Political Psychology, Vol. 20, No. 1, 1999 Personality Profiles and Political Parties Gian Vittorio Caprara and Claudio Barbaranelli University of Rome "La Sapienza" Philip G. Zimbardo Stanford University This paper explores relationships between basic personality profiles of voters and their political party preferences. The Italian political system has moved recently from previously extreme, ideologically distinctive parties to form complex coalitions varying around more centrist orientations. Significant evidence was found for the utility of the Five-Factor Model of Personality in distinguishing between voters' expressed preferences, even given this greater subtlety in proposed values and agendas. More than 2,000 Italian voters who self identified as having voted for new center-left or center-right political coalitions differed systematically in predicted directions on several personality dimensions measured by the Big Five Questionnaire. In the context of the model, center-right voters displayed more Energy and slightly more Conscientiousness than center-left voters, whose dominant personality characteristics were Agreeableness (Friendliness) and Openness; Emotional Stability was unrelated to either group. This relationship between individual differences in personality and political preferences was not influenced by the demographic variables of voters' gender, age, or education. Thus, personality dimensions proved to be stronger predictors of political preference than any of these standard predictor variables. Implications are discussed regarding links among personality, persuasion, power, and politics. KEY WORDS: personality profiles; political parties; Five-Factor model; voters. Politics involves systems of external rules and implicit principles of power management for achieving leader or party goals, ideally for the communal good. Personality involves systems of distinctive self-regulatory mechanisms and structures for guiding cognitive, affective, and motivational processes toward achieving individual and collective goals, while preserving a sense of personal identity (Bandura, 1997; Caprara, 1996; Mischel & Shoda, 1995). How these societal and individual systems might be related has long been a source of speculation and 175 0162-895X © 1999 International Society of Political Psychology Published by Blackwell Publishers, 350 Main Street, Maiden, MA 02148, USA, and 108 Cowley Road, Oxford, OX4 1JF, UK. 176 Caprara, Barbaranelli and Zimbardo serious concern for philosophers, political scientists, psychologists, and ordinary citizens. Although it is easy to think of these entities as existing in totally different realms, operating at different levels and with different operational structures, there are vital commonalities that suggest a more dynamic interaction between politics and personality. Political parties advocate beliefs and values that legitimate the socioeconomic conditions in which people live, and which they aspire to achieve. Thus, they can exert enormous influence on the quality of the daily life of individual citizens, even shaping basic perspectives of options, goals, attitudes, and values. However, political parties are not simply sociological entities, but rather are creations of, and collections of, people who themselves operate as individual and social entities. Citizens bring to the political arena needs and aspirations for personal and social well-being that determine their choice of political party, and, in turn, may influence the agendas and behavior of politicians. In democracies, both politicians and the people they serve set conditions and constraints on each other's aspirations. Studying the relationships between personality and politics is complicated by all the inherent difficulties in establishing broad person-behavior-situation reciprocal interactions. Nevertheless, it is important to clarify the extent to which voters' personal dispositions (beliefs, goals, habitual behavior patterns) and political agendas are mutually interdependent. It becomes a matter of empirical research to determine what is general and what is contextual in the relationship between the personalities of individual voters and the ideological positions and agendas of particular political parties. This task seemed easier in earlier times, when the creative team of researchers at the University of California, Berkeley, could develop a psychodynamically focused theory about how the needs and values of those who were characterized as "authoritarian personalities" meshed with their choice of extreme political identification (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950). Despite the theoretical and methodological limitations of that pioneering research (Christie & Jahoda, 1954), it generated intellectual enthusiasm about the ways in which personality constructs might be related to, and enhance, our understanding of political behavior. Research emerged from a host of theoretical perspectives that proposed connections between political behavior and various individual difference constructs from personality and social psychology, such as tender-mindedness and tough-mindedness (Eysenck, 1954), conservatism (McClosky, 1958), alienation (Seeman, 1959), conservatism/dogmatism (Rokeach, 1960), anomy (Srole, 1965), and power motivation (Browning & Jacob, 1964; Winter, 1973). This line of research seemed to hold much promise regarding politics and personality inquiry (Greenstein & Lerner, 1969). However, in the absence of a general theory of personality or consensual agreement about its standardized assessment, research focused on multiple individual constructs without being guided by an integrated conceptual vision (see Personality Profiles and Political Parties 177 Brewster-Smith, 1968; Knutson, 1973). Focusing on the operation of personality traits in isolation gave way to subsuming their impact under the broader study of social attitudes and the power of situational variables as influencing all social behavior, including political action (Zimbardo, Ebbesen, & Maslach, 1977). The resurgence of interest in personality and politics beginning in the late 1970s focused on the analysis of political leadership (see Hermann, 1977, 1986; Simonton, 1990; Tetlock, 1983; Winter, 1987). A variety of individual difference characteristics, such as cognitive style, motivation, intelligence, and value orientation, were assessed using different methods and were linked to a variety of political performances and criteria. It is surprising to us that so little of this body of research investigated relationships between personality and the political preferences of citizens, as anticipated by Di Renzo (1974). We feel that the time is now ripe to pursue the provocative links between patterns or profiles of personality traits of citizen-voters and their particular political behaviors. How do the public policies and promotional propaganda of political parties, especially the rhetoric of political campaigning, affect the kinds of individuals who will come to endorse or reject them? How do the personality patterns of voters create matches or mismatches with the "image management" of political candidates? These are but a few of the questions raised by reflecting on the converging or diverging paths on which politicians, political parties, and personalities of voters may be plotted. A recently developed consensual standard for assessing a limited, fundamental set of personality traits, the Five-Factor Model of Personality (FFM), offers a valuable tool to aid such investigations. In addition, the availability of statistical analysis techniques for determining power effects (as effect sizes) of predicted personality-politics links, while controlling for the many sociological and status variables that usually confound such interpretations, helps contribute to enhancing our knowledge about the relationships between personality and politics. After briefly outlining the utility of the FFM for this type of investigation, we discuss how the new political situation in Italy, as in many democracies worldwide, poses a critical challenge for linking personality to political party preference because of the rise of political coalitions that coalesce around centrist positions instead of diverging around formerly ideological extreme positions. The Five-Factor Model of Personality In recent years, scholars seeking a consensual lexicon to describe personality, in alliance with researchers aiming to identify the basic components of personality structure, have developed the FFM as a common framework for organizing personality descriptors and traits. Although the explanatory value of the model is still under discussion, the robustness of a host of findings across methods, populations, and researchers represents a unique and encouraging event in personality psychology. 