Back to Journal Cover Page


CURRENT RESEARCH IN SOCIAL PSYCHOLOGY


Volume 8, Number 20
Submitted: June 16, 2003
First Revision: June 19, 2003
Second Revision: July 4, 2003
Accepted: July 7, 2003
Publication Date: July 7, 2003

The False Polarization Effect in Explanations of Attitudinal Behavior

Girts Dimdins
Henry Montgomery
Stockholm University

Ivars Austers
University of Latvia

ABSTRACT

Supporters and opponents of Latvia’s EU membership rated attitudinal behavior of EU supporters and opponents on a number of causal explanation scales from their own perspective and simulated perspectives of both groups. From a target's in-group perspective, both groups rated the causes of behavior as more stable, controllable and rational, and less subjective and less influenced by others than the respective group did from their own perspective. Although the results were not consistent for all rating scales and all perspectives, the study demonstrates that a false polarization effect occurs in explanations of attitudinal behavior. Directions for future research are discussed.

[275]
---------------
[276]

THEORETICAL BACKGROUND

Lee Ross (1977) notes that ''Individuals must, for the most part, share a common understanding of the social actions and outcomes that affect them, for without such consensus, social interaction would be chaotic, unpredictable, and beyond the control of the participants'' (p. 179). For the past several decades, psychologists and other researchers have focused on the construal of social reality (see Hewstone, 1989 for a review). A significant portion of these studies deals with causal attribution in an intergroup context (Hewstone, 1990; Deschamps, 1983). These studies consistently have found that individuals prefer in-group-serving attributions relative to out-group-serving attributions (Hewstone, 1990).

The definition of ''group-serving'' has varied from study to study, depending on which attributes are used. The oldest classification of intergroup attributions, based on Heider's (1958) model of interpersonal attributions, distinguishes between internal (dispositional) and external (situational) attributions. The basic hypothesis is that more internal attributions would be made for in-group members' socially desirable behavior and more external attributions for socially undesirable behavior. The opposite should hold true for out-group members' behaviors. Several classic studies using this classification (Taylor & Jaggi, 1974; Duncan, 1976; Rosenberg & Wolfsfeld, 1977; Hewstone & Ward, 1985) have provided mixed support for this hypothesis. The classification itself has been criticized for several methodological drawbacks (see Miller, Smith, & Uleman, 1981).

An improved model was proposed by Islam and Hewstone (1993), who suggest measuring attributions along the continuums of causal locus, stability, controllability (Weiner, 1986), and globality (Abramson, Seligman, & Teasdala, 1978). The hypothesis for this model is that positive in-group behavior and negative out-group behavior would be attributed to internal, stable, controllable (by the actor), and global causes; negative in-group and positive out-group behavior should be attributed to more external, unstable, uncontrollable, and specific causes. A number of successful attribution studies have used this classification (Islam & Hewstone, 1993; Wilder, Simon & Faith, 1996; Lee & Robinson, 2000; Austers, 2002). Again, the support for the hypothesis has been mixed.

Malle (1999, 2001) suggests another alternative to the internal/external classification. Central to the model of folk explanations of behavior (FEB) is the folk concept of intentionality (Malle & Knobe, 1997). Within the model, explanations that interpret an agent's behavior as unintentional are considered cause explanations. Explanations that interpret an agent's behavior as intentional are divided into three major groups. Causal history of reasons contains factors from an agent's personal history that cause the intentional behavior without the agent being aware of them. Reasons are factors that the agent considered when forming the intention to act, and enabling factors are factors that clarify how it was possible that the agent completed the intended action. Reasons are further divided into desires, beliefs, and valuing. Whereas the FEB model has been used mostly in explanations of individual behavior (Malle, Knobe, O'Laughlin, Pearce, & Nelson, 2000), it can also be successfully applied in analyzing the explanations of group behavior (O'Laughlin & Malle, 2002).

A number of studies of group attributions have been carried out that ask the participants to take the perspective of their in-group or out-group members. Kemdal and Montgomery (2001) found that animal experimenters and animal rights activists could take each other's perspective, resulting in a reversed actor-observer effect. Austers (2002) and Austers and Montgomery (2001) found similar results. Robinson, Keltner, Ward, and Ross (1995) tested some aspects of group attribution when they asked their participants to rate the basis of their own political judgments and that of their in-group and out-group members. The respondents indicated that they personally had been less influenced by ideology or political orientation than either their peers or opponents. These findings suggest a presence of the false polarization effect in the context of intergroup attributions.

The false polarization effect (see Pronin, Puccio, & Ross [2002] for a review of this phenomenon) is better known from the studies of group attitudes. Generally, the effect can be defined as an overestimation of the expected group score on a certain attribute in comparison with the actual group score on that attribute, which has been found across a number of different contexts (Robinson, Keltner, Ward, & Ross, 1995; Keltner & Robinson, 1997; Robinson & Friedman, 1995; Rouhana, O'Dwyer, & Morrison Vaso, 1997; Diekman, Eagly, & Kulesa, 2002). These findings show that the false polarization effect is a very robust phenomenon, influencing various types of ratings. There is no reason why the same effect should not appear in attribution studies, but very few published papers address this issue. Austers (2002) asked ethnic Latvian and Russian schoolteachers to rate positive and negative behaviors from their own and out-group perspective on a number of attributional dimensions. The results indicated some false polarization effect in predicting out-group's responses, but not consistently. To our knowledge, the paper of Robinson et al. (1995) provides the most convincing evidence of the false polarization effect in the context of group attributions. However, in their study the effect is reported on a small number of context-specific dimensions. We designed this study to explore the presence of the false polarization effect on more general attributional dimensions, including ''traditional'' dimensions familiar from the previous research.

[276]
---------------
[277]

This study tests whether the false polarization effect would be present in explanations of causes of attitudinal behavior when these explanations are given from in-group and out-group perspectives. By attitudinal behavior we understand general, non-specific behavior that expresses the actor's attitude towards a certain issue, in other words--acting in line with one's attitudes. We chose attitudinal behavior instead of more specific group behaviors because it allowed us to construct simple, context-independent and unambiguous stimuli for our study.

