Candidates As Engendered Products: Prototypes in Political Person Perception

Basil G. Englis, Rutgers University
Greta Eleen Pennell, Rutgers University
ABSTRACT - Current models of voting behavior consider voting as partly based on an evaluation of the candidate as a person. Such candidate-centered approaches to voting direct our attention voting as a social judgment. Thus, voters' prototypes of the "ideal" politician become important in considering how individual candidates will be judged and how votes will be cast. Moreover, a better understanding of voters' perceptions of the "ideal" politician can be useful strategic input for the positioning of political candidates. Previous studies of the prototypes associated with ideal candidates have not provided a sufficiently detailed examination of the content of voter prototypes (content domains) nor have they considered potential differences in prototypes as a function of level of office or candidate gender. The findings of the present research emphasize the difference between the prototypes of ideal candidates when gender is left unspecified, as compared with the prototypes for ideal political candidates whose gender is specified as either male or female. The discussion considers the implications of the results for the positioning of female politicians.
[ to cite ]:
Basil G. Englis and Greta Eleen Pennell (1993) ,"Candidates As Engendered Products: Prototypes in Political Person Perception", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 612-619.

Advances in Consumer Research Volume 20, 1993      Pages 612-619

CANDIDATES AS ENGENDERED PRODUCTS: PROTOTYPES IN POLITICAL PERSON PERCEPTION

Basil G. Englis, Rutgers University

Greta Eleen Pennell, Rutgers University

ABSTRACT -

Current models of voting behavior consider voting as partly based on an evaluation of the candidate as a person. Such candidate-centered approaches to voting direct our attention voting as a social judgment. Thus, voters' prototypes of the "ideal" politician become important in considering how individual candidates will be judged and how votes will be cast. Moreover, a better understanding of voters' perceptions of the "ideal" politician can be useful strategic input for the positioning of political candidates. Previous studies of the prototypes associated with ideal candidates have not provided a sufficiently detailed examination of the content of voter prototypes (content domains) nor have they considered potential differences in prototypes as a function of level of office or candidate gender. The findings of the present research emphasize the difference between the prototypes of ideal candidates when gender is left unspecified, as compared with the prototypes for ideal political candidates whose gender is specified as either male or female. The discussion considers the implications of the results for the positioning of female politicians.

The growing dominance of electronic media in the marketing of political candidates has dramatically affected the manner in which voters are exposed to politicians. One hundred, or even fifty, years ago voters generally only read the speeches of politicians as printed in pamphlets and in newspapers, or they read articles written about their issue positions and speeches. Only in the realm of local politics did voters have any form of more "direct" contact with politicians. Even then politicians were usually seen in a crowded setting C a smoke-filled auditorium or waving from the back of a departing train. And, at the national level of politics only a small fraction of the electorate ever actually heard or saw politicians. The advent and proliferation of radio acted to "humanize" political information, transforming politicians into people and enabling voters to make inferences about candidates as people. For example, when a politician increased the volume or cadence of his or her speech a listener might infer that the speaker was determined, whereas hesitations and stutters might convey a lack of certainty (cf. Englis 1992).

Today, television dominates voter "contact" with politicians (e.g., Graber 1984; Swerdlow 1988). In bringing politicians into the intimacy of America's homes, television directs voters' attention to the personal characteristics and qualities of politicians. It allows viewers to examine subtle changes in facial expression and posture. Television also provides "behind the scenes" (although often staged) views of politicians at home with their families, on vacation, or at social gatherings. As a result, personal attributes (i.e., personality, physical appearance) become inescapable aspects of politicians as people. Although these media bites "humanize" political leaders, the access provided to voters also highlights the need for marketers to "manage" mass-mediated voter impressions. The recent case of President Bush's gastric disturbance in Japan highlights the risks of "uncontrolled" television coverage.