178 Caprara, Barbaranelli and Zimbardo The dimensions of Extroversion (or Surgency, Energy), Agreeableness (or Friendliness), Conscientiousness, Emotional Stability (or Neuroticism), and Openness to Experiences (or Intellect, Culture) represent a point of convergence of the psycholexical and questionnaire approaches to the study of personality. Advocates of the FFM argue that it subsumes most of the traditional trait taxonomies and provides a comprehensive and reasonably sufficient summary of major individual differences (Digman, 1990; John, 1990; McCrae, 1989; Ostendorf & Angleitner, 1992; Wiggins, 1996). The strength of the FFM derives from its pragmatic value of representing a well-substantiated and agreed-upon framework for describing personality (Caprara, Barbaranelli, & Livi, 1994). Insofar as it provides a common language for research and assessment in personality psychology, it provides a useful mapping of individual differences. Furthermore, insofar as the FFM identifies the main dimensions underlying the reports and ratings that people make of their own and of others' personalities, it may better focus investigations of the relationships between these dimensions and relevant social outcomes. It is evident that the FFM does not provide a sufficiently fine-grained description of personality because more than five dimensions are needed to capture the multifaceted aspects of individuality and the complex interactions among multiple combinations of traits that give rise to the uniqueness of personalities. It also seems evident that we cannot totally rely on latent dimensions extracted from large populations of respondents to capture the dimensions that underlie the constellations of beliefs and behaviors of single individuals. However, the same factors that result from the aggregation of individual difference data across multiple respondents may provide a valuable compass to map onto a common reference framework the constellations of beliefs and habitual behaviors within given populations. Doing so aids the exploration of their influence on relevant social outcomes, such as political choice. The Contemporary Italian Political System Italy, like the United States and other democratically organized societies, is undergoing a remarkable political transition in which political parties previously identified as extremely divergent on the ends of continuums of political opposition now "regress toward the mean." New coalitions have formed, and continue to evolve, that mesh prior political antagonists into pragmatically organized entities, under new banners, broadly appealing slogans, and contingently varying policies. Before the early 1990s, the conservative-right could be differentiated from the liberal-left around sociological variables such as gender (men were overrepre-sented among left voters), age (older), income or socio-economic status (SES) (higher), and occupation (more professionals and white-collar workers). In contrast, the new Italian coalitions cut across most such traditional boundary markers. Personality Profiles and Political Parties 179 Recently, the political power of the Christian Democrats, Republicans, Liberals, Socialists, and Communists suddenly collapsed, after 40 years of ruling Italy in various combinations. In their place, two main coalitions formed: center-left and center-right. To the left side of this central position went some of the former Christian Democrats, some ex-Socialists, ex-Republicans, and all the ex-Communists (renamed Partito Democratico della Sinista, PDS, and Partito della Rifondazione Comunista, PRC). To the right migrated other Christian Democrats, Socialists, ex-Liberals (under a new party title, Forza Italia), and all the heirs of the Neo-Fascists (under the banner of Allianza Nazionale). Furthermore, a separatist movement independent from the other parties (the Lega Nord) captured a significant portion of votes, mostly in northern Italy. The center-right prevailed in the national elections of 1994, but its period of instability ended with new elections in 1996, in which the center-right ("Polo delle liberta") had a slight popular majority, but the center-left ("Ulivo") prevailed with greater parliamentary representation (because of an electoral system in which a slight majority of votes did not allow the center-right to achieve a parliamentary majority). The most basic ideological and political propaganda differences between these two new coalitions (each filled with many former political "enemies") can be summarized as the center-left expressing greater concern about issues of social welfare and equity (i.e., distributive justice) while the center-right emphasizes its concerns for individual freedom, economic deregulation, and self-ownership of business. If the center-left were now "community-oriented liberals," the new center-right were "free market-oriented libertarians." In the 1996 electoral campaign, the center-right's appeal was its power to enact innovative approaches to Italy's economic problems, using dynamic entrepreneurial strategies to reward individual initiative. By contrast, the center-left campaigned around issues of broadening people's rights, increasing well-being and quality of life, along with promoting full employment, health care, social security, and education. The voters' primary concerns were channeled around issues of high taxes and unemployment, according to our surveys (Caprara, Calo', & Barbaranelli, 1997). Personality Profiles Predicting Political Party Preferences We believe that despite the substantial overlap between these political powers on many dimensions, the central discriminating features of their political profiles could be mapped onto the personality taxonomy provided by the FFM. Our exploratory hypothesis was that adult voters who chose the center-right political party in the recent Italian elections would be, on average, those highest on the personality factor of Energy, but low on Agreeableness. In contrast, center-left voters would reveal the opposite pattern of dominant traits, with Agreeableness being most prominent. They could also be expected to be high on Openness to Experiences or Culture (or Intellect), given the traditionally greater involvement among "intellectuals" and "intelligentsia" with more Leftist political philosophy. 