Correspondence concerning this article should be sent to: Hiroshi Nonami, School of Sociology, Kwansei Gakuin University, 1-1-155 Uegahara, Nishinomiya, Hyogo 662-8501, Japan. (Email: firstname.lastname@example.org)
Abstract: The present study investigated the social determinants required to stimulate consistency in minority members’ attitudes and behavior. The experiment was designed to examine the consistency of individuals in the context of an intergroup belief cross wherein the majority belief (or minority) in a categorical in-group shifted to the minority (or majority) in the out-group. The results indicated that the members of a majority within an in-group consistently preserved their behavioral intentions and beliefs even though they were positioned as a minority within an out-group. That is, the minority members in the whole intergroup context maintained consistency diachronically as a result of support from the majority of their in-group. The theoretical implications on the consistency of minority members, such as whistle-blowers in companies, are discussed.
However, if an individual consistently deviates from the majority, their deviation will be more remarkable, and their risk of isolation will increase. Despite this, examples of consistent active minorities who asserted their deviant opinions to the majority are observed frequently in our society (Moscovici, 1976). Thus, investigation is needed into the social determinants that stimulate this consistent behavior among individuals who are minorities in intergroup contexts.
In our society, a belief that is held by the majority in one group is frequently supported only by the minority in another group. For example, the evaluation of dishonest acts can differ across companies or among Japanese consumers (Okamoto & Konno, 2006). Although some acts are approved by the majority in one company, they are rejected as an injustice by the majority in other companies or by consumers. Thus, an individual who evaluates the act as an injustice will belong to the minority in the former company, but will be a member of the majority in the latter. In this context, support for a belief from the majority or minority reverses between the in-group and the out-group. The behavioral intentions of individuals in this context are thought to be determined based on the support from the majority (or minority) within both the categorical in-group and the out-group. For example, a whistle-blower who informs consumers of an injustice in their company is a minority in the company, but belongs to the majority in society as a whole (Okamoto, Wan, & Honda-Howard, 2006). In contrast to a whistle-blower, people generally intend to be members of the majority in their company, even if they belong to the minority in out-groups.
The above studies investigated the effect of behavioral consistency of minority members in terms of their influence on other members within the group. However, the decision to maintain consistent behavior should be affected by both the in-group and the out-group within an intergroup context, even though their consistency effectively alters only the members of the in-group. Nonami (2006) defined an intergroup belief cross as the context in which the ideological majority and minority is reversed between the in-group and out-group; this context can be used to examine the influence of support from the majority or minority in both the in-group and the out-group on the attitudinal and behavioral consistency of individuals. Although Baker and Petty (1994) investigated the influence of support from the majority or minority on the attitude of individuals, their study manipulated only the majority or minority within the in-group as an independent variable. The present study investigates the social determinants underlying the consistent behavior of minority members in the context of an intergroup belief cross.
In an intergroup belief cross, differences in numerical size between the categorical in-group and out-group are important because unequal group size will affect one's position in the intergroup context as a whole. For instance, a personal belief consistent with the majority in a smaller categorical in-group might be positioned as a minor opinion having only minority supporters among total members of the in-group and the out-group if the belief is supported by only a minority in the larger out-group. According to Yasuno (2001, 2006), people maintain consistent opinions on the basis of the support from people close to them, such as family members or friends, even if their opinions are minor in society as a whole. If people are admitted to an in-group majority, they might preserve their attitudinal or behavioral consistency, even if they are located as a minority in the out-group, because support from in-group members is regarded as more important to individuals than support from out-group members (Allen & Wilder, 1975; Martin, 1992). Moreover, members of small-scale groups are said to identify highly with their in-group (Simon & Brown, 1987). Thus, it is predicted that the support from the majority, even in a small in-group, is an important factor in individuals maintaining their attitudinal and behavioral consistency.