This emphasis on politicians as "people" has in part arisen from the properties of broadcast media, and in part by a shift in the nature of political campaigns. Increasingly, "style becomes substance" (McGinniss 1969, p. 30) as "modern politicians are created and marketed [for public consumption] much like a bar of soap" (Solomon 1992, p. 29). In much the same way as personal advertisements promote people as products (Hirschman, 1987), politicians are commodified and voters are lured by presumably desirable constellations of personal characteristics. Today it is people, more than ideologies or issues, that are the products of politics. The basis for voters' evaluations, attitudes, and eventual voting behavior has shifted from the traditional arena of parties and issues to that of social judgment. This paper examines the personal characteristics that voters associate with "ideal" politicians and considers the importance of these perceptions in the positioning and marketing of political candidates.

Models of Voting Behavior

Since the 1940s, three general models of voting behavior have dominated political science: these emphasize either the party (party-centered), policies (policy-centered), or the person (candidate-centered) (e.g., Campbell, Converse, and Miller 1960; Kirkpatrick, Lyons, and Fitzgerald 1975; Popkin et al. 1976). Party-centered models dominated from the 1940s through the 1960s and conceptualized voting as a choice between parties that was dictated by long-standing party loyalties. Thus, party label acted like a brand name by representing both a mark of affiliation with one group over another, and a general set of expectancies regarding future policies. As long as important differences do not exist between the product offered (platform) and the consumer's own needs, party is assumed to be the basis of voter decision-making. However, historical analyses (e.g., Stokes and Miller 1962) reveal that party membership data alone have rarely been sufficient to predict voting behavior in national elections.

Policy-centered models gained prominence during the 1960's and 1970's when specific issues (i.e., the Vietnam War, Watergate) polarized the electorate along other than traditional party lines. These models conceptualize voting as the result of a voter's evaluations of each candidate's positions on important issues, and how well the candidates' positions match the voter's own views (e.g., Fiorina, 1981). Policy-centered models assume that voters are in a high-involvement choice situation and that actual differences between the issue positions of the candidates are of central importance in determining voting behavior.

However, unless issues loom large for the voter either because of their personal relevance or the nature of the election, there is little incentive to make the complex judgments often required in differentiating between candidates' issue positions and policies (cf. Popkin et al. 1976). Candidate-centered voting models conceive of voting as a choice between candidates as people, rather than as a choice between policies or parties. It is assumed that voters either have little reason for incurring the cost of analyzing true policy differences, or that the differences are so small as to be virtually nonexistent, or finally, that issue involvement is so low that choice is based on candidate attributes rather than issue position. Several early studies suggest that voter choice has become progressively more candidate centered (e.g., Kirkpatrick, Lyons, and Fitzgerald, 1975; Popkin et al, 1975). Thus, voters are in the position of appraising candidates' personal qualities in judging their ability to carry out the demands of office, and the personal characteristics that candidates convey to voters become important factors in the social judgment assumed to underlie voting.

Prototype Theory and Voting as Social Judgments

The contexts, processes, and consequences of social judgments have been widely studied under the rubric of social cognition. Much of this research has focused on the ways in which information about people is selectively gathered, organized and used in social life. Cantor and Mischel (1977) argue that much social information is coded "in terms of . . . a few simple cognitive categories" (p. 6) which are organized into prototypes. Such social prototyping acts as a kind of "personnel file," providing ". . . a convenient precis of the one with whom we wish to deal" (Klapp 1972, p. 4). Social prototypes act as "cognitive reference points" forming the basis for inferences and other forms of reasoning (Rosch 1975; 1981).

As noted by Lakoff (1987), the content of prototypes can be non-stereotypical, and can include rare, abstract, ideal cases that are then used in categorization and decision making. Thus, people can readily describe their ideal home, mate, or boss even though the ideal case has never been encountered in reality. The use of prototypes has been examined in several judgment areas relevant to the present study. For example, occupational prototypes have been shown to be organized around the traits and abilities needed for their enactment (Ostrom and Lingle 1977). Recent consumer research has applied a prototype framework to understanding how consumers use product attributes to represent product categories (Sujan 1985), and how consumers use patterns of consumption activities to represent lifestyle groups (Englis and Solomon 1992a,b).

