The Two Psychologies of Conflict Resolution

Home , Conflict (process)

Group Processes & G Intergroup Relations P 1999 Vol 2(2): 99–118 I R The Two Psychologies of Conﬂict Resolution: Differing Antecedents of Pre-Experience Choices and Post-Experience Evaluations

Tom R. Tyler New York University Yuen J. Huo University of California, Los Angeles E. Allan Lind Duke University

The literature on the ‘myth of self-interest’ model of perceived human motivation suggests that people believe that both they and others are more motivated by self-interest than is actually the case. Four studies are reported which test one implication of the myth of self-interest: the psychology of pre-experience preferences and post-experience evaluations will differ. We hypothesize that people arrive at pre-experience preferences for decision-making procedures based upon the belief that they want to maximize their self-interest. Further, they will define their self-interest in material terms. Consequently, they choose procedures that they believe promise them the best material outcomes. However, post-experience evaluations are based upon a different factor – the quality of the treatment received during the course of the procedure. The results of all four studies support the suggestion that the psychology of preference and of evaluation differ as predicted. The findings suggest that preference and choice should be viewed as reflecting different psychological processes. keywords choice, satisfaction, self-interest, the relational model

Miller and Ratner (1996, 1998) have recently Author’s note suggested that there is a discrepancy between Address correspondence to Tom Tyler, the actual and the assumed power of self-inter- Department of Psychology, New York est in human motivation. The essence of their University, 6 Washington Place, Room 550, argument is that people overestimate the inﬂu- New York, NY 10003, USA. ence of self-interest on attitudes and behaviors, [email: [email protected]]

Copyright Ą 1999 SAGE Publications (London, Thousand Oaks, CA and New Delhi) [1368-4302(199904)2:2; 99–118; 007953] Group Processes & Intergroup Relations 2(2) a discrepancy between the real and the per- The development of research on the ‘myth of ceived causes of behavior that Miller and Rat- self-interest’ provides a basis for predicting ner label ‘the myth of self-interest’. Our goal when each set of concerns is more or less here is to use the idea of the myth of self- important. Specifically, the myth of self-interest interest to help resolve a problem which has suggests that people’s preferences among pro- existed for years within the social psychology of cedures might be more heavily influenced by conflict resolution: how to reconcile the instru- instrumental concerns than are their post hoc mental and relational models of reactions to evaluations of those same procedures. Prior to conflict resolution procedures. experiencing a conflict resolution procedure, Two models have been put forward to explain people, believing themselves to be concerned people’s evaluations of conflict resolution pro- about the favorability of their outcomes, might cedures: the instrumental and the relational. choose among conflict resolution procedures The instrumental model suggests that when based upon estimates of the likelihood of dealing with others people are concerned about obtaining favorable outcomes through each the favorability of their outcomes, and define procedure. However, after they actually ex- outcomes in terms of material gains and losses perience a procedure people may evaluate it (see e.g. Brett, 1986; Kurtz & Houlden, 1981). relationally, because the experience of disre- The relational model argues that people are spectful treatment is so distasteful and carries concerned about the identity implications of such negative identity-relevant imagery that how they are treated in the course of the conflict even a positive material outcome cannot over- resolution experience, and draw identity infor- come these negative aspects of the experience. mation from the treatment they receive from In other words, there may be two psycholo- others (Lind & Tyler, 1988; Tyler & Lind, 1992). gies of disputing. The first is a self-interested Each model assumes a single explanation for psychology of disputing, which people think pre-experience choices among and post-experi- reflects their desires. This psychology shapes ence evaluations of disputing procedures and pre-experience preferences and choices among experiences, and each model can claim some procedures. The second is a relational psychol- support in the research literature. ogy of disputing. This model manifests itself in There are studies that clearly show strong post-experience evaluations and in behavioral links between positive reactions to procedures reactions to conflict resolution decisions. and a cluster of instrumental variables, such as the perceived favorability of their outcomes and The ‘myth of self interest’ estimates of control over those outcomes (e.g. Brett & Goldberg, 1983; Houlden, LaTour, A core feature of cultural ideology is its depic- Walker, & Thibaut, 1978; Kurtz & Houlden, tion of the nature of human motivation (Miller 1981). Other studies show just as clearly that & Ratner, 1996, 1998). Miller and Ratner use there are strong links between the quality evidence from a variety of sources to suggest of the treatment received from others, i.e. dig- that within American society there is a wide- nified or respectful treatment, the considera- spread belief that all people are motivated to tion of one’s views, etc., and positive reactions act based upon their self-interest. In a series of to procedures (e.g. Lind, Erickson, Friedland, laboratory studies, they demonstrate that peo- & Dickenberger, 1978; Lind et al., 1990; Tyler, ple overestimate the influence of personal self- Rasinski, & Spodick, 1985). In a previous analy- oriented interests upon their own and others’ sis of this body of evidence, Lind and Tyler attitudes and behaviors (Miller & Ratner, (1988) suggested that both self-interest and 1998). In other words, people believe that both relational concerns play a role in reactions to they themselves and others shape their behav- procedures, but even if this analysis is correct, it iors to maximize their individual self-interest. remains unclear when one concern or the Further, people define their self-interest in other predominates. terms of material gains and losses. The concept

100 Tyler et al. the two psychologies of conflict resolution of self-interest itself does not indicate what in turn can lead to post-experience evaluations people value when they deal with others. How- that are driven by an altogether different set of ever, theories of self-interest have typically concerns. focused upon material goods. And the self- The studies just mentioned, along with many interest that is the focus of the Miller and other studies of evaluations of disputing pro- Ratner work is the exchange of material goods. cedures and experiences (see Lind & Tyler, Hence, while it might be argued on theoretical 1988, and Tyler, Boeckmann, Smith, & Huo, grounds that being treated with dignity and 1997, for reviews) show that selfish, material, respect (i.e. ‘relational concerns’; Lind & Tyler, outcome concerns generally are less important 1988; Tyler & Lind, 1992) is a valuable and to people than are the concerns about the desired outcome, people themselves think of social relationships that exist or arise within the outcomes in more material terms. This is con- disputing context. In particular, this body of sistent with the myth of self-interest, which research has shown that people place great suggests to people that they value material gains weight on such things as whether they are and losses. treated politely and with respect, whether the Because choosing a conflict resolution proce- conflict resolution process allows them some dure, or indicating one’s preference a priori for fundamental dignity, and whether their views using one procedure or another, is an exercise and needs are considered. Tyler and Lind in predicting both one’s own and one’s oppo- (1992) argue, in the course of proposing a nent’s behavior and reactions, work on the ‘relational model’ of authority and conflict myth of self-interest would predict a strong resolution, that these quality of treatment vari- effect for the perceived favorability of the pro- ables are important because they are seen as cedure’s outcome. That is, we expect people to indicative of whether a sound social relation- choose, or to express a preference for, pro- ship exists with other disputants or third parties cedures that they think will give them what they within the disputing context (see also Lind, in want. A closely related idea, especially given the press). The relational model argues that people conflicting desires that people can anticipate in are concerned about their standing in groups an interpersonal conflict, is that people will and other social relations, since their member- pick procedures that give them the most con- ship in such social entities is an important trol over the outcome, i.e. that allow the person aspect of their social self (Smith & Tyler, 1997). to dictate that his or her chosen outcome People use their treatment by others as one prevail over the chosen outcome of the other, index of social status and, as a consequence, supposedly equally selfish, party to the dispute. focus upon treatment by others when reacting to their personal experiences. Three aspects of The relational model treatment are particularly relevant – evidence of neutrality, trust in the motives of others, and A substantial and growing body of research treatment with respect and dignity (labeled shows that people are not so self-interested as generally as ‘relational concerns’; Tyler, 1989b; the myth of self-interest portrays them as being. Tyler & Lind, 1992). For example, research on reactions to real- A reconsideration of the literature on self- world experiences with police or courts shows interest versus relational concerns in the evalu- that people are more concerned about issues of ation of conflict resolution procedures reveals morality and procedural fairness than about that some of the strongest support for self- getting the best material outcome for them- interest antecedents of procedural evaluations selves (Lind et al., 1990; Tyler, 1990). Hence, is seen in studies that examine a priori prefer- the factors that really affect attitudes and behav- ences for one conflict resolution procedure ior may be substantially different from those over another (e.g. Kurtz & Houlden, 1981), that people, using the myth of self-interest, while the strongest support for relational ante- think will affect attitudes and behavior, and this cedents is seen in studies that examine post hoc