In our study we tried to use all the attributional dimensions reported in earlier literature. In addition to the four variables of causal locus, controllability, stability, and globality, we included a number of explanatory dimensions taken from Malle's FEB model. The items constructed on the basis of the FEB model represented factors differentiating among the major groups of explanations within the model, as well as various types of explanations within each group. The differentiating factors were intentionality (distinguishing cause explanations from all other explanations) and awareness (distinguishing causal history of reasons from reason explanations). The specific explanations included in the questionnaire were various types of causal history of reasons and reason explanations. (In an earlier unpublished study, where we asked participants to give free-response explanations of attitudinal behavior, more than 95% of the explanations fell into these two categories.)

We included two items measuring the evaluation of the target behavior in positive-negative terms, to control whether the group serving bias occurs in the expected direction (i.e., each target behavior is seen as more positive from its in-group perspective than out-group perspective). Finally, we also included an item asking the participants to estimate the distribution of people sharing their opinions and those with opposite opinions in the society. This was another control question because stronger in-group favoritism can be expected among the members of minority groups (Hewstone, Rubin, & Willis, 2002; Mullen, Brown, & Smith, 1992).

We predicted that when imagining how specific causes for a given target behavior are rated from the target's in-group and out-group perspectives, participants would give significantly higher (or lower) ratings than the respective in-group and out-group members themselves. Theoretically, the mean differences should occur in the group-serving direction. However, because we were using general descriptions of two groups' attitudinal behavior rather than descriptions of positive and negative behavior, it was difficult to make specific predictions about the direction of mean differences. Although one would expect that each target behavior should be seen more favorably from the target's in-group perspective than the out-group perspective, it does not necessarily mean that in-group behavior is seen as explicitly positive, and out-group behavior as negative. In other words, from the previously published group attribution studies we could not predict with certainty how in-group favoritism should manifest itself on each of the attributional variables. Our general prediction was that the mean differences from opposing perspectives should be in opposite directions for the same behavior. For example, if the causes for supporter behavior are rated as highly stable from the supporter perspective (implying that stability is seen as a favorable attribute) they should be rated as relatively unstable from the opponent perspective, and vice versa.

After a pilot study of several topics eliciting political attitudes among Latvian students, we decided that Latvia's membership in the European Union is a controversial political issue worthy of study. The topic provides an excellent context for examining perspective taking in explanations of group behavior. Over the last two years, the debate on the issue has grown with the opinions of both supporters and opponents highly salient and well represented in mass media. Although traditionally EU membership supporters have been in a slight majority in Latvia, the opinion polls predict a close race in the referendum planned for autumn 2003.

METHOD

Sample

One hundred forty-one students (110 women) at the University of Latvia participated in the study. The participants were undergraduate students of education. The mean age of the participants was 19 years. Majority of respondents (82%) were ethnic Latvians.

Questionnaire

First, we asked the participants to indicate their age, gender, ethnicity, and general attitude towards joining the European Union (pro vs. against). We also asked whether they believed (a) that EU supporters were in the considerable majority in Latvia, (b) that there was approximately equal number of opponents and supporters, or (c) that the opponents were in the majority.

[277]
---------------
[278]

In the second part of the questionnaire, we provided descriptions of two opposite behaviors and asked participants to rate various explanations of these behaviors from their own perspective as well as from the perspective of both EU supporters and EU opponents. The opposite behaviors were described with the following statements: (a) ''There are many people in Latvia who actively support Latvia entering the European Union''; (b) ''There are many people in Latvia who actively protest against Latvia entering the European Union''. Each description was written on top of a separate sheet.

We asked the participants to rate the causes of both target behaviors from their own perspective, EU opponent, and EU supporter perspective by answering a number of questions. All but the last two questions offered various explanations for the behavior, and they were constructed to cover both ''traditional'' attribution dimensions and the FEB coding scheme. The survey items are shown in Appendix A. The participants rated to what extent in their opinion each explanation accounted for the target behavior. The last two questions asked the participants to evaluate the target behavior in positive-negative terms. All ratings were made on a five-point Likert scale.

Correspondingly, for group perspectives, half of the participants were asked the following: ''How would people who support (oppose) Latvia entering the European Union answer the question . . .''. For the other half, the following question was asked: ''How would University of Latvia students who support (oppose) Latvia entering the European Union answer the question . . . ''. The two different wordings were used to control for the possible effects of how the group perspectives were defined.

The order of items and target behaviors was counterbalanced across the questionnaires. The participants first rated each target behavior from their own perspective. Then they rated each target from both group perspectives in random order. To reiterate, each participant made ratings of two target behaviors from three perspectives, answering the same set of questions six times.

In the third part of the questionnaire, we asked the participants to indicate their agreement/disagreement with seven items measuring their support to EU membership (see Appendix A). Afterwards, the participants rated the same items from both group perspectives.

RESULTS

First, we checked for the effects of questionnaire type (wording for the group perspective: supporters/opponents in general vs. LU students) and the effects of perceived balance (supporters in majority vs. opponents in majority vs. both groups equal). We included each of these variables as a between-subjects factor in a 4-way ANOVA with participant attitude (supporter vs. opponent) as another between-subjects variable and target behavior (supporter vs. opponent) and perspective (self vs. supporter vs. opponent) as within-subjects variables. We ran the ANOVA on all variables in our study both on raw data and the indexes reported below. We found some significant effects involving both variables, but the directions of mean differences varied from analysis to analysis. Because the effects of both variables did not seem to be systematic, we do not report them.

We based our data analysis on planned pairwise comparisons between the rating means from own perspective and the corresponding means from both group perspectives. We used a one-tailed t-test where we had clear predictions about the directions of mean differences (i.e., attitude ratings) and a 2-tailed t-test for the other comparisons (i.e., causal explanation ratings). Such a direct test of the false polarization effect increased the power of the analysis in comparison with a full ANOVA. For each comparison, we computed effect size (Cohen's d) to provide an estimate of the magnitude of mean differences. According to conventional estimates suggested by Cohen (1988), an effect size of about 0.2 can be seen as small, an effect size of about 0.5 as medium, and an effect size of 0.8 or higher as large.