To the extent that voting has become more candidate centered, a prototype framework should be useful for examining the cognitive organization of voters' perceptions of ideal politicians. Indeed, in the early 1980s systematic research was undertaken to identify candidate characteristics that might be related to voter's choices (e.g., Abelson et al. 1982 Kinder and Abelson, 1981; Sullivan et al. 1984). For example, in 1980 the pre-presidential National Election Survey protocol was modified to include structured questions concerning the traits and emotions voters associated with the candidates. Analyses showed that 60% of the variance in preference for Carter or Reagan was accounted for by trait and emotion measures alone (Kinder and Abelson 1981; also Abelson, Kinder and Fiske 1982). As Kinder and Abelson (1981) note:

"Trait judgments and affective reactions are not merely rationalization for party allegiance... the substantial autonomy of traits and affects from party attachments contributes to our general case. The choice of a president is very much a choice between two people - in particular, the traits they convey and the feelings they elicit." (p. 23)

Clearly, perceptions of candidates as people contribute to voting behavior independently of more "traditional" factors such as party identification or the candidate's stand on the issues of the day. Although these findings strongly suggest that social judgment is an important component of voters' decision making, they do not shed much light on voters' conceptions of what attributes define the "ideal politician."

One group of researchers (Kinder et al. 1980) used an open-ended protocol to assess the characteristics voters associated with ideal and "anti-ideal" presidents, and measured personality traits (e.g., honesty, recklessness) and areas of performance (e.g., provides strong leadership, breaks the law). The results showed that association of positive and negative characteristics with different candidates was related to voter support. However, other content domains (e.g., physical appearance) are important in social judgment, and there is empirical evidence to suggest that voter perceptions of politician attributes are partly dependent on level of office. One study (Adams 1974) showed that a national office (such as president) is seen as a more symbolic role than is a local office (where voters may expect a concrete outcome). And, attributes associated with groups under-represented in politics (i.e., women or people of color) were found to be liabilities for candidates in national but not local elections. In the present study we examine differences in politician prototypes as a function of level of office, and we expand the range of content domains.

Gender has long been considered an important factor in social judgment. [Whereas "sex" is typically used to denote the biological factors that differentiate men and women, "gender" is used to indicate that male and female represent socially contructed categories.] Gender has been shown to influence the overall evaluation of female candidates (e.g., Mend, Bell and Bath 1976), and physical attractiveness acts as a liability for female but not for male candidates (e.g., Bowman 1984; Sigelman et al. 1986). However, little attention has been paid to the question of how the characteristics associated with female politicians compare with those associated with males, and how each compares with voter prototypes when gender is unspecified.

Since the great majority of office holders have been male, voter prototypes of ideal politicians should over-represent "male" characteristics. Phenomena which co-occur frequently become tightly linked in people's minds (Tversky and Kahneman 1974). Taken to the extreme, one well-understood or easy-to-perceive aspect of a category may come to stand for the entire category (Lakoff 1987). In the case of political prototypes the high level of covariation between being a politician and being male may have "masculinized" the role of politician. The "engendering" of social roles is found in research on highly sex-segregated occupations. Such jobs take on the "Y gendered attributes associated with the sex of their workforce" (Williams 1989, p. 2). For example, secretaries (99% women) and are expected to be emotionally sensitive and nurturing whereas airplane pilots (99% men) are assumed to be emotionally reserved and detached (Williams, 1989). The extent to which the prototypes of the ideal and ideal male politician overlap with one another and differ from ideal female prototypes provides one indication of how each level of office is engendered. Comparison of the prototypes for female politicians with those for males, as well as to the prototype for an ideal politician when gender is unspecified can provide strategic input to how female candidates may best be positioned for election into particular offices.