101 Group Processes & Intergroup Relations 2(2) evaluations of experiences with one procedure ple experience with their conflict resolution or another. This pattern of support for the two experiences may come in part from their own models of evaluation makes sense if one choices, or more specifically from the criteria assumes that actually experiencing a given pro- they use to choose a way of resolving the con- cedure or process is more likely to engage the flict. If procedures are chosen on one set of social identity processes that are assumed to criteria, but evaluated on another, it is not underlie relational concerns. In this respect, surprising that the choices fail to result in much actual experience is different than is ‘simulat- satisfaction for those with the original disputes. ing’ one’s reactions to an as-yet-unexperienced conflict resolution. This argument seems quite Study 1 reasonable, since relational concerns are pre- cisely the sort of social considerations that are The first study examines preference in a situa- neglected in the set of beliefs that constitute tion in which people are choosing among vari- ‘the myth of self interest’. ous types of behavior which they might engage If the psychology of a priori choice and in to resolve a social conflict. Two stages of preference differs from the psychology of post conflict resolution are identified: pre-experi- hoc evaluation in the way we are suggesting, it ence preferences among alternative procedures would explain another common finding in the for resolving a conflict, and the post-experience study of conflict resolution. Studies of conflict evaluations of one’s actual experiences with a resolution suggest that people often end up conflict resolution procedure and its results. feeling dissatisfied with the behavioral choices This study will examine the influence of two they make about how to deal with their conflicts psychological antecedents on preference and with others, even when those choices appear to evaluation. Those antecedents are: instrumen- be freely made. Lind, Huo, and Tyler (1994), tal judgments of outcome favorability and out- for example, find that, irrespective of how they come control; and relational judgments about try to resolve conflicts, people’s post-experience the quality of treatment by others. Prior to a evaluations of their choices are on average procedure participants are asked about negative. whether a procedure, if chosen, is likely to lead People often seem to choose behaviors, and to a favorable outcome and/or to dignified then regret their choices, feeling post-experi- treatment by others. Following their actual ence regret and dissatisfaction with their dis- experience using a procedure participants are pute resolution experiences. Consistent with asked whether the procedure in fact led to a this argument, field studies of mediation have favorable outcome and whether, during the found a persistent discrepancy between choice procedure, they were in fact treated with dig- and post-experience evaluation (MacCoun, nity. By examining the strength of the relation- Lind, & Tyler, 1992; Tyler, 1989a). The studies ship between instrumental and relational judg- reviewed by MacCoun et al. and Tyler show that ments and pre-experience preferences and it is difficult to induce conflicting parties to post-experience evaluations, we can test the choose mediation, a procedure that lessens propositions we advance above. third party control and that makes the sure This study focuses on the interpersonal con- attainment of any given outcome problematic, flicts that occur in the participant’s work and but that also engenders feelings of social personal life. Participants are first asked to engagement and involvement. However, once consider a hypothetical interpersonal conflict experienced, mediation procedures tend to be and to indicate how they think they would evaluated quite positively by the disputing par- prefer to deal with that conflict. Seven potential ties, and to produce decisions that the parties approaches are presented to the participants: are likely to find satisfactory and voluntarily ignoring the conflict, giving in to the other accept. person, exerting social influence, trying to per- We argue that the discontent that many peo- suade, negotiating, seeking the services of a

102 Tyler et al. the two psychologies of conflict resolution mediator, and seeking the services of an arbi- cedures were compared: (1) variables trator. Since participants may not be familiar associated with the favorability of the outcomes with these various procedures, each is pre- expected or experienced under the procedure sented in a short paragraph that explains what in question and (2) variables associated with the the procedure involves. With each procedure quality of the treatment the participant the psychological basis of preferences was estab- received or expected to receive from author- lished by examining the relationship between ities. the expected consequences associated with In the study participants responded to a two- using that procedure (e.g. obtaining a favorable part questionnaire. In the first part of the outcome; receiving dignified treatment) and questionnaire, participants were asked to con- preferences for using that procedure (I would sider a hypothetical scenario in which they were like to use this procedure to resolve a conflict of in conflict with another person. Several features this type). of the scenario were manipulated: (1) the rela- Participants were then asked to give post- tionship of the disputants (friend vs. acquaint- experience evaluations of a recent real conflict ance); (2) the issues in dispute (money vs. in their lives. The psychological basis of these insult); and (3) the similarity of the ethnic evaluations was examined by asking the partici- background of the other disputant to that of pants retrospective questions about what hap- the participant (the same vs. different). Partici- pened during the enactment of the procedure pants were asked to evaluate seven different (e.g. ‘I received a favorable outcome’; ‘I was ways that they might deal with the dispute: treated with dignity’) and their post-procedural ignoring it, giving in, using friends to pressure evaluations (‘I would use this procedure again the other person (social influence), trying to in a similar situation’). persuade the other person, negotiating a mutually acceptable solution, using mediation, and Method using arbitration. The procedure was a within- Participants There were 774 participants in subject factor – each participant considered all this study. Of these, 324 were undergraduates at of the seven procedures. the University of California at Berkeley, who In the second part of the study participants participated either in partial fulfillment of a were asked to recall an actual interpersonal requirement for an introductory psychology conflict in which they had been involved and to class (n ϭ 183) or for payment (n ϭ 141). answer questions regarding that conflict. The another 206 were undergraduates at the Chi- recall instructions included the ethnicity nese University of Hong Kong, who completed manipulation, with half of the participants the questionnaires as part of a course require- asked to recall a conflict with another person of ment; and 244 were students at the University of the same ethnic background, one-half with a Osnabrück, who also completed the question- different ethnic background. The ‘ethnicity of naires as part of a course requirement. the other party condition’ to which the participant was assigned in the second part of the Design This study had two parts: a pre-ex- study was the same as that to which he or she perience assessment of hypothetical conflict had been assigned in the first part of the study. resolution preferences and post-experience However, their relationship to the other party evaluations of real conflict experiences. In the and/or the issues in dispute may have differed first part, participants considered hypothetical from those presented in the hypothetical sce- disputes and indicated how they would respond nario. to them. In the second part, participants described how they actually dealt with real disputes Questionnaire and how they subsequently evaluated their Preference-related measures Participants were first disputing experience. In each case two antece- asked about their preferences for using each of dents of reactions to conflict resolution pro- the seven approaches to the conflict they were

