Journal of Experimental Social
Psychology, 36, 275-303
(2000).
© Academic Press
Recruitment
of Exemplars as Reference Points in Social Judgments
University of North Florida and Instytut Psychologii Polskiej Akademii Nauk
Krzysztof Konarzewski
Instytut Psychologii Polskiej Akademii Nauk
Michael A. Motes
University of North Florida
Michael Motes is now at Texas Christian
University.
We wish to thank Pawel Grzelak, Alexa Smith,
Zuzanna Smolenska, Magda Wahl, and Linda Yeldell for their help at various
stages of data collection and Gary Brase, Tom Harlow, Linda Henkel, John
Skowronski, Eliot Smith, and Michael Zarate for their thoughtful comments on
earlier versions of the manuscript.
Address reprint requests and correspondence
to Jerzy J. Karylowski, Department of Psychology, University of North Florida,
4567 St. Johns Bluff Road, South, Jacksonville, FL 32224-2645. E-mail:
jurek@unf.edu
Four experiments examined the recruitment of individual-person exemplars as reference points in social judgments. Making a judgment regarding one person facilitated making the same judgment regarding another, particularly when the two targets were of the same sex and similar age. For the category of sex, this category-specific facilitation was strongest for participants who were highly traditionally sex-typed. Compared to initial judgments regarding specific persons, initial judgments regarding social prototypes were less effective in facilitating subsequent judgments regarding other persons. This occurred despite greater perceived similarity between prototype-person pairs than between person-person pairs. In general, the predicted effects occurred less consistently for self than for other familiar-person exemplars. Overall, the results provide support for a privileged role of exemplars in social judgments and for the relevance of the social categories of sex and age in determining which specific exemplars are used in making judgments regarding particular targets.
Recruitment of Exemplars as Reference Points in Social Judgments
It is often assumed that when making judgments about a target person, we access not only what we know regarding attributes and behaviors of the target but also our knowledge regarding related social entities. Most research in this area has focused on involvement of mental representations of classes of people, such as nationality-specific, race-specific, or gender-specific stereotypes or prototypes (see Hamilton & Sherman, 1994; Kunda & Thagard, 1996; Mackie & Hamilton, 1993; Macrae, Stangor, & Hewstone, 1996 for recent reviews of stereotyping literature). However, the role of cognitive representations of specific individuals (exemplars) as points of reference in making social judgments has also been proposed (Andersen & Baum, 1994; Andersen & Cole, 1990; Andersen, Glassman, Chen, & Cole, 1995; Holyoak & Gordon, 1983; E. R. Smith & Zarate, 1992). As is the case with prototypes, when activated as social reference points, exemplars may serve at least two types of functions. They may be used as stand-ins for absent, poorly developed, and/or relatively inaccessible representations of the target, thus leading to analogical social judgments (cf., Beike & Sherman, 1994) or as anchors, standards of comparison against which the target is judged (cf., Biernat, Manis, & Nelson, 1991; Biernat & Manis, 1994).
Evidence from nonsocial areas indicate that exemplars are used as points of reference in a variety of judgment tasks including judgments of magnitude (Holyoak, 1978), color (Rosch, 1975; Sun, 1983), vertical position (L. B. Smith, Cooley, & McCord, 1986), and distance (Holyoak & Mah, 1982; Sadalla, Burroughs, & Staplin, 1980). Within social psychology, several lines of research provide, mostly indirect, evidence for the involvement of representations of someone other than the target person in judgments regarding that target. For instance, research on attitudinal assimilation - contrast effects (Hovland & Sherif, 1952), false consensus (Ross, Greene, & House, 1977, see also Marks and Miller, 1987 for a review), attributive projection (Holmes, 1968; Lemon & Warren, 1976; Shrauger & Patterson, 1974), and attributive contrast (Dunning and Cohen, 1992) is often used to support the view that the self plays a role in judgments regarding others. Conversely, the impact of our perceptions of others on how we see ourselves has been postulated in research stemming from social comparison theory (Festinger, 1954; Suls & Wills, 1991) and from various theories of identification (Bandura, 1969; Freud, 1921; Sears, Rau, & Alpert, 1965). In the same vein, research shows that information concerning significant others (parents, siblings, teachers and peers) may constitute an integral part of a chronically accessible self-concept (Aron, Aron, Tudor, & Nelson, 1991; Lord, Desforges, Chacon, Pere, & Clubb, 1992; McGuire & McGuire, 1988).
Recent studies suggest that similar processes may occur without involvement of self as either target or reference point. In particular, Andersen and her associates (Andersen & Baum, 1994; Andersen & Cole, 1990; Andersen, et al., 1995) demonstrated that people may incorporate characteristics of significant others into their perceptions of new targets. Results of those experiments indicate that familiarity facilitates recruitment of exemplars to be used as reference points in social judgments. Furthermore, the effect of familiarity seems to be mediated by chronic accessibility (Andersen, et al., 1995). Because representations of highly familiar others tend to be highly accessible, they are more likely to serve as social reference points. Not surprisingly, contextual accessibility can also facilitate recruitment of exemplars as reference points both for representations of highly familiar others (Andersen et al., 1995) and for newly formed representations of previously unfamiliar persons (Lewicki, 1985, 1986).
Although these studies have taken important steps toward establishing the role of exemplars (other than the self) as reference points in social judgments, they do not allow for direct and systematic comparisons between effects due to involvement of representations of highly familiar others and similar effects due to involvement of the self and (except for Andersen & Cole 1990, Experiment 3) due to the involvement of social prototypes. Such comparisons are important not only because they could provide a direct link between this new line of research and previous research pointing to the role of self and social prototypes as social reference points but also because they could help to rule out the possibility that the effects obtained for exemplars were mediated by activation of self-knowledge and/or by activation of social prototypes. Because representations of others are likely to share elements with representations of the self (Markus, J. Smith, & Moreland, 1985; Lemon & Warren, 1976; Shrauger & Patterson, 1974) and with generalized social representations, including prototypes (see Hamilton & Sherman, 1994; Kunda & Thagard, 1996; Mackie & Hamilton, 1993; Macrae, Stangor, & Hewstone 1996; Sedikides, Schopler, & Insko, 1998, for recent reviews), such a possibility should be carefully considered. Furthermore, existing experiments provide little guidance regarding what, other than chronic accessibility, determines the choice of particular exemplars as points of reference in judgments regarding specific targets. In other words, little is known about the rules people use to recruit exemplars as reference points. We believe that the concept of natural social categories provides a useful point of departure for research in this area.
Natural Social Categories
How social information is organized and what categories people use to simplify the richness and complexity of their social worlds are questions that have been asked in social perception and group dynamics research for several decades now. More recently, the importance of social categorization has re-emerged in research on social identity (Turner, 1987), self-identity (McGuire & McGuire, 1988), and stereotyping (Hamilton & Sherman, 1994; Macrae, Stangor, & Hewstone, 1996; Mackie & Hamilton, 1993; Tajfel, 1981; Sedikides, Schopler, & Insko, 1998).
Factors that can affect the choice of features used in categorization of people include feature salience and accessibility (Fiske, 1980; Higgins & King, 1981), and personal relevance of the feature to the observer (Erber & Fiske, 1984; Hansen & Hansen, 1988). In general, features that are easily observable, universal, and closely related to social roles provide powerful cues for social categorization. In this vein, it has been postulated that sex and age (Brewer, 1988; E. R. Smith & Zarate, 1992) and also race (E. R. Smith & Zarate, 1992) constitute natural categories and that perceivers tend to spontaneously use those categories in processing information about people. However, existing empirical evidence for the role of sex (Brewer & Lui, 1989; Pendry & Macrae, 1996; Stangor, Lynch, Duan, & Glass, 1992; Taylor, Fiske, Etcoff, & Ruderman, 1978; Zarate & E. R. Smith, 1990), age (Brewer & Lui, 1989) and race (Stangor, Lynch, Duan, & Glass, 1992; Zarate & E. R. Smith, 1990) as natural social categories is not without challenge. In particular, Kunda and Thagard (1996) have recently argued that because those studies looked only at subjects' responses to presentations of multiple, previously unknown targets, the generalizability of their results is limited. Furthermore, because subjects in most of the experiments were explicitly asked to categorize, those experiments were not well suited to provide evidence for sex, age, and race as natural social categories. The involvement of social categories in processing information about individuals who are not total strangers and in situations when categorization is not explicitly mandated remains to be demonstrated.