The above prediction is also consistent with optimal distinctiveness theory (Brewer, 1991; Pickett, Silver, & Brewer, 2002). This theory asserts that two motives affect in-group bias and social identity: assimilation and differentiation. Individuals who are part of a minority within the in-group are said to evoke their need for assimilation, and wish to conform to the larger group. They are thought to have two choices in the context of an intergroup belief cross. First, they can conform to the majority within their in-group by changing their opinion. Second, they can move from their in-group to an out-group to acquire support from the majority for their original opinion. The first option is expected to occur more frequently than the second, because movement to a categorical out-group brings about a higher cost to the individual. Thus, it is predicted that individuals who are supported by only a minority of their in-group would change their attitude and behavior to conform to the majority within that in-group. Conversely, when individuals are supported in their original beliefs by the majority of their in-group, their need for assimilation would be satisfied and their original attitude and behavior would be maintained, even if the size of their in-group is smaller than that of out-groups.
Accordingly, individuals in the majority within their in-group will consistently preserve their attitude or behavior regardless of the support from the majority or minority in the out-group. Even if they are positioned as members of a minority in the whole intergroup context, the consistency will be maintained. In contrast, the intention to maintain consistency among members of a minority within the in-group is thought to be threatened even if they are placed as a majority in the whole intergroup context. That is, individuals whose opinion is congruent with the minority within their in-group will shift their attitude or behavior regardless of the support from the majority or minority in the larger out-group. The present study examined the following two hypotheses in the context of an intergroup belief cross consisting of groups differing in size.
When individuals are supported by the majority of their in-group, their behavioral intention will be maintained consistently regardless of whether they have support from the majority or the minority within the larger out-group. In other words, even if they are located as a minority in the intergroup context as a whole, their diachronic consistency will be high.
Individuals who are supported by a minority of their in-group will change their opinion even if that opinion was congruent with the majority of the intergroup context as a whole. That is, their diachronic consistency will be low regardless of the support from the majority or minority within the larger out-group.
Nonami (2006) manipulated social category within real groups (two departments of a school within a university) to investigate the consistency of minority opinions within intergroup belief crosses. In addition to the effect of the majority within the in-group, the results revealed an effect of the out-group majority on the diachronic consistency of individuals. However, the manipulations in that study were considered inadequate due to the effects of the superordinate category common to both the in-group and the out-group, that is, being members of the same school weakened the boundary between the groups. The present study adopted the minimum group paradigm to manipulate the social categorization of the in-group and out-group (e.g., Tajfel, Flament, Billig, & Bundy, 1971; Billig & Tajfel, 1973). This procedure is expected to sufficiently manipulate the intergroup belief cross while excluding any effects of superordinate category.
To manipulate the intergroup belief crosses, the current experiment required two groups of unequal size, with the requirement that members of the small group consider the opinions in both groups to determine their own opinions. The context of companies in our society could be defined as an example of this situation because members of a company are thought to pay attention to the minority or majority in other companies as well as in their own company to make group decisions. The present experiment employed the scenario method in which all participants play the role of a staff member of a company.
Data from 173 undergraduate participants (60 male and 113 female) was analyzed.
A novel topic was chosen to ensure the scenario was independent of any prior information or beliefs that the participants may have held. An essay (approximately 260 words in Japanese) which explained the merits and demerits of a “new kind of yogurt” was constructed. Merits were described as follows: “a new kind of lactobacillus is valued medically because it activates human digestion, resulting in prevention of adult-onset diseases. This yogurt might result in large profits for your company.” In contrast, demerits were described as follows: “the lactobacillus requires hard preservatives because of its delicacy. A yogurt with these preservatives might result in significant costs to control the quality and harm the image of your company.” In a preliminary analysis, 50 undergraduate students (they were recruited as another sample that was separated from the above 173 participants) read the essay. The proportion of students who agreed with the yogurt did not differ significantly from the proportion who disagreed (26 students agreed, and 24 students disagreed, χ2(1) = 0.08, ns).
The consecutive scenes by which to measure the diachronic consistency in the scenario were prepared according to the following procedures. First, the opinions on the yogurt expressed in free-form writing by the 50 students in the preliminary test described above were assessed, and 32 opinions (16 positive and 16 negative opinions) were selected. Second, the persuasiveness of those 32 opinions was estimated by another group of students (N = 32) using a 5-point scale. Based on these ratings, eight positive opinions and eight negative opinions were chosen for the main experiment based on the equivalency in their rated persuasiveness (mean of positive eight opinions = 3.34, and mean of negative eight opinions = 3.28, t(14) = 0.52, ns). An example of a positive opinion was, “this kind of epoch-making yogurt must be commercialized at once ... ,” and a typical negative opinion was, “the economical regular yogurt is better than the new yogurt that includes hard preservatives.” These 16 opinions were presented consecutively across four pages of the stimulus booklet, and the participants’ behavioral intentions were measured across the pages to define diachronic consistency. The order of the pages was counterbalanced.