180 Caprara, Barbaranelli and Zimbardo Emotional Stability should be equally distributed across both political orientations, thus being an irrelevant personality dimension in their differentiation. It was difficult to predict which political orientation would be marked by greater Conscientiousness. It was part of the propaganda of the center-right, asserting that only they had the energy, vision, and also the persistence needed to lead Italy in new directions. However, Conscientiousness could also be an attribute of those on the center-left, if the concept is interpreted as being reliable in its commitment to people, and following through on their promises of a better quality of life in Italy. By statistically controlling for demographic variables such as age, gender, and education, we hoped to ascertain whether these variables had any effect on political orientation of our examined population, as found earlier with national data showing slight preferences of youth and women for the center-right coalition (Calvi & Vannucci, 1995). We also wanted to determine whether any interactions between personality and demographic variables were significant. Finally, we intended to assess the differential impact of personality and relevant demographic variables on political choice. Our predictions were developed from an analysis of the actual contents of political programs and propaganda presented by the center-left and center-right coalitions during the 1996 national election campaign. Previous research by Di Renzo (1963) found that Italian left-oriented politicians tended to be more open-minded than Italian right-oriented politicians. However, these results are more than 30 years old, reflecting a political situation very different from the one currently functioning in Italy. More recent studies have examined different populations as well as different contexts for relating Openness to political views. Trapnell (1994) and McCrae (1996) demonstrated that, at least in Western societies, the more people are open-minded the more they are politically left-oriented, while the less they are open-minded the more they are politically right-oriented. McCrae noted that "variations in Openness are the major psychological determinant of political polarities" (1996, p. 325) because "openness predisposes individuals toward liberal political views" (p. 327). Our predictions are also based on assumptions about the ways in which individuals' dominant personality traits guide their perceptions of media messages, as well as their decisions about the kind of experiences and people with which they will become involved (in this regard, see Driscoll, Hamilton, & Sorrentino, 1991; Shaller, Boyd, Yohannes, & O'Brien, 1995; Webster & Kruglanski, 1994). We reasoned that "energetic-dominant" people should be more attracted by leaders and parties strongly emphasizing individualism and self-ownership (center-right), whereas "agreeable-friendly" people should find more congenial political agendas emphasizing solidarity and collective well-being (center-left). Those open to new ideas and experiences should be more sympathetic to political programs emphasizing education, multiculturalism, and tolerance of diversity (center-left). Personality Profiles and Political Parties 181 Thus, the present research examined the interplay between the "public persona" of these two new political coalitions, in terms of their electoral rhetoric and image construction, and the "private persona" of the personality profiles of voters who chose to affiliate with one or the other coalition. We should add that the match between voter personalities and political party orientation is also influenced by the personalities of political leaders, or more accurately, by voter perceptions of the candidates' personalities (see Jones & Hudson, 1996). Those perceptions may be veridical, or "constructed" and "managed" by the campaign staff or public relations units of political parties. The present study did not go beyond investigating the first case of the association between personality of voters and political orientation. Method A large number of voters, self-identified as having expressed their preference in the 1994 Italian election for either the center-right or center-left coalitions, were individually administered a personality inventory that generates personality profiles encompassing the five factors of the FFM. Additional demographic data were collected that might be related to party preference, personality, or their linkages; these data were statistically partialed out in subsequent analyses. Participants Participants were recruited by about 400 psychology majors as part of a course assignment in Personality Psychology at the University of Rome. Each student, acting as a research assistant, was briefed on the general aims of the research, instructed in how to administer the personality inventory (the Big Five Questionnaire, BFQ), and requested to collect six inventories. They were also required to collect data that would be equally distributed by political party choice, gender, and age. Educational and occupational data were also collected to establish the diversity and representativeness of this sample of voters. The total number of participants was 2,442. From this sample, 158 respondents were excluded because they voted for a political coalition different from the two main coalitions about which we are making predictions (i.e., center-left and center-right), 195 more were excluded because they did not vote for any coalition, and 59 were excluded because they failed to report their past voting behavior. Of the remaining 2,030 participants, nearly half represented each political coalition, with 1,020 from the center-right and 1,010 from the center-left. Overall, the sample was middle-aged (mean of 43.5 years, but with a large standard deviation of 17.2), with gender balanced (50.3% males, 49.6% females). Table I presents the demographic characteristics of our sample categorized by gender, occupation, and education, according to party affiliation. Our sample, while quite large, is a convenience sample and not a probabilistic, random sample. Nevertheless, any possible biasing effects due to differential distribution of key demographic 182 Caprara, Barbaranelli and Zimbardo Table I. Demographic Characteristics of the Sample Gender Center-Ri ght Center-Left N % N % Males 538 52.85 482 47.86 Females 480 47.15 525 52.14 Total 1,018 1,007 ition Professional 146 14.50 184 18.47 Technician 90 8.94 67 6.73 Skilled worker 17 1.69 27 2.71 Unskilled worker 49 4.87 45 4.52 Employee 190 18.87 212 21.29 Housewife 128 12.71 105 10.54 Student 219 21.75 225 22.59 Retired 151 15.00 114 11.45 Unemployed 17 1.69 17 1.71 Total 1,007 996 Education Elementary school 110 10.87 95 9.48 Junior high school 175 17.29 152 15.17 Senior high school 608 60.08 572 57.09 College/university 119 11.76 183 18.26 Total 1,012 1,002 characteristics (such as gender, age, education) were controlled for statistically, as explained below. Although the demographic features of the typical voter for each coalition are surprisingly comparable, several statistically significant differences emerged. Relative to center-left voters, center-right voters were slightly older [means of 42.6 vs. 40.8 years, respectively; F(l, 2025) = 4.46, p < .05], with a slight prevalence of females [%2(1) = 5.03, p < .05], more housewives and retired persons, but fewer professionals [%2(9) = 25.80, p < .01]. The center-left drew a greater percentage of college and university graduates [%2(3) = 117.33, p < .001], as we had anticipated. Measures Personality assessment. All participants completed the BFQ (Caprara, Barbaranelli, Borgogni, & Perugini, 1993; Caprara et al., 1994). This questionnaire was developed on the basis of findings derived from the first lexical study on Italian Personality Profiles and Political Parties 183 personality adjectives. The BFQ was developed using a rational-based or "top down" approach (Burisch, 1984; Comrey, 1988): Once the Big Five were identified as the higher-order, most recurrent factors of personality, facets or subdimensions were identified from scanning the pertinent literature, and phrase-items were produced for measuring these constructs. In comparison to other questionnaires for assessing the FFM (e.g., NEO-PI-R; Costa & McCrae, 1992), the