Present Experiments
The present studies were designed to provide new experimental evidence for the involvement of exemplars as social reference points and to examine the role of social category information in determining which specific exemplars become activated in making judgments regarding particular targets. By bringing together research on social categorization and on the role of prototypes as reference points in social judgments, two lines of research that have been developing in relative isolation from each other, we hoped to provide unique contributions to both.
Our experiments were designed to allow comparisons between effects involving representations of familiar others, representations of the self, social category representations (prototypes), and representations of people in general. Such comparisons were intended to demonstrate the involvement of familiar-other exemplars in a way that rules out activation of self, social prototypes, or general social knowledge as an alternative interpretation. In examining the role of social category information, we focused on sex and age: categories that are often postulated as likely to be processed automatically (cf., Brewer, 1988; E. R. Smith & Zarate, 1992).
Our experiments addressed the issue of activation of social reference points using a priming methodology. In Experiments 1 and 1a we tested a prediction that social judgments about acquaintances are facilitated by previous activation (priming) of highly familiar exemplars, particularly for target pairs matched with respect to a chronically accessible social category. More specifically, we expected that making a judgment regarding a highly familiar target should speed up making the same judgment regarding a less familiar target, particularly if the two targets are of the same sex and similar age. The exemplar model maintains that relevant exemplars are recruited when making a judgment about a relatively unfamiliar target. If so, making the same judgment about a familiar target, especially one matching on sex and age, on the immediately preceding trial should facilitate such exemplar recruitment. A relevant exemplar has been activated directly by the preceding question. Hence, judging the less familiar target should be facilitated by the heightened accessibility of this exemplar.
In Experiment 2 and 2a, which were based on the task-facilitation paradigm developed by Klein, Loftus, & Burton (1989, Experiment 2), we attempted to extend this category-specific target-priming(1) effect to a situation in which an acquaintance served as a priming target and a highly familiar exemplar served as a test target. We predicted that making a judgment regarding a relatively unfamiliar target should facilitate making the same judgment regarding a highly familiar target of the same sex and similar age. This prediction relied on the following: if answering a question regarding the relatively unfamiliar Person A (priming target), brings to mind the highly familiar Person B as a reference point (or one of several reference points), then we should be particularly fast at subsequently answering the same question regarding Person B (test target).
Experiment 1
Overview
The main goal of Experiment 1 was to demonstrate facilitating effects of priming familiar-person exemplars on subsequent judgments regarding relatively unfamiliar targets that varied systematically with respect to sex and age. Effects of priming were assessed by examining latencies of judgments regarding preferences. The target of the first (priming) judgment was always a highly familiar person (self, parent, or a friend) and the target of the second (test) judgment was either a relatively unfamiliar person (acquaintance) or a general social prototype (most people). The most people prototype was included to explore the role of significant-other exemplars and the role of self as reference points in making judgments regarding people in general.
An additional goal of Experiment 1 was to examine the role of social category information depending on individual differences in the personal relevance of the category in question. More specifically, we examined the importance of matching prime and test targets with respect to sex, depending on the participant's own sex-role identity. We expected that traditionally sex-typed participants (masculine men and feminine women) should exhibit a stronger category-specific facilitation for the category of sex than participants who are not traditionally sex-typed. This prediction was based on the notion that attributes and social categories perceived by individuals as highly self-relevant are more likely to play a role in the perception of others (Carpenter, 1988; Fong & Markus, 1982; Markus, J. Smith, & Moreland, 1985; Lewicki, 1984).
Method
Participants
One hundred eleventh-graders from a large high school in Warsaw, Poland participated in the experiment in exchange for a small monetary payment. There were 50 women and 50 men. Masculinity and Femininity scores on the Polish adaptation (Konarzewski, 1994) of the 24-item Personal Attributes Questionnaire (Spence, Helmreich, & Stapp, J., 1974), collected a few weeks earlier, were available for all participants.
Procedure and Materials
Participants were tested in a school computer lab individually or in groups of up to nine students.
Preliminary task. Participants entered names, initials, or other unique labels identifying four males and four females that they personally knew. Two of the males and two of the females were peers and the remaining four were a generation older. In addition, one of the two males and one of the two females in each age category were highly familiar (best male friend, best female friend, mother, and father) and the remaining four people were relatively unfamiliar (acquaintances).
Judgments of preferences. The preliminary task was immediately followed by judgments of preferences. There were 150 trials divided into 6 blocks, each consisting of 20 experimental and 5 control trials. The first block of 25 trials served as a practice block. Each of the 150 trials consisted of three questions: a priming (or priming control) question, a test question, and a filler question. Both priming question and test question called for judgments of target-person's preferences involving pairs of activities (e.g., [A] Read a novel / [B] Watch a movie, [A] Work in an office / [B] Take care of old people, etc.) with the same pair of activities being used for both questions. Judgments were made on a scale ranging from 1 ( [A] much more enjoyable) to 5 ( [B] much more enjoyable)(2). Filler questions, separating consecutive trials, consisted of subtraction problems involving two-digit numbers with correct answers between 1 and 5. A total of 300 activities divided into 150 pairs were used in the experiment(3). The pairs were presented in a constant random order with a different pair of activities used in each trial.
The priming question always involved one of the four highly familiar targets whose names were entered to the computer during the preliminary task or yourself. Each of the five priming targets was used in five of the 25 trials in each block. The target named in the test question was always either one of the four relatively unfamiliar persons whose names were entered to the computer during the preliminary task or most people. Each of those five targets was used in five trials in each block.
The order of priming targets was counterbalanced across subjects, except for the first 25 (practice) trials, in which the order was semi-random and constant for all subjects. The order of test targets was constant for all subjects and semi-random within each block, with the restriction that the same target person could not appear in the two adjacent trials. Thus, with the first block excluded, the remaining 125 trials constituted a completely balanced 5 (priming target) x 5 (test target) x 5 (replication) within-subject design with 5 versions of the counterbalancing order for priming targets as a between-subject variable.
Items were displayed on the screen with one of the eight target-names, the word Yourself, or the phrase Most People always appearing directly below a pair of activities. Pressing any of the response keys, 1 through 5, on the computer keypad resulted in the termination of the display. This was immediately followed by the second preference judgment item or by a math problem item. Response latencies were recorded by a software clock. Adjacent trials were separated by a semi-random graphic display lasting approximately 1.5 s. In addition, at the end of each block participants were asked to take a few moments to relax and to press a key when they were ready to continue. In making preference judgments, participants were asked to "guess which of the two activities would this person enjoy more [...] disregarding whether or not the person has the necessary training and experience". They were asked to answer each question as it appeared on the screen and to work "at the fastest comfortable pace".
Results
For the purpose of data analysis, male and female participants were split into two groups based on sex-specific medians for the Femininity - Masculinity difference score on PAQ. Female participants with Femininity - Masculinity scores above the median for all female participants (N = 21) and male participants with Femininity - Masculinity scores below the median for all male participants (N = 23) were assigned to the High Traditionally Sex-Typed group, whereas the remaining participants were assigned to the Low Traditionally Sex-Typed group.
Test latencies below 500 ms (1.64 %) were excluded from the analysis. Also, test latencies that were more than 3.5 standard deviations above the participant's mean (1.12 %) were truncated to that value.
Effects of Priming Matched and Mismatched Exemplars on Latencies of
Judgments about Acquaintance
As can be seen in Figure 1(4), for each of the four test targets involving specific acquaintances, the shortest latencies occurred in trials in which the prime matched the test target for both sex and age and the longest latencies occurred when the prime and target were fully mismatched. To test our prediction that same-sex, similar-age exemplars facilitate judgments regarding acquaintances more than exemplars mismatched with respect to sex, age, or both, test latencies were classified into four categories depending on whether the test target did or did not match the priming target with respect to sex and age. The latencies were standardized separately for each participant within each of the four test targets involving specific acquaintances. Such standardization allowed us to express latencies for a given priming target / test target combination in comparison to latencies of responding to the same test questions preceded by different primes. Trials involving most people as test target were excluded from this analysis.
A 5 (priming target) x 2 (sex matching) x 2 (age matching) x 2 (participant's sex) x 2 (sex-typing) mixed model ANOVA revealed main effects of both sex matching, F (1, 96) = 46.76, p < .001, and age matching, F (1, 96) = 28.77, p < .001. Consistently with unstandardized data presented in Figure 1, mean standardized latencies were shortest when the priming target and the test target were matched with respect to sex and with respect to age (M = -.32). Mean standardized latencies were longest (M = .24) when the two targets were mismatched with respect to both categories.