103 Group Processes & Intergroup Relations 2(2) considering: ignoring the situation and avoid- experience with a same or other ethnicity other ing contact with the other person; giving in to (each participant received the same level of the the other person; using their power and their ‘ethnicity of other person’ manipulation on the influence over friends and family to pressure two parts of the experiment). Two aspects of the other person; trying to persuade the other the dispute were established in the question- person that they are right; negotiating to find a naire. First, participants indicated whether or compromise that both parties will feel is accept- not the person they dealt with was a friend able; seeking the assistance of an impartial third (‘Prior to the dispute, I felt very close to this party for suggestions; and seeking the assistance person’). Second, participants were asked to of an impartial third party to make a binding indicate the cause of the dispute (Did this decision. We assessed the participants’ prefer- dispute occur because: ‘The other person was ence for using each procedure by asking them rude or impolite to you’ and/or because ‘The to indicate ‘how likely’ they thought it was that other person refused to give you money or they ‘would try to deal with the problem’ in other things they owed you’). each of these seven ways. Participants were asked to evaluate their feel- Participants then rated each of the seven ings ‘during and after’ the dispute. In partic- procedures on a series of psychological dimen- ular, they were asked to evaluate the favorability sions. Instrumental judgments were measured of their outcome, and the quality of their treat- by answers to the question: ‘How likely is it that ment during the procedure. To assess outcome each method would lead to the solution you favorability, participants were asked to rate wanted?’ (1: ‘very unlikely’ to 7: ‘very likely’, their agreement or disagreement that: ‘The respectively). Relational judgments about the outcome of the dispute gave me what I wanted’ quality of people’s treatment by others were and ‘The outcome was very favorable to me’ measured by answers to five questions asking: (Cronbach’s alpha ϭ .83). Quality of treatment ‘How likely is it that the problem would be was established by asking whether the partici- resolved in a dignified way’; ‘. . . that issues pant agreed or disagreed with statements that: would be brought into the open’; ‘. . . that ‘issues were brought into the open’; ‘the dis- relevant facts would be considered’; ‘that their pute was resolved in a dignified way’; ‘my rights views would be considered’; and ‘. . . that their as a person were protected’; ‘my views were rights as a person would be protected?’ These considered and taken into account’; and ‘what I five items were combined into a single scale wanted was considered’ (Cronbach’s alpha ϭ (Cronbach’s alpha ϭ .90). .85) Two dependent variables were assessed: ‘How Two dependent variables were examined. likely is it that you would use the procedure to First, participants were asked how likely they deal with the problem?’ and ‘How likely is it would be to try to use the same procedure in that the amount of conflict with the other the future if they had a dispute of a similar type. person, now or in the future, would be Second, they were asked whether the conflict reduced?’ by using the procedure. they experienced changed their relationship with the other person ‘for the better’. Post-experience evaluations Participants were next asked to think about a ‘real interpersonal Analytic considerations Altogether, 774 par- dispute’ that they had recently experienced. ticipants completed questionnaires, with each They were asked to indicate what they did in giving information about seven procedures. resolving their own conflict. Participants were This produced 5418 procedural evaluations. told at this point that ‘we are interested in what For the preference analysis these 5418 proce- you actually did to resolve the dispute, not what dural evaluations were used as the basis for you would hypothetically do’. analysis. Each evaluation was weighted by One aspect of the actual dispute was manipu- 0.142857 (1/7) to restore the sample size to 774 lated: participants were asked to recall an and to avoid inflating the significance of statis-

104 Tyler et al. the two psychologies of conflict resolution tical tests (see Lind et al., 1994, p. 282, for a to address this issue. In the regression analysis discussion of this methodology). Post-experi- outcome favorability and quality of treatment ence evaluations used the sample of 774 partici- were entered as independent variables. They pants as the unit of analysis.1 were used to predict the dependent variables: preference for dealing with the problem in a Results particular way and ratings of the likelihood of Dissatisfaction In keeping with the finding of increased conflict. The manipulated character- most prior research on dispute resolution, this istics of the conflict presented to the partici- study also finds that people typically prefer to pants (for example, the conflict with a friend or settle their disputes through negotiation and acquaintance), as well as participant nationality persuasion (see Lind et al., 1994, for a more were included as controls.2 The results are detailed discussion of this point). In this study shown in Table 1. people are generally found to express post- The results shown in Table 1 indicate that experience dissatisfaction with the manner in preferences for using a particular procedure to which they tried to resolve their conflicts, even deal with a conflict are most strongly influ- when they use their preferred methods. On enced by judgments concerning the likelihood average, participants ended up expressing neg- that using the procedure will lead to a favorable ative feelings in the post-experience ratings. An outcome (beta ϭ .54, p < .001). Assessments overall affect scale indicated that mean affect about whether or not the procedure is likely to was negative for subgroups of participants using lead to high quality relational treatment by each of the seven procedural choices. In other others appear to have a lesser influence (beta ϭ words, no matter how participants had chosen .17, p < .001). A statistical comparison of the to resolve their conflict, their average post- magnitude of the correlations linking prefer- experience feeling was negative. This supports ence to instrumental and relational concerns is the suggestion that people are dissatisfied with significant (t(771) ϭ 5.34, p < .001; see Blalock, the procedures they choose to resolve disputes. 1972, p. 407, for a description of a procedure for comparing the magnitude of two correla- Pre-experience preferences The first concern tions using a t test). of this study is with the psychological ante- Judgments about the likelihood of conflict cedents of preferences for different methods of with others are strongly affected by judgments conflict resolution. Regression analysis was used about the likelihood the respondent would

Table 1. Preferences among anticipated procedures: Study 1

Want to deal with problem Conﬂict that will occur if using this procedure the procedure is used

Beta weights Expectation of favorable outcome .54*** .16*** Expectation of treatment with dignity .17*** .27*** Dispute characteristics Friend? .01 .04 Same ethnicity? .00 Ϫ.01 Topic of dispute Ϫ.01 .00 Nationality US vs. Hong Kong Ϫ.03 Ϫ.09* Germany vs. Hong Kong Ϫ.07* Ϫ.02 adj. R2 45%*** 16%***

* p < .05; ** p < .01; *** p < .001. Note: Entries are beta weights.

105 Group Processes & Intergroup Relations 2(2) receive high quality relational treatment (beta occurred? Again, regression analysis was used to ϭ .27, p < .001), and are also significantly examine this relationship. The results are influenced by the expected favorability of the shown in Table 2. The first question asked is respondent’s outcomes (beta ϭ .16, p < .001). whether or not the disputant would use the The magnitude of these two relationships is not same procedure again in the future in a similar significantly different. situation. The results indicate that both rela- Dispute characteristics – whether the dispute tional concerns (beta ϭ .30, p < .001) and involved a friend or a person of the same ethnic instrumental concerns (beta ϭ .27, p < .001) background, or issues of honor or money – had influenced judgments about what people would little influence on either of the dependent do in the future. These two relationships are variables. Similarly, the ethnic background of not significantly different in their magnitude. the disputants had little influence. These find- In contrast to pre-experience preference judg- ings suggest that people make choices among ments, instrumental concerns were not the procedures based primarily upon their sense of dominant factor. They were equal in impor- which procedures will lead to favorable out- tance to relational concerns in post-experience comes. Hence, before they enter into a conflict analyses. resolution procedure people act as if they are Prior to experiencing the procedure people the self-interest seekers described by the cul- indicated that they felt that the quality of treat- tural ‘myth of self-interest’. In terms of prefer- ment by others would be as strong as the ences, self-interest is not a myth. Interestingly, favorability of the outcomes they received in the impact of the myth of self-interest is not shaping future conflict. After the experience, ubiquitous: the respondents recognized that however, their feelings about future conflict relational issues would play at least as strong a were only influenced by the quality of treat- role in shaping future conflict as would out- ment by others. Outcome favorability had no come favorability. Nonetheless, their views significant influence on social climate. A com- about which procedure to use were primarily parison of the magnitude of these relationships shaped by instrumental judgments. indicates that quality of treatment was signifi- How do people evaluate their dispute resolu- cantly more influential (t(771) ϭ 5.17, tion experiences after those experiences have p < .001).

Table 2. Post-experience evaluations of procedures: Study 1

Would you use this Conﬂict procedure again? has lessened

Beta weights Favorability of outcome .27*** .03 Quality of treatment .30*** .34*** Dispute characteristics Same/diff. background Ϫ.02 Ϫ.04 Close to other person? Ϫ.05 .10** Cause – rudeness? Ϫ.08* Ϫ.06 Cause – money? Ϫ.07* .00 Nationality US vs. Hong Kong .10* Ϫ.10* Germany vs. Hong Kong .03 Ϫ.23** R2 31%*** 18%***

* p < .05; ** p < .01; *** p < .001. Note: Entries are beta weights.