BFQ was developed with the following goals: (a) to be more parsimonious in the number of facets that referred to each primary dimension, and in the number of phrases produced; (b) to be, as far as possible, coherent with the definition of the five factors and of their facets, as they are referred to in the literature; and (c) to provide a measure of social desirability by means of a Lie (L) scale. The BFQ contains five domain scales and 10 "facet" scales, plus a Lie scale designed for measuring the Social Desirability response set and the tendency to distort meanings of the scores. For each of the 132 items in the questionnaire, respondents indicate the extent to which they agree or disagree with the personal relevance of the item on a 5-point scale (1, very false for me; 5, very true for me). Table II presents short definitions of the domain and facet scales in the BFQ. Table II. Definition of BFQ Domains and Facets Global Domains Facets Energy: Level of activity, vigor, sociability, talkativeness, need to excel, persuasiveness, competitiveness Friendliness: Concern and sensitiveness Dynamism: Activity and enthusiasm Dominance: Assertiveness and self- Cooperativeness: Altruism, empathy, generosity, unselfishness Politeness: Kindness, civility, docility, and trust Scrupulousness: Dependability, orderliness, and precision Perseverance: Capability of fulfilling one's own tasks and commitments, tenaciousness, confidence toward others and their needs Conscientiousness: Self-regulation in both its proactive and inhibitory aspects Emotional Stability: Capability of controlling one's emotional reactions; absence of negative affects; psychological adjustment Openness: Broadness of one's own cultural interests; tolerance of diversity; exploration of novelty persistence Emotion Control: Absence of anxiety, depression, and vulnerability; mood stability Impulse Control: Capability of controlling irritation, discontent, and anger Openness to Culture: Intellectual curiosity, interest in being informed, appreciation of culture Openness to Experiences: Openness to novelty; tolerance of values; interest toward diverse people, habits, and life-styles 184 Caprara, Barbaranelli and Zimbardo The psychometric properties of the BFQ have been validated on large samples of Italian respondents (Caprara et al., 1993; Caprara et al., 1994) and in cross-cultural comparisons in five nations (Barbaranelli, Caprara, & Maslach, 1997; Caprara, Barbaranelli, Bermudez, Maslach, & Ruch, 1999). The construct validity of the BFQ scales has been demonstrated by high correlations with analogous scales in the NEO-PI on both Italian and American samples (Barbaranelli et al., 1997; Caprara et al., 1993). The five domain scales of the BFQ correlated substantially with the corresponding domain scales of NEO-PI in two independent samples of Italian and American adults. The convergent correlations were .66, .65, .65, -.80, and .62 for Italians (n = 467), and .75, .74, .63, -.80, and .65 for Americans (n = 203), respectively, for Energy versus Extroversion, Friendliness versus Agreeableness, Conscientiousness, Emotional Stability versus Neuroticism, and Openness versus Openness to Experiences. Although the BFQ is similar in many ways to the NEO-PI, we believe the BFQ has a number of advantages that enhance its research utility (see Caprara et al., 1993). The basic psychometric characteristics of the BFQ were confirmed in this sample. Reliability coefficients (Cronbach's as) of the 10 facet scales ranged from .60 to .82; as of the five domain scales ranged from .73 to .88. Correlations among the 10 facet scales ranged from -.15 to .61; correlations among the five domain scales ranged from .09 to .40. Political party orientation. Political orientation was measured directly by asking participants which coalition they had voted for in the former political election. This is a measure of past voting behavior because the previous Italian political elections took place in 1994, and the research data were collected in 1996. Statistical Analyses The aims of the statistical analyses were twofold. First, we wanted to examine the profiles of the two groups and test for significant differences in personality as related to political affiliation, when the impact of demographic variables on personality was held constant. Second, we wanted to examine the differential impact of personality and demographic characteristics on political affiliation. In both cases, gender, age, and education were taken into account because of their expected relationship with political orientation, and also with personality. Although the traditional impact of these variables on personality is well known, it is worth noting that Italian females tended to be more conservative than males, and that the left voters tended to be younger and better educated in our sample. Differences among our two voter classifications on each of the five domain scales and on the 10 facet scales of the BFQ were examined with a multivariate analysis of covariance (MANCOVA). Gender (males vs. females) and political orientation (center-left vs. center-right) were the design variables, with scores on the five BFQ domain scales as dependent variables, using age and educational level as covariates. This complex data analysis design was adopted because demographic Personality Profiles and Political Parties 185 differences in the composition of the two samples of voters necessitated various statistical controls to untangle the background status variable effects from our hypothesized relationships between personality and political party preference. The impact of personality and demographic characteristics on political choice was examined with two logistic regressions conducted with a stepwise-forward approach. In both cases, political choice was considered as the dependent variable; in the first regression, the independent variables were the five BFQ domain scales and the demographic variables (gender, age, education); in the second regression the independent variables were the 10 BFQ facet scales and the same demographic variables. Results BFQ Domain Scales MANCOVA. The MANCOVA results revealed significant multivariate effects due to gender [F(5, 1939) = 44.86, p < .001], to political orientation [F(5, 1939) = 44.86,p < .001], and to the covariates [F(10, 3876) = 32.95,/? < .001]. However, the interaction among gender and political orientation did not result in significant outcomes in either the multivariate test [F(5, 1939) = .96, p = .43] or any univariate test. Examination of the effect size1 of design variables and covariates indicates that 5% of the variance in personality trait differences is explained by political orientation, 8% by age and education, and 11% by gender. Next, we present separately the significant results due to political preference and to demographic differences in personality traits that emerged on the univariate tests. Personality-political preference effects. As predicted, significant relationships were found between three of the five factors of the FFM and the political orientation of the participants in our investigation. Those voters identified as supporting the center-right were significantly more dominant than the center-left in their BFQ scores on the factor of Energy [F(l, 1942) = 21.64, p < .001; effect size = .013]. In addition, they also showed a slight effect of greater Conscientiousness [F(l, 1942) = 5.58, p < .05], but with a minimal effect size (= .003). We had made no prediction about this personality domain. The interesting contrast is seen in the highly significant show of Friendliness (or Agreeableness) among the center-left voters [F(l, 1942) = 20.07, p< .001; effect size = .011], as well as their greater degree of Openness (or Intellect) [F(l, 1942) = 19.80, p < .001; effect size = .014]. We had not expected to find any differential effects on the dimension of 1 Throughout this article, the measure used to assess effect size was T|2. This coefficient is interpreted as the proportion of the total variability in the dependent variable that is accounted for by variation in the independent variable. It is the ratio of the between-groups sum of squares to the total sum of squares. These effect sizes are relatively modest, although highly significant, revealing the complexity of the many contributing variables in the phenomena under consideration. 