Research Overview and Hypotheses

One of the primary limitations of previous research concerning political prototypes has been the treatment of politicians as a relatively homogeneous group, with little consideration of the effect of politician gender or level of office on the content and elaboration of prototypes. This study was designed to test four hypotheses concerning the effects of politician gender and level of office on political person prototypes.

H1: Since there is a strong historic association between males and political office, the prototypes for the ideal when gender is unspecified and for the ideal male politician will be similar in content and will differ from those associated with ideal female politicians. A corollary to this hypothesis is that we do not expect to find differences between the prototypes of the ideal and male political figures.

H2: Since voters have less exposure to and experience with female politicians, there should be a narrower range of characteristics (i.e., less elaboration and greater sterotypy) associated with female prototypes than for either the gender-unspecified ideal or male politicians.

H3: Since there is more media attention paid to the presidency than other levels of office, presidential prototypes should be more elaborated than either the gubernatorial or legislative prototypes. Moreover, we do not expect a significant difference in the degree of elaboration of the latter two prototypes.

H4: Greater familiarity with and exposure to the presidency should result in a prototype that displays the greatest level of variation of content (less stereotypy). Since gubernatorial office and the presidency are both executive roles, the prototypes for these offices should be similar and distinct from the prototype of the ideal legislator (irrespective of gender).

METHOD

Sample. One hundred seventy nine undergraduates (average age 21 yr.) enrolled in an introductory marketing course received class credit for their participation in the study. Five respondents were dropped from all analyses due to incomplete questionnaires or extreme levels of response (>3 standard deviations above the sample mean; see Mosteller and Tukey 1977, pp. 16-25 on contaminated distributions). The final sample (N = 174) was comprised of 59 males and 115 females. Eighty percent of the sample was registered to vote and 50% had voted in the last election. Seventy-six percent described themselves as moderate on most political issues and nearly half of the registered voters in the sample (48.8%) were registered Republicans. The sample was generally a suburban, upper middle class group (average family income $71,500) with 66.6% having fathers with a managerial or professional occupation). [Owing to the higher than average SES of our sample, the findings may not generalize to all Americans and may be limited to the prototypes held by college-educated, upper-middle class voters-a considerable block of voters.]

Procedure. Respondents were told that the study was concerned with the characteristics, qualifications, abilities, and behaviors voters associate with ideal political figures. They were given an open-ended questionnaire concerned with ideal political figures and encouraged to write whatever came to mind for each question, and to feel free to use their own words and phrases in their descriptions. Recent work in social cognition and gender stereotypes has led researchers to conclude that these belief systems are composed of more than personality traits (Ashmore and Del Boca 1981), and that the multiple descriptive domains are only loosely interconnected (cf. Ashmore and Del Boca 1986; Deaux 1985). Thus, we included traits, physical appearance, abilities, and emotions as content domains in our assessment of political person prototypes. The questions were designed to encourage respondents to imagine or picture their ideal politician and were worded as follows: "Please try to imagine the ideal president [or legislator or governor]. When you think of this person what are the traits or personality characteristics that come to mind? List and/or describe these characteristics and traits in the space below." A similar format was used for the physical appearance, ability, and emotion content domains. Thus, the design was a between-subjects factorial manipulation of candidate gender (male, female, or unspecified) and level of office (president, governor, or state legislator/representative); respondents were assigned to one of the resulting nine conditions.

Prototype elaboration was measured by a simple count of the number of items mentioned in response to each question. Prototype content was coded with a data-driven scheme designed to stay as close as possible to the idiographic data. Initial content coding yielded 187 trait categories, 172 physical appearance categories, 155 abilities and 92 emotions. A strong decision rule was used to collapse categories on the basis of semantic similarity (thesaurus lists of synonyms). Items similar in meaning to the others in the group and representative of the larger category were combined. Then, using a "scree slope" approach within each content domain, the point of discontinuity in the frequency of category use was identified as a cut-off point. On this basis, the number of categories was reduced to 28 traits, 26 physical appearance attributes, 17 abilities, and 14 emotions.