106 Tyler et al. the two psychologies of conflict resolution

Discussion ate seven different ways that they might deal The findings of Study 1 are consistent with the with the dispute: ignoring it, giving in, using argument advanced by the myth of self-interest. friends to pressure the other person (social As predicted by our analysis of the implications influence), trying to persuade the other person, of the myth of self-interest and the relational negotiating a mutually acceptable solution, model of authority, people’s preferences were using mediation, and using arbitration. Each more strongly guided by their instrumental participant considered all seven possible pro- judgments than were their post-experience cedures, each of which was explained in a brief evaluations of their experiences. paragraph. In the second part of the study participants Study 2 were asked to recall a recent personal interaction with a professor. They were then asked Study 2 is based upon samples of students from questions about that interaction. Participants the United States and Japan. The study has two were asked to focus on the last time they parts. The first part replicates Study 1 by asking discussed issues such as getting help with work, participants to respond to a hypothetical sce- settling a conflict with the professor, discussing nario in which they are involved in a dispute grades, or seeking help to resolve a conflict with with another person. In the scenario, as in others. Study 1, the nature of (1) the dispute (money or insult), (2) the relationship among the dis- Questionnaire putants (friends or acquaintances), and (3) the Pre-experience preferences Participants were first similarity of their social backgrounds (similar, asked about their preferences for using each of different) were varied. In the second part of the the seven approaches to the hypothetical con- study students were asked to recall a real inter- flict outlined in Study 1. Participants then rated action with a professor or teacher and to each of the seven procedures on the same respond to a series of questions about that psychological dimensions used in Study 1. interaction. These questions assessed the basis Instrumental judgments were indexed by esti- of post-experience evaluations of the experi- mates of how likely it is that each method would ence. lead to the solution the participant wanted. Relational judgments reflected the likelihood Method that each method would lead to a high quality Participants The participants in Study 2 were of treatment. Expected quality of treatment was college students. In the United States, 181 stu- indexed using the same five questions used in dents completed the questionnaire as partial Study 1 (Cronbach’s alpha ϭ .90). The two fulfillment of a course requirement. In Japan, dependent variables measured in Study 1 were 165 students also completed the questionnaire again assessed and were used in this study. as partial fulfillment of a course requirement. Post-experience evaluation Participants were Design Participants responded to a two-part asked to evaluate their feelings ‘during and questionnaire. In the first part of the ques- after’ their recent dispute with a professor. tionnaire, participants were asked to consider a Outcome favorability was assessed by asking hypothetical scenario in which they were in participants ‘how favorable’ the outcome was to conflict with another person. The character- them. Quality of treatment was assessed by istics of that scenario were varied in: (1) the asking participants to respond to five questions: relationship of the disputants (friends vs. ‘How seriously were your views considered?’; acquaintances); (2) the issues in dispute ‘How much respect was shown for your rights?’; (money vs. insult); and (3) the similarity of the ‘How much consideration was given to what you person’s background to their own (similar vs. needed and wanted?’; ‘How politely were you different). Each participant was asked to evalu- treated?’; and ‘To what extent did the discus-

107 Group Processes & Intergroup Relations 2(2) sion center on facts, not on irrelevant issues?’ cedures. In the regression analysis outcome (Cronbach’s alpha ϭ .82). favorability and quality of treatment were The dependent variable was the willingness entered as independent variables. They were to use the procedure again. Participants were used to predict the dependent variables: partici- asked two questions, which were averaged to pant’s preference for dealing with the problem form an index of willingness to use the proce- in a particular way and ratings of the likely dure again: ‘If I had it to do over again, would I future conflict involved (Table 3). take this problem to the professor again?’ and As in Study 1, the results of the Study 2 ‘If a friend were in the same class, would I analysis suggest that participants’ decisions suggest that he/she take problems to the same among possible dispute resolution procedures professor?’ (Cronbach’s alpha ϭ .84). were primarily influenced by their evaluations of the likelihood of gaining a favorable outcome through the procedure (beta ϭ .47, Analytic considerations A total of 346 partici- p < .001). There was also an apparently lesser pants were interviewed, with each giving infor- influence of expectations about the quality of mation about seven procedures. This leads to treatment that would be received from each 2422 procedural evaluations. As in Study 1, for procedure (beta ϭ .16, p < .01). A comparison the preference analysis these 2422 procedural of the strength of these relationships indicates evaluations were weighted by 0.142857 (1/7) to that the connection with expectations of favor- produce a sample size of 346. Participants also able outcome is significantly stronger (t(343) ϭ reacted to their own experience with the pro- 2.67, p < .05). Nationality had no direct influ- fessor, however it was resolved. The sample of ence on preferences (beta ϭ .02, ns).3 346 participants was used as the basis for this Further, judgments about the likelihood of analysis. gaining a favorable outcome shaped people’s expectations about how much conflict will Results occur (beta ϭ .23, p < .001), while judgments Preference The first concern of this study is about the quality of treatment had no direct with the psychological antecedents of prefer- effect (beta ϭ .09, ns) These two relationships ences for different methods of conflict resolu- were not significantly different in magnitude. tion. Regression analysis is used to examine the Again, nationality had no direct influence (beta relationship between judgments about the ϭ .01, ns). Further, the characteristics of the likely consequences of using particular pro- dispute – i.e. whether it involved a friend or a cedures and preferences for using those pro- person of the similar background, or issues of

Table 3. Preferences among anticipated procedures: Study 2

Want to deal with problem Conﬂict that will occur if using this procedure the procedure is used

Beta weights Expectation of gain .47*** .23*** Expectation of treatment with dignity .16** .09 Dispute characteristics Friend? Ϫ.01 .01 Same ethnicity? .01 .01 Topic of dispute .00 .01 Nationality .02 .01 R2 35%*** 9%***

* p < .05; ** p < .01; *** p < .001. Note: Entries are beta weights.

108 Tyler et al. the two psychologies of conflict resolution honor or money – had little influence on any of that these findings support those of Study 1, the dependent variables. where the pre- and post-experience settings How do people evaluate their experiences were comparable. after those experiences have occurred? Again, regression analysis was used to examine this Discussion relationship. The results are shown in Table 4. The findings of Study 2 are consistent with The dependent variable is whether or not the those of Study 1 and with the argument disputant would use the same procedure again advanced by the myth of self-interest and the in a similar situation in the future. The results relational model. Again, people’s preferences indicate that both relational concerns (beta ϭ before they have an experience are more .50, p < .001) and instrumental concerns (beta strongly guided by their instrumental judg- ϭ .18, p < .001) influence judgments about ments than are their post-experience evalu- what people would do in the future, and that ations of their experiences. As in Study 1, a relational concerns were the stronger determi- discontinuity is seen between the factors shap- nant of returning to the professor with a similar ing the willingness to use a procedure and the problem. The difference between the two cor- factors shaping the willingness to use a social relations is significantly different in magnitude procedure again after experiencing it once. (t(343) ϭ 3.81, p < .001). In contrast to pre- Quality of treatment becomes more important experience preference judgments, instrumen- after people have had personal experience with tal concerns are not the dominant factor. They a procedure. are significantly less important than are relational concerns. Study 3 One important caution in interpreting the Study 2 data, which is not an issue in the Study 1 Studies 1 and 2 compare preferences and evalu- data, is that the situations participants are con- ations in a situation in which people are choos- sidering when making pre-experience and post- ing among alternative ways of handling disputes experience judgments are different. In making and problems. While choice among procedures pre-experience assessments about the desirabil- is examined in early studies of procedural jus- ity of procedures participants were considering tice (Thibaut & Walker, 1975) and is important disputes in their everyday life. However, when in studies of negotiation behavior, many proce- making post-experience evaluations of an dural justice studies examine not choices experience, they were evaluating an experience among procedures, but the choice of whether in which they had dealt with a university pro- or not to use one particular procedure in a fessor or instructor. Hence, the situations are specified sequence (see e.g. Lind, Kulik, not the same. Nonetheless, it is encouraging Ambrose, & de Vera Park, 1993; Lind et al., 1990). For example, in legal disputing, a plain- tiff must first decide whether or not to engage Table 4. Post-experience evaluations of procedures: the legal system at all in his or her dispute, then Study 2 he or she must decide whether or not to use Would you use this negotiation to resolve the dispute, then if nego- procedure again? tiation fails he or she must decide whether or not to go to mediation and/or trial. Study 3 Beta weights extends our test of the hypotheses to such a Favorability of outcome .18*** setting. Study 3 examines procedural choices Quality of treatment .50*** Nationality .16*** and procedural evaluations among people who R2 43%*** are deciding whether to sue someone who has caused them bodily injury. * p < .05; ** p < .01; *** p < .001. We expect the same basic discontinuity to Note: n ϭ 346. play out in the sequential choice of procedures.