Ninety-two participants (65%) identified themselves as supporters of Latvia's EU membership, and forty-nine (35%) identified themselves as opponents. The top row of Table 1 shows the means of strength of support to EU membership from own and both imagined group perspectives. The index was calculated as the average of the seven items in the third part of the questionnaire. The index values can vary from 1 to 5, higher score corresponding to stronger support to Latvia's EU membership. The false polarization effect is clearly visible in the results. Both groups overestimated the support displayed by EU supporters and the opposition displayed by EU opponents.

[278]
---------------
[279]

Table 1. Mean attitude ratings from own perspective and the imagined group perspectives

 

Ratings from own perspective

Ratings from the supporter perspective

Ratings from the opponent perspective

 

By supporters
(n = 92)

By opponents
(n = 49)

By supporters

By opponents

By supporters

By opponents

Support to EU membership

3.88
(0.55)

2.32
(0.56)

4.30***
(0.61)

4.29***
(0.65)

1.91**
(0.85)

1.94**
(0.74)

Effect size

   

0.72

0.68

0.56

0.57

Evaluation of target behavior:

           

Supporter behavior

3.53
(0.67)

2.77
(0.76)

4.13***
(0.73)

3.96***
(0.87)

1.96***
(0.82)

2.36**
(0.91)

Effect size

   

0.85

0.55

1.02

0.54

Opponent behavior

2.63
(0.81)

3.52
(0.82)

1.99***
(0.86)

2.13***
(0.81)

3.85*
(1.06)

3.80
(0.97)

Effect size

   

0.76

0.62

0.34

0.31

Note. Standard deviations are given in parentheses. Asterisks indicate that rating from the imagined group perspective is significantly different (1-tailed t-test) from the corresponding rating from own perspective. Means in columns 3 and 4 are compared to means in column 1; means in columns 5 and 6 are compared to means in column 2. The effect size (Cohen's d) for each comparison is given below the respective mean in columns 3-6. * p < .05. ** p < .01. *** p < .001.

[279]
---------------
[280]

We calculated an index measuring the evaluation of target behavior as the average of two items: liking/disliking the behavior, and the belief that the behavior benefits the interests of the Latvian society. Cronbach's Alphas for both items were calculated separately for each perspective and target behavior (because each of target/perspective combinations represents a separate repeated measurement using the same two items). All but one (0.54) were above 0.6. Table 1 shows the means of the evaluative index. The index value can range from 1 (unfavorable evaluation) to 5 (favorable evaluation). Again, both groups showed a strong false polarization effect when evaluating supporter behavior from both perspectives and opponent behavior from the supporter perspective. The mean difference was not significant when opponent behavior was evaluated from the opponent perspective, but the means were in the expected direction.

Our study replicated the previous findings showing a false polarization effect when estimating group attitudes. Our main interest, however, was whether the same pattern of results would be found in estimating group ratings of causal explanations. Table 2 shows the mean ratings of stability of causes and ratings of actors' control over the causes of their behavior. All ratings were made on a 1 to 5 Likert type scale; larger values correspond to higher stability and control.

Table 2. Mean ratings for the traditional dimensions of causal attribution from own perspective and the imagined group perspectives

 

Ratings from own perspective

Ratings from the supporter perspective

Ratings from the opponent perspective

 

By supporters
(n = 92)

By opponents
(n = 49)

By supporters

By opponents

By supporters

By opponents

Causes are stable

           

Supporter behavior

2.96
(0.91)

2.94
(0.84)

3.76***
(0.97)

3.84***
(0.99)

2.49*
(1.04)

3.02
(0.88)

Effect size

   

0.85

0.92

0.47

0.09

Opponent behavior

2.87
(0.90)

3.17
(0.96)

2.59
(1.00)

2.86
(1.04)

3.59*
(1.16)

3.61*
(0.86)

Effect size

   

0.29

0.01

0.39

0.48

Actors control the causes of their behavior

           

Supporter behavior

3.12
(0.94)

3.13
(1.00)

3.82***
(1.00)

3.86***
(1.02)

2.53**
(1.08)

2.94
(0.94)

Effect size

   

0.72

0.75

0.58

0.20

Opponent behavior

2.82
(1.06)

3.29
(1.06)

2.46*
(0.95)

2.84
(1.12)

3.64
(1.09)

3.59
(1.22)

Effect size

   

0.36

0.02

0.33

0.26

Note. Standard deviations are given in parentheses. Asterisks indicate that rating from the imagined group perspective is significantly different (2-tailed t-test) from the corresponding rating from own perspective. Means in columns 3 and 4 are compared to means in column 1; means in columns 5 and 6 are compared to means in column 2. The effect size (Cohen's d) for each comparison is given below the respective mean in columns 3-6. * p < .05. ** p < .01. *** p < .001.

[280]
---------------
[281]

For two dimensions--causes in situation vs. actor and the globality of the causes of target behavior--we found no significant mean differences. Therefore the means of these two variables are not shown in Table 2 (but the means are reported in Table B3 and Table B6 in Appendix B). The false polarization effect appears in estimations of stability of causes of behavior. The strongest overestimation occurs in predicting supporters' ratings of supporter behavior where both groups rated the causes as significantly more stable than the supporters did from their own perspective. In other cases, the significant mean differences are in the same direction. (From the in-group perspective, the causes are seen as more stable.) Interestingly, the responses of supporter and opponent participants themselves do not differ for supporter behavior (t[138] = 0.12, p = .90) or opponent (t[136] = -1.83, p = .07) behavior. Nevertheless, both groups expected other supporters and opponents to be biased towards their in-groups.

We found similar results also for ratings of the actors' control over the causes of their behavior. Again, the strongest overestimation of group position occurs when rating the supporter behavior from supporter perspective. For supporter behavior, both groups did not differ in their ratings from own perspective (t[138] = -0.03, p = .98). For opponent behavior, opponents gave higher control ratings than supporters did (t[139] = -2.51, p < .05). Nevertheless, the overestimation of group position was stronger and more frequent for the supporter behavior than the opponent behavior.

To summarize, of the four traditional attribution dimensions, the false polarization effect appears in the stability of causes and the actors' control of the causes of attitudinal behavior. For both ratings, the effect was stronger for supporter behavior than for opponent behavior, and the effect was stronger from the imagined supporter perspective, than the opponent perspective.

To reduce the number of variables, we ran a factor analysis on the items based on the FEB coding scheme. Sums of ratings for each item from all three perspectives for both target behaviors were entered into the analysis. A principal component analysis with Varimax rotation yielded two factors, accounting for 50% of the total variance. The results of the factor analysis are shown in Table 3.