The booklets, consisting of the essay and the above four scenes, were printed on two colors of paper: white and blue. All participants received booklets in both colors in their classroom during a social psychology lecture. First, to manipulate the in-group and out-group on the basis of social category, all participants were told that they would be assigned as staff members to commercialize a “new kind of yogurt” for either a small-scale food company employing approximately 100 staff or a large-scale food company employing approximately 300 staff. Second, a trivial test was carried out in which participants estimated the number of candies in a transparent bottle (this was presented as an important task for food company staff). Based on their overestimation or underestimation (nobody estimated the number correctly) they were asked to open and read only the white or blue booklet, respectively. In fact, all participants were actually assigned to the small-scale company scenario, regardless of the paper color. During the experiment, they could observe the color of the booklets (white or blue) that the students around them had opened. The assignment procedure and the observable white or blue papers were designed to increase the reality of in-group and out-group for the participants, and highlight their in-group identity.
After being assigned a colored booklet to open, participants read the essay about a “new kind of yogurt” on the first page of the booklet. On the next page, they expressed their initial attitudes toward the commercialization of the yogurt indicating either their agreement or disagreement. In addition, they evaluated the yogurt on five items using 7-point scales, such as “strongly bad (1)–strongly good (7),” and “strongly undesirable (1)–strongly desirable (7).”
Three charts were presented on the third page that illustrated the results of fictitious marketing research about the yogurt. One chart indicated the percentage of approval and disapproval among staff of the small-scale company as the in-group, the second showed the analogous percentages for the large-scale “out-group” company, and the final chart showed the total percentages for both companies grouped together. Based on these charts, participants were expected to understand that their own initial attitude declared on the second page was supported by either the majority or minority at the in-group and the out-group.
Next, each of the eight positive or negative opinions selected as a result of the above-described preliminary analyses was presented across four pages as opinions expressed by staff members of the small-scale and large-scale company. That is, participants read four opinions (one agreement and one disagreement from their in-group and the other two opinions from the out-group). On each page, participants responded to items measuring their intention to insist on their own opinion.
Following the four pages presenting the opinions, the five items used for measuring the initial evaluations of the yogurt were represented on the final page. Upon completion of all pages, the true purpose of the experiment was explained to all participants.
In the present experiment, a 2 (in-group support: majority or minority in the small-scale company) × 2 (out-group support: majority or minority in the large-scale company) design was adopted. On the pretext of fictitious marketing research, the percentage of approval and disapproval for the yogurt in the small-scale company was represented as 76% and 24% in the charts described above. The majority was in agreement half of the time and in disagreement the other half of the time. For the larger company, the percentages were similar, 78% and 22%, and again agreement and disagreement were each in the majority half of the time. This manipulation is same as that in Nonami (2006). Figure 1 illustrates the experimental conditions in the current study.
On the basis of their initial attitude toward the “new kind of yogurt,” each participant's view was coincident with either the majority or minority, and the participants were counted as members of the majority or minority within the in-group and the out-group, respectively. For example, those students who initially indicated agreement and were informed that the majority for both the in-group and the out-group agreed with the yogurt, were counted as participants of the in-group majority support/out-group majority support cell (case 1 in Figure 1), as were those participants who initially indicated disagreement and were informed that the majority in both groups disagreed. The number of participants assigned to each of the four experimental cells ranged from 30 to 57.
After reading both positive opinions and negative opinions about the “new kind of yogurt” on each of the four pages, participants responded to three items such as “If I was asked to present my opinion to executives of my company, I would insist on either agreement (or disagreement) with this yogurt.” These questions were repeated across the four pages to measure the diachronic consistency of the participants’ behavioral intentions to assert their opinions. For each item they first expressed their attitude (agreement or disagreement) toward the yogurt, and then estimated their behavioral intention on a 7-point scale ranging from weakly (1) to strongly (7). If the first attitudinal category on each item changed from initial indication of agreement or disagreement that participant expressed before reading the scenario, the intention score on the 7-point scale was reversed to a negative score.