109 Group Processes & Intergroup Relations 2(2)

If people are making choices using the cultural injury, how much money did you think you ideology represented by the ‘myth of self-inter- might be able to get?’ They were also asked est’, they should make their choices about ‘How did you think you would be treated by the whether to pursue a claim in instrumental people you would have to deal with to get terms. They should believe that their satisfac- compensation?’ These two items represented tion will depend upon whether they receive the measures of instrumental and relational resources from the other person, not on how expectations, respectively. (The dollar estimate that person treats them. However, once they was transformed using a logarithmic transfor- have been involved in a claiming experience, mation to normalize it. Previous research has people should evaluate that experience by shown that the relationship between dollar esti- considering the quality of their treatment by mates or dollar reports of outcomes and any others. psychological variable tend to be minimal unless a transformation of this sort is used; see e.g. Lind et al., 1993; Lind, Greenberg, Scott, & Methods Welchans, 1998.) As a measure of the extent to The data for Study 3 were taken from a national which the participants had activated the tort survey of 2555 people who were either injured claiming system as a dispute resolution proce- seriously enough to miss a day of their normal dure, we computed a ‘claiming procedure activity or who had filed some sort of claim as a index’, which is a count of the number of result of an injury. The 2555 people were identi- claiming options (e.g. direct claiming from the fied using a screening survey of approximately source of the injury, contacting an attorney, 10,000 households in the United States. (See filing a lawsuit) activated by the respondent. Hensler et al., 1991, for a complete description Participants who did in fact claim and who of the survey design, methods, and findings.) had received some monetary compensation Part of the interview administered to these were also asked two questions about the out- participants asked about their thoughts and come of the dispute and one question about the actions with respect to claiming compensation fairness of the treatment they had experienced, for their injury and their feelings about any a relational issue. The two outcome questions claiming experience. asked ‘What was the total payment for all the The pre-experience analyses reported here claims and lawsuits against other people or are based on 305 of these participants: those organizations?’ (using the logarithm of the who reported that they had considered claim- amount in the analyses) and, to tap a more ing for compensation from some person, organ- subjective assessment of outcome, ‘How ization, or institution and who also offered an adequate did you think the amount you estimate of how much they would obtain in received was?’ The post-experience fairness compensation for their injury. The post-experi- question asked ‘How did you feel you were ence analysis is based upon 223 participants treated by the different people in the process: who received a monetary settlement and who very fairly, somewhat fairly, somewhat unfairly, could therefore report their satisfaction with or very unfairly?’ The dependent variable in the the outcome of the claim. post-experience analyses was the respondent’s rating in response to a question that asked ‘At Measures The participants were asked if they this point, how do you feel about the process had ever thought about or talked about claim- you had to go through to try to get compensa- ing compensation for their injury (claims to tion from the person or group responsible for their own insurers or workers compensation the injury?’ claims were not included because most of these claims are not disputed). If they responded Results affirmatively, they were asked ‘When you first The results of the pre-experience report data considered trying to get compensation for this analyses are given in Table 5. As can be seen,

110 Tyler et al. the two psychologies of conflict resolution

Table 5. The pre-experience antecedents of claiming Table 6. Post-experience satisfaction (log dollars choice: Study 3 gained): Study 3

Expected outcome .12* Favorability of outcome Ϫ.06 (log of expected outcome) (log of dollars received) Expected quality of treatment .05 Quality of treatment received .54*** from others adj. R2 2%* adj. R2 33%*** * p < .05; ** p < .01; *** p < .001. Note: Entries are beta weights. * p < .05; ** p < .01; *** p < .001. Note: Entries are beta weights. the choice of whether and how much to invoke the context of real-world choices of substantial claiming procedures is significantly linked to policy importance. the expected outcome of the claim, but not to Study 3 also has an additional importance. In perceptions about the likely treatment that Studies 1 and 2 people were not evaluating the would be received in pursuing the claim. How- same conflict during the pre-conflict and post- ever, the difference in the importance of these conflict stages. They were evaluating the same instrumental and relational indices is not sig- type of conflicts, but not the same specific nificant (t(303) ϭ 1.22, ns), suggesting that conflicts. However, in Study 3 people were both factors have some role in shaping claiming evaluating the same conflict at both stages of decisions. the analysis. Hence, it is reassuring that the The analyses of the post-experience reports findings of Study 3 are parallel to those of involved two regression analyses, both predict- Studies 1 and 2. ing overall satisfaction with the claiming pro- Further, Studies 1 and 2 compare hypothet- cedures experienced. Table 6 reports the ical conflicts, about which participants make regression using the log dollars received and choices, and real conflicts, about which they the quality of treatment rating; Table 7 reports make evaluations. While there is no a priori the regression using the perceived adequacy of reason to imagine that people will be more or outcome and the quality of treatment rating. As less instrumental in hypothetical cases, Study 3 can be seen from the tables, in both analyses utilizes a real conflict for both types of judg- the stronger regression coefficients were those ment. Hence, it is further reassuring that Study associated with the quality of treatment varia- 3 finds a pattern like that of Studies 1 and 2. ble. In both sets of analyses, a comparison of the magnitude of the association shows that Study 4 quality of treatment is significantly more important (Table 6: t(220) ϭ 6.64, p < .001; Table 7: Like Study 3, Study 4 examines the choice of t(220) ϭ 8.62, p < .001). whether or not to move forward with a claim or

Table 7. Post-experience satisfaction (perceived Discussion adequacy of outcome): Study 3 The results of Study 3 replicate those of Studies 1 and 2. Choice is instrumentally based, while Favorability of outcome .08 post-experience evaluation is dominated by (perceived adequacy of outcome) relational concerns about the quality of the Quality of treatment received .90*** treatment received. In this case, however, the from others type of choice being studied is not a choice adj. R2 93%*** among alternatives. Instead, people are choosing whether or not to pursue a course of action. * p < .05; ** p < .01; *** p < .001. In addition, Study 3 shows the phenomenon in Note: Entries are beta weights.