Table 3. Results of Principal Components Analysis

   

Component

Item

h2

1

2

Target behavior is intentional [a]

.600

.303

.713

Actors are aware of the causes of their behavior [a]

.580

-.077

.758

Target behavior is determined by care for the country's future [a]

.589

.030

.767

Target behavior is determined by actors' own interests [b]

.392

.610

.145

Target behavior is determined by actors' emotions [b]

.533

.728

-.051

Target behavior is determined by actors' desires [b]

.615

.631

.466

Target behavior is determined by actors' valuations of EU [b]

.538

.709

.187

Target behavior is determined by actors' personality traits [b]

.471

.680

.090

Target behavior is determined by the actors' group memberships [b]

.344

.572

.130

Target behavior is determined by actors' beliefs [b]

.597

.537

.556

Target behavior is determined by accepting others' opinions

.252

.432

-.256

Eigenvalues for unrotated solution

 

3.82

1.69

Variance explained before rotation (%)

 

34.73

15.38

Note. [a] These items were used to calculate the Rationality index. [b] These items were used to calculate the Subjectivity index.

[281]
---------------
[282]

Reliability analysis was then conducted for sets of variables with high loading on the same factor. Cronbach's alpha for these sets of variables was calculated separately for ratings of each target from each perspective (because each of these ratings is a separate repeated measurement using the same set of items). Both factors were transformable into reliable indexes. The first index was calculated as an average of three variables: intentionality of the target behavior, actors' awareness of causes of their behavior, and the extent to what behavior is caused by actors' care about the country's future. We labeled this the Rationality index because all items are related to rational causes of behavior. All but one (0.58) Alphas for this index were above 0.60. The index values can vary from 1 to 5, higher scores corresponding to higher rationality. The second index was calculated as an average of 7 items, all of which seem to be related to actors' inner states and psychological characteristics. Therefore, we labeled it the Subjectivity index. (The belief item, which had similar loading on both factors, was included in Subjectivity index because it increased its reliability, and it decreased the reliability of the Rationality index.) Again, all but one (0.57) Alphas for this index were above 0.60. The index values can vary from 1 to 5, higher scores corresponding to higher subjectivity. One item (extent to which the behavior is caused by accepting others' opinions) that could not be included in any of the indexes was analyzed separately. It was measured on a five-point Likert scale, where higher score indicate more influence from accepting others' opinions. Table 4 shows the corresponding means.

Table 4. Mean ratings for explanations derived from the FEB coding scheme from own perspective and the imagined group perspectives

 

Ratings from own perspective

Ratings from the supporter perspective

Ratings from the opponent perspective

 

By supporters
(n = 92)

By opponents
(n = 49)

By supporters

By opponents

By supporters

By opponents

''Rationality'' index

           

Supporter behavior

3.29
(0.78)

3.13
(0.70)

3.99***
(0.83)

3.86***
(0.85)

2.67**
(0.91)

2.96
(0.82)

Effect size

   

0.87

0.70

0.57

0.22

Opponent behavior

2.94
(0.78)

3.45
(0.83)

2.68*
(0.93)

2.83
(0.83)

3.79*
(0.94)

3.84*
(0.83)

Effect size

   

0.30

0.14

0.38

0.47

''Subjectivity'' index

           

Supporter behavior

3.68
(0.52)

3.81
(0.51)

3.46**
(0.54)

3.38**
(0.59)

3.54*
(0.62)

3.67
(0.68)

Effect size

   

0.42

0.54

0.48

0.23

Opponent behavior

3.80
(0.53)

3.79
(0.53)

3.68
(0.61)

3.80
(0.57)

3.36***
(0.64)

3.54*
(0.59)

Effect size

   

0.21

0.00

0.73

0.45

Accepting others' opinions

           

Supporter behavior

3.13
(1.04)

3.55
(0.88)

2.75*
(1.02)

2.98
(1.15)

3.33
(1.21)

3.22
(1.12)

Effect size

   

0.37

0.14

0.21

0.33

Opponent behavior

3.20
(1.11)

3.02
(0.92)

3.52*
(1.05)

3.29
(1.14)

2.64*
(1.13)

3.06
(1.03)

Effect size

   

0.30

0.08

0.39

0.04

Note. Standard deviations are given in parentheses. Asterisks indicate that rating from the imagined group perspective is significantly different (2-tailed t-test) from the corresponding rating from own perspective. Means in columns 3 and 4 are compared to means in column 1; means in columns 5 and 6 are compared to means in column 2. The effect size (Cohen's d) for each comparison is given below the respective mean in columns 3-6. * p < .05. ** p < .01. *** p < .001.

[282]
---------------
[283]

The pattern of means for the Rationality index resembles those discussed above. For supporter behavior, means of supporter and opponent groups did not differ significantly, t(136) = 1.17, p = .25. However, both groups displayed a false polarization effect when rating the behavior from the imagined supporter perspective, and supporter participants showed the same tendency when estimating the position of opponents. For opponent behavior, both groups differed in their ratings, t(139) = -3.60, p < .001. Three of the four group perspective means showed a false polarization effect. The results suggest that the participants saw rationality as a favorable cause of attitudinal behavior.

The pattern is less clear for the Subjectivity index. The means of supporter and opponent participants from own perspective did not differ for supporter (t[134] = -1.38, p = .17) or opponent (t[139] = 0.09, p = .93) behavior. Supporter and opponent participants saw the causes of both target behaviors as less subjective from both targets' in-group perspectives (supporter behavior from supporter perspective and opponent behavior from opponent perspective). However, no false polarization effect in the opposite direction occurred when the causes were rated from targets' out-group perspectives.

Finally, regarding acceptance of others' opinions as the cause of attitudinal behavior, there was some evidence of a false polarization effect, but the mean differences were relatively small. From own perspective, the means differed significantly for the supporter behavior (t[136] = -2.38, p < .05), but not opponent behavior (t[136] = 0.94, p = .35). With one exception, all of the mean differences between own and group perspectives were consistently in the same direction, assigning less influence by others from in-group perspective and more from out-group perspective.