186 Caprara, Barbaranelli and Zimbardo Emotional Stability between these two groups of voters, and indeed, we found none [F(l, 1942)= 1.44, p = . 23]. Figure 1 shows the personality profiles of the two voter groups, expressed as T-scores for each of the five domains of the BFQ. It is noteworthy that the differences in personality traits due to political partisanship are still significant after controlling for all of the demographic variables. (An additional ANCOVA combining education and occupation as a composite variable, excluding students, also failed to modify any of the results relating personality and political orientation.) Demographic data differences. There were significant univariate test effects of the two covariates of age and education with several of the personality dimensions measured by the BFQ. Specifically, age was negatively associated with Energy ((3 = -.12, p < .001) and with Openness ((3 = -.18, p < .001). But older participants were more emotionally stable than younger ones, as revealed in the positive association between age and Emotional Stability ((3 = .07, p < .05). Higher levels of education were shown to be positively associated with almost the full range of personality factors, notably Energy ((3 = .07, p < .01), Friendliness ((3 = .11, p < .001), Emotional Stability ((3 = .09,/? < .05), and Openness ((3 = .26, p < .001). (However, as academicians, we are not happy to note the failure of higher education to be associated with greater Conscientiousness, a null effect for which we have no ready explanation.) Gender differences emerged for all scales. Males showed higher scores than females in Energy [F(l, 1942) = 64.04, p < .001; effect size = .033], Conscientiousness [F(l, 1942) = 10.08, p < .01; effect size = .004], Emotional Stability [F(l, 1942) = 97.65,/? < .001; effect size = .054], and Openness [F(l, 1942) = 9.13, p < .05; effect size = .002]. Females were significantly higher than males only in the personality domain of Friendliness [F(l, 1942) = 23.99, p< .001; effect size = .013]. However, none of these effects modified any of the predicted, significant effects found between political party identification and specific personality traits. Logistic regression. None of the demographic variables entered into the final regression equation, which was highly significant [%2(4) = 109.61, p < .001]. The only variables that had significant impact on political preference were four of the BFQ domain scales, namely Energy (r = -.12, p < .001), Friendliness (r = .06, p < .001), Openness (r = .13, p < .001), and Conscientiousness (r = -.05, p < .001). The equation allowed the correct classification of 61.4% of center-left voters and 57.6% of center-right voters, with an overall effectiveness rate of 59.5%. These results are quite surprising because only personality had a significant impact on political preference, whereas none of the demographic variables had an impact on party choice behavior once the personality effects were extracted. 188 Caprara, Barbaranelli and Zimbardo BFQ Facet Scales MANCOVA. The MANCOVA showed significant multivariate effects due to gender [F(10, 1934) = 33.42, p < .001], to political orientation [F(10, 1934) = 9.94,/? < .001], and to the covariates [F(20, 3866) = 26.92, p < .001]. Again, the interaction among gender and political orientation did not result in significant outcomes on either the multivariate test [F(5, 1934) = 1.34, p = .19] or any univariate test. Analysis of the effect size of design variables and covariates showed that 5% of the variance of the BFQ facet scales is explained by political orientation, 11% by age and education, and 14% by gender. Next, we present separately the significant results due to political preference and to demographic differences in personality traits that emerged on the univariate tests. Personality-political preference effects. Significant relationships were found between eight of the 10 BFQ facet scales and the political orientation of the participants in our investigation. Those voters identified as supporting the center-right showed significantly higher scores on the BFQ facet scale of Dominance [F(l, 1942) = 41.50, p < .001; effect size = .02]. They also showed a slight effect of Scrupulousness [F(l, 1942) = 4.02, p < .05; effect size = .002] and of Perseverance [F(l, 1942) = 5.27, p < .05; effect size = .003] relative to those voters whose political preference was center-left. We found that center-left voters showed highly significant scores on Cooperativeness [F(l, 1942) = 17.31, p < .001; effect size = .01], Politeness [F(l, 1942) = 13.73, p < .001; effect size = .01], Impulse Control [F(l, 1942) = 6.72, p < .01; effect size = .00], Openness to Culture [F(l, 1942) = 13.42, p < .001; effect size = .01], and Openness to Experiences [F(l, 1942) = 23.52, p < .001; effect size = .011]. Figure 2 shows the profiles of the two voter groups whose T-scores are arrayed on the 10 BFQ facet scales. Again, it is noteworthy that the differences in personality facets due to political partisanship are still significant after controlling for our demographic variables. Demographic data differences. There were significant univariate test effects of the two covariates of age and education with several of the facet scales of the BFQ. Specifically, age was negatively associated with Dynamism ((3 = -.08,p < .001), Dominance ((3 = -.09, p < .001), Perseverance ((3 = -.11, p < .001), Openness to Culture ((3=-.09,p < .001), and especially Openness to Experiences ((3 = -.29,p < .0001). Older participants were higher than younger voters in Scrupulousness ((3 = .12, p < .001), Emotion Control ((3 = .08, p < .001), and Impulse Control ((3 = .05, p < .05). Higher levels of education were shown to be positively associated with almost the full range of BFQ facets, notably with Dynamism ((3 = .06, p < .01), Cooperativeness ((3 = .12, p < .001), Politeness ((3 = .08, p < .001), Perseverance ((3 = .11, p < .01), Emotion Control ((3 = .08, p < .001), Impulse Control ((3 = .07, p < .01), Openness to Culture ((3 = .25, p < .001), and Openness to Experiences ((3 = .18, p< .001). 190 Caprara, Barbaranelli and Zimbardo Gender differences emerged for all facet scales except Scrupulousness and Openness to Culture. Males showed higher scores than females in Dynamism [F(l, 1942) = 4.19, p < .05; effect size = .002], Dominance [F(l, 1942) = 127.22, p < .001; effect size = .058], Perseverance [F(l, 1942) = 8.46, p < .01; effect size = .004], Emotion Control [F(l, 1942) = 185.40, p < .001; effect size = .083], Impulse Control [F(l, 1942) = 26.30, p < .001; effect size = .013], and Openness to Experiences [F(l, 1942) = 7.32, p < .01; effect size = .004]. Females were significantly higher than males on the two facet scales of Cooperativeness [F(l, 1942) = 18.41, p < .001; effect size = .009] and Politeness [F(l, 1942) = 25.39, p<. 