111 Group Processes & Intergroup Relations 2(2) grievance. Study 4 examines the views of a court: the likelihood of winning, and their sample of employees asked about their recent judgments about the overall quality of their experiences in their workplace. relationship with their work supervisor. Employees were also asked about a recent Method personal experience with their supervisor. They Participants Study 4 is based upon interviews were asked about the favorability of their out- with a sample of 409 employees in Chicago. come from and the quality of their treatment Participants were drawn from a random sample during that experience. These judgments were of adults in the city of Chicago. The participants used to predict the degree to which they actu- were chosen in two ways. The largest group (n ally thought about or tried to go to court about ϭ 303) were participants who met three cri- the decisions made by their supervisor during teria: (1) they were working at least 20 hours a their actual personal experience. week; (2) they had a supervisor; and (3) they had a recent personal experience with their Questionnaire Employees were presented supervisor. Recent personal experience was with two hypothetical scenarios: they were fired established at the beginning of each interview without explanation and they were fired after a by asking participants to recall an incident in dispute with their supervisor. In each case, they which they had talked to their supervisor about were asked to estimate the likelihood that they ‘getting help in solving problems, or making would go to court to try to get their jobs back. decisions, or settling disagreements about how Employees were also asked to estimate the like- work should be done, or to discuss issues of pay, lihood that if they went to court in each case, promotion, work hours, or similar issues, or to they would be successful in getting their job help resolve a dispute with a customer, a co- back. worker, or with the supervisor themselves’. Of In addition, employees were asked to evalu- those contacted who met these criteria, 75 ate the quality of treatment they had generally percent were successfully interviewed. A second received from their supervisor. Six items were sample of 106 participants, drawn from the used. Participants were asked whether their same sampling frame, were screened in the supervisor generally: ‘got the information same way, but were asked to think of a recent needed to make good decisions’; ‘treated peo- experience in which they ended up feeling ple politely’; ‘showed concern for people’s ‘angry or upset’. Of participants meeting these rights’; ‘allowed people to state their views criteria, 73 percent were successfully inter- before making decisions’; ‘considered people’s viewed. views’; and ‘tried to be fair’. These items were used to create a general quality of treatment Design Participants were asked about two scale (Cronbach’s alpha ϭ .90). issues. First, they were asked to consider two Employees were also asked about their recent hypothetical scenarios. In the first scenario they personal experience with their supervisor. Par- considered a situation in which they ‘were fired ticipants were asked about how they were trea- by their company without any explanation or ted during this experience using a 13-item opportunity to discuss the issue with their quality of treatment scale (Cronbach’s alpha ϭ supervisor’. In the second scenario, they con- .94). The scale asked employees whether: ‘they sidered a situation in which they were fired were treated politely’; ‘their rights were respec- because they ‘had a dispute with their work ted’; ‘good information was gathered’; ‘issues supervisor over the quality of their work’. Fol- were brought into the open’; ‘their supervisor lowing each scenario employees were asked to was honest’; ‘they were given a chance to estimate the likelihood that they would go to speak’; ‘their views were considered’; ‘fair court to sue their company. The study exam- procedures were used’; ‘their supervisor tried ined the impact of two factors on decisions to be fair’; ‘their supervisor tried to explain about whether people felt they would go to decisions’; ‘their supervisor gave honest

112 Tyler et al. the two psychologies of conflict resolution explanations’; ‘their supervisor tried to take Table 9. Factors shaping whether employees account of their needs’; and ‘they were treated considered suing after a personal experience with fairly’. their supervisor: Study 4 Following their description of their experi- Expectation of success if .17* ence employees were asked about whether they go to court thought about going to court to try to complain Quality of treatment by supervisor .40*** about their supervisor’s decision. Two ques- during experience tions were asked: whether they thought about 2 complaining and whether they actually did adj. R 18%*** complain. These two items were correlated (r ϭ .46), so they were combined into a single index. * p < .05; ** p < .01; *** p < .001. Note: Entries are beta weights for an equation in Finally, employees were asked whether they which all terms are entered at the same time. thought that, if they went to court, they would have been effective in overturning the supervisor’s decision. dealing with their supervisor had no influence upon whether or not they thought they would Results contest their firing when thinking about hypo- According to the argument we have outlined, thetical situations in which they were imagining employee’s responses to the hypothetical sce- being fired (Scenario 1: beta ϭ .03, ns; Scenario narios outlining circumstances in which they 2: beta ϭ .04, ns). These differences in magni- went to court in response to being fired under tude are significant in the case of Scenario 2 two distinct sets of circumstances should be (t(138)ϭ1.97, p < .05), but not in the case of shaped by judgments about the likelihood that Scenario 1 (t(138)ϭ1.49, ns). they would win if they went to court. Table 8 Table 9 presents the results of a regression shows the findings in the case of both scenarios. analysis in a situation in which employees are In each case, employees are significantly influ- considering whether or not to sue their com- enced by their judgments about whether or not pany after they have had a personal experience they would win if they went to court to contest with their supervisor. In contrast to the hypo- their firing (Scenario 1: beta ϭ .51, p < .001; thetical situation already outlined, when Scenario 2: beta ϭ .64, p < .001). In contrast, employees were asked about their reactions to a the general quality of their treatment when past personal experience with their supervisor, they reacted in terms of their treatment by their Table 8. Factors shaping whether employees supervisor during that experience (beta ϭ .40, considered suing in response to hypothetical firing p < .001) more strongly than in terms of scenarios: Study 4 whether they felt that they would win if they went to court (beta ϭ .17, p < .05). This Scenario 1: Scenario 2: difference in magnitude of the correlations was Fired without Fired after a ϭ explanation dispute significant (t(138) 3.88, p < .001). In this situation, in which employees are reacting to an Expectation of success .51*** .64*** experience, they focused more strongly upon if go to court the quality of their treatment than they did General quality of .03 .04 upon their expectations of gain. treatment by supervisor General discussion adj. R2 26%*** 40%*** The studies reported above, and the literature * p < .05; ** p < .01; *** p < .001. on reactions to conflict resolution procedures, Note: Entries are beta weights for an equation in reveal a seeming paradox: people’s choices and which all terms are entered at the same time. their evaluations appear to be out of line, with

113 Group Processes & Intergroup Relations 2(2) preferences and choices being responsive to they have personal feelings about their experi- different factors than those that shape evalu- ences. These personal feelings are a reflection ations of disputing experiences. This paradox is of their post-experience feelings, in which they difficult to explain using either social exchange are more concerned about issues of justice, theory or the relational model of procedure morality, and social relationships than is sug- alone. Both models assume that choices and gested by the model embodied in the ‘myth of evaluations are determined by the same psycho- self-interest’ (Tyler & Lind, 1992). Conse- logical factors. However, the apparent paradox quently, people evaluate their experiences can be explained by adding to the conceptual in terms of the quality of their treatment by mix the processes involved in the ‘myth of self- others. interest’, which argues that people’s prefer- Why focus on quality of treatment? Tyler and ences are shaped by different psychological Lind (1992; Lind & Tyler, 1988) argue that dynamics than are their evaluations. people are concerned about their inclusion The ‘myth of self-interest’ model articulated status within social groups. Their treatment by by Miller and Ratner (1996, 1998) hypothesizes others is a cue which people use to assess their that people’s thoughts about their own motiva- social status (Smith & Tyler, 1997). Social status tions are responsive to the cultural ideology of is important because people use it to under- self-interest. As a consequence, when people stand their social identity and, through that think about the type of people that they believe understanding, to construct their feelings of themselves to be, when they try to ‘model’ their self-worth and self-esteem (Hogg & Abrams, own psychology, they think that they will experi- 1988; Turner, 1991; Turner, Hogg, Oakes, ence the greatest benefit when they do those Reicher, & Wetherell, 1987). Implicit in this things which most directly advance their own account of the findings we report above is the personal self-interest. In the case of interper- idea that social considerations, such as those sonal conflict, people believe that they will be emphasized by the relational model, are under- happiest if they choose the procedures that will considered in the pre-experience calculus. That maximize their own personal gain. Further, is, people are unaware of the degree to which they think of personal gain in material terms, as they utilize their experiences with others to a gain or loss of material resources. Hence, in construct their social identity, and hence place preference-expressing situations like those too little emphasis upon how they will be trea- examined in Studies 1 and 2, people should ted by others. prefer a forum for dispute resolution which One of the core arguments of attribution they believe will lead to the material outcomes theory is that people make the ‘fundamental they want. In choice situations like those exam- attribution error’. They put too little weight ined in Studies 3 and 4 they should move upon situational factors when making infer- forward with grievances when they think that ences about the causes of the behavior of others they are likely to prevail. (Nisbett & Ross, 1980). Interestingly, studies of When making choices about procedures they the actor–observer bias suggest that this tend- have not yet experienced, people by definition ency is stronger when people are making infer- cannot rely upon their personal experience to ences about the behavior of others (Fiske & guide them. Hence, they must model their own Taylor, 1991). That is, people are more able to likely reactions, relying upon their ideologies see the influence of social forces on their own about what type of person they are. When they behavior. However, the findings outlined here do this, they make choices that are based upon suggest that people are insensitive to another the belief that they will be happiest if they gain aspect of the social situation – the impact of the the most for themselves. quality of treatment by others on their feelings. But the actual experience of a disputing People do not recognize that a key input into procedure involves something quite different. their post-experience satisfaction is the quality Once people have experienced a procedure, of the treatment they receive from others.