RESULTS

Prototype Elaboration

Separate analyses were conducted for the elaboration scores within each prototype content domain. A 3 (politician gender) x 3 (level of office) x 2 (sex of voter) x 2 (voter/non-voter) between-subjects ANOVA was the full model for each analysis. The strongest effects were obtained for the emotion component of the prototypes. There was a significant effect of politician gender [F (2, 141) = 5.113, p < .007], where prototypes were more elaborated for the gender-unspecified ideal (M = 3.07) and male politicians (M = 3.35) than for the female politician (M = 2.63, p < .003). There was also a significant main effect for sex of respondent [F (1, 141) = 8.658, p < .02]. Females provided more descriptors of how their ideal political figure would make them feel (M = 3.20) than did males (M = 2.66). The ANOVA also revealed a significant interaction involving sex of respondent and voting behavior [F (1, 141) = 6.640, p < .048]. In this interaction males and females who did not vote in the last election had equally well-elaborated prototypes (M = 2.7 and 3.0, respectively). However, male voters, generated fewer responses to this question (M = 2.6) than did female voters (M = 3.45, p < .005).

Analyses of the physical appearance characteristics revealed a significant level of office by sex of voter interaction (F(2,142) = 13.35, p < .05). This interaction was carried primarily by the greater elaboration of the physical appearance prototype for the ideal president by female voters (M = 6.08) than by male voters (M = 4.0, p < .0003). Males provided significantly fewer descriptors of their ideal president's physical appearance (M = 4.0, p < .05) than they did for their ideal legislators (M = 5.46) or governors (M = 4.78). Females, on the other hand, listed more physical appearance attributes for an ideal president (M = 6.08, p < .05) than for an ideal legislator (M = 5.03) or governor (M = 5.1).

There were no significant main effects or interactions in prototype elaboration for traits or abilities. However, one planned comparison for level of office showed that a greater number of traits was used to describe the ideal president (M = 6.186, p < .04) than for either legislator (M = 5.245) or governor (M = 5.417).

Prototype Content

Quantitative as well as qualitative approaches were used to examine the content of politician prototypes. Contingency tables were constructed to examine whether or not content category use (e.g., the specific patterns of traits mentioned) differed as a function the independent variables in the design (level of office, politician gender, and so on). Category use was contingent upon politician gender (TraitsX2(54) = 96.66, p < .0003; Physical Appearance X2(50) = 130.88, p < .0001; Abilities X2 (32)= 47.84, p < .034; Emotions X2 (26)= 45.43, p < .011). In addition, the descriptors used across the three levels of office differed significantly in content in all of areas except physical appearance (Traits X2 (54)= 79.76, p < .013; Abilities X2 (32)= 51.53, p < .016; Emotions X2 (26) = 39.03, p < .049). There were no effects of sex of voter and voting experience. Although category usage was contingent overall upon level of office and politician gender, there were also several commonalities in politician prototypes. In order to provide a sense of general prototype content, the following summarizes the most commonly used descriptors overall.

In general, ideal political figures are seen as possessing the traits of honesty, warmth and charm, and being knowledgeable and open-minded. They are also intelligent, strong, and confident. The prototypic appearance is of someone in their 40's or 50's, who is well-dressed, tall (over six feet), and attractive but not too attractive. Many respondents were quite specific in terms of what constitutes appropriate attire, including not just "suits" but "Brooks Brothers" or "navy business suit" in their descriptions. Hair color and style, along with facial features (e.g., eye color, smiles, wrinkles) were also frequently mentioned. The most prominent ability area related to communication skills C public speaking, writing well, and giving clear answers; related abilities included being good at working with a wide range of personalities, being a skilled manager and negotiator, and keeping in touch with and trying to meet the needs of the people. In addition, the ideal politician is seen as persuasive, organized, good at listening, and able to make decisions. Another ability area included having a broad understanding and knowledge of the issues in general and within specific areas such as defense, budget, and social problems. The emotions associated with an ideal politician included general feelings of happiness and security C feeling safe (protected and secure), trusting, comfortable (at ease, relaxed) and self-confident. Others involved belief or faith in the politician, a feeling of being "in good hands," and the feeling that an ideal politician is someone to look up to who elicits feelings of pride.

