshio Yamagishi et al.

nationality, on the one hand, and to a non-collectivist population, on the other. The results indicate that the findings are generalizable in both respects; the findings originally observed in the minimal group situation with Japanese participants were replicated in situations involving an enduring category of nationality with non-collectivist, Australian participants. As predicted by the first two hypotheses addressed in this study, the commonality of the knowledge about shared group membership had a positive effect when the participants played PD games with ingroup members, but not when they played the same games with outgroup members. In contrast, the expected effect of culture was not observed. The differential effects of the knowledge manipulation summarized above existed both among Australian participants and Japanese participants. These findings provide us with an important insight concerning the cross-cultural generalizability of the group heuristic. Yamagishi and colleagues (Yamagishi et al., unpubl. data, 2003) argue that generalized exchanges with ingroups are a universal aspect of human group life, and thus humans share similar intuitive beliefs about ingroups as a place for generalized exchanges. On the other hand, they also argue that some aspects of the group heuristic may be culture specific reflecting the differential importance of ingroup cooperation in different cultures. The present study is a first step toward clarifying the universality and cultural specificity of the group heuristic. Based on a rather intuitive view of collectivist and individualist cultures, we predicted that expectations of generalized exchanges would be stronger and, thus, the effects of knowledge manipulation would be stronger, among Japanese participants than among Australian participants. Our study did not support this prediction, suggesting that the core operation of the group heuristic is rather universal. Similarly consistent across the two nationality groups of participants in this study is the lack of effect of group identification measures on the level of cooperation. Except for the unexpected negative effect of the similarity measure of group identity on the level of cooperation among Australian participants, identity measures were not correlated with the level of cooperation in any of the within-subject conditions. This findings is not surprising, however, given that similar lack of correlation has been frequently reported in Prisoner’s Dilemma experiments in which real money rather than symbolic points were used as incentives (Yamagishi & Kiyonari, 2000; Buchan, 2002). Our study also reveals interesting cross-cultural differences. Particularly, the overall level of cooperation was much higher among Australian participants than among Japanese participants; the overall average cooperation of Australian participants was 1.6 times as much as that of Japanese participants. While this result may seem counter-intuitive, it is consistent with previous cross-cultural studies using Prisoner’s Dilemma or similar games indicating that Japanese participants are less likely to cooperate with strangers than participants from individualistic cultures (Toda et al., 1978; Yamagishi, 1988a; Kim & Son, 1998; Buchan et al., 2002). Yamagishi (1988a, b) and Yamagishi and Yamagishi (1994) argues that the high level of mutual cooperation frequently observed in a collectivist culture, in particular in the Japanese culture, is a product of collectivist social institutions that monitor and control individual members’ behavior and sanctions against non-cooperative behavior (see also Hagen & Choe, 1998; Huff & Kelley, 2003). Thus, in a group artificially created in the laboratory where no formal or informal institutions monitor and sanction individuals’ behavior, participants from a collectivist culture tend to expect a lower level of cooperation from others and they themselves cooperate at a lower level than those from an individualist culture. The latter are more used to forming voluntarily based cooperation without monitoring and sanctioning (Putnam, 1994; Yamagishi, 1998). © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

187

We observed an unexpected outgroup-favoring pattern in cooperation, and the results of additional analyses of postexperimental questionnaire items and expectations revealed that this unexpected outgroup-favoring behavior is based on fairness concerns among Australians and outgroup-enhancing stereotypes and a higher level of expectations of cooperation from the outgroup (i.e. Australians) among Japanese participants. The unexpected outgroupfavoring cooperation pattern observed in this study warns us of the need to be cautious in applying the theory that has been derived from laboratory studies to the more ‘realistic’ environment.

Acknowledgments The present study was supported by grants from Japan Society for the Promotion of Science to the first author. He is grateful to the Center for Advanced Studies in the Behavioral Sciences for providing an opportunity to prepare this manuscript. Work in Australia was supported by a grant from the Faculty of Science, Technology and Engineering at La Trobe University, Melbourne to Margaret Foddy and Michael Platow.