001; effect size =.012]. Logistic regression. As was found with the domain scales, none of the demographic variables entered into the final regression equation for the facet scales, which was significant [%2(5) = 113.24, p < .001]. The only variables that had significant impact on political preference were five of the BFQ facet scales: Dominance (r = -.12, p < .001), Cooperativeness (r = .03, p = .05), Perseverance (r = -.03, p = .02), Openness to Culture (r = .04, p < .01), and Openness to Experiences (r = .09, p < .001). The equation allowed the correct classification of 61.4% of center-left voters and 56.8% of center-right voters, with an overall effectiveness rate of 59.4%. These results clearly support the generalization derived from our analysis of the BFQ domain scales: Only personality had a significant impact on political preference, whereas none of the demographic variables modified in any way this significant personality-politics linkage. Discussion and Conclusions Across a large, diverse sample of Italian voters, specific personality profiles were predicted and found to be associated significantly with preferences for either of two contemporary political coalitions. These new political coalitions are composed of heterogeneous arrays of former political adversaries functioning as expedient, pragmatic electoral entities. Despite considerable overlap in the demographic structure of supporters of both coalitions, those that endorsed the platform of the center-right coalition were characterized as especially high on the Energy personality dimension of the FFM, slightly positive on Conscientiousness, but with low trait scores on Friendliness and Openness. Exactly the opposite personality profile characterized those citizens who preferred the center-left coalition, with high degrees of Friendliness (Agreeableness) and Openness. The fifth factor of Emotional Stability played no role in political party preference, as we had expected given its irrelevance to any aspect of the ideology, leadership style, party platform, or propaganda of either coalition. Our analyses at the facet level allow further specification of the differences among the two groups of voters. Both facets of Friendliness and both facets of Openness differentiated between the two groups, as did the Dominance facet of Personality Profiles and Political Parties 191 Energy, whereas the level of the other Energy facet (Dynamism) was almost equal in the two groups. In particular, the BFQ Dominance items that most differentiated the two groups were: "I'm willing to apply myself to the very end just to excel," "I'm always sure of myself," "Nothing is obtained in life without being competitive" (where center-right voters were higher than their rivals), and "I don't like work environments where there's a lot of competition" (where center-left voters outperformed center-right voters). We highlight again the fact that these relationships between personality traits and political party identification were independent of any apparent influences of age, gender, or education, when they were statistically controlled. However, one may question whether the BFQ Openness scale is just another index of political ideology; if so, its value as a correlate of political affiliation would be compromised. We can disentangle this potential confounding by examining the content of the BFQ Openness items that differentiated the two groups. They are: "I prefer to read rather than engage in a sports activity" (where center-left voters outperformed their rivals), "I'm not interested in television programs which are too serious," "I don't devote much time to reading," "I don't think knowing history serves much," "I don't waste time acquiring knowledge that's not strictly related to my field of interest," "Life-styles and customs of other peoples have never interested me," and "I don't know what pushes people to behave differently from the norm" (where center-right voters outperformed their rivals). With the exception of the last item, none of these items refers to tolerance of values, politics, or liberalism. It is, therefore, reasonable to construe the BFQ Openness scale as indexing something other than political orientation or political ideology. This result provides further support for our conclusion that new personality assessment instruments, such as the BFQ, deserve wider use in facilitating the systematic exploration of linkages between personality variables and political choices. We encourage their further use in exploring the rich intermediate process-level of social-cognitive dynamics, such as party propaganda, candidate image design, and the schema, perceptions, and persuasability of voters. The remainder of this paper outlines some constraints on our conclusions in light of recent criticisms of the limitations of the FFM and the nonrandom nature of our sample. We then present some suggestions for rethinking the personality-politics linkages in terms of more dynamic, multidirectional, bicausal models. Finally, we consider ways to widen the scope of future investigations by using new analyses of voters' perceptions of the personality of political candidates, along with a cost-benefit analysis of information-gathering and retention by voters of political parties' persuasive communications and media messages designed to focus or bias voting decisions. 192 Caprara, Barbaranelli and Zimbardo Enhancing the Utility of the FFM Recent reviews of the FFM have raised both specific and general criticisms about its validity and usefulness in personality research. Critics have noted some limitations, such as overgeneralizability, "folk psychology" development status, lack of truly orthogonal factors, different trait names for the five factors in different measuring systems, and conceptualization that defies disconfirmation.2 Counter-replies have challenged some of these criticisms while incorporating some of the cogent insights, where possible, into more coherent conceptions of this new "human compass" that attempts to map individual differences in personality onto a common reference structure (see Barrick & Mount, 1991; Briggs, 1992; Caprara, 1996; McCrae & John, 1992; Ozer & Reise, 1994). It is beyond the scope of this article to deal with this substantial literature, rather recognizing the potential constructive value of the current heated debate between advocates and adversaries of this approach to personality assessment. Readers are reminded of the limitations posed by the "convenience" sample used here for its inexpensive utility rather than the more costly, but scientifically appropriate, random administration of the personality scales and determination of voting behavior. In defense, we can only point to the large size of the sample, its diversity in terms of age, education, and occupation, and its gender balance, along with the results showing that some of these demographic variables "behaved" in line with prior data collected from random surveys of the Italian electorate in 1994 (see Calvi & Vannucci, 1995). We should also mention a recent empirical failure of personality factors to be related to political orientation (Mehrabian, 1996). The only significant effects found were a positive correlation between Conservatism and Conscientiousness and a negative correlation of Conservatism with Intellect, which fit in general with our reported findings. However, two features of that research limit its validity and generalizability, namely small sample sizes (fewer than 100 respondents in any of the studies) and indirect measurement of political orientation by self-report scales of Conservatism and Libertarianism rather than by actual political party affiliation of voters, as in the present study. Personality, Beliefs, Persuasion, and Schema Activation Our research is based on a relatively simple model in which voter personality traits and political party preferences are correlated, either because particular personality traits guide the selection by individuals of certain kinds of experiences, or because political party values and ideologies select personalities from the general distribution as followers or "true believers." However, we are aware that 2 Readers are referred to the following critiques for a