Analyses of content within the gender of politician and level of office conditions also revealed many subtle differences in the traits, physical appearance, abilities and emotions used to describe ideal political figures. The following highlights a few of these differences, especially as they vary as a function of gender and level of office. Tables 1 through 4 provide a detailed listing of the content of each prototype. The percentages included in these tables were computed on the basis of the total number of responses generated by all subjects within each cell of each table.

Politician Gender

As predicted, the content of the female politician prototypes differed in important ways from male or gender-unspecified prototypes. Although being intelligent and knowledgeable were frequently associated with gender-unspecified and male prototypes, these traits were not part of the female prototype. [An exception was the case of the female president which included intelligent but not knowledgeable.] It is interesting to note, that although intelligence was not part of the female legislator or governor prototype, being well-educated was part of both. It appears that although intelligence may be sufficient in most cases, that general trait requires some type of credential in women to be validated. A similar pattern was found for the trait "confident." The prototypical female political figure is seen as conservatively dressed, as are all presidential prototypes. Voters were more likely to mention facial features when asked about an ideal female politician. Keeping in touch with the needs of the public, organization, and being able to get the job done and accomplish goals were abilities commonly found in all prototypes except those for female politicians. Although persuasive abilities were scattered across the prototypes, all female targets included this ability. Finally, the emotions respondents anticipated experiencing in relation to their ideal female political person were more likely to involve respect and a sense that the person was a role-model or someone to look up to.

Level of Office. Responses to the question regarding the abilities of the various political figures provided the clearest distinction between the three levels of office. Presidents and governors were described as being good at negotiations and at decision making. Despite the different branches of government represented by governors and legislators, these two offices shared many features including people skills and keeping in touch with the needs of the public. A possible explanation for this is that both represent a more "local" level of office.

DISCUSSION

In this study we examined the elaboration and content of nine political person prototypes. The most striking pattern in the present results is that degree of elaboration of politician prototypes is generally constant across level of office and politician gender, but that the content of these prototypes is dependent upon both variables. Prototypes were more well-elaborated among female respondents and among those with prior voting experience (i.e., more highly involved respondents), than for males and non-voters. This result may be partly due to women's greater verbal fluency. And, those with prior voting experience have most likely given more thought to the qualities they seek in a politician as part of their decision making when they vote. This higher level of involvement may be reflected in the more extensive prototypes among voters. Despite the differential elaboration of prototypes by men and women and by voters and non-voters, these groups did not differ significantly in the content used to describe their ideal political figures.

Our examination of prototype content provided strong support for our hypotheses that the category of political persons is not homogeneous and that female political figures are perceived very differently from either the gender-unspecified ideal or male politician. This result supports our assumption that prior treatment of political candidates as a homogeneous group is too simplistic for understanding how political candidates are perceived by voters. The findings clearly indicate that the prototype of an president is different from that of an ideal legislator or governor. This study represents a preliminary foray into differentiating the larger category of "politician" into meaningful subordinate categories. Future research should consider how those differences arise as a function of different responsibilities of office (e.g., legislative vs. executive) and as a function of the perceived closeness of the office to the respondent. For example, governors and legislators may be seen as being in direct contact with the average citizen whereas the president is relatively isolated from such contact. Research is also needed which examines the effects of media exposure on the shaping of political prototypes.