The only other significant effect was the predicted Sex Matching x Sex-Typing interaction, F (1, 96) = 6.54, p < .02. As is shown in Figure 2, the obtained pattern was consistent with predictions: the effect of sex matching was considerably larger for participants who were highly traditionally sex-typed, F (1, 42) = 30.32, p < .001, eta2 = .42, than for participants who were not, F (1, 54) = 13.57, p < .001, eta2 = .20. The Age Matching x Sex-Typing interaction did not approach significance, F < 1, indicating the specific nature of the significant Sex Matching x Sex-Typing interaction. Thus, the results show that priming judgments involving familiar person exemplars facilitate test judgments about acquaintances more if the priming and test targets are of the same sex and similar age. Moreover, the relevance of the sex category is greater for participants who are traditionally sex-typed.
A closer examination of mean latencies presented in Figure 1, provides an opportunity for comparing the role of self and the role of significant-other exemplars in facilitating judgments about acquaintances. The obtained pattern provides no evidence for a privileged status of the self and, in fact, suggests that than other-person exemplars might be more effective in facilitating judgments about acquaintances. Specifically, planned contrasts revealed that both the effect of sex matching and the effect of age matching were significant for each of the non-self priming targets (all ps < .05) but not for the self (both ps > .1). In addition, judgments regarding same-sex young acquaintance were significantly faster following priming judgments involving same-sex friend than following priming judgments involving the self, F (1, 98) = 4.16, p < .05. This last finding is particularly relevant because it involves a direct comparison between the self and the other-person exemplar that both match the test target on sex and age. However, because neither the Target x Sex Matching nor the Target x Age Matching interaction was significant in the omnibus ANOVA, these results should be treated with caution. In any case, the overall pattern of results makes it highly unlikely that the observed priming effects for other-person exemplars could have been mediated by self-knowledge activation.
Controlling for Similarity of Judgments Regarding Priming and Test
Targets
To address the possibility that the priming effects obtained for latencies were simply due to participants' tendency to make more similar judgments in the case of trials involving matched targets than in the case of trials involving mismatched targets, we first conducted a corresponding analysis on standardized absolute differences between numerical responses (judgments) to priming and test questions. As expected, test judgments were the most similar to priming judgments (standardized absolute difference were the smallest) when primes matched the test target with respect to both sex and age (M = -.391) and were the least similar when prime and test targets were completely mismatched (M = .362). As in the case of latency data, the main effect of sex matching, F (1, 96) = 104.25, p < .001, and the main effect of age matching, F (1, 96) = 28.76, p < .001, were significant. Interestingly, the Sex Matching x Sex-Typing interaction did not approach significance, F (1, 96) < 1. Thus, it appears that, in making judgments regarding preferences, individuals who are not traditionally sex-typed assume, perhaps correctly, as much sex-category based similarity as individuals who are highly traditionally sex-typed. This might simply reflect the social reality of sex-related differences in interests and preferences. However, the significant Sex-Typing x Sex Matching interaction for latency data suggests that traditional sex-typing is associated with greater use of sex-specific social reference points.
Analysis of covariance performed on latencies, with standardized absolute differences between numerical responses to priming and test questions as covariates, replicated both the main effect of sex matching, F (1, 95) = 4.99, p < .03, and the main effect of age matching, F (1, 95) = 8.21, p < .005. In addition, the Sex Matching x Sex-Typing interaction also remained significant, F (1, 95) = 5.83, p < .02. Significant effects of sex matching and age matching with similarity of responses controlled provide evidence that those effects were not entirely due to associated effects involving similarity of responses. This is consistent with the reasoning that because both similar and dissimilar members of a category may be used as social reference points in making judgments regarding a target, latency effects, although weaker, should occur even with the similarity of judgments controlled.
Effects of Exemplar Priming on Latencies of Judgments about Most People
A separate 5 (priming target) x 2 (participant's sex) x 2 (sex typing) mixed-model ANOVA performed on latencies of test questions involving the most people target did not produce any significant effects. Furthermore, an examination of means presented in Figure 1, revealed no indication of a privileged status of the self (as compared to familiar-other exemplars) in facilitating judgments regarding most people. In fact, those judgments were slightly (non significantly) slower following priming judgments involving the self than following priming judgments involving any other exemplar, except for the opposite-sex parent.
Discussion
Overall, our results provide support for the prediction that judgments regarding relatively unknown persons are facilitated by prior activation of familiar-person exemplars, particularly for exemplars that are matched with respect to sex and age. Furthermore, as predicted, the relevance of sex-category matching depends on the individual's sex role orientation: the effect is stronger for individuals who are highly traditionally sex-typed than for individuals who are not. Finally, we found no indication of a privileged role of the self (as compared to familiar-other exemplars) in making judgments about acquaintances or about people in general. If anything, self appears to be playing a somewhat lesser role.
It should be noted that the demonstrated effect of making judgments about familiar persons on subsequent judgments regarding acquaintances of the same sex and similar age could also be explained without assuming any particular role played by exemplar activation. Specifically, the effect could be mediated by activation of sex-specific and age-specific prototypes rather than by exemplar activation. For instance, in trying to decide which of the two activities my friend Joe might enjoy more, I may ask myself "would most young guys enjoy more activity A or activity B?" This, in turn, would facilitate subsequent judgments regarding an acquaintance, particularly if that acquaintance was also a young man. In this respect, the present experiment provides evidence for the recruitment of gender-specific and age-specific reference points, but remains inconclusive as to the nature of the representations involved. The primary goal of Experiment 1a was to address the possibility that the effects of priming obtained in Experiment 1 were mediated by the activation of category-specific prototypes.
Experiment 1a
Overview
The design of Experiment 1a was essentially identical to the design of Experiment 1, except that two prototypes were substituted for two of the exemplars used as priming targets in Experiment 1. Specifically, instead of opposite-sex friend and same-sex parent, most men my generation and most women parents' generation were used for male participants whereas most women my generation and most men parents' generation were used for female participants. Thus, the design involved an intentional confounding between sex and age for priming targets: both college-age targets were of the same sex as the participant and both older targets were of the opposite sex. At the same time, for two of the test targets, same-sex peer acquaintance and opposite-sex older acquaintance, the design allowed for a direct comparison between the effects of exemplar priming versus prototype priming for matched and mismatched primes (the remaining two exemplars used as test targets, opposite-sex peer acquaintance and same-sex older acquaintance, were retained from the original design as fillers and were not included in the analysis).
We reasoned that if the effects of exemplar priming found in Experiment 1 were, indeed, mediated by activation of sex-specific and age-specific prototypes, activating such prototypes directly (as primes) should prove even more effective. On the other hand, a finding that priming of category-specific exemplars facilitates subsequent judgments regarding acquaintance more than does priming of category-specific prototypes, would provide evidence against such alternative interpretation, thus underscoring a privileged role of exemplars.
A secondary goal of Experiment 1a was to provide data regarding relative importance of exemplars and prototypes for judgments about people in general.
Method
Participants
Forty- five University of North Florida undergraduates participated in the experiment in return for extra course credit. There were 32 women and 13 men.
Procedure and Materials
Participants were tested in individual rooms. They used a five-button response box to enter their judgments of preferences. Except for the modifications already described and except for the language, all other aspects of the experimental procedure and materials were identical to those in Experiment 1.
Results
Test latencies below 500 ms (4.96 % (5)) were excluded from the analysis. Also, test latencies that were more than 3.5 standard deviations above the participant's mean (<1 %) were truncated to that value. Because preliminary analysis revealed no significant effects involving subjects' sex, sex was not included in the final analysis.
Effects of Exemplar and Prototype Priming on Judgments about
Acquaintances
As can be seen in Figure 3, test judgments for the same-sex, young acquaintance were the fastest following priming judgments involving same-sex friend and were the second fastest following priming judgments involving the self. Thus the two matched exemplars emerged as the most effective primes for that target, Similarly, test judgments for the opposite-sex, older acquaintance, were fastest following priming judgment involving opposite-sex parent, again a matched exemplar. In addition, in both cases, test judgments were the slowest following priming judgments that involved prototypes that did not match the test target.
In order to formally assess the effects of priming judgments involving matched and mismatched exemplars and prototypes, test latencies for judgments regarding same-sex peer acquaintance and opposite-sex older acquaintance were classified into four categories depending on whether the priming target was an exemplar or a prototype and whether or not it matched the test target with respect to sex and age. As in the previous experiment, the latencies were standardized separately for each participant within each of the test targets. Because preliminary analysis revealed that the conclusions remained unchanged regardless of whether responses for the self or for the same-sex friend were used for the college-age exemplar, responses for those two kinds of primes were averaged out for the final analysis.
