Perceived variability and stereotype change Author Miles Hewstone, Jürgen Hamberger Publication Year 2000 Type Journal Article Abstract This article explores the neglected issue of how attempts to change perceived variability might positively, or adversely, affect attempts to change central-tendency measures of stereotyping. Two studies investigated the impact of group variability on stereotype change. Study 1 crossed pattern of disconfirming information (concentrated/dispersed) with a direct manipulation of the group's perceived variability (high/low); Study 2 crossed pattern with an indirect manipulation of the group's perceived variability (no subgrouping/subgrouping). Both studies yielded a reliable interaction between pattern of stereotype-disconfirming information and perceived group variability. A central-tendency measure of stereotyping (and a measure of dispersion in Study 2) was weaker in the dispersed than concentrated conditions only when variability was low. Disconfirmers were also perceived as more typical in the dispersed than concentrated conditions when variability was low. Mediational analyses showed that the relationship between stereotyping and perceived typicality was bidirectional rather than typicality driving stereotyping. These results may help us to design interventions that achieve multiple types of generalized change in the way outgroups are so often negatively perceived. Keywords perceived variability, central-tendency measures, group variability, stereotype disconfirming information Journal Journal of Experimental Social Psychology Volume 36 Pages 103-124 Type of Article Journal Article DOI 10.1006/jesp.1999.1398 Full text The following is an excerpt of the intervention methodology. For more information, please see the full text of the article on the publisher's website or through your institution's library. STUDY 1 Method Overview. This study investigated the effects of pattern of information and a direct manipulation of variability on the perceived typicality of members of the ‘‘physics students’’ group and stereotyping of the group as a whole. Participants and design. The participants were 68 undergraduate students (28 males and 40 females) at Cardiff University [...]. The participants were randomly assigned to one of the four cells of a 2 (perceived group variability: low/high) X 2 (pattern of disconfirming information: concentrated/dispersed) between-subjects design. Each condition contained 17 participants and approximately equal numbers of males and females. Procedure. [...] To manipulate perceived variability, participants were told that we had developed a ‘‘similarity index,’’ a scale ranging from 1 to 100 with higher numbers denoting greater similarity within each student group. The participants were then provided with preliminary results (in fact, fictional), in the form of three graphs, of a survey of student groups in Wales showing this similarity index for five student groups (Architecture, Law, Physics, Business, and Sociology). The three graphs (histograms for ‘‘attitudes,’’ ‘‘personality,’’ and ‘‘overall similarity’’) indicated that physics students were either more similar (low perceived variability) or less similar (high perceived variability) to each other than were the other student groups. To check on this manipulation, participants were asked, immediately after looking at the graphs, to rate how similar the members within each student group were to each other (‘‘not at all,’’ 1; ‘‘very,’’ 7). Participants were then told that they would learn about eight members of the physics students group in turn, and each group member was presented on a separate page, with the order of group members randomized across booklets. [...] In the concentrated condition the stereotype-inconsistent information was all assigned to two of the eight members (who each exhibited six inconsistent and no other behaviors); the remaining six members each showed two stereotype-consistent and four irrelevant behaviors. In the dispersed condition the stereotype-inconsistent information was spread across six members of the group such that each one slightly disconfirmed the group stereotype (each member displayed two stereotype-inconsistent behaviors, one stereotype-consistent behavior, and three stereotype-irrelevant behaviors); the remaining two members each displayed three stereotype-consistent behaviors and three stereotype-irrelevant behaviors. [...] Dependent measures. After reading the six sentences about each group member, participants rated how typical he was of physics students in general (‘‘not at all typical,’’ 1; ‘‘very typical,’’ 7) before reading about the next member. Participants then rated on a 100-mm scale how characteristic each of six traits was of physics students in general (‘‘not at all,’’ 0; ‘‘very,’’ 100); two traits were stereotype consistent (‘‘interested in new technology,’’ ‘‘hard working’’), two inconsistent (‘‘fashion conscious,’’ ‘‘involved in student politics’’), and two irrelevant (‘‘sociable,’’ ‘‘generous’’). STUDY 2 Method Overview. This study investigated the effects of pattern of information and an indirect manipulation of perceived variability (subgrouping) on the perceived typicality of members of the ‘‘science students’’ group and both central-tendency and dispersion measures of the extent to which the group as a whole was stereotyped. Participants and design. The participants were 64 undergraduate ‘‘arts’’ students (21 males and 43 females) at Cardiff University [...]. The participants were randomly assigned to one of the four cells of a 2 (perceived variability: no-subgrouping/subgrouping) X 2 (pattern of disconfirming information: concentrated/dispersed) between-subjects design. Each condition contained 16 participants and approximately equal numbers of males and females. Stimulus materials and procedure. The materials and procedure for Study 2 followed closely that of Study 1, with the following changes. First, the target outgroup which had to be rated was not ‘‘physics’’ students but ‘‘science’’ students because we had to choose a group for which participants could list subgroups. Due to this modification arts students were chosen as participants. [...] In the subgrouping condition participants were first shown a pie-chart with the first-year intake of science students for several faculties of the University [...] to illustrate how a larger group (‘‘science students’’) could be split up into subgroups. [...] Participants were then asked to think more about the ‘‘science students’’ group and to identify any types or subgroups (space was given for eight), in addition to those shown in the chart, they thought existed within the group of ‘‘science students’’ and to list them. These groups could be formal (e.g., engineering majors) or informal (e.g., ‘‘anoraks,’’ ‘‘nerds’’). Participants were asked to imagine they were describing the subgroup to a friend. After completing this task, participants were asked to think about the groups in the chart as well as those they had listed and to write a few lines of description about what each subgroup was like and how it was different from other subgroups. Immediately following this manipulation, participants were asked to rate how similar to or different from each other they though science students were as a group on a 7-point scale (‘‘they’re all completely different from one another,’’ 1; ‘‘they’re pretty much alike,’’ 7). In the no-subgrouping condition, participants rated the similarity of the science students without prior manipulation. [...] Dependent measures. The perceived typicality of each group member and rated characteristics of the group as a whole were measured in the same way as in Study 1. Participants were also asked to use slashes to mark where on each 100-mm line the most extreme science students would fall. The difference between the rating of the most and least extreme member of the group (full range) was calculated for each of the two consistent and inconsistent traits. Type of Prejudice/Bias Minimal/Artificial Country United States Method Lab Setting College/University Google ScholarDOIBibTeX