In addition, the absolute differences between scores evaluating the “new kind of yogurt” on five items before and after reading all opinions were calculated to indicate cognitive change. Finally, identification with the in- and out-group was measured using four items, such as “I feel that the success of this company is my own success” using 7-point scales.
There were no sex differences across cells (χ2(3) = 5.50, ns), and also no significant difference between the percentage of participants who expressed initial agreement or disagreement (χ2(3) = 1.61, ns). After the percentage of approval and disapproval with the “new kind of yogurt” in the in-group and out-group was presented in the charts, all participants indicated whether their own initial position was in accordance with either the majority or the minority of the in- and out-group. They also indicated whether they belonged to the majority or the minority in the whole context including both the in-group and the out-group. None of the participants wrongly indicated their attitudinal position in either the two groups or the context as a whole. The means of the four items measuring identification with each group, in- and out- (α = 0.83 and α = 0.85, respectively), were 5.33 and 2.05, respectively. Identification with the small-scale company as the in-group was higher than with the large-scale company as the out-group, and no significant main effects or interaction were found in the in-group support (2) × out-group support (2) ANOVA.
Consistent behavioral intentions
Figure 2 shows the mean responses to the three items that measured the diachronic consistency of the intention to insist on an opinion across the successive four pages of the scenario. A three-way mixed ANOVA [in-group support (2) × out-group support (2) × succession of scenes (4)] on this data revealed a main effect of in-group support, F(1,169) = 10.74, p < 0.01. Regardless of the support from the out-group (either the minority or the majority), participants indicated consistently higher intentions to insist on their opinion across all four pages when their attitude was in accordance with the majority of the in-group as compared to the situation when they were supported by only a minority of the in-group (M = 4.48, SD = 2.26 in the former, and M = 3.11, SD = 3.11 in the latter). In addition, a main effect of succession was found, F(3,507) = 17.84, p < 0.001. Intentions to assert an opinion were higher on the first and second pages than on the third and fourth pages (M = 4.51, SD = 2.37 for the first, M = 3.91, SD = 2.96 for the second, M = 3.40, SD = 3.39 for the third, and M = 3.34, SD = 3.66 for the fourth page). The main effect of out-group support was not significant, and there were no significant interactions.
The participants’ evaluation of the “new kind of yogurt” was measured with the same five items before and after reading all opinions. The absolute difference between the means before and after is defined as the degree of cognitive change, and shown in Figure 3. An in-group support (2) × out-group support (2) ANOVA revealed a significant main effect of in-group support, F(1,169) = 5.57, p < 0.05. In addition to this, there was a significant interaction, F(1,169) = 4.75, p < 0.05.
Analysis of the simple main effect of out-group support revealed this was not significant among participants whose attitudes were consistent with the majority of the in-group, F(1,169) = 1.23, ns. In contrast, the effect of out-group support was slightly significant for those whose attitudes accorded with the in-group minority, F(1,169) = 3.85, p < 0.07. In other words, participants could maintain their own evaluations when they were supported by the majority of their in-group, regardless of support from the majority or minority of the out-group (M = 0.53, SD = 0.56 in the former, and M = 0.37, SD = 0.62 in the latter). But among participants with only minority support from their in-group, their degree of cognitive change was increased slightly by support from the minority of the out-group compared with support from the majority of the out-group (M = 0.85, SD = 0.69 in the former, and M = 0.55, SD = 0.76 in the latter).
There was no simple main effect of in-group support among participants whose attitudes were consistent with the majority of the out-group, F(1,169) < 1.00, ns. However, there was a significant effect for those who were consistent with the out-group minority, F(1,169) = 9.35, p < 0.01. Participants in the out-group majority showed less cognitive change regardless of support from either the in-group majority or minority. Cognitive change was greater in the out-group minority when their attitudes were consistent with the minority of the in-group, compared with when they were consistent with the majority of the in-group (M = 0.85, SD = 0.69 in the former, and M = 0.37, SD = 0.62 in the latter).