DISCUSSION

On a number of rating dimensions for explanations of attitudinal behavior, we found significant differences between the actual ratings of the participants and their estimations of others’ ratings on the same dimension. Our study confirms that the false polarization effect is a robust phenomenon, which can affect not only estimation of group attitudes but also explanations of attitudinal behavior. On average the magnitude of the false polarization effect was moderate--for most of the reported significant mean differences, the effect size was around medium, according to Cohen's (1988) estimates. However, it should be noted that we found such differences on most of the variables in our study, and that most of the time these differences were in the expected direction. This leads us to believe that, although moderate in effect size, our results are indicators of a robust psychological tendency.

At the same time, the effect was not present on all attributional dimensions. We did not find the false polarization effect in ratings of globality of causes and the person-situation attributions. Moreover, there were no group-serving effects for these two scores at all. One explanation for this finding is that in our study we did not compare positive and negative behaviors by both groups. Although all four dimensions have been shown to yield group-serving biases for comparisons of positive and negative in-group and out-group behaviors (c.f., Islam & Hewstone, 1993), the same effects may be less pronounced when explaining non-specific attitudinal behavior. At the same time, the participants showed clear false polarization effect in group-serving direction on other dimensions where group-serving biases in ratings from own perspective were not pronounced. This fact suggests another possible explanation: whereas stability and control as concepts have positive connotations, globality of causes and the person-situation distinction per se have neither positive nor negative connotations. One may speculate that it is the evaluative connotation of a rating dimension, which triggers the false polarization effects in estimating in-group and out-group ratings, and this perhaps contributes to group-serving biases in general. Asking the participants to evaluate attributional dimensions in positive-negative terms in the future studies could provide more information about the nature of the false polarization bias and about the group-serving effects in intergroup attributions in general.

From the practical point of view, our findings illustrate how the false polarization effect can result in overestimation of group differences. On a number of variables we found very small or insignificant differences between the actual ratings of supporter and opponent participants; however both groups expected much stronger differences when taking the perspective of their in-group and out-group members. If the parties in a controversial issue assume more grounds for disagreement or conflict than is actually the case, it may impede rational discussions between the groups and cause ungrounded pessimism about the possibility of negotiations. Moreover, the false polarization effect may put additional pressure on group leaders/representatives, who may be reluctant to make concessions in negotiations, fearing discontent of their in-group members. Awareness of the false polarization effect may contribute to conflict prevention and resolution in various intergroup contexts.

[283]
---------------
[284]

In the wider context of attribution research, the results of our study provided some additional information about causal explanations of attitudinal behavior. Stability, control, and rationality were seen as more positive causes of attitudinal behavior (more often ascribed to an in-group target). Subjectivity and influence by other people were seen as less positive. These findings are similar to the results reported by Kenworthy and Miller (2002); they found that more rationality and less externality and emotionality were attributed to in-group attitudes in comparison with out-group attitudes.

Another possible direction for the future research might be using similar perspective-taking studies with specific positive and negative in-group and out-group behaviors. Such designs would allow formulating more precise hypotheses about the expected directions of mean differences and acquire clearer results.

We did not find systematic and interpretable effects of how the groups were defined in our study (supporters/opponents in general vs. student supporters/opponents). However, both definitions were relatively general, relating to a large group whose members' individual opinions were unknown to the participants. The question remains if one would find false polarization effect if the groups were relatively small, and their individual members familiar to the respondents. A study by Kemdal and Montgomery (2000) suggests that under such conditions the false polarization effect may disappear. The effect of group characteristics on the false polarization effect is another prospective direction for future research.

REFERENCES

Abramson, L. Y., Seligmann, M. E. P., & Teasdale, J. D. (1978). Learned helplessness in humans: Critiques and reformulation. Journal of Abnormal Psychology, 18, 317-334.

Austers, I. (2002). Ability to take an out-group's perspective in explaining positive and negative behaviors. Scandinavian Journal of Psychology, 43, 407-412

Austers, I., & Montgomery, H. (2001). Explaining an out-group's political behavior from four different perspectives. Journal of Baltic Psychology, 2, 23-33.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press.

Deschamps, J.-C. (1983). Social Attribution. In J. Jaspars, F. D. Fincham, & M. Hewstone (Eds.), Attribution theory and research: Conceptual, developmental and social dimensions (pp. 223-240). London: Academic Press.

Diekman, A. B., Eagly, A. H., & Kulesa, P. (2002). Accuracy and bias in stereotypes about the social and political attitudes of women and men. Journal of Experimental Social Psychology, 38, 268-282.

Duncan, B. L. (1976). Differential social perception and attribution of intergroup violence: Testing the lower limits of stereotyping of blacks. Journal of Personality and Social Psychology, 34, 590-598.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

Hewstone, M. (1989) Causal attribution: From cognitive processes to collective beliefs. Oxford: Basil Blackwell.

Hewstone, M. (1990). The ''ultimate attribution error''? A review of the literature on intergroup causal attribution. European Journal of Social Psychology, 20, 311-335.

Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias. Annual Review of Psychology, 53, 575-604.

Hewstone, M., & Ward, C. (1985). Ethnocentrism and causal attribution in Southeast Asia. Journal of Personality and Social Psychology, 48, 614-623.

[284]
---------------
[285]

Islam, M. R., & Hewstone, M. (1993). Intergroup attributions and affective consequences in majority and minority groups. Journal of Personality and Social Psychology, 64, 936-950.

Keltner, D., & Robinson, R. J. (1997). Defending the status quo: Power and bias in social conflict. Personality and Social Psychology Bulletin, 23, 1066-1077.

Kemdal, A. B., & Montgomery, H. (2000). Seeing it from the other side: Perspectives and attitudes towards animal experimentation. Unpublished manuscript, Stockholm University.

Kemdal, A. B., & Montgomery, H. (2001). Explaining own and others' behavior in a controversial issue: Animal experimentation. Journal of Social Psychology, 141, 693-713.

Kenworthy, J. B., & Miller, N. (2002). Attributional biases about the origins of attitudes: Externality, emotionality, and rationality. Journal of Personality and Social Psychology, 82, 693-707.

Lee, F., & Robinson, R. J. (2000). An attributional analysis of social accounts: Implications of playing the blame game. Journal of Applied Social Psychology, 30, 1853-1879.

