HUMAN RESOURCE ASSIGNMENT

topic_6_art_5.pdf

Home >Human Resource Management homework help >HUMAN RESOURCE ASSIGNMENT

Research article

Looking for performance in personality inventories: The primacy of evaluative information over descriptive traits

SYLVAIN CARUANA1,2*, RÉGIS LEFEUVRE1 AND PATRICK MOLLARET1 1Cognition, Health & Socialization Laboratory (C2S), University of Reims Champagne-Ardenne, Reims, France; 2CDE Consultants, Reims, France

Abstract

Three experiments were designed to demonstrate that job performance inferences from personality inventories rely more on the agentic or communal value conveyed by the items compared with the Big-Five traits they are supposed to describe. In the first two experiments, the participants had to predict the job performances of fictitious job applicants based on their responses to a personality inventory. In Experiment 1, the information on personality was held constant, such that the applicants’ responses varied solely on their agentic, communal, or purely descriptive orientation. In Experiment 2, the social value of the responses again varied as well as the information about the applicants’ personality (agreeable vs. conscientious). The results showed that the agentic profiles were the most predictive of the performance, regardless of the personality factors. In Experiment 3, we reversed the procedure. The participants filled out a personality inventory in the place of a more or less successful employee. The results here showed that the information about the performance had the greatest impact on the agentic items, independent of the personality factors measured. These results confirm the relevance of social judgment models in personality research. Copyright © 2014 John Wiley & Sons, Ltd.

Personality inventories are widely used in evaluative settings to recruit individuals and to make predictions about their performances (Kuncel, Ones, & Sackett, 2010; Rothstein & Goffin, 2006). Individuals and professionals are able to infer the performance of individuals from personality profiles (Dunn, Mount, Barrick, & Ones, 1995). Nevertheless, because many items from personality inventories display both evaluative and descriptive informa- tion about individuals (Backström, 2007; Backström, Björklund, & Larsson, 2009), the interpretation of the link between personality scores and the prediction of perfor- mance remains unclear. Broadly speaking, the evaluative part of personality inventories relies on the positive or negative connotation of the items, whereas their descriptive part concerns the behavioral tendency evoked by the items. For example, being efficient and productive at work (a Revised NEO Personality Inventory (NEO PI-R) item) conveys both evaluative information (the item is positive) and descriptive information (the item describes a tendency to behave in a conscientious way). The aim of the present contribution was (1) to dissociate the descriptive informa- tion from the evaluative information, which is confounded in most items from personality inventories, and (2) to show that the evaluative dimension prevails in the inference of performance at work, which is in line with an evaluative approach to social judgment.

THE TWO EVALUATIVE DIMENSIONS

Research has shown that everyday social judgments are constructed around two fundamental evaluative dimensions (Abele & Wojciszke, 2007, 2013; Bi, Ybarra, & Zhao, 2013; Dubois & Beauvois, 2005; Fiske, Cuddy, & Glick, 2007; Fiske, Cuddy, Glick, & Xu, 2002; Judd, James-Hawkins, Yzerbyt, & Kashima, 2005; Woike, Lavezzary, & Barsky, 2001; Ybarra et al., 2008). Communion, the first dimension, explains most of the variance in social judgments (Abele & Wojciszke, 2007; Ybarra, Park, Stanik, & Lee, 2012) and evaluates the degree to which individuals are oriented toward others (Cislak, 2013). Communion is operationalized by adjectives referring to warmth, morality, or sociability, which can communicate a positive evaluation (e.g., warm, sociable, and honest) or a negative one (e.g., cold, aggressive, and dishonest). Agency, the second dimension, refers to the promotion of the self and dominance. Agentic traits are defined as self-profitable traits (Peeters, 1992), meaning that they are more useful for the self than for others. Beyond the usefulness for the self in daily interactions, it has been proposed that agentic behaviors are the most adaptive for use in professional life (Beauvois & Dubois, 2009; Dubois & Beauvois, 2005). Indeed, as shown by Dubois (2010), individuals with the most valued occupations are positively judged on agentic traits (e.g., dynamic, active, and competent). Conversely, an individual

*Correspondence to: Sylvain Caruana, Laboratoire C2S (Cognition, Santé, Socialisation), Université de Reims Champagne-Ardenne, UFR Lettres et Sciences Humaines, 57, rue Pierre Taittinger, 51096 Reims Cedex, France. E-mail: [email protected]

European Journal of Social Psychology, Eur. J. Soc. Psychol. 44, 622–635 (2014) Published online 26 June 2014 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ejsp.2034

judged negatively on agency (e.g., defeatist, lazy, or naive) is likely to be associated with the lowest social status. Judgments on communion or agency are not deemed to describe how individuals or groups really are or their genuine psychological tendencies. Instead, they may reflect the system of intergroup relations and rationalize stereotypes (Fiske et al., 2002), justify the inequality of economic resources by ascribing the positive agentic traits to affluent individuals (Oldmeadow & Fiske, 2007, 2010), and communicate the socio-economic status of individuals (or even animals; see Dubois & Beauvois, 2011).