114 Tyler et al. the two psychologies of conflict resolution

The findings of our studies and our explana- determining their preferences people may tion of those findings resolve the apparent abandon their insights from their own experi- paradox about the differing antecedents of ence in favor of the conventional wisdom that choice and evaluation, but an interesting ques- outcomes will predominate. tion remains. Why do people, who after all have A third explanation is that people simply are in the past generally experienced some mani- not very good at intuiting the causes of their festation of the procedures in question, con- own happiness or unhappiness. They may tinue to determine their preferences based on emerge from disputes in which others treated instrumental concerns? Why don’t their prefer- them with dignity and consideration feeling ences flow directly from their knowledge that in good and from disputes in which others treated the past they have reacted most strongly to how them poorly feeling bad, but they may not make they were treated by others? In other words, the connection to the underlying psychology of how can the ‘myth’ of self-interest be main- relational judgments. Thus when procedures tained? We might expect that people would chosen to maximize self-interest end up being learn from their experience and change the disappointing in terms of relational experi- way they determine their preferences. ences, people may be aware that they are dissat- There are several possible explanations for isfied, but be uncertain about why dissatisfac- the maintenance of the myth of self-interest. tion is occurring. They may simply know they One explanation suggests that people may are unhappy, and they may begin to search know why they are unhappy, but do not feel their ideology for explanations of their unhap- able to act differently. People may feel that they piness. And their ideology may tell them that have to be responsive to the competitive nature they are unhappy because the outcomes they of the social environment in which their con- received are not good enough. Hence, people flict is occurring. They may feel that they must may maintain the myth of self-interest because protect themselves from others, who would take they have difficulties understanding their advantage of their softness regarding instru- experience. Future research may be able to mental issues (cf. Kelley & Stahelski, 1972). differentiate between these several explana- Hence, the competitive environment may tions of the effect; for the moment it is impor- encourage the emergence of a defensive ori- tant to demonstrate the phenomenon that entation, which promotes a focus on instru- underlies the discontinuity between preference mental issues. If so, then the key issue regarding and evaluation. the myth of self-interest is the belief that other One interesting line of research that suggests people are instrumental in their orientation. insights about the ability to maintain the myth A second explanation for the findings has to of self-interest is work on hindsight biases. This do with a tension between cultural and personal research shows that people make mistakes in views about how decisions should be made. their recall of what they thought would happen, People may understand that their happiness in even if they are motivated to be accurate a given conflict situation depends on the quality (Dawes, 1988; Fischhoff, 1980). Hence, after of interpersonal treatment, but they may lack the Berlin wall came down, people say that they confidence in their beliefs. In our individualis- expected that event to occur before it hap- tic culture, people are told that the way one pened. Of course, if they had been interviewed obtains happiness is through obtaining the out- about the event before it happened they would comes one wants – the ‘myth of self-interest’. As have estimated its likelihood of occurring to be we noted earlier, participants might intuit that low. While this research makes clear that peo- happiness actually comes from the quality of ple distort their image of the past, it does not treatment by others, more than from obtaining completely explain the findings of these stud- desired outcomes, but they may not trust their ies. If people distorted their views about the own judgment when finally pushed to choose a pre-experience past to be consistent with their procedure. When they come to the point of post-experience judgments, we would expect

115 Group Processes & Intergroup Relations 2(2) them to decide that they make their choices out people choose the high paying, but unfair, of a desire to gain high quality of treatment. organization to work for, but subsequently There is no evidence that people make such evaluate their work environment in terms of the distortions. quality of their treatment by others, then they As has been noted, the findings of these will experience job dissatisfaction. Choosing for studies suggest that preferences for methods of instrumental reasons, but evaluating in terms of dispute resolution, and reactions to these meth- quality of treatment, will inevitably lead to dis- ods, should not be treated as resulting from a satisfaction in a wide variety of interpersonal single psychological model. One implication of and organizational settings. the myth of self-interest is that preference and More generally, the basic phenomenon we evaluation might develop from two distinct psy- see in the studies reported here may call into chological mechanisms. This discrepancy flows question a great deal of economic and political from the ‘myth’ aspect of the myth of self- ideology about the validity of expressed prefer- interest – that people’s ideology about motiva- ences. Economic analyses of policy generally tion is not consistent with their true motiva- assume that people’s choices or their expressed tions. Hence, the paradox being outlined only preferences, if met, will maximize their perso- occurs where people have an inaccurate image nal satisfaction. An important normative justifi- of their own motivation. These findings help to cation of market capitalism is that the way to explain the puzzling literature on the psychol- maximize human happiness is to allow people ogy of mediation. By demonstrating that prefer- free choice to the greatest extent possible. The ence and evaluation are not based on the same line of thinking expressed here suggests that psychological model the research reported choices and preferences are arguably discon- here may provide an explanation for the diffi- nected from what people actually experience as culty of getting people to participate in media- most satisfactory. tion programs, in spite of the high approval Similarly, important parts of democratic the- ratings of those who do participate. This in turn ory rest on the assumption that people will be suggests that there may be psychological sup- most satisfied with leaders and policies they port for mandatory mediation programs – only freely choose themselves. Might it be the case, by mandating participation in such procedures we are forced to wonder, that people are so can people be led into experiences that they often dissatisfied with political leaders because will ultimately regard as quite satisfactory (see, they mismodel their post-election experiences Lind, 1997). with those leaders? In fact, political scientists These studies are also important because have demonstrated a powerful incumbency they support relational model predictions effect, in which people are continually very about when relational concerns will be impor- favorable in their ratings of new Presidents, tant. They suggest that they will be especially whose popularity inevitably declines during important after experiences, not before them. their term in office. It may be that such Presi- People’s concerns about relationships are acti- dents are chosen based upon the expected vated in the course of their efforts to resolve a favorability of their policies, but evaluated dispute, when they experience favorable or through the manner in which they treat unfavorable treatment from others. citizens. Irrespective of why they are occurring, the Clearly these are very important issues. What findings outlined have implications far beyond is needed is an effort to identify the situations interpersonal conflict. If the preferences gov- within which preference, choice, and evalu- erning behavioral choices are more instrumen- ation converge, and situations in which they tal in character than are post-experience feel- differ. In the latter type of situations the prob- ings, then people may well open themselves to lems outlined above may emerge. In such set- dissatisfaction in many areas of their life. Con- tings, the psychology of evaluations and the sider, for example, a job choice situation. If psychology of preference may diverge, opening