Malle, B. F., & Knobe, J. (1997). The folk concept of intentionality. Journal of Experimental Social Psychology, 33, 101-121.

Malle, B. F. (1999). How people explain behavior: A new theoretical framework. Personality and Social Psychology Review, 3, 23-48.

Malle, B. F. (2001). Folk explanations of intentional action. In B. F. Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions and intentionality: Foundations of social cognition (pp. 265-286). Cambridge, MA: MIT Press.

Malle, B. F., Knobe, J., O'Laughlin, M. J., Pearce, G. E., & Nelson, S. E. (2000). Conceptual structure and social functions of behavior explanations: Beyond person-situation attributions. Journal of Personality and Social Psychology, 79, 309-326.

Miller, F. D., Smith, E. R., & Uleman, J. (1981). Measurement and interpretation of situational and dispositional attributions. Journal of Experimental Social Psychology, 17, 80-95.

Mullen, B., Brown, R., & Smith, C. (1992). In-group bias as a function of salience, relevance, and status: An integration. European Journal of Social Psychology, 22, 103-122.

O'Laughlin, M. J., & Malle, B. F. (2002). How people explain actions performed by groups and individuals. Journal of Personality and Social Psychology, 82, 33-48.

Pronin, E., Puccio, C., and Ross, L. (2002). Understanding misunderstanding: Social psychological perspectives. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 636-665). Cambridge: Cambridge University Press.

Robinson, R. J., & Friedman, R. A. (1995). Mistrust and misconstrual in union-management relationships: causal accounts in adversarial contexts. International Journal of Conflict Management, 3, 312-327.

Robinson, R. J., Keltner, D., Ward, A., & Ross, L. (1995). Actual versus assumed differences in construal: ''Naïve realism'' in intergroup perception and conflict. Journal of Personality and Social Psychology, 68, 404-417.

Rosenberg, S. W., & Wolfsfeld, G. (1977). International conflict and the problem of attribution. Journal of Conflict Resolution, 21, 75-103.

[285]
---------------
[286]

Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 173-220). New York: Academic Press.

Rouhana, N. N., O'Dwyer, A., Morrison Vaso, S. K. (1997). Cognitive biases and political party affiliation in intergroup conflict. Journal of Applied Social Psychology, 27, 37-57.

Taylor, D. M.., & Jaggi, V. (1974). Ethnocentrism and causal attribution in a South Indian context. Journal of Cross-Cultural Psychology, 119, 162-171.

Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer Verlag.

Wilder, D. A., Simon, A. F., & Faith, M. (1996). Enhancing the impact of counterstereotypic information: Dispositional attributions for deviance. Journal of Personality and Social Psychology, 71, 276-287.

APPENDIX A. THE SURVEY INSTRUMENT

First Part of the Questionnaire (Cover Page)

By answering the questions in this survey you are taking part in a study of political attitudes carried out by Stockholm University and University of Latvia reserachers. Your responses are confidential and will be used for research purposes only.

The aim of the study is to explore how people evaluate the behavior of supporters and opponents of the European Union from their own and imagined point of view. Therefore we ask you to answer a number of similar questions several times. It is very important that you answer all the questions.

Please, tick the appropriate response or fill in the information!

Your age: [participant instructed to indicate the age]

Sex: M F [participant instructed to check one]

Ethnicity: [participant instructed to indicate the ethnicity]

Generally you:

support Latvia's membership of the European Union

oppose Latvia's membership of the European Union

[participant instructed to check one]

In your opinion:

EU supporters in Latvia are in a significant majority; EU opponents are in a minority

there are approximately equal number of EU supporters and opponents in Latvia

EU opponents in Latvia are in a significant majority; EU supporters are in a minority

[participant instructed to check one]

In the rest of the questionnaire, please, answer the questions by circling the response that corresponds to your opinion!

[end of cover page]

[286]
---------------
[287]

Second Part of the Questionnaire

Description of supporter behavior:

There are many people in Latvia who actively support Latvia entering the European Union

Description of opponent behavior:

There are many people in Latvia who actively protest against Latvia entering the European Union

Note. In the second part of the questionnaire, each of the target behaviors was rated on 17 items, provided below. Each target behavior appeared three times in the questionnaire, and correspondingly was rated from three perspectives: from own perspective, and from two imagined group perspectives--supporter perspective and opponent perspective.

Items for rating causal explanations and evaluating target behavior from own perspective:

In your opinion, to what extent do the causes of the supporters'/opponents' behavior originate in the situation rather than within themselves?

[participant instructed to indicate the response on a 5-point scale where 1 = not in the least and 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior stable (invariable)?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent do the supporters/opponents control their behavior?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent do the causes of the supporters'/opponents' behavior influence their behavior in all situations (also in those not related to Latvia's EU membership)?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent is the supporters'/opponents' behavior intentional?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the supporters/opponents aware of the causes of their behavior?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their personality traits?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their own interests?

[1 = not in the least; 5 = to the highest extent]

[287]
---------------
[288]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their membership in various social groups?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their care about the country's future?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by accepting other people's opinions?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their beliefs?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their desires (aspiration for a specific result)?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their subjective liking or disliking for various aspects of the EU?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent are the causes of the supporters'/opponents' behavior determined by their emotions?

[1 = not in the least; 5 = to the highest extent]

In your opinion, to what extent does the supporters'/opponents' behavior correspond to the interests of the Latvian society?
[participant instructed to indicate the response on a 5-point scale where

-2 = does not correspond at all;
-1 = does not correspond rather than corresponds;
0 = is neutral;
+1 = corresponds rather than does not correspond;
+2 = fully corresponds]

To what extent do you like or dislike the supporters'/opponents' behavior?
[participant instructed to indicate the response on a 5-point scale where

-2 = dislike very much;
-1 = dislike rather than like;
0 = neither like nor dislike;
+1 = like rather than dislike;
+2 = like very much]

[288]
---------------
[289]

Note: These items appeared twice in the questionnaire: once for the supporter behavior and once for the opponent behavior.

Items for rating causal explanations and evaluating target behavior from the group perspectives:

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior originate in the situation rather than within themselves?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are stable (invariable)?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the supporters/opponents control their behavior?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior influence their behavior in all situations (also in those not related to Latvia's EU membership)?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the supporters'/opponents' behavior is intentional?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the supporters/opponents are aware of the causes of their behavior?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their personality traits?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their own interests?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their membership in various social groups?