A 2 (test target: same-sex peer acquaintance or opposite-sex older acquaintance) x 2 (priming target type: exemplar or prototype) x 2 (matching: matched or mismatched) repeated measures ANOVA revealed a main effect of prime type, F (1, 44) = 6.16, p < .02, and a main effect of matching, F (1, 44) = 20.07, p < .001. Neither effect was qualified by an interaction. Consistently with unstandardized data presented in Figure 3, standardized latencies revealed that test judgments following matched primes were faster, M = -.25, than test judgments following mismatched primes, M = .31 Also, judgments following exemplar primes were faster, M = -.09, than judgments following prototype primes, M = .15. This last result speaks against the possibility that the effects of exemplar priming found in Experiment 1 (and replicated in the present experiment) were mediated by activation of sex-specific and age-specific prototypes and provides an additional evidence for a privileged status of exemplars as reference points in judgments regarding acquaintances. In addition, the absence of interaction between prime type and matching shows that the facilitating effect of target matching is not restricted to exemplar primes but that it occurs for prototype primes as well. In fact, even though on average exemplars facilitated judgments about acquaintances more than prototypes, matched prototypes were marginally more effective, M = -.12, than mismatched exemplars, M = .22, F(1, 44) = 3.78, p < .06.
A possibility should be considered that greater effectiveness of exemplars as compared to prototypes was simply due to a closer match between priming targets and test targets for exemplar than for prototype primes. Even though exemplar and prototype primes were equivalent with respect to sex and age, exemplar primes could have, at least theoretically, provided a closer match to test targets with respect to other, uncontrolled, attributes. Results of a corresponding ANOVA performed on standardized absolute differences between numerical responses (judgments) to priming and test questions, speak against such possibility. As expected, standardized differences between test judgments and priming judgments were significantly smaller for matched, M = -.45, than for mismatched target pairs, M = .38, F(1, 44) = 82.91, p < . 001. More interestingly, those differences were also significantly smaller for trials with prototype primes, M = -.19, than for trials with exemplar primes, M = .12, F(1, 44) =10.83, p < . 002.(6) Thus, acquaintances turned out to be perceived as less similar to exemplars than to prototypes, yet judgments regarding those acquaintances were facilitated more by previous judgments involving exemplars than by previous judgments involving prototypes.
As expected, analysis of covariance performed on test latencies with judgment differences used as (specific) covariates, replicated the significant effect of prime type, F (1, 43) = 5.87, p < . 02. Consistently with results of Experiment 1, the effect of matching also remained significant, F(1, 43) = 13.10, p < . 002.
Effects of Exemplar and Prototype Priming on Judgments about Most People
A separate one-way repeated measures ANOVA performed on latencies of test responses involving most people with priming target as an independent variable revealed no significant effect of priming target, F(4, 176) < 1. Thus, our conclusion regarding particular effectiveness of exemplar priming for judgments about acquaintances does not generalize to judgments about people in general.
A corresponding analysis performed on standardized absolute differences between numerical responses (judgments) to priming and test questions, revealed a significant effect of priming target, F(4, 176) = 15.05, p < .001. Post hoc comparisons showed that, as it was the case with judgments about acquaintances, judgments about people in general were more similar to judgments about sex/generation-specific prototypes than to judgments about exemplars, all ps < .01.
Discussion
Overall, the results provide strong evidence against the interpretation of Experiment 1 results in terms of prototype activation. Instead, they offer support for the privileged status of familiar-person exemplars. Despite the fact that, in terms of their content, test judgments about acquaintances turned out to be less similar to priming judgments about exemplars than to priming judgments about prototypes, they were facilitated more by the by the former than by the latter.
Consistently with results of Experiment 1, same-sex similar-age exemplars facilitated test judgments more than exemplars mismatched with respect to sex and age. In addition, matched prototypes facilitated judgments more than mismatched prototypes. This last result indicates that target-matching effect is not restricted to exemplar targets. Not surprisingly, activation of sex-specific and age-specific information might speed up judgments about persons also when such activation does not involve exemplars directly.
A potential criticism that applies both to Experiment 1 and to the present experiment is that because familiar-person exemplars were directly activated during the priming task, the spontaneous nature of exemplar recruitment during the testing task could not be established. In Experiment 2, we used a more subtle approach in which familiar-person exemplars were not directly primed.
Experiment 2
In this experiment, five highly familiar persons (self, male friend, female friend, and two parents) were used as test targets and four acquaintances (two peers and the other two a generation older, one male and one female in each age group) were used as priming targets. In addition, the design employed control priming judgments that did not involve any particular target. Specifically, in 5 out of 25 trials in each block, participants were asked to judge on a 5-point scale "which of the two activities requires more experience and training". Judgments regarding most people were not included in the design.
Even though exemplars of relatively unfamiliar others (acquaintances) seem unlikely to be used as points of reference in making judgments regarding highly familiar persons we, nevertheless, expected to find reliable effects of priming judgments. This was based on the reasoning that making a judgment regarding preferences of a relatively unfamiliar target would activate exemplars of highly familiar, significant persons, particularly those of the same sex and age. Consequently, making a judgment regarding preferences of the first (unfamiliar) target should facilitate a subsequent judgment regarding preferences of the second (familiar) target, again, particularly when they are of the same sex and similar age. Thus, even though we expected to find in Experiment 2 similar effects of priming judgments as we did in Experiment 1, (except for effects involving individual differences, which in the present experiment were not assessed), the postulated process underlying the effects was somewhat different. In Experiment 1, our predictions were based on the assumption that the process of exemplar recruitment was taking place primarily during the testing task: familiar-person exemplars directly primed by the first question are subsequently recruited as a reference points when answering the second question. On the other hand, predictions for the present experiment were based on the assumption of exemplar recruitment occurring primarily during the priming task: familiar-person exemplars are spontaneously recruited when answering the priming question about an unfamiliar person, and this speeds up subsequent judgments regarding those familiar-person exemplars.
Method
Participants
One hundred and five University of North Florida undergraduates participated in the experiment in return for extra course credit. There were 78 women and 27 men.
Procedure and Materials
Participants were tested in individual rooms. Except for the modifications described in the Overview, and except for the language, all other aspects of the experimental procedure and materials were identical to those in Experiment 1.
Results
Latencies of responses to priming and test questions shorter than 500 ms (< 1 %) were excluded from the analysis. Also, test latencies that were more than 3.5 standard deviations above the participant's mean (1.0 %) were truncated to that value.
As is shown in Figure 4, for each of the five test targets latencies were longest in the case of control primes and the shortest in the case of primes that matched the test target with respect to both sex and age. With the control primes excluded, we expected the longest latencies to occur in trials with primes mismatched with respect to both age and sex. This was true for every test target except the self. When self was the test target, latencies were actually somewhat longer for trials involving priming target mismatched with respect to sex or age, but not both (opposite-sex young acquaintance and same-sex older acquaintance) than for trials involving a fully mismatched target (older acquaintance of an opposite sex). Because preliminary analysis revealed no reliable effects involving subjects' sex, sex was not included in the final analysis.
Effects of Priming Judgments about Matched and Mismatched Acquaintances
on Latencies of Test Judgments
To provide statistical evidence for the predicted effects of priming judgments, test latency data for each test target, excluding control trials, were classified into four categories depending on whether the priming target did or did not match the test target with respect to sex and age. The latencies were standardized separately for each participant within each of the five test targets. Trials involving control primes were excluded from this analysis. A 5 (test target: self, same-sex friend, opposite-sex friend, same-sex parent, or opposite-sex parent) x 2 (sex matching: sex matched or sex mismatched) x 2 (age matching: age matched or age mismatched) repeated-measures ANOVA performed on standardized test latencies revealed a main effect of sex matching, F (1, 100) = 24.48, p < .001, and a main effect of age matching, F (1, 100) = 10.46, p < .002. Neither effect was qualified by a significant interaction. Consistently with unstandardized data presented in Figure 4, across the five test targets, responses to the test items were fastest when the priming target and the test target were matched with respect to sex and with respect to age (M = -.17) and responses were slowest (M = .13) when the two targets were mismatched with respect to both categories. These results are consistent with our reasoning that making judgments about relatively unfamiliar persons would increase accessibility of exemplars involving significant, highly familiar persons of the same sex and similar age. (7)
Similarity of Judgments Regarding Priming and Test Targets
As in Experiment 1, a corresponding analysis performed on standardized absolute differences between numerical responses to priming and test questions revealed both the main effect of sex matching and the main effect of age matching, F (1, 104) = 185.06, p < .001, and F (1, 104) = 7.24, p < .01, respectively. Test judgments were the most similar to priming judgments (standardized absolute difference were the smallest) in the case of primes that matched the test target with respect to both sex and age (M = -.37). They were the least similar (standardized differences were the largest) in the case of primes that were completely mismatched (M = .27). As in previous experiments, analysis of covariance performed on standardized test latencies with standardized judgment differences used as (specific) covariates, replicated both the main effect of sex matching, F (1, 103) = 8.35, p < .01, and the main effect of age matching F(1, 103) = 4.82, p < .05. This indicates that the effects of priming judgments obtained for latency data were, at least partially, independent of similar effects obtained for similarity of judgments.