End notes 1. Testing the effect of the group heuristic theory involves allocation of tangible and substantial rewards. Studies involving evaluation of ingroup and outgroup members and not actual allocation of tangible rewards are not appropriate for this purpose. Even studies involving allocation of money may not be useful for this purpose if each decision (such as a choice of a matrix) does not involve substantial rewards, as heuristics are not likely to be triggered in such situations (cf. Kiyonari et al., 2000). 2. Australians are usually found to have higher ratings on measures of individualism than are Japanese. See for example, Kashima et al. (1995). 3. This version of the Prisoner’s Dilemma game maintains its definitive structure but is more understandable by participants, and also generalizes to exchange settings. 4. Belongingness (‘How strongly did you have a sense of being a part of the Australian/Japanese group of participants?’); rivalry (reverse item, ‘How strongly did you have a sense of rivalry to the other group?’); typicality (‘How accurate would it be if you were described as a typical Australian/Japanese?’); awareness (‘How often do you acknowledge the fact that you are an Australian/Japanese?’); pride (‘How good would you feel if you were described as a typical Australian/Japanese?’); attachment (‘To what extent do you feel attached to Australia/ Japan?’); definition (‘How strongly do you identify yourself as an Australian/Japanese?’). 5. The knowledge manipulation did not affect Japanese participants’ expectations for Australians’ cooperation level, and thus we pooled the two knowledge conditions in this analysis.

References Adorno, T. W., Frenkel-Brunswick, E., Levinson, D. J. & Sanford, R. N. (1950). The Authoritarian Personality. New York: Harper. Amancio, L. (1989). Social differentiation between ‘dominant’ and ‘dominated’ groups: Toward an integration of social stereotypes and social identity. European Journal of Social Psychology, 19, 1– 10. Billig, M. & Tajfel, H. (1973). Social categorization and similarity in intergroup behaviour. European Journal of Social Psychology, 3, 27–52. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

188

Toshio Yamagishi et al.

Brewer, M. B. (1981). Ethnocentrism and its role in interpersonal trust. In: M. B. Brewer & B. Collins, eds. Scientific Inquiry in the Social Sciences, pp. 214–231. San Francisco, CA: Jossey Bass. Brewer, M. B. (1999). The psychology of prejudice: Ingroup love or outgroup hate. Journal of Social Issues, 55, 429–444. Brewer, M. B. & Kramer, R. M. (1986). Choice behavior in social dilemmas: Effect of social identity, group size, and decision framing. Journal of Personality Social Psychology, 50, 593–604. Brown, R. (2000). Social identity theory: Past achievements, current problems and future challenges. European Journal of Social Psychology, 30, 745–778. Buchan, N. R., Croson, R. T. A. & Dawes, R. M. (2002). Swift neighbors and persistent strategies: A cross-cultural investigation of trust and reciprocity in social exchange. American Journal of Sociology, 108, 168–206. Cosmides, L. (1989). The logic of social exchange: Has natural selection shaped how humans reason? Studies with the Wason selection task. Cognition, 31, 187–276. Cosmides, L. & Tooby, J. (1992). Cognitive adaptations for social exchange. In: J. H. Barkow, L. Cosmides & J. Tooby, eds. The Adapted Mind: Evolutionary Psychology and the Generation of Culture, pp. 163–228. New York: Oxford University Press. de Cremer, D. & van Vugt, M. (1999). Social identification effects in social dilemmas: A transformation of motives. European Journal of Social Psychology, 29, 871–893. Deaux, K. (2000). Models, meanings and motivations. In: D. Capozza & R. Brown, eds. Social Identify Processes: Trends in Theory and Research, pp. 1–14. London: Sage. Diehl, M. (1989). Justice and discrimination between minimal groups: The limits of equity. British Journal of Social Psychology, 28, 227–238. Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117–140. Foddy, M. & Hogg, M. (1999). Impact of leaders on resource consumption in social dilemmas: The intergroup context. In: M. Foddy, M. Smithson, S. Schneider & M. Hogg, eds. Resolving Social Dilemmas, pp. 309–330. Philadelphia, PA: Psychology Press. Hagen, J. M. & Choe, S. (1998). Trust in Japanese interfirm relations: Institutional sanctions matter. Academy of Management Review, 23, 589–600. Hartstone, M. & Augoustinos, M. (1995). The minimal group paradigm: Categorization into two versus three groups. European Journal of Social Psychology, 25, 179–193. Hewstone, M., Argyle, M. & Furnham, A. (1982). Favouritism, fairness and joint profit in long-term relationships. European Journal of Social Psychology, 12, 283–295. Hofstede, G. (1980). Culture’s Consequences. Beverly Hills: Sage. Hogg, M. A. (1992). The Social Psychology of Group Cohesiveness: From Attraction to Social Identity. New York: Harvester. Hogg, M. A. & Abrams, D. (1988). Social Identifications: A Social Psychology of Intergroup Relations and Group Processes. London: Routledge. Huff, L. & Kelley, L. (2003). Levels of organizational trust in individualist versus collectivist societies: A seven-nation study. Organizational Science, 14, 81–90. Insko, C. A. & Schopler, J. (1998). Personal control, entitativity, and evolution. In: C. Sedikides, J. Schopler & C. A. Insko, eds. Intergroup Cognition and Intergroup Behavior, pp. 75–107. Mahwah, NJ: Lawrence Erlbaum. Insko, C. A., Schopler, J. & Sedikides, C. (1998). Personal control, entitativity, and evolution. In: C. Sedikides, J. Schopler & C. A. Insko, eds. Intergroup Cognition and Intergroup Behavior, pp. 109–122. Mahwah, NJ: Lawrence Erlbaum. Jin, N. & Yamagishi, T. (1997). Group heuristics in social dilemma. Japanese Journal of Social Psychology, 12, 190–198 (in Japanese with an English abstract). Jin, N., Yamagishi, T. & Kiyonari, T. (1996). Bilateral dependency and the minimal group paradigm. Japanese Journal of Psychology, 67, 77–85 (in Japanese with an English abstract). Karau, S. J. & Williams, K. D. (1993). Social loafing: A meta-analytic review of social integration. Journal of Personality Social Psychology, 65, 681–706. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