[1 = not in the least; 5 = to the highest extent]

[289]
---------------
[290]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their care about the country's future?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by accepting other people's opinions?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their beliefs?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their desires (aspiration for a specific result)?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their subjective liking or disliking for various aspects of the EU?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the causes of the supporters'/opponents' behavior are determined by their emotions?

[1 = not in the least; 5 = to the highest extent]

How would people who support/oppose Latvia entering the European Union answer the question to what extent the supporters'/opponents' behavior corresponds to the interests of the Latvian society?

[-2 = does not correspond at all;
-1 = does not correspond rather than corresponds;
0 = is neutral;
+1 = corresponds rather than does not correspond;
+2 = fully corresponds]

How would people who support/oppose Latvia entering the European Union answer the question to what extent they like or dislike the supporters'/opponents' behavior?

[-2 = dislike very much;
-1 = dislike rather than like;
0 = neither like nor dislike;
+1 = like rather than dislike;
+2 = like very much]

Note: These items appeared four times in the questionnaire--once for each possible combination of target behavior (supporter or opponent) and group perspective (supporter or opponent). For half of the participants, the wording in all items was ''University of Latvia students'' instead of ''people''.

[290]
---------------
[291]

Third Part of the Questionnaire

Instruction for the own perspective:

Please indicate to what extent you agree or disagree with the following statements. Indicate your response by circling the appropriate number next to each statement!

Instruction for the group perspectives:

How would people who support/oppose Latvia entering the European Union agree or disagree with the following statements.? Indicate your response by circling the appropriate number next to each statement!

Items for measuring the strength of support to Latvia's EU membership:

Latvia should enter the European Union.
[participant instructed to indicate the response on a 5-point scale where

1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

The European Union membership will do more harm than good to Latvia. *

[1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

Latvia will only gain by remaining outside the European Union. *

[1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

I am personally against Latvia's membership into the European Union. *

[1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

[291]
---------------
[292]

It would be better for majority of Latvians if Latvia became a member of the European Union.

[1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

European Union membership threatens the sovereignty of Latvia. *

[1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

I personally support Latvia's membership into the European Union.

[1 = strongly disagree;
2 = disagree rather than agree;
3 = neither agree nor disagree;
4 = agree rather than disagree;
5 = strongly agree]

Note: These items appeared three times in the questionnaire, and correspondingly were rated from three perspectives: own, supporter, and opponent. Asterisked items are reverse-scored. Reliability measures (Cronbach's Alpha): for ratings from own perspective, Alpha = .90, for ratings from the supporter perspective, Alpha = .78, for ratings from the opponent perspective, Alpha = .82.

APPENDIX B. CORRELATION MATRIXES WITH MEANS AND STANDARD DEVIATIONS FOR ALL VARIABLES

Table B1. Attitude towards EU membership

Supporter participants

 

n

M

SD

(1)

(2)

(1) Own perspective

90

3.88

0.55

   

(2) Supporter perspective

85

4.30

0.61

.39**

 

(3) Opponent perspective

87

1.91

0.85

.03

-.48**

Opponent participants

 

n

M

SD

(1)

(2)

(1) Own perspective

46

2.32

0.56

   

(2) Supporter perspective

47

4.29

0.65

-.34*

 

(3) Opponent perspective

49

1.94

0.74

.33*

-.71**

* p < 0.05, ** p < 0.01.

[292]
---------------
[293]

Table B2. Evaluation of the target behavior

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

92

3.53

0.67

         

(2) SB, supporter persp.

91

4.13

0.73

.14

       

(3) SB, opponent persp.

91

1.96

0.82

.08

-.33**

     

(4) OB, own perspective

92

2.63

0.81

-.07

-.13

.19

   

(5) OB, supporter persp.

91

1.99

0.86

.03

-.32

.43**

.30**

 

(6) OB, opponent persp.

91

3.85

1.06

-.27*

.30**

-.41**

.06

.35**

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

48

2.77

0.76

         

(2) SB, supporter persp.

49

3.96

0.87

-.06

       

(3) SB, opponent persp.

49

2.36

0.91

.23

-.22

     

(4) OB, own perspective

48

3.52

0.82

-.01

.09

.23

   

(5) OB, supporter persp.

49

2.13

0.81

.18

-.43**

.30*

.17

 

(6) OB, opponent persp.

48

3.80

0.97

.02

.68**

.01

.37**

-.28

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[293]
---------------
[294]

Table B3. Situation vs. actor ratings

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

92

3.42

0.97

         

(2) SB, supporter persp.

92

3.48

0.95

.29**

       

(3) SB, opponent persp.

91

3.11

1.09

.31**

.00

     

(4) OB, own perspective

92

3.46

1.02

.06

.08

.33**

   

(5) OB, supporter persp.

92

3.14

1.13

.08

-.04

.42**

.23*

 

(6) OB, opponent persp.

91

3.57

1.05

-.04

.12

-.06

-.03

.07

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

48

3.46

0.92

         

(2) SB, supporter persp.

49

3.63

1.11

.34*

       

(3) SB, opponent persp.

49

3.55

0.89

.46**

.29*

     

(4) OB, own perspective

49

3.39

1.00

.28

.47**

.60**

   

(5) OB, supporter persp.

49

3.43

1.12

.05

-.09

-.10

-.10

 

(6) OB, opponent persp.

49

3.47

0.98

.36*

.54**

.41**

.51**

-.07

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[294]
---------------
[295]

Table B4. Stability ratings

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

92

2.96

0.91

         

(2) SB, supporter persp.

91

3.76

0.97

.29**

       

(3) SB, opponent persp.

92

2.49

1.04

.30**

.12

     

(4) OB, own perspective

91

2.87

0.90

.01

-.04

.19

   

(5) OB, supporter persp.

91

2.59

1.00

.19

-.10

.48**

.32**

 

(6) OB, opponent persp.

91

3.59

1.16

-.03

.09

.16

.19

.06

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

48

2.94

0.84

         

(2) SB, supporter persp.

49

3.84

0.99

.15

       

(3) SB, opponent persp.