Discussion
Results of the present experiment extend the category-specific effect of priming judgments found in Experiment 1 to a task involving acquaintances as priming targets and highly familiar persons as test targets. As expected, making a judgment about an acquaintance facilitated subsequent judgment about a highly familiar target (best friend, parent, or the self), particularly when the two targets were of the same sex and similar age. This finding is consistent with the notion that when making judgments about relatively unfamiliar persons, people activate mental representations of highly familiar significant persons, again, particularly those of the same sex and similar age.
As in previous experiments, the predicted pattern of results appeared less consistently for the self than for other familiar-person exemplars. However, because none of the interactions between priming target and matching approached significance, this observation has to be treated with caution.
As it was the case in Experiment 1, results of the present experiment could be interpreted in terms of prototype activation rather than in terms of the activation of familiar person exemplars. Such activation of a sex-specific and generation-specific prototypes could, presumably, facilitate judgments regarding highly familiar persons used as test targets, particularly if the test target was of the same sex and similar age as the priming target. The purpose of Experiment 2a was to address the possibility that the effects of exemplar priming obtained in Experiment 2 could have been mediated by the activation of category-specific prototypes. As in the case of Experiment 1a, this was accomplished by including direct priming of category-specific prototypes in the design.
Experiment 2a
Overview
The design of Experiment 2a was essentially identical to the design of Experiment 2, except that prototypes were substituted for three of the priming targets used in that experiment. Specifically, instead of opposite-sex peer acquaintance and same-sex older acquaintance, most men my generation and most women parents' generation were used for male participants whereas most women my generation and most men parents' generation were used for female participants. In addition, instead of requires more experience and training, most people was used as a control priming target. As in Experiment 1a, the design involved an intentional confounding between sex and age for priming targets: both college-age targets were of the same sex as the participant and both older targets were of the opposite sex. Importantly, for three of the test targets: self, same-sex friend, and opposite-sex parent, the design allowed for a comparison between the effects of exemplar priming and the effects of prototype priming for matched and mismatched primes (the remaining two test targets, were retained from the original design as fillers and were not included in the analysis).
Our reasoning here was essentially the same as our reasoning in Experiment
1a. We speculated that if the effects of exemplar priming found in Experiment 2
were mediated by activation of sex/generation-specific prototypes, activating
such prototypes directly (as primes) should prove even more effective.
Conversely, a finding that test judgments are facilitated more by priming
judgments involving acquaintances than by priming judgments involving exemplars
would speak against such alternative interpretation.
Method
Participants
Forty- five University of North Florida undergraduates participated in the experiment in return for extra course credit. There were 36 women and 9 men.
Procedure and Materials
Participants were tested in individual rooms. They used a five-key response box to enter their judgments of preferences. In order to avoid accidental premature responding, the keys remained unoperational during the first 500 ms following a stimulus presentation. Except for the modifications already described and except for the language, all other aspects of the experimental procedure and materials were identical to those in Experiment 2.
Results
Effects of Priming Judgments about Matched and Mismatched Acquaintances and Prototypes on Latencies of Test Judgments
Test latencies that were more than 3.5 standard deviations above the participant's mean (<1 %) were truncated to that value and collapsed across replications. As can be seen in Figure 5, test judgments for the same-sex friend were fastest following priming judgments involving same-sex acquaintance. In the same vein, test judgments for the opposite-sex parent, were fastest following priming judgment involving an opposite-sex older acquaintance. Again, the pattern was less clear for test judgments involving the self. The speed of those judgments was about the same following priming judgments involving same-sex acquaintance as following priming judgments involving a matched stereotype. Finally, for each of the three diagnostic test targets, judgments were the slowest following priming judgments that involved prototypes that did not match the test target.
Standardized latencies for test judgments regarding self, same-sex friend, and opposite-sex parent, excluding latencies for trials with control primes, were classified into four categories depending on whether the priming target was an exemplar or a prototype and whether it did or did not match the test target. A 3 (test target: self, same-sex friend, or opposite-sex parent) x 2 (priming target type: exemplar or prototype) x 2 (matching: matched or mismatched) repeated-measures ANOVA revealed a main effect of prime type, F (1, 44) = 11.75, p < .001, and a main effect of matching, F (1, 44) = 51.40, p < .001. Neither effect was qualified by a significant interaction. Consistently with unstandardized data presented in Figure 5, the pattern of mean standardized test latencies indicates that priming judgments involving matched targets (M = -.16) facilitated test judgments more than priming judgments involving mismatched targets (M = .16). Furthermore, priming judgments involving acquaintances (M = -.07) facilitated test judgments more than priming judgments involving prototypes (M = .08). This last result speaks against the possibility that the effects of exemplar priming reported in Experiment 2 could have been mediated by activation of sex-specific and age-specific prototypes. However, it should also be noted that, as in Experiment 1a, the effect of target matching occurred for both types of primes. In fact, judgments primed by matched prototypes were faster not only compared to judgments primed by mismatched prototypes, F (1, 44) = 40.53, p < .001, but also, to a lesser extent, compared to judgments primed by mismatched acquaintances, F (1, 44) = 7.81, p < .01.
As would be expected, test judgments primed by control prototypes (most people) were faster than test judgments primed by mismatched prototypes, F (1, 44) = 26.77, p < .001. In addition, they were marginally slower than test judgments primed by matched prototypes, F (1, 44) = 2.42, p < .15. Also, judgments primed by control primes were significantly slower than judgments primed by matched exemplars, F (1, 44) = 8.61, p < .005, and marginally faster than judgments primed by mismatched exemplars F (1, 44) = 2.59, p < .15.
Similarity of Judgments Regarding Priming Targets and Test Targets
Results of a corresponding ANOVA performed on standardized absolute differences between numerical responses (judgments) to priming and test questions revealed a significant main effect of matching, F (1, 44) = 68.94, p < .001 and a marginally significant effect of prime type, F (1, 44) = 2.82, p < .1. As expected, test judgments were more similar to priming judgments when the two targets were matched, M = -.17, than when they were mismatched, M = .22. Also, differences between priming and test judgments were, on average, smaller in the case of prototype priming targets, M = .00, than in the case of exemplar priming targets, M = .06. However, this marginally significant effect of prime type was qualified by a significant Prime Type x Matching interaction, F (1, 44) = 5.57, p < .05. The difference between prototype primes and exemplar primes was significant in the case of matched targets, F (1, 44) = 8.90, p < .005, but not in the case of mismatched targets, F (1, 44) < 1. (8)
Analysis of covariance performed on test latencies with judgment differences used as (specific) covariates, replicated the significant effect of prime type, F(1, 43) = 14.54, p < .001. Consistently with results of our previous experiments, the effect of matching also remained significant, F(1, 43) = 19.15, p < .001.
Discussion
Results of the present experiment closely parallel results of Experiment 1a. Even though, in terms of their content, test judgments turned out to be less similar to priming judgments involving acquaintances than to priming judgments involving category-specific prototypes (at least in the case of matched targets), they were facilitated more by the former than by the latter. This pattern of results provides support for the interpretation of Experiment 2 results in terms of the notion that when making judgments about relatively unfamiliar persons, people activate mental representations of highly familiar significant persons, particularly those of the same sex and similar age. An alternative interpretation of the results of Experiment 2, that is in terms of the mediation of prototypes, is inconsistent with the results of the present experiment.
General Discussion
The results of our experiments provide strong support for a privileged role of exemplars in judgments of preferences and for the relevance of the social categories of sex and age in determining which specific exemplars are used in making judgments regarding particular targets. Our results show that, compared to making a control judgment not involving any specific person, making a judgment regarding one person facilitated making the same judgment regarding another. This was the case particularly when the two targets were of the same sex and similar age. The effect of sex- and age-category matching remained statistically significant with similarity of the two judgments controlled.
The effects of priming judgments are robust. They occurred across two different populations (US college students and Polish high school students) and for both highly familiar and relatively unfamiliar priming targets. However, as predicted, facilitation due to sex-category matching was affected by individual differences in sex-role identification, with traditionally sex-typed individuals exhibiting the stronger effect.