THE FIVE DESCRIPTIVE DIMENSIONS

Personality assessment is grounded on a different research tradition, which identifies the major dimensions on which indi- viduals differ from each other in terms of general personality tendencies. The Five-Factor Model (FFM) currently dominates the field of personality research. This model is considered to be a predictive and universal summary of personality along five dimensions (Conscientiousness, Extraversion, Agreeable- ness, Openness, and Emotional stability) that are assumed to be endogenous and stable across time (McCrae & Costa, 2006). These five dimensions are measured with several focused behavioral descriptions and tendencies, which operationalize each personality factor (e.g., “I am a very active person” operationalizes Extraversion). The items can take the form of an attitude (e.g., “I’m efficient and productive at work”), a behavioral frequency (e.g., “I often try new and for- eign foods”), or a propensity to act in a particular way (e.g., “I laugh easily” and “being active”). Respondents are instructed to sincerely indicate the degree to which each item is a good descriptor of themselves. The responses are then computed in a global score for each personality factor. Nevertheless, it appears that the more valorized items offer the opportunity to present oneself in a favorable way (Backström et al., 2009; Backström, Björklund, & Larsson, 2011) and to make a good impression regarding performance abilities.

DESCRIPTIVE VERSUS EVALUATIVE APPROACH TO PERFORMANCE PREDICTION

Personality inventories that measure the FFM (e.g., Alter Ego, Caprara, Barbaranelli, Borgogni, & Perugini, 1993; NEO PI-R, Costa & McCrae, 1992) have been widely used to predict professional performance (Barrick & Mount, 1991, 2005; Barrick, Mount, & Judge, 2001; Hough, Eaton, Dunnette, Kamp, & McCloy, 1990; Judge, Higgins, Thoresen, & Barrick, 1999). Conscientiousness and Emotional stability appear to predict professional performance regardless of the nature of the job (Barrick & Mount, 1991, 1993; Barrick et al., 2001; Chiaburu, Oh, Berry, Li, & Gardner, 2011; Judge et al., 1999; Kuncel et al., 2010; Salgado, 1997), whereas Extraversion and Openness predict performance only for specific occupations, such as sales representative performance or training proficiency, respectively (Barrick & Mount, 1991; Salgado, 1997). In con- trast, Agreeableness is not directly linked to performance (Barrick & Mount, 1993; Barrick et al., 2001; Judge et al.,

1999; Kuncel et al., 2010). It becomes predictive only in very specific cases where interdependence among co-workers is absolutely decisive (Mount, Barrick, & Stewart, 1998). This descriptive approach to performance prediction dominates the field of research.

Few studies have investigated the role of the two evaluative dimensions in the inference of job performance. In a study by Abele (2003), correlation and regression analyses showed a reciprocal link between agency and career success. Individuals reporting higher agency also reported higher career success; the opposite was also true. In another experiment, Abele, Rupprecht, and Wojciszke (2008) administered arbitrary feedbacks of success versus failure to participants engaged in performance tasks. Independent of their actual performance, the participants who believed that they succeeded in the task considered themselves to be more agentic than those who failed. Importantly, the communal dimension was unaffected by the feedback manipulation. In fact, agency self-evaluations are de- termined by an individual’s perception of success (see also Wojciszke, Baryla, Parzuchowski, Szymkow, & Abele, 2011; Wojciszke & Sobiczewska, 2013). Agentism self-attributions are also affected by status. For example, Moskowitz, Suh, and Desaulniers (1994) showed that individuals who experienced a supervisor role described themselves as more agentic than individuals who experienced a supervisee role. Hence, self- descriptions seem to be influenced by the social context and the nature of the tasks. Additionally, it appears that the evaluation of others on agency and communion is partly determined by the social context and the nature of the task and the relationships (Cottrell, Neuberg, & Li, 2007). While everyday relationships are evaluated on the communal dimen- sion (Abele & Bruckmüller, 2011), agency takes primacy in more evaluative settings. Abele and Brack (2013) showed that participants searched for agency in others only when they shared goals or exchanged relationships with them, whereas they searched for communion when they were committed to independent goals. In a similar vein, Dubois and Aubert (2010) showed that an agentic target was more chosen than a communal one under professional settings, and conversely, the communal target was more chosen under a friendly con- text. Finally, Cislak (2013) found that having power or status leads to a greater search for agency in subordinates.

These results confirm that agency emerges from performance stakes because of its ability to distinguish productive individuals from unproductive ones in professional settings. In the personal- ity literature, performance and competence (i.e., productivity) are recurrently assessed with scales completed by the manager (e.g., Hogan, Rybicki, Motowidlo, & Borman, 1998; Tsui, Pearce, Porter, & Tripoli, 1997). Because these scales aim to quantify the success versus failure of employees, or at least their ability to succeed or fail in their jobs, only agentic items should be related to performance inferences in recruitment settings.