49

3.02

0.88

.26

.05

     

(4) OB, own perspective

47

3.17

0.96

-.09

.37*

.32*

   

(5) OB, supporter persp.

49

2.86

1.04

.14

-.19

.09

-.08

 

(6) OB, opponent persp.

49

3.61

0.86

.24

.34*

.12

.42**

.10

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[295]
---------------
[296]

Table B5. Control ratings

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

92

3.12

0.94

         

(2) SB, supporter persp.

92

3.82

1.00

.33**

       

(3) SB, opponent persp.

91

2.53

1.08

.09

.05

     

(4) OB, own perspective

92

2.82

1.06

.19

.21*

-.08

   

(5) OB, supporter persp.

92

2.46

0.95

-.16

-.23*

.27**

.26*

 

(6) OB, opponent persp.

92

3.64

1.09

.11

.34**

-.03

.12

-.11

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

48

3.13

1.00

         

(2) SB, supporter persp.

49

3.86

1.02

.03

       

(3) SB, opponent persp.

49

2.94

0.94

.32*

-.10

     

(4) OB, own perspective

49

3.29

1.06

.17

.21

.27

   

(5) OB, supporter persp.

49

2.84

1.12

.02

-.16

.23

.08

 

(6) OB, opponent persp.

49

3.59

1.22

.23

.40**

.12

.41**

-.16

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[296]
---------------
[297]

Table B6. Globality ratings

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

92

3.28

0.82

         

(2) SB, supporter persp.

92

3.33

0.88

.25*

       

(3) SB, opponent persp.

92

3.13

0.92

.01

.10

     

(4) OB, own perspective

91

3.10

0.96

.18

.05

.09

   

(5) OB, supporter persp.

92

3.41

0.92

.03

-.02

.11

.42**

 

(6) OB, opponent persp.

92

3.16

0.76

.21*

.22*

.19

.07

.08

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

48

3.25

0.91

         

(2) SB, supporter persp.

49

3.06

1.05

.23

       

(3) SB, opponent persp.

49

3.35

0.97

.31*

.14

     

(4) OB, own perspective

49

3.14

0.89

.16

.24

.04

   

(5) OB, supporter persp.

49

3.31

0.94

.25

-.04

.20

.12

 

(6) OB, opponent persp.

48

3.19

0.94

.11

.44**

.00

.35*

-.14

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[297]
---------------
[298]

Table B7. Rationality index

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

92

3.29

0.78

         

(2) SB, supporter persp.

91

3.99

0.83

.34**

       

(3) SB, opponent persp.

91

2.67

0.91

.06

-.14

     

(4) OB, own perspective

92

2.94

0.78

.22*

.23*

.08

   

(5) OB, supporter persp.

91

2.68

0.93

.04

-.25*

.36**

.30**

 

(6) OB, opponent persp.

90

3.79

0.94

-.06

.46**

-.05

.09

-.04

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

46

3.13

0.70

         

(2) SB, supporter persp.

49

3.86

0.85

.37*

       

(3) SB, opponent persp.

47

2.96

0.82

.21

-.08

     

(4) OB, own perspective

49

3.45

0.83

.41**

.34*

.30*

   

(5) OB, supporter persp.

49

2.83

0.83

.24

-.06

.08

-.07

 

(6) OB, opponent persp.

48

3.84

0.84

.40**

.47*

.16

.60**

-.04

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[298]
---------------
[299]

Table B8. Subjectivity index

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

91

3.68

0.52

         

(2) SB, supporter persp.

90

3.46

0.54

.39**

       

(3) SB, opponent persp.

89

3.54

0.62

.43**

.40**

     

(4) OB, own perspective

92

3.80

0.53

.48**

.21*

.44**

   

(5) OB, supporter persp.

89

3.68

0.61

.50**

.26*

.48**

.57**

 

(6) OB, opponent persp.

89

3.36

0.64

.35**

.41**

.22*

.40**

.06

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

45

3.81

0.51

         

(2) SB, supporter persp.

47

3.38

0.59

.31*

       

(3) SB, opponent persp.

49

3.67

0.68

.48**

.41**

     

(4) OB, own perspective

49

3.79

0.53

.56**

.32*

.56**

   

(5) OB, supporter persp.

49

3.80

0.57

.48**

.17

.60**

.78**

 

(6) OB, opponent persp.

48

3.54

0.59

.44**

.36*

.41**

.60**

.44**

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[299]
---------------
[300]

Table B9. Accepting others' opinions ratings

Supporter participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

91

3.13

1.04

         

(2) SB, supporter persp.

92

2.75

1.02

.14

       

(3) SB, opponent persp.

92

3.33

1.21

.25*

.04

     

(4) OB, own perspective

91

3.20

1.11

.18

-.01

.06

   

(5) OB, supporter persp.

92

3.52

1.05

.25*

.02

.11

.38**

 

(6) OB, opponent persp.

92

2.64

1.13

-.02

.45**

-.16

.11

.13

Opponent participants

 

n

M

SD

(1)

(2)

(3)

(4)

(5)

(1) SB, own perspective

47

3.55

0.88

         

(2) SB, supporter persp.

49

2.98

1.15

.14

       

(3) SB, opponent persp.

49

3.22

1.12

.42**

.02

     

(4) OB, own perspective

47

3.02

0.92

.20

-.19

.20

   

(5) OB, supporter persp.

49

3.29

1.14

.27

-.32*

.23

.27

 

(6) OB, opponent persp.

49

3.06

1.03

.09

.30*

-.12

.25

.16

SB = supporter behavior; OB = opponent behavior

* p < 0.05, ** p < 0.01.

[300]
---------------
[301]

AUTHORS' BIOGRAPHIES

Girts Dimdins is currently a doctoral student at the Department of Psychology, Stockholm University. His research is focusing on the role of perspective taking in intergroup perception. E-mail: gds@psychology.su.se

Henry Montgomery is a professor at the Department of Psychology, Stockholm University. His research deals with mental processes and structures associated with people's values, attitudes and decisions. A particular interest is the role of perspectives in people's interpretations of actions and action alternatives. E-mail: hmy@psychology.su.se

Ivars Austers is an associate professor at the Department of Psychology, Faculty of Education and Psychology, University of Latvia. His research interests are social cognition and political psychology. E-mail: austers@lanet.lv

[301]
---------------
[302]