Group-based cooperation

189

Karp, D., Jin, N., Yamagishi, T. & Shinotsuka, H. (1993). Raising the minimum in the minimal group paradigm. Japanese Journal of Experimental Social Psychology, 32, 231–240. Kashima, Y., Yamaguchi, S., Kim, U., Choi, S. C., Gelfand, J. M. & Yuki, M. (1995). Culture, gender and self: A perspective from individualism-collectivism research. Journal of Personality Social Psychology, 69, 925–937. Kim, Y. & Son, J. (1998). Trust, cooperation and social risk: A cross-cultural comparison. Korea Journal, 38, 131–153. Kiyonari, T. (2002). Expectations of a generalized exchange system and ingroup favoritism: An experimental study of bounded reciprocity. Japanese Journal of Psychology, 73, 1–9 (in Japanese with an English abstract). Kiyonari, T., Tanida, S. & Yamagishi, T. (2000). Social exchange and reciprocity: Confusion or a heuristic? Evolution and Human Behavior, 21, 411–427. Kollock, P. (1997). Transforming social dilemmas: Group identity and cooperation. In: P. Danielson, ed. Modeling Rational and Moral Agents, pp. 186–210. Oxford: Oxford University Press. Kramer, R. M. (1991). Intergroup relations and organizational dilemmas: The role of categorization processes. Research in Organizational Behavior, 13, 191–228. Kramer, R. M. & Brewer, M. B. (1984). Effects of group identity on resource use in a simulated commons dilemma. Journal of Personality Social Psychology, 46, 1044–1057. Kramer, R. M. & Goldman, L. (1995). Helping the group or helping yourself? Social motives and group identity in resource dilemmas. In: D. A. Schroeder, ed. Social Dilemmas: Perspectives on Individuals and Groups, pp. 46–67. Westport, CT: Praeger. Kramer, R. M., Pommerenke, P. & Newton, E. (1993). Effects of social identity and interpersonal accountability on negotiator decision making. Journal of Conflict Resolution, 37, 633–654. Leyens, J-P., Yzerbyt, V. & Schadron, G. (1994). Stereotypes Social Cognition. London: Sage. Makimura, Y. & Yamagishi, T. (2002). Ongoing group interaction, ingroup favoritism, and reward allocation. A study of ingroup favoritism with on-going interactions: A reward allocation experiment. Japanese Journal of Psychology, 73, 488–493 (in Japanese with an English abstract). Markus, H. R. & Kitayama, S. (1991). Culture and self: Implications for cognition, emotion and motivation. Psychological Review, 98, 224–253. Meade, R. D. (1985). Experimental studies of authoritarian and democratic leadership in four cultures: American, Indian, Chinese and Chinese-American. High School Journal, 68, 293–295. Mullen, B., Brown, R. & Smith, C. (1992). Ingroup bias as a function of salience, relevance and status: An integration. European Journal of Social Psychology, 22, 103–122. Nisbett, R., Peng, K., Choi, I. & Norenzayan, A. (2001). Culture and systems of thought: Holistic versus analytic cognition. Psychological Review, 108, 291–308. Oakes, P. J. & Turner, J. C. (1990). Is limited information processing the cause of social stereotyping. European Review of Social Psychology, 1, 111–135. Platow, M. J., Harley, K., Hunter, J. A., Hanning, P., Shave, R. & O’Connell, A. (1997). Interpreting ingroup-favouring allocations in the minimal group paradigm. British Journal of Social Psychology, 36, 107–118. Putnam, R. D. (1994). Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press. Rubin, M. & Hewstone, M. (1998). Social identity theory’s self-esteem hypothesis: A review and some suggestions for clarification. Personality Social Psychology Review, 2, 40–62. Singh, R., Choo, W. M. & Poh, L. L. (1999). In-group bias and fair-mindedness as strategies of selfpresentation in intergroup perception. Personality Social Psychology Bulletin, 24, 147–162. Smith, P. B. & Bond, M. H. (1999). Social Psychology Across Cultures, 2nd edn. Boston, MA: Allyn and Bacon. Tajfel, H., Billig, M., Bundy, R. & Flament, C. (1971). Social categorization in intergroup behaviour. European Journal of Social Psychology, 1, 149–178. Tajfel, H. & Turner, J. (1986). The social identity theory of intergroup behavior. In: S. Worchel & W. G. Austin, eds. Psychology of Intergroup Behavior, pp. 7–24. Chicago, IL: Nelson Hall. © Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005