Importantly, several studies have documented the structural representation of the FFM across the two evaluative dimen- sions. For some authors, agency and communion work as two supra-ordinate evaluative factors of the Big Five (Blackburn, Renwick, Donnelly, & Logan, 2004; Digman, 1997; McCrae et al., 2008) in which conscientiousness ap- pears as the most agentic dimension and agreeableness is the most communal one. However, from an item perspective

Personality, evaluative information, and performance inference 623

analysis, most items from personality inventories may convey evaluative information (Backström et al., 2009, 2011; Wojciszke, 1994), which can be encoded according to both agency and communion (McAdams, Hoffman, Mansfield, & Day, 1996; Woike, Gershkovich, Piorkowski, & Polo, 1999; Woike et al., 2001; Woike & Polo, 2001; Wojciszke, 1994).

Our objective was to demonstrate that the descriptive infor- mation by the FFM is orthogonal to the evaluative informa- tion. We hypothesized that removing the agentic part of the personality dimensions should also exclude performance from personality scores. In other words, we aimed to show that performance is inferred from the evaluative part of the personality scores. As an implication, conscientiousness could appear to be predictive of performance only because of its over-representation of agentic items.

OVERVIEW OF THE PRESENT STUDY

We conducted three experiments designed to show that infer- ences of professional performance rely more on the social value of the items than on the personality factors these items are supposed to operationalize. The experiments were pre- ceded by a pilot study designed to measure the agentic or com- munal value communicated by the items of two well-known personality inventories based on the FFM. The objective was to obtain a new inventory in which each of the five dimensions was illustrated by the same proportion of agentic and commu- nal items. This was, in fact, the only way to orthogonalize evaluative and descriptive information conveyed by the items of the personality inventories.

In Experiments 1 and 2, the participants had to predict task performance and contextual performance (organizational citizenship behavior (OCB)) of a fictitious applicant. The re- sponses of the fictitious applicant to our personality inventory were the only piece of information provided to the participants. In Experiment 1, the responses varied only on agency or communion, whereas the information about the five personality factors was held constant. In Experiment 2, the applicant profiles varied both in social value (agency vs. communion) and in personality (conscientiousness vs. agreeableness). Based on research that showed that the agentic self-descriptions corre- spond most closely to the expectations in an occupational context, we postulated that inferences about task performance would vary according to the level of agency conveyed by the profiles (Experiments 1 and 2) and not according to the descrip- tive personality factors (Experiment 2). Moreover, as the contex- tual performance (organizational citizenship) is associated with altruistic motives and co-worker assistance (Borman & Motowidlo, 1997; Tsui et al., 1997), we expected this variable to be influenced by both the agentic and communal information conveyed by the profiles.

In Experiment 3, the paradigm was reversed: the partici- pants had to infer the personality of a target from information about the target’s high versus low job performance. As all the personality dimensions of our inventory had the same evaluative implications, we expected that a high- performance would convey a strong agentic profile within all the personality factors.

PILOT STUDY

As the social value of the items in the personality inventories had never been measured, the pilot study assessed the agentic and communal dimensions of the NEO PI-R (sample 1) and Alter Ego items (sample 2). Our main objective was to construct a questionnaire that crossed the Big Five with the two evaluative dimensions. To build this questionnaire, we used two well-known personality inventories (the Alter Ego, Caprara et al., 1993, French translation, Caprara, Barbaranelli, & Borgogni, 1997; and the NEO PI-R, Costa & McCrae, 1992, French translation, Costa, McCrae, & Rolland, 1998).

Participants

Sample 1

One hundred and fifty-three participants (85 men, 65 women, 3 non-specified; mean age: 34.8 years; SD: 12.5) were interviewed at their workplace. Our sample was composed of managers and supervisors (26.80%), employees and techni- cians (46.40%), and students (21.57%). The remaining 5.23% were classified as “Other occupations.” Eighty-five participants were asked about the items’ agentic content, and 68 participants were asked about the items’ communal content.

Sample 2

Fifty-two participants (21 men, 31 women; mean age: 33.21 years; SD: 11.1) were interviewed at their workplace. The sample consisted of diversified occupational sectors (sales, accountants, clerical, and industry production workers) and education levels. Twenty-nine participants were asked about the items’ agentic content, and 23 participants were asked about the items’ communal content.

Procedure and Materials

The questionnaire was presented as a study investigating the relevance of personal information that individuals communi- cate in social or professional networks (e.g., Hall, Pennington, & Lueders, in press; Marcus, Machilek, & Schütz, 2006; Utz, 2010). The participants were asked about the evaluative conse- quences of describing themselves with items drawn from the personality inventories, in either an agentic or a communal context. They were instructed to rate a series of statements to create a “personal profile” for either Facebook (a social network that operationalizes the communal context) or Viadeo (a professional network that operationalizes the agentic context). Facebook was briefly described as a network that al- lows users to develop friendly relations with other members. The participants had to rate each item on a 7-point scale. For each item, the question was “Does this particular item give the impression of someone who has the sort of qualities needed to attract a large number of friends on Facebook (+3) or someone who is devoid of those qualities (�3)?” Viadeo was described as a professional network designed to promote professional contacts. In this context, the question was “