The partial correlation coefficients for cognitive change, in-group identity, and behavioral intentions (the means across four pages) were calculated. Only the correlation between cognitive change and behavioral intention was significant, r(173) = –0.55, p < 0.001. In-group identity did not correlate with either intention (r(173) = 0.08, ns) or cognitive change (r(173) = 0.07, ns).
Since Moscovici et al. (1969), researchers have demonstrated that minority members can alter the attitudes and behaviors of the majority when they display a consistent behavioral style. The present study investigated the social determinants for minority members to maintain behavioral consistency.
Consistency in behavioral intentions to assert their own opinions was higher and maintained diachronically among participants who were supported by the majority of their in-group, even when their attitude was congruent with the minority in the whole intergroup context. This is consistent with the first hypothesis. Moreover, they indicated less cognitive change as long as they were supported by the in-group majority; their final beliefs were consistent with their original cognitive views, regardless of the support from the out-group (majority or minority).
In contrast, participants whose beliefs were congruent with the in-group minority decreased the consistency in their intentions, irrespective of whether their attitudinal position accorded with the majority or the minority of the out-group. These results support the second hypothesis. However, those in the in-group minority showed fewer cognitive changes when their attitude was congruent with the majority of the out-group, compared with when they were congruent with the out-group minority. The pattern in which participants had support from the in-group minority and the out-group majority was defined as a superficial compliance because they decreased their behavioral intentions to assert their own opinions, but they maintained their original cognitive evaluation. In contrast, participants who belonged to the minority within both the in-group and the out-group showed increased cognitive changes and decreased behavioral consistency. This pattern might represent psychological mobility from the minority belief to the majority.
These results are consistent with optimal distinctiveness theory (Brewer, 1991; Pickett et al., 2002). Individuals whose attitudes were in accordance with a minority within their in-group were thought to evoke the need for assimilation, and changed their cognitions or behaviors to conform to the majority. However, individuals who belong to a majority of their in-group and to a minority in the out-group maintained the diachronic consistency of their behavioral intentions despite their position as a minority within the whole intergroup context. Thus, the position as a minority of the in-group is thought to be more effective in stimulating the need for assimilation than being positioned as a minority of the out-group.
Optimal distinctiveness theory (Brewer, 1991; Pickett et al., 2002) also predicts that the need for differentiation might be evoked among people whose attitudes are consistent with the majority of the in-group or the out-group. In particular, differentiation would be aroused when people are supported by the majority of the in-group. However, it may be only when people decide their opinions on personal tasks that they evoke the need for differentiation. In the present experiment, participants were concerned with group strategic decisions based on their role in a company: this is defined as a group task. In addition, in order to exclude the effects of the superordinate category found in Nonami (2006), the boundary between the in-group and out-group was emphasized by the description of a context consisting of two different companies. This disrupted their mobility from the in-group to the out-group. In this situation, it would be difficult to express opinions that differed from the in-group majority, perhaps preventing the need for differentiation from being evoked. In contrast to the present study, Nonami (2006) reported that in addition to the effect of being positioned as the majority of the in-group, congruency with the majority of the out-group was also effective in preserving attitudinal and behavioral consistency in an ideological task (evaluation of a bio-ethical problem) in which individuals decided their opinions privately. Further research is needed to elucidate the difference between group and personal decision-making tasks on the consistency of minority members.
The results of the present study revealed that an intergroup belief cross in which a majority in the in-group is positioned as the minority in the out-group is an important factor in individuals maintaining their behavioral consistency within an intergroup context. Conversely, behavioral consistency decreased in another type of intergroup belief cross context in which the minority within the in-group reverses to the majority in the out-group. Thus, the position as the majority of the categorical in-group is thought to underlie the consistency of minority members in society. In other words, a minority in society can maintain their behavioral consistency on the basis of support from the majority within their categorical in-group. This also suggests that many people will comply with the majority of their in-group, even if they are positioned as a minority within society as a whole. Even in a company that has group norms against the zeitgeist (Pérez et al., 1995; Sanchez-Mazas, 1996), except for a few whistle-blowers, people will generally follow the majority within their company.