This last finding, while consistent with previous results documenting the impact of the personal importance of an attribute on social categorization (Erber & Fiske, 1984, Hansen & Hansen, 1988), constitutes one of the few examples that shows the effect of an individual difference variable on a process-oriented measure of social cognition (cf., Moskowitz, 1993). Although, our experiments did not allow for testing similar constraints for matching with respect to age, we would expect individual differences in age-related schemata to produce a pattern of results analogous to the pattern found for individual differences in sex-role identification in Experiment 1. Furthermore, features other than sex and age might be used as a basis of social categorization, and these features may have an effect on the recruitment of exemplars as points of reference in social judgments. Previous social categorization research suggests that race (E. R. Smith & Zarate, 1992; Zarate & E. R. Smith, 1990) and personal relationships (Sedikides, Olsen, & Reis, 1993) are two probable candidates.
Experiments 1a and 2a provided an opportunity for direct comparisons between effects due to the activation of exemplars and effects due to the activation of age-specific and sex-specific prototypes. Results of those comparisons strongly support the notion that, when making judgments about individuals, people are more likely to rely on exemplars than on prototypes. Strikingly, the advantage of exemplars as primes was demonstrated even when test targets were seen as more similar to prototypes than to exemplars used as priming targets.
However, it should be noted that sex-specific and age-specific prototypes used in our experiments were still quite general. Yet, there is considerable evidence that perceivers often distinguish between subtypes within the broader social categories determined by sex (e.g., Clifton, McGrath, & Wick, 1976) or age (e.g., Brewer, Dull, & Lui, 1981). How individual-person exemplars would compare to such highly specific prototypes remains an open question.
In interpreting the effects of priming judgments, we assumed that both the facilitation effects of priming familiar targets (Experiments 1 & 1a) and the facilitation effects of priming unfamiliar targets (Experiments 2 & 2a) were due to an involvement of familiar-person exemplars as reference points. This presumes that two somewhat different hypothetical processes were responsible the effects: a spontaneous recruitment of familiar person exemplars taking place during a priming task involving unfamiliar targets, and a prompted recruitment of such recently primed exemplars occurring during the testing task. It is, however, prudent to consider the possibility that the facilitation effect of priming unfamiliar targets could have been, at least partially, due to unfamiliar-person exemplars assuming the role of reference points in subsequent judgments regarding highly familiar targets. Such possibility seems theoretically unlikely (cf., Andersen & Cole; 1990, Andersen, at al., 1995) yet given the fact that our unfamiliar targets were repeatedly primed during the experiment and considering that only a small number of targets was used, not impossible. In any case, this would not alter our two basic conclusion: (a) exemplars play a privileged role in judgments of preferences and (b) social categories of sex and age influence the use of specific exemplars for making judgments regarding particular targets.
Although our results provide clear evidence for the importance of person exemplars as points of reference in social judgments, the present studies were not designed to answer questions regarding various functions that social reference points may serve. Future research will focus on recruitment of exemplars as reference points in relation to their function. In particular distinguishing between reference points that are used as stand-ins for absent, poorly developed, and/or relatively inaccessible representations and reference points used as anchors, standards of comparison against which the target is judged will need to be addressed. . The first type of function, which is based on analogical judgments, is likely to lead to assimilation effects. The second, based on anchoring, is likely to lead to contrast effects. The "shifting standards model" developed by Biernat and her associates (Biernat & Kobrynowicz, 1999; Biernat & Manis, 1994; Biernat, Manis, & Nelson, 1991), which accounts for both assimilation and contrast effects associated with category-specific social reference points, is particularly relevant here. However, unlike the present work, the model is silent regarding the form of the representations involved (i.e., exemplars vs. prototypes).
In general, our results are consistent with findings reported by Andersen and her associates (Andersen and Cole, 1990; Andersen, et al., 1995) which indicate that familiarity and chronic accessibility play a role in the recruitment of exemplars as points of reference in performing social judgments. However, they also clearly demonstrate that those are not the only factors. In particular, our results indicate that the self, arguably the most familiar and the most chronically accessible exemplar that we have in our cognitive arsenal (Rogers, 1981), not only does not enjoy any special preference as a social reference point but, if anything, seems to be somewhat less effective in facilitating social judgments.
A relative compartmentalization of the self compared to representations of others (cf., Cantor & Kihlstrom, 1989; Cantor, Markus, Niedenthal, & Nerius, 1986; Karylowski & Skarzynska, 1992; Linville & Carlston, 1994; Markus & Wurf, 1987; Wyer & Srull, 1989, chap. 13) might be one way to explain this phenomenon. In addition, our tendency to see ourselves as unique (Brewer, 1991; Codol, Jarymowicz, Kaminska-Feldman, & Szuster-Zbrojewicz, 1989; Snyder & Fromkin, 1980) makes us less inclined to apply general social categories (including categories of sex and age) to self than to others. Consequently, we may see ourselves more in terms of nonmembership in various outgroups than in terms of membership in any particular ingroup (Simon, 1993), thus reducing the chances of self emerging as a relevant point of reference for targets of the same sex, age, or whatever other common attributes are particularly salient in categorizing those targets.
Our finding that the self does not enjoy a privileged status as a reference point in social judgments, challenges an implicit assumption that has accompanied much of the empirical research in this area. For instance, research on attitudinal assimilation-contrast effects (Hovland & Sherif, 1952), false consensus (Marks and Miller, 1987; Ross, Greene, & House, 1977), attributive projection (Holmes, 1968; Lemon & Warren, 1976; Shrauger & Patterson, 1974), and attributive contrast (Dunning & Cohen, 1992) has typically been interpreted in terms of the special status of self as a point of reference in social judgments. Yet, what appeared to emerge in those studies as, mostly indirect, evidence for the role of self was, in fact, still mostly indirect, evidence for the role of chronically accessible exemplars. Even the most recent series of studies designed to provide more direct, experimental evidence for the role of self as a reference point (Dunning & Hayes, 1996) included comparisons between the self and representations of distant acquaintances but not comparisons between the self and representations of highly familiar others, such as parents or friends.
Of course, it is possible that for some types of social judgments, or under special circumstances, the self does indeed play a special role. Future research might show that predicting affective responses to emotion-provoking situations and life events, especially situations and life events with which the participant has had some experience, may put self at a privileged position as a potential reference point. Also, making judgments regarding attributes that constitute a particularly central or particularly salient aspect of ones self-knowledge, could give the self a special edge (cf., Karylowski, 1990; Lemon & Warren, 1976; Shrauger & Patterson, 1976). Finally, focusing attention on the self, will make self-knowledge more accessible, thus increasing the likelihood that the self, rather than another exemplar will be used as a point of reference in social judgments (cf., Karylowski & Skarzynska, 1992). However, a different set of circumstances could lead to other exemplars acquiring a privileged position, for example, when making judgments involving attributes with respect to which a specific exemplar, other than the self, is particularly salient.
As a final comment, we would like to note that the type of priming effects that were demonstrated in the present experiments are different from the type of priming described in most of the existing social cognition literature in at least two respects. First, because the priming item and the test item refer, in our experiments, to two different (social) concepts, e.g., Person A and Person B, priming effects obtained in these experiments should be classified as semantic (Neely, 1977) or associative priming (Carlston & E. R. Smith, 1996). In contrast, as noted by Carlston and E. R. Smith (1996) almost all social-psychological research on priming involves effects of prior exposure to the same concept. Research on the effects of priming category labels representing attitude objects on the activation of trait concepts, or evaluations, associated with those objects (e.g., Bargh, Chaiken, Raymond, & Hymes, 1996; Devine, 1989; Dovidio, Evans, & Tyler, 1986; Fazio, Sanbonmatsu, Powell, & Kardes, 1986) constitutes a notable exceptions. In addition, existing demonstrations of social priming using process-oriented measures, such as latencies, are limited to priming in which the prime, test, or (usually) both represent attributes (e.g., traits) rather than objects such as persons or groups (Higgins, 1989, 1996; Bargh, 1989). On the other hand, the priming effects obtained in present experiments employed person names as both priming items and test items, thus providing an opportunity to apply priming methodology to the study of how persons (rather than attributes) are organized into social categories. Because of their unobtrusive, non-reactive nature, and because they can easily incorporate ideographically selected (participant-specific) as well as nomothetically selected (generic) targets, the priming methodology used in the present experiments is well suited for future social categorization research.
References
Andersen, S. M. & Baum, A. (1994). Transference in interpersonal relations: Inferences and affect based on significant-other representations. Journal of Personality, 62, 459-497.