fuller appreciation of the important conceptual and empirical issues they pose: Block (1995), McAdams (1992), Pervin (1994), and Tellegen (1993). Personality Profiles and Political Parties 193 in reality only a multicausal, feedback model can begin to capture the dynamic interaction among the key variables and catalytic processes operating in the contemporary political arena of democratic countries. We propose further that such analyses must also include multiple levels, including intrapsychic, social, and systems levels of variables and processes. For example, it is important to understand the set of dynamics in the initial organization of a political party or coalition because the party selects political candidates who have particular personality traits they deem desirable to voters. Party propaganda and media-controlled information dissemination help to create, construct, or modify the personality images of candidates in the voters' minds. Voters try to get optimal information about the current political scene at minimal information-processing cost, relying on well-defined schemata to provide simplistic, heuristic short cuts to establishing their preferences. But voter personalities, beliefs, and values also bias the ways in which they process available political information and their sense of match/mismatch with particular political candidates' personalities and ideologies (Tetlock, 1983). During an election campaign, some of these variables will be shifting and modified by feedback from polls of party/candidate popularity (see Crewe & King, 1994). Thus, even this cursory overview underscores the necessity of raising the general level of complexity of future investigations into the fascinating dynamics at work in the arena of political psychology (see Bean & Mughan, 1989). Future researchers can add to our fuller appreciation of these complex transactions by taking account first of the voter's sense of identity and concerns for presenting a desirable image to others (see Funder, Kolar, & Blackman, 1995). In addition, the role of personality disposition functions simultaneously with the belief and value systems maintained by voters. Such systems are part of the motivational-cognitive network that directs information acquisition, integration, and retrieval about political parties and political leaders (Fiske & Taylor, 1991; Greenwald, 1980). Those processes should also influence voter perception of the personality of political leaders, which in turn fosters either greater identification and convergence or mismatches with them (see Simonton, 1990). "Perceptions of leadership quality depend upon personality traits" of leaders as judged by followers, according to Jones and Hudson's recent analysis (1996, p. 229). They argued that leaders must present the public with a set of traits contributing to the belief that they can lead in a "businesslike" fashion. Public ratings of leaders are affected by changes in their perceived personality traits, as is the party's electoral support (Crewe & King, 1994). The modern media's role in presenting, and even creating, political images cannot be overstated. Thus, political parties spend enormous amounts of money on image manipulation that typically shows their candidate as effective and energetic and/or sympathetic, friendly, and willing to listen to the needs of the voters. Our previous research has shown that voters in Italy and the United States simplify their personality judgments of the major political candidates in ongoing 194 Caprara, Barbaranelli and Zimbardo election campaigns by restricting the usual five factors (which they used for self-rating and ratings of nonpolitical public comparison figures) to a combination of only two or three factors (Caprara, Barbaranelli, & Vicino, in press; Caprara, Barbaranelli, & Zimbardo, 1997). These collapsed, simplified "politician's factors" are Energy/Innovation (blending Energy and Openness) and Honesty/Trustworthiness (blending Agreeableness, Conscientiousness, and Emotional Stability). Such simplified personality perceptions of political candidates may derive from a cognitively efficient strategy that voters adopt non-consciously to code the mass of complex information bombarding them daily during a political election, and to guide their eventually dichotomous decision about voting for or against particular candidates. Because center-right voters were more energetic and center-left voters were more friendly, there was a complementary matching process among the voters' personalities and the political leaders' personalities in the recent Italian national election. Information Costs and Communication Processing Another way to think about the linkages between voters' personalities and beliefs and politicians' personalities is in terms of schematic information processing that is fostered by political party (or coalition) propaganda and advertisements. The costs to voters of gathering and meaningfully organizing information relevant to their voting decision are reduced by forming well-defined, simplifying schemata of both leaders and parties (Jones & Hudson, 1997). The political parties reduce the transaction costs of electoral participation by sending out low-cost signals to voters, so that mere party affiliation alone provides considerable information about the candidates' position on the political spectrum. Moreover, these parties will have already chosen leaders with particular personality trait patterns, believed to be appealing and desirable to their intended constituency (Winter, 1987). So voters are really encouraged to engage in simplified heuristic, or peripheral, information processing by using well-defined schematic representations that undercut more complex, systematic information processing (Petty & Cacioppo, 1986; Zimbardo & Leippe, 1994). Research is needed that analyzes the nature of the persuasive messages political parties generate in terms of their "personality" appeals to voter personalities and candidate personalities. We hope that our findings, along with the conceptual analyses sketched above, will rekindle the interest of psychologists from the fields of personality, social psychology, and cognitive psychology in the nature of the contributions they can make collectively to the broad realm of political behavior. As social scientists, it is imperative that we better understand and unravel the complexities in these vital transactions between political parties, leaders, voters, and the mediating and situational processes that interrelate them. As citizens, we also need to become better informed about how to mindfully cast our votes for politicians and parties on the basis of systematic analyses of their platform and ideological values, rather Personality Profiles and Political Parties 195 than engaging in simplistic peripheral processing of media-created images of political candidates' personalities. AUTHOR'S ADDRESS Correspondence concerning this article should be addressed to Gian Vittorio Caprara, Department of Psychology, University of Rome "La Sapienza," Via dei Marsi 78, 00185 Rome, Italy. E-mail: caprara@axrma.uniromal.it REFERENCES Adomo, T., Frenkel-Branswik, E., Levinson, D., & Sanford, R. (1950). The authoritarian personality. New York: Harper. Bandura, A. (1997). Self efficacy: The exercise of control. New York: Freeman. Barbaranelli, C, Caprara, G. V., & Maslach, C. (1997). Individuation and the five factor model of personality traits. European Journal of Psychological Assessment, 13, 75-84. Barrick, M. R., & Mount, M. K. (1991). The big five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44, 1-26. Bean, C, & Mughan, A. (1989). Leadership effects in parliamentary elections in Australia and Britain. American Political Science Review, 83, 1165-1180. Block, J. (1995). A contrarian view of the five-factor approach to