Important differences between ideal female politicians and the gender unspecified and male ideal politicians were also evident. Of particular note was the omission of numerous traits, abilities, and emotions that were pervasive in descriptions of the gender-unspecified and male ideals. These omissions may not indicate that people adjust what they believe to be ideal when they think of female politicians. These omissions, especially when compared to the more abstract ideal, may instead reflect the salience of particular features in characterizing a particular social category. For example, we do not think that because knowledgeable was excluded from the ideal female president prototype that people do not feel a female president needs to be informed or that knowledgeable is somehow less important for a female president than a male or the ideal president. We are suggesting that when people think about their ideal female president, for example, being knowledgeable simply does not occur to them.

TABLE 1

THE SIX MOST FREQUENTLY OCCURRING TRAITS IN EACH POLITICAL PERSON PROTOTYPE

TABLE 2

THE SIX MOST FREQUENTLY OCCURRING PHYSICAL APPEARANCE ATTRIBUTES IN EACH POLITICAL PERSON PROTOTYPE

TABLE 3

THE SIX MOST FREQUENTLY OCCURRING ABILITIES IN EACH POLITICAL PERSON PROTOTYPE

TABLE 4

THE FIVE MOST FREQUENT EMOTIONS ASSOCIATED WITH EACH POLITICAL PERSON PROTOTYPE

Such selective misperception may result in a disadvantage for female candidates, who may as a consequence be perceived as being further from the ideal. Conversely, there are instances in which the female prototype is closer to the gender-unspecified ideal than the male prototype. Highlighting this proximity is potentially an important factor in enhancing candidate success. Exploring the multiplicity of ways in which the female prototype differs from and is similar to the ideal is a first step in understanding how female candidates can be positioned within the constellation of traits, appearance, abilities, and emotions associated with ideal political leaders.

REFERENCES

Abelson, Robert P., Kinder, Donald R., Peters, Mark D., Fiske, Susan T. (1982), "Affective and Semantic Components in Political Person Perception." Journal of Personality and Social Psychology, 42, 619-630.

Adams, William C. (1974), "Candidate Characteristics, Office of Election, and Voter Responses," Paper presented at the 1974 annual spring meeting of the National Capital Area Political Science Association, Washington, D.C.

Ashmore, Richard D., and Del Boca, Francis K. (1981), "Conceptual Approaches to Stereotypes and Stereotyping," in Cognitive Processes in Stereotyping and Intergroup Behavior, D. L. Hamilton ed., Hillsdale, NJ: Lawrence Erlbaum.

Ashmore, Richard D. (1986), "Gender Stereotypes," in The Social Psychology of Female-Male Relations: A Critical Analysis of Central Concepts, Richard D. Ashmore and Francis K. Del Boca eds., New York: Academic Press.

Bowman, Ann. (1985), "Physical Attractiveness and Electability: Looks and Votes," Women and Politics, 4, 55-65.

Campbell, Angus, Converse, Philip E., Miller, W. E., and D. E. Stokes (1960). The American Voter. NY: Wiley & Sons.

Cantor, Nancy and Walter Mischel (1979), "Prototypes in Person Perception," Advances in Experimental Social Psychology, Vol. 12, in Leonard Berkowitz,ed., New York: Academic Press, 4-52.

Deaux, Kay (1984), "From Individual Differences to Social Categories: Analysis of a Decade's Research on Gender," American Psychologist, 39, 105-116.

Englis, Basil G. (1992). "The Role of Affect in Political Advertising: Voter Emotional Responses to the Nonverbal Behavior of Politicians," in Advertising and Consumer Psychology, Eddie Clark, Timothy Brock and David Stewart eds., Hillsdale, NJ: Lawrence Erlbaum.

Englis, Basil G. and Michael R. Solomon (1992). "Consumption Constellations: Implications for Advertising Strategy." Proceedings of the Society for Consumer Psychology, 23-30.