Andersen, S. M. & Cole, S. W. (1990). Do I know you?: The role of significant others in general social perception. Journal of Personality & Social Psychology, 59, 384-399.
Andersen, S. M., Glassman, N. S., Chen, S., & Cole, S. W. (1995). Transference in social perception: The role of chronic accessibility in significant-other representations. Journal of Personality & Social Psychology, 69, 41-57.
Anderson, N. H. (1968). Likeableness ratings of 555 personality-trait words. Journal of Personality and Social Psychology, 9(3), 272-279.
Aron, A., Aron, E. N.. Tudor, M., Nelson, G. (1991). Close relationships as including other in the self. Journal of Personality & Social Psychology, 60, 241-253.
Bandura, A. (1969). Social learning theory of identification processes. In D. A. Goslin (Ed.), Handbook of socialization theory and research (pp. 213-262). Chicago: Rand McNally.
Bargh, J. A. (1989). Conditional automaticity: Varieties of automatic influence in social perception and cognition. In: J. S. Uleman & J. A. Bargh (Eds.) Unintended thought (pp. 3-51). New York: Guildford.
Bargh, J. A., Chaiken, S., Raymond, P., Hymes, C. (1996). The automatic evaluation effect: Unconditional automatic attitude activation with a pronunciation task. Journal of Experimental Social Psychology, 32, 104-128.
Beike, D. R. & Sherman, S. J. (1994). Social inference: Inductions, deductions, and analogies. In R. S. Wyer, Jr. & T. K. Srull (Eds.), Handbook of social cognition (second edition). Hillsdale, N.J.: Lawrence Erlbaum.
Biernat, M. & Kobrynowicz D. (1999). A shifting standards perspective on the complexity of gender stereotypes and gender stereotyping. In: Swann, W. B., Jr., & Langlois, J. H. (Eds.) Sexism and stereotypes in modern society: The gender science of Janet Taylor Spence. Washington, DC: APA.
Biernat, M., & Manis, M. (1994). Shifting standards and stereotype-based judgments. Journal of Personality & Social Psychology, 66, 5-20.
Biernat, M., Manis, M., & Nelson, T. E. (1991). Stereotypes and standards of judgment. Journal of Personality & Social Psychology, 60, 485-499.
Brewer, M. B. (1988). A dual process model of impression formation. Advances in Social Cognition, 1, 1-36.
Brewer, M. B. (1991). The social self: On being the same and different at the same time. Personality & Social Psychology Bulletin, 17, 475-482.
Brewer, M. B., Dull, V., & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes. Journal of Personality and Social Psychology, 41, 656-670.
Brewer, M. B., & Lui L. L. (1989). The primacy of age and sex in the structure of person categories. Social Cognition, 7, 262-274.
Cantor, N. & Kihlstrom, J. F. (1989). Social intelligence and cognitive assessments of personality. In: R. S. Wyer & T. K. Srull (Eds.). Advances in social cognition: Vol. 2. Social intelligence and cognitive assessments of personality (pp. 1-59). Hillsdale, NJ: Erlbaum.
Cantor, N., Markus, H., Niedenthal, P., & Nerius, P. (1986). On motivation and the self-concept. In: R. M. Sorrentino & E. T. Higgins (Eds.). Handbook of motivation and cognition: Foundations of social behavior. New York: Guildford.
Carlston, D. E. & Smith, E. R. (1996). Principles of mental representation. In E. T. Higgins & A. W. Kruglanski (Eds.) Social psychology: Handbook of basic principles. New York: Guildford.
Carpenter, S. L. (1988). Self-relevance and goal-directed processing in the recall and weighting of information about others. Journal of Experimental Social Psychology, 24, 310-332.
Clifton, A. K., McGrath, D.,& Wick, B. (1976). Stereotypes of woman: A single category? Sex Roles, 2, 135-148.
Codol, J. P., Jarymowicz, M., Kaminska-Feldman, M., & Szuster-Zbrojewicz, A. (1989). Asymmetry in the estimation of interpersonal distance and identity affirmation. European Journal of Social Psychology, 19, 11-22.
Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality & Social Psychology, 56, 5-18.
Dovidio, J. F., Evans, N., & Tyler, R. B. (1986). Racial stereotypes: The contents of their cognitive representations. Journal of Experimental Social Psychology, 22, 22-37
Dunning, D. & Cohen, D. L. (1992). Egocentric definitions of traits and abilities in social judgment. Journal of Personality & Social Psychology, 63, 341-355.
Dunning, D. & Hayes, A. F. (1996). Evidence for egocentric comparison in social judgment. Journal of Personality & Social Psychology, 71, 213-229.
Erber, R. M. & Fiske, S. T. (1984). Outcome dependency and attention to inconsistent information. Journal of Personality & Social Psychology, 47, 709-726.
Fazio, R. H. (1990). A practical guide to the use of response latency in social psychological research. In: C. Hendrick & M. S. Clark (Eds.) Research methods in personality and social psychology. Newbury Park, Ca: Sage.
Fazio, R. H. & Williams, C. J. (1986). Attitude accessability as a moderator of the attitude-perception and attitude-behavior relations: An investigation of the 1984 presidential election. Journal of Personality and Social Psychology, 51, 505-514.
Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation of attitudes. Journal of Personality & Social Psychology, 50, 229-238.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140.
Fiske, S. T. & Linville, P. W. (1980). What does the schema concept buy us? Personality & Social Psychology Bulletin , 6, 543-557.
Fong, G. & Markus, H. (1982). Self-schemas and judgments about others. Social Cognition, 3, 191-204.
Freud, S. (1921). Group psychology and the analysis of the ego. (Transl. New York: International Psychoanalytical Press, 1922).
Hansen, C. H. & Hansen, R. D. (1988). Finding the face in the crowd: An anger superiority effect. Journal of Personality & Social psychology, 54, 917-924.
Hamilton, D. L. & Sherman, S. J. (1994). Stereotypes. In R. S. Wyer, Jr. & T. K. Srull (Eds.), Handbook of social cognition (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Higgins, E. T. (1989). Knowledge accessibility and activation: Subjectivity and suffering from unconscious sources. In: J. S. Uleman & J. A. Bargh (Eds.) Unintended thought (pp. 75-123). New York: Guildford.
Higgins, E. T. & King, G. (1981). Accessibility of social constructs: Information-processing consequences of individual and contextual variability. In N. Cantor & J. F. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp.69-122). Hillsdale, NJ: Erlbaum.
Higgins, E. T. (1996). Knowledge activation: Accessibility, applicability, and salience. In E. T. Higgins & A. W. Kruglanski (Eds.) Social psychology: Handbook of basic principles. New York: Guildford.
Holmes, D. S. (1968). Dimensions of projection. Psychological Bulletin, 69, 248-268.
Holyoak, K. J. (1978). Comparative judgments with numerical reference points. Cognitive Psychology, 10, 203-243.
Holyoak, K. J. & Gordon, P. C. (1983). Social reference points. Journal of Personality & Social Psychology, 44, 881-887.
Holyoak, K. J., & Mah W. A. (1982). Cognitive reference points in judgments of symbolic magnitude. Cognitive Psychology, 14, 328-352.
Hovland, C. I., & Sherif, M. (1952). Judgmental phenomena and scales of attitude measurement: Item displacement in Turnstone scales. Journal of Abnormal and Social Psychology, 47, 822-832.
Huston, D. A. & Fazio, R. H. (1989). Biased processing as function of attitude accessibility: Making objective judgments subjectively. Social Cognition, 7, 51-66.
Karylowski, J. J. (1990). Trait prototypicality and the accessibility of trait related information in self-other similarity judgments. Journal of Personality & Social Psychology, 58, 975-983.
Karylowski, J. J. & Skarzynska, K. (1992). Asymmetry effect in self-other similarity judgments depends on priming. Social Cognition, 10, 235-254.
Kihlstrom, J. F., Cantor, N., Albright, J. S., Chew, B. R., Klein, S. B., & Niedenthal, P. M. (1988). Information processing and the study of the self. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 145-178). New York: Academic Press.
Klein, S. B., Loftus, J., & Burton, H. A. (1989). Two self-reference effects: The importance of distinguishing between self-descriptiveness judgments and autobiographical retrieval in self-refererent encoding. Journal of Personality & Social Psychology, 56, 853-865.
Konarzewski, K. (1994). Polska adaptacja kwestionariusza do badania identyfiakcji z rola zwiazana z plcia. Warszawa: Instytut Psychologii Polskiej Akademii Nauk
Kuder, G. F. (1981). Kuder Preference Record Vocational Form CP. Monterey, Ca: CTB Macmillan/Mc-Graw-Hill School Publishing.