personality description. Psychological Bulletin, 117, 187-215. Brewster-Smith, M. (1968). A map for the analysis of personality and politics. Journal of Social Issues, 24, 15-28. Briggs, S. (1992). Assessing the five factor model of personality description. Journal of Personality, 60, 254-293. Browning, R., & Jacob, H. (1964). Power motivation and the political personality. Public Opinion Quarterly, 28, 75-90. Burisch, M. (1984). Approaches to personality inventory construction: A comparison of merits. American Psychologist, 35, 863-872. Calvi, G., & Vannucci, A. (1995). L'elettore sconosciuto [The unknown voter]. Bologna, Italy: II Mulino. Caprara, G. V. (1996). Structures and processes in personality psychology. European Psychologist, 1, 14-26. Caprara, G. V., Barbaranelli, C, Bermudez, J., Maslach, C, & Ruch, W. (1999). Multivariate methods for the comparison of factor structures in cross-cultural research: An illustration with the Big Five Questionnaire. Manuscript submitted for publication. Caprara, G. V., Barbaranelli, C, Borgogni, L., & Perugini, M. (1993). The Big Five Questionnaire: A new questionnaire for the measurement of the five factor model. Personality and Individual Differences, 15, 281-288. Caprara, G. V., Barbaranelli, C, & Livi, S. (1994). Mapping personality dimensions in the big five model. European Journal of Applied Psychology, 44, 9-16. Caprara, G. V., Barbaranelli, C, & Vicino, S. (in press). Personalita' epolitica [Personality and politics]. Giornale Italiano di Psicologia. Caprara, G. V., Barbaranelli, C, & Zimbardo, P. G. (1997). Politicians' uniquely simple personalities. Nature, 385, 493. 196 Caprara, Barbaranelli and Zimbardo Caprara, G. V., Calo', G., & Barbaranelli, C. (1997). Le determinant! dell'intenzione di voto: Un contributo empirico [The determinants of voting intentions: An empirical contribution]. Micro e Macro Marketing, 6, 273-301. Christie, R., & Jahoda, M. (Eds.) (1954). Studies on the scope and method of the authoritarian personality. Glencoe, IL: Free Press. Comrey, A. L. (1988). Factor-analytic methods of scale development in personality and clinical psychology. Journal of Consulting and Clinical Psychology, 56, 754-761. Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources. Crewe, I., & King, A. (1994). Did Major win? Did Kinnock lose? In A. Heath, R. Jowell, J. Curtice, & B. Taylor (Eds.), Labour's last chance? The 1992 election and beyond (pp. 125-147). Aldershot, UK: Dartmouth. Digman, J. M. (1990). Personality structure: Emergence of the five factor model. Annual Review of Psychology, 41, 417^140. Di Renzo, G. (1963). Personalita' e potere politico [Personality and political leadership]. Bologna, Italy: II Mulino. Di Renzo, G. (Ed.) (1974). Personality and politics. New York: Doubleday. Driscoll, D. M., Hamilton, D. L., & Sorrentino, R. M. (1991). Uncertainty orientation and recall of person-descriptive information. Personality and Social Psychology Bulletin, 17, 494-500. Eysenck, H. (1954). The psychology of politics. London: Routledge and Kegan Paul. Fiske, S., & Taylor, S. (1991). Social cognition. New York: McGraw-Hill. Funder, D. C, Kolar, D. C, & Blackman, M. C. (1995). Agreement among judges of personality: Interpersonal relations, similarity, and acquaintanceship. Journal of Personality and Social Psychology, 69, 656-672. Greenstein, F., & Lerner, M. (1969). A source book for the study of personality and politics. Chicago: Markham. Greenwald, A. (1980). The totalitarian ego: Fabrication and revision of personal history. American Psychologist, 35, 603-618. Hermann, M. (Ed.) (1977). The psychological examination of political leaders. New York: Free Press. Hermann, M. (1986). Political psychology. San Francisco: Jossey-Bass. John, O. P. (1990). The "big five" factor taxonomy: Dimensions of personality in natural language and in questionnaires. In L.A. Pervin (Ed.), Handbook of personality: Theory and research (pp. 66-100). New York: Guilford. Jones, P., & Hudson, J. (1996). The quality of political leadership: A case study of John Major. British Journal of Political Science, 26, 229-244. Jones, P., & Hudson, J. (1997). Why do political parties (political firms) exist? Unpublished manuscript, University of Bath, England. Knutson, J. (1973). Handbook of political psychology. San Francisco: Jossey-Bass. McAdams, D. P. (1992). The five factor model in personality: A critical appraisal. Journal of Personality, 60, 329-361. McClosky, H. (1958). Conservatism and personality. American Political Science Review, 52, 27^15. McCrae, R. R. (1989). Why I advocate the five factor model: Joint analysis of the NEO-PI and other instruments. In D. M. Buss, & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 237-245). New York: Springer Verlag. McCrae, R. R. (1996). Social consequences of experiential openness. Psychological Bulletin, 120, 323-337. McCrae, R. R., & John, O. P. (1992). An introduction to the five factor model and its applications. Journal of Personality, 60, 175-215. Personality Profiles and Political Parties 197 Mehrabian, A. (1996). Relations among political attitudes, personality and psychopathology assessed with new measures of Libertarianism and Conservatism. Basic and Applied Social Psychology, 78,469-491. Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of personality: Reconceptualizing situations, dispositions, dynamics and invariance in personality structure. Psychological Review, 102, 246-268. Ostendorf, F., & Angleitner, A. (1992). On the generality and comprehensiveness of the five-factor model of personality: Evidence of five robust factors in questionnaire data. In G. V. Caprara, & G. Van Heck (Eds.), Modern personality psychology (pp. 73-109). London: Harvester Wheatsheaf. Ozer, D. J., & Reise, S. P. (1994). Personality assessment. Annual Review of Psychology, 45, 357-388. Pervin, L. A. (1994). A critical analysis of current trait theory. Psychological Inquiry, 5, 103-113. Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: Springer-Verlag. Rokeach, M. (1960). The open and closed mind: Investigations into the nature of belief systems and personality systems. New York: Basic Books. Seeman, M. (1959). On the meaning of alienation. American Sociological Review, 24, 783-791. Shaller, M., Boyd, C, Yohannes, J., & O'Brien, M. (1995). The prejudiced personality revised: Personal need for structure and formation of erroneous group stereotypes. Journal of Personality and Social Psychology, 68, 544-555. Simonton, D. (1990). Personality and politics. In L. Pervin (Ed.), Handbook of personality (pp. 670-692). New York: Guilford. Srole, L. (1965). A comment on "Anomy." American Sociological Review, 30, 757-762. Tellegen, A. (1993). Folk concepts and psychological concepts of personality and personality disorder. Psychological Inquiry, 4, 122-130. Tetlock, P. (1983). Cognitive style and political ideology. Journal of Personality and Social Psychology, 45, 118-126. Trapnell, P. D. (1994). Openness versus intellect: A lexical left turn. European Journal of Personality, 8, 273-290. Webster, D. M., & Kraglanski, A. W. (1994). Individual differences in need for cognitive closure. Journal of Personality and Social Psychology, 67, 1049-1062. Wiggins, J. S. (Ed.) (1996). The five-factor model of personality: Theoretical perspectives. New York: Guilford. Winter, D. (1973). The power motive. New York: The Free Press. Winter, D. (1987). Leader appeal, leader performance, and the motive profiles of leaders and followers: A study of American presidents and elections. Journal of Personality and Social Psychology, 52, 196-202. Zimbardo, P. G., Ebbesen, E. B., & Maslach, C. (1977). Influencing attitudes and changing behavior (2nd ed.). Reading, MA: Addison-Wesley. Zimbardo, P. G., & Leippe, M. (1994). The psychology of attitude change and social influence. New York: McGraw-Hill.