However, as noted above, the current study also showed a superficial compliance consisting of smaller cognitive changes and a decline in the intention to assert their original opinions among people who were supported from the in-group minority and the out-group majority. This suggests people maintain attitudinal consistency on the basis of support from the majority in the out-group despite their position as a minority within the in-group. Research on whistle-blowing in society has frequently reported this kind of compliance in group decisions within companies (Okamoto & Konno, 2006; Okamoto et al., 2006). When these individuals express attitudinal consistency on the behavioral level, they will be whistle-blowers and a minority within the company. Further investigations into the psychological predictors of behavioral consistency for a minority in a group decision-making context are needed.
In addition, the present study revealed that cognitive changes increased and behavioral consistency decreased when individuals were in the minority of both the in-group and the out-group. In this case, their psychological mobility from the minority to the majority is interpreted as a conversion behavior resulting from the majority influences of the in-group and out-group. This finding is consistent with the results of Baker and Petty (1994), who reported that the conversion process was aroused by individuals’ accordance with the minority or discordance with the majority in their in-group. In the present study, the conversion process was stimulated by accordance with the minority (and discordance with the majority) of both the in-group and the out-group.
In the present study, cognitive changes were found only when people's attitudes accorded with the minority in both the in-group and the out-group. But compliance (smaller cognitive changes and decline of intentions) was observed among those whose attitudes accorded with the in-group minority and the out-group majority. In the context of the Baker and Petty (1994) study in which participants were informed only of the supportive (or non-supportive) distribution of members within the in-group without information on the out-group, people decided their attitudes and behaviors based only on the distributions within the in-group. However, it is thought that they would determine their attitudes and behaviors on the basis of the distributions in both the in-group and the out-group when informed about both distributions. It is suggested that whether their accordance with the minority stimulates the conversion process depends on where the minority is located: in the in-group or the out-group. In future studies, this theoretical implication should be investigated in more detail to clarify the effect of support from the in-group or the out-group on the conversion and compliance process.
One limitation of the current experiment was that there was the effect of succession of scenes on behavioral intentions. Across the successive four pages behavioral intentions were lower on the latter two pages than on the first two pages. Because the order of opinions was counterbalanced across the pages, this effect was not derived from the differences in opinions across the pages. Rather, it was likely evoked by the number of opinions that participants had read on the scenario. With each successive page they were aware of more numerous (positive and negative) opinions on the topic (a “new kind of yogurt”). Thus, they were thought to deliberate on the topic as they read, and changed their opinions. This kind of effect is thought to be inevitable when the scenario method is employed with successive trials that require participants to read various opinions, even though this design is indispensable for measuring the diachronic consistency of behavior. Thus, this procedure needs to be improved in future.
The context consisting of two differently sized companies that was presented to manipulate the intergroup belief cross should be noted as another limitation. It is possible that some participants interpreted the difference in size as a difference in status between the groups. That is, the small company could be confused with a low-status company. Simon and Hamilton (1994) reported that group size and status interactively affect self-stereotyping and in-group homogeneity of minority members. Members of a high-status minority are said to increase self-stereotyping relative to members of a high-status majority. Thus the behavioral and attitudinal consistency of minority members could be predicted to depend on the status of the group. As a more serious limitation, it should also be noted that the intergroup belief cross in the current study was defined as a game situation in which members of a company make decisions on the basis of predictions about decision making by people in other companies (Okada, 1996). In such situations, in addition to the differences in group status or boundaries between the in-group and out-group and others, decisions are influenced by various motivations: to maximize benefits for themselves, the benefits to the in-group, or the benefits of all groups. To examine the implications of the current study for future research, the experimental situation in which the intergroup belief cross is manipulated should be reconsidered.
Moscovici's (1976) genetic model advocates that the diachronic behavior of individuals results in innovation within the in-group. This view opposes the functionalist model that emphasizes the influence of a group on individuals. The former stresses the interdependence between individuals and the group, while the latter underlines the dependence of individuals on the group. The present study found that the in-group majority reinforces the consistency of the minority within an intergroup context. This suggests that individuals depend on the majority within the in-group to innovate the social system. Thus, the genetic model and functionalist model appear to be linked. Research into the social determinants of the consistency of a minority will theoretically consolidate Moscovici's paradigm, and explain the decision making process of people in minority groups, such as whistle-blowers.