Englis, Basil G. and Michael R. Solomon (1992b), "To Be and Not to Be?: Reference Group Stereotyping and The Clustering of America," Rutgers University, unpublished manuscript.

Fiorina, M. D. (1981). Retrospective Voting in American National Elections. New Haven, CT: Yale University Press.

Graber, Doris A. (1984). Mass Media and American Politics, 2nd ed. Washington, DC: Congressional Quarterly Press.

Hirschman, Elizabeth C. (1987), "People as Products: Analysis of a Complex Marketing Exchange," Journal of Marketing, 51 (January), 98-108.

Kinder, Donald R. and Robert P. Abelson (1981), Appraising presidential candidates: Personality and affect in the 1980 campaign. Paper presented at the 1981 annual meeting of the American Political Science Association.

Kinder, Donald R., Mark D. Peters, Robert P. Abelson, and Susan T. Fiske (1980), "Presidential Prototypes," Political Behavior, 2, 315-337.

Kirkpatrick, S. A., W. Lyons, and M. R. Fitzgerald (1975), "Candidates, Parties, and Issues in the American Electorate." American Politics Quarterly, 3, 247-283.

Klapp, Orrin E. (1972), Heroes, Villians, and Fools: Reflections of the American Character, San Diego: Aegis.

Lakoff, George (1987), Women, Fire, and Dangerous Things: What Categories Reveal About the Mind, Chicago: University of Chicago Press.

McGinniss, Joe (1969), The Selling of the President 1968, New York: Trident.

Mend, Michael R., Tony Bell and Lawrence Bath (1976), "Dynamics of Attitude Formation Regarding Women in Politics," Experimental Study of Politics, 5, 25-39.

Ostrom, Thomas M., and J. H. Lingle (1977), Thematic Constructions in Person Impressions. Paper presented in a symposium titled "New Directions in the Analysis of Impression Formation," American Psychological Association Convention.

Popkin, S., J. W. Gorman, C. Phillips, and J. A. Smith(1976). "What have you done for me lately? Toward an investment theory of voting," The American Political Science Review, 70, 779-805.

Rosch, Eleanor (Eleanor Heider) (1975) "Cognitive Reference Points." Cognitive Psychology, 7, 532-547.

Rosch, Eleanor (1981) "Prototype Classification and Logical Classification: The Two Systems" in New Trends in Cognitive Representation: Challenges to Piaget's Theory, E. Scholnick ed., Hillsdale, NJ: Lawrence Erlbaum, pp. 73-86.

Sigelman, Carol K., Dan B. Thomas, Lee Sigelman, and Frederick D. Ribich (1986). "Gender, Physical Attractiveness, and Electability: An Experimental Investigation of Voter Biases." Journal of Applied Social Psychology, 13, 229-258.

Solomon, Michael R. (1992), "Celebritization and Commodification in the Interpersonal Marketplace," Rutgers University, unpublished manuscript.

Stokes, D. E. and W. E. Miller (1962), "Party Government and the Saliency of Congress." Public Opinion Quarterly, 26, 531-546.

Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgments," Journal of Consumer Research, 12 (June), 31-46.

Sullivan, Denis G., Roger D. Masters, John T. Lanzetta, Basil G. Englis, and Gregory J. McHugo (1984). The Effect of President Reagan's Facial Displays on Observers' Attitudes, Impressions, and Feelings About Him. Paper presented at the 1984 annual meeting of the American Political Science Association, Washington, DC.

Swerdlow, Joel L. (ed.) (1988), Media Technology and the Vote, Boulder, CO: Westview Press.

Tversky, Amos & Kahneman, Daniel, (1974) "Judgment under Uncertainty: Heuristics and Biases." Cognitive Psychology, 5, 207-232.

Williams, Christine, L. (1989), Gender Differences at Work: Women and Men in Nontraditional Occupations, Berkeley: University of California Press.

----------------------------------------