116 Tyler et al. the two psychologies of conflict resolution the possibility of dissatisfaction with freely cho- organizations (Vol. 1, pp. 81–90). Greenwich, CT: sen procedures. Hence, we need a mapping of JAI Press. the domain of the myth of self-interest, which Brett, J. M., & Goldberg, S. (1983). Mediator– identifies those areas in which people are un- advisors: A new third-party role. In M. Bazerman & R. Lewicki (Eds.), Negotiating in organizations (pp. aware of the true basis of their own feelings. 165–176). Beverly Hills, CA: Sage. Dawes, R. M. (1988). Rational choice in an uncertain Notes world. San Diego, CA: Harcourt, Brace, Jovanovich. Fischhoff, B. (1980). For those condemned to study 1. Weighting the sample size to adjust for multiple the past. In R. A. Schwede & D. W. Fiske (Eds.), ratings only deals with one of two problems New directions for methodology of behavioral science (pp. introduced by using within-participant ratings. 79–93). San Francisco, CA: Jossey-Bass. Each participant provided seven observations, so Fiske, S. T., & Taylor, S. E. (1991). Social cognition. the observations are not entirely independent. (2nd ed.). New York: McGraw-Hill. Because this is true, it is important to treat the Hensler, D. R., Marquis, M. S., Abrahamse, A. F., statistical analyses with caution. However, the Berry, S. H., Ebener, P. A., Lewis, E. G., Lind, conclusions of the studies are based upon E. A., MacCoun, R. J., Manning, W. G., Rogowski, comparisons of the importance of different J. A., & Vaiana, M. E. (1991). Compensation for factors within the same analytic approach. accidental injuries in the United States. Santa Monica, Assessments of relative importance should be less CA: Rand. influenced by nonindependence than Hogg, M. A., & Abrams, D. (1988). Social explorations of whether the importance of a identifications. New York: Routledge. factor is significantly different than zero. Houlden, P., LaTour, S., Walker, L., & Thibaut, J. 2. The possibility of interactions between nationality (1978). Preferences for modes of dispute and the importance placed upon instrumental or resolution as a function of process and decision relational judgments was also tested. However, control. Journal of Experimental Social Psychology, 14, interactions were not found in any of the studies 13–30. reported here. Huo, Y. J., Smith, H.-J., Tyler, T. R., & Lind, E. A. 3. As in Study 1, nationality was not found to (1996). Superordinate identification, subgroup interact with the importance placed upon identification, and justice concerns. Psychological receiving favorable outcomes or high quality of Science, 7, 40–45. treatment. Kelley, H. H., & Stahelski, A. (1972). Social interaction basis of cooperators’ and competitors’ beliefs about others. Journal of Personality and Social Acknowledgments Psychology, 16, 190–197. Kurtz, S. T., & Houlden, P. (1981). Determinants of The research reported here was supported by procedural preferences of post court-martial National Science Foundation grants SES-9113863 military personnel. Basic and Applied Social and SES-9113752 and by the support of the Psychology, 2, 27–43. American Bar Foundation. Thanks are due to Eileen Lind, E. A. (1997). Testimony on the ‘Alternative Young, Pui Lau, and Annie Chen for their assistance Dispute Resolution and Settlement Act’ (H.R. in conducting Study 1. Note that other publications 2603, 105th Congress, 1st Session). Testimony, based on other aspects of the data sets reported here House Judiciary Committee, Subcommittee on include: Tyler et al. (1997); Lind, Tyler, and Huo Courts and Intellectual Property, 9 October. (1997); Lind et al. (1994); Huo, Smith, Tyler, and Lind, E. A. (in press). Fairness heuristic theory: Lind (1996); Hensler et al. (1991). Justice judgments as pivotal cognitions in organizational relations. In J. Greenberg & R. References Cropanzano (Eds.), Advances in organizational behavior. San Francisco, CA: New Lexington. Blalock, H. M., Jr. (1972). Social statistics. New York: Lind, E. A., Erickson, B. E., Friedland, N., & McGraw-Hill. Dickenberger, M. (1978). Reactions to procedural Brett, J. M. (1986). Commentary on procedural models for adjudicative conflict resolution: A justice papers. In R. Lewicki, M. Bazerman, & B. cross-national study. Journal of Conflict Resolution, Sheppard (Eds.), Research on negotiation in 22, 318–242.

117 Group Processes & Intergroup Relations 2(2)

Lind, E. A., Greenberg, J., Scott, K. S., & Welchans, esteem and group-oriented behaviors. Journal of T. D. (1998). The winding road from employee to Experimental Social Psychology, 33, 146–170. complainant: Situational and psychological Thibaut, J., & Walker, L. (1975). Procedural justice. determinants of wrongful termination claims. Paper Englewood Cliffs, NJ: Erlbaum. presented to the Academy of Management Turner, J. C. (1991). Social influence. Milton Keynes: meetings, August, San Diego, CA. Open University Press. Lind, E. A., Huo, Y. J., & Tyler, T. R. (1994). . . . And Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S., justice for all: Ethnicity, gender, and preferences & Wetherell, M. S. (1987). Rediscovering the social for dispute resolution procedures. Law and Human group. Oxford: Blackwell. Behavior, 18, 269–290. Tyler, T. R. (1989a). The quality of dispute Lind, E. A., Kulik, C. T., Ambrose, M., & de Vera resolution processes and outcomes. Denver Park, M. V. (1993). Individual and corporate University Law Review, 66, 419–436. dispute resolution. Administrative Science Quarterly, Tyler, T. R. (1989b). The psychology of procedural 38, 224–251. justice: A test of the group value model. Journal of Lind, E. A., MacCoun, R. J., Ebener, P. A., Felstiner, Personality and Social Psychology, 57, 830–838. W. L. F., Hensler, D. R., Resnik, J., & Tyler, T. R. Tyler, T. R. (1990). Why people obey the law. New (1990). In the eye of the beholder: Tort litigant’s Haven, CT: Yale. evaluations of their experiences in the civil justice Tyler, T. R., Boeckmann, R. J., Smith, H. J., & Huo, system. Law and Society Review, 24, 953–996. Y. J. (1997). Social justice in a diverse society. Boulder, Lind, E. A., & Tyler, T. R. (1988). The social psychology CO: Westview. of procedural justice. New York: Plenum. Tyler, T. R., & Lind, E. A. (1992). A relational model Lind, E. A., Tyler, T. R., & Huo, Y. (1997). of authority in groups. Advances in Experimental Procedural context and culture: Variation in the Social Psychology, 25, 115–191. antecedents of procedural justice judgments. Tyler, T. R., Lind, E. A., Ohbuchi, K., Sugawara, I., & Journal of Personality and Social Psychology, 73, Huo, Y. J. (1998). Conflict with outsiders: 767–780. Disputing within and across cultural boundaries. MacCoun, R., Lind, E. A., & Tyler, T. R. (1992). Personality and Social Psychology Bulletin, 24, Alternative dispute resolution in trial and 137–146. appellate courts. In D. Kagehiro & W. S. Laufer Tyler, T. R., Rasinski, K., & Spodick, N. (1985). The (Eds.), Handbook of law and psychology. New York: influence of voice on satisfaction with leaders: Springer-Verlag. Exploring the meaning of process control. Journal Miller, D. T., & Ratner, R. K. (1996). The power of of Personality and Social Psychology, 48, 72–81. the myth of self-interest. In L. Montada & M. J. Lerner (Eds.), Current societal concerns about justice. Paper received 6 August 1998; revised version accepted 4 New York: Plenum. December 1998. Miller, D. T., & Ratner, R. K. (1998). The disparity between the actual and assumed power of self- interest. Journal of Personality and Social Psychology, Biographical notes 74, 53–62. Nisbett, R., & Ross, L. (1980). Human inference: tom r. tyler is in the Department of Psychology, Strategies and shortcomings of social judgment. New York University. Englewood Cliffs, NJ: Prentice-Hall. yuen j. huo is in the Department of Psychology, Smith, H. J., & Tyler, T. R. (1997). Choosing the University of California at Los Angeles. right pond: The influence of the status of one’s e. allan lind is at Fuqua School of Business, Duke group and one’s status in that group on self- University.

118