190

Toshio Yamagishi et al.

Thibaut, J. W. & Kelley, H. H. (1959). The Social Psychology of Groups. New York: Wiley. Toda, M., Shinotsuka, H., McClintock, C. G. & Stech, F. J. (1978). Development of competitive behavior as a function of culture, age, and social comparison. Journal of Personality Social Psychology, 36, 825–839. Triandis, H. C. (1995). Individualism and Collectivism. Boulder, CO: West View. Turner, J. C. (1975). Social comparison and social identity: Some prospects for intergroup behavior. European Journal of Social Psychology, 5, 5–34. Turner, J. C., Brown, R. J. & Tajfel, H. (1979). Social comparison and group interest in ingroup favoritism. European Journal of Social Psychology, 9, 187–204. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D. & Wetherell, M. S. (1987). Rediscovering the Social Group: A Self-Categorization Theory. Oxford: Blackwell. van Vugt, M. & de Cremer, D. (1999). Leadership in social dilemmas: The effects of group identification on collective actions to provide public goods. Journal of Personality Social Psychology, 76, 587–599. Wit, A. P. & Wilke, H. A. (1992). The effect of social categorization on cooperation in three types of social dilemmas. Journal of Economic Psychology, 13, 135–151. Yamagishi, T. (1988a). The provision of a sanctioning system in the United States and Japan. Social Psychology Quarterly, 51, 32–42. Yamagishi, T. (1988b). Exit from the group as an individualistic solution to the public good problem in the United States and Japan. Journal of Experimental Social Psychology, 24, 530–542. Yamagishi, T. (1998). Trust: The Evolutionary Game of Mind Society. Tokyo: Tokyo University Press (in Japanese). Yamagishi, T., Jin, N. & Kiyonari, T. (1999). Bounded generalized reciprocity: Ingroup boasting and ingroup favoritism. Advances Ingroup Processes, 16, 161–197. Yamagishi, T., Jin, N. & Miller, A. S. (1998). Ingroup favoritism and culture of collectivism. Asian Journal of Social Psychology, 1, 315–328. Yamagishi, T. & Kiyonari, T. (2000). The group as the container of generalized reciprocity. Social Psychology Quarterly, 63, 116–132. Yamagishi, T. & Yamagishi, M. (1994). Trust and commitment in the United States and Japan. Motivation and Emotion, 18, 129–166.

© Blackwell Publishing Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2005