Kunda, Z. & Thagard, P. (1996). Forming impressions from stereotypes, traits, and behaviors: A parallel constraint-satisfaction theory. Psychological Review, 103, 284-308.
Lemon, N. & Warren, N. (1976). Salience, centrality, and self-reference of traits in construing others. British Journal of Social and Clinical Psychology, 13, 119-124.
Lewicki, P. (1984). Self-schema and social information processing. Journal of Personality & Social Psychology, 47, 1177-1190.
Lewicki, P. (1985). Nonconscious biasing effects of single instances on subsequent judgments. Journal of Personality & Social Psychology, 48, 563-574.
Lewicki, P. (1986). Nonconscious social information processing. New York: Academic Press.
Linville, P. & Carlston, D. E. (1994). Social cognition of the self. In P. G. Devine, D. L. Hamilton, & T. M. Ostrom (Eds.), Social cognition: Impact on social psychology. San Diego, CA: Academic Press.
Lord, C.G. (1980). Schemas and images as memory aids: Two modes of processing social information. Journal of Personality & Social Psychology, 38, 257-269.
Lord, C. G., Desforges, D. M., Chacon, S., Pere, G., & Clubb, R. (1992). Reflections on reputation in the process of self-evaluation. Social Cognition, 10, 2-29.
Mackie, D. & Mackie, D. L. (1993). Affect, cognition, and stereotyping: Interactive processes in group perception. San Diego, CA: Academic Press Inc.
Macrae, C. N., Stangor, C., & Hewstone, M. (1996). Stereotypes and stereotyping. New York: Guildford.
Marks, G., & Miller, N. (1987). Ten years of research on the false consensus effect: An empirical and theoretical review. Psychological Bulletin, 102, 72-90.
Markus, H., Smith, J., & Moreland, R. L. (1985). Role of self-concept in the perception of others. Journal of Personality & Social Psychology, 49, 1494-1512.
Markus, H., & Wurf, E. (1987). The dynamic self-concept: A social psychological perspective. Annual Review of Psychology, 38, 299-338.
McGuire, W. J., & McGuire, C. V. (1988). Content and process in the experience of self. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 97-144). New York: Academic Press.
Moskowitz, G. B. (1993). Individual differences in social categorization: The influence of personal need for structure on spontaneous trait inferences. Journal of Personality & Social Psychology, 65, 132-142.
Neely, J. H. (1977). Semantic priming and retrieval from lexical memory: Roles of inhibitionless spreading activation and limited-capacity attention. Journal of Experimental Psychology: General, 106, 226-254.
Pendry, L. F. & Macrae, C. N. (1996). Personality & Social Psychology Bulletin, 22, 249-256.
Rogers, T. B. (1981). A model of the self as an aspect of human information processing system. In N. Cantor & J. F. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp. 193-214). Hillsdale, NJ: Erlbaum.
Rosch, E. (1975). Cognitive reference points. Cognitive Psychology, 7, 532-547.
Ross, L., Greene, D., & House, P. (1977). The "false consensus effect": An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13, 279-301.
Sadalla, E. K., Burroughs, W. J., & Staplin L. J. (1980). Reference points in spatial cognition. Journal of Experimental Psychology: Human Learning and Memory: 6, 516-528.
Sears, R. R., Rau, L., & Alpert, R. (1965). Identification and child rearing. Stanford: Stanford University Press.
Sedikides, C., Olsen, N., & Reis, H. T. (1993). Relationships as natural categories. Journal of Personality & Social Psychology, 64, 71-82.
Sedikides, C., Schopler, J., & Insko, C. (Eds.). (1998). Intergroup cognition and intergroup behavior. Mahwah, NJ: Erlbaum.
Shrauger, J. S. & Patterson, M. B. (1974). Self-evaluation and the selection of dimensions for evaluating others. Journal of Personality, 42, 569-585.
Simon, B. (1993). On the asymmetry of the cognitive construal of ingroup and outgroup: A model of egocentric social categorization. European Journal of Social Psychology, 23, 131-147.
Smith, E. R. & Henry, S. (1996). An in-group becomes part of the self: Response time evidence. Personality & Social Psychology Bulletin, 22, 635-642.
Smith, E. R. & Zarate, M. A. (1992). Exemplar-based model of social judgment. Psychological Review, 99, 3-21.
Smith, L. B., Cooley, N. J., & McCord, C. (1986). What is "high"?: The development of reference points for "high" and "low". Child Development, 57, 583-602.
Snyder, C. R., & Fromkin, H. L. (1980). Uniqueness: The human pursuit of difference. New York: Plenum.
Spence, J., Helmreich, R., & Stapp, J. (1974). The personality attributes questionnaire: A measure of sex-role stereotypes and masculinity-femininity. JSAS Catalog of Selected Documents in Psychology, 4, 43, (MS no. 617).
Staats, S. & Skowronski, J. J. (1992). Perceptions of self-affect: Now and in the future. Social Cognition, 10, 415-431.
Stangor, C., Lynch, L., Duan, C., & Glass, B. (1992). Categorization of individuals on the basis of multiple social features. Journal of Personality & Social Psychology, 62, 207-218.
Suls, J. & Wills, T. A. (1991). Social Comparison: Contemporary theory and research. Hillsdale, NJ: Lawrence Erlbaum.
Sun, R. K. (1983). Perceptual distances and the basic color term encoding sequence. American Anthropologist, 85, 387-391.
Tajfel, H. (1981). Human groups and social categories : studies in social psychology. New York: Cambridge University Press.
Taylor, S. E., Fiske, S. T., Etcoff, N. L., & Ruderman, A. J. (1978). Categorical and contextual bases of person memory and stereotyping. Journal of Personality & Social Psychology, 36, 778-793.
Winer, B. J. (1971). Statistical principles in experimental design. (2nd. ed.). New York: McGraw-Hill.
Winter, L. & Uleman, J. S. (1984). When are social judgments made? Evidence for the spontaneousness of trait inferences. Journal of Personality & Social Psychology, 47, 237-252.
Wyer, R. S., & Srull, T. K. (1989). Memory and cognition in its social context. Hillsdale, NJ: Erlbaum.
Zarate, M. A. & Smith, E. R. (1990). Person categorization and stereotyping. Social Cognition, 8, 161-185.
1. We are referring to this type of social priming as target-priming to distinguish it from attribute-priming in which previous activation of a specific attribute (e.g., helpful) facilitates subsequent activation of the same or related attribute (e.g., friendly). We will return to this distinction in General Discussion.
2. It should be noted that most research using response latency as an indicator of accessibility and/or cognitive processing relies on dichotomous (yes/no) responses rather than on a 5-point Likert-type scale (however, see Fazio & Williams, 1986; Huston & Fazio, 1989; Karylowski, 1990; Staats & Skowronski, 1992 for some exceptions).The dichotomous scale is normally preferred because it reduces the error variance of response latencies by minimizing the contribution of time required to locate and to press the appropriate response key. The reason we used a Likert-type scale in Experiments 1, 2, & 3 was because we wanted to be able to better control for similarity of judgments regarding priming and test targets. See Fazio (1990) for a further argument on acceptability of using latency measures with a Likert-type scale.
3. .Most of the activities were adopted from Kuder Reference Record Vocational Form CP (Kuder, 1981).
4. In this and subsequent experiments latencies were transformed to logarithms prior to data analysis. This was done to reduce positive skew, as recommended by Winer, 1971, p.200. For the purpose of clarity, when reporting mean unstandarized latencies, we present means transformed back to the original scale.
5. This high percentage of very fast responses was apparently due to an oversensitivity of response boxes used in the present experiment.
6. The mean unstandardized absolute differences between priming and test judgments were as follows: M = .92 for trials with matched exemplar primes, M = .82 for trials with matched prototype primes, M = 1.69 for trials with mismatched exemplar primes, and M = 1.41 for trials with mismatched prototype primes.
7. A comparison between average latencies for trials with fully mismatched primes (M = 2.869) and latencies for trials with control primes (M = 3.416), F (1, 100) = 184.25, p < .001, indicates that in addition to a category-specific effects, non-specific priming has also occurred. It seems likely that, in judging which of two activities required more experience and training, participants focused on different aspects of those activities than they did in answering preference questions. If so, longer latencies in answering a subsequent question regarding preferences might have been simply due to the extra time required to focus on preference-relevant aspects of the two activities.
8. The mean unstandardized absolute differences were as
follows: M = 1.08 for matched exemplar primes, M = .84 for matched prototype
primes, M = 1.43 for mismatched exemplar primes, and M = 1.50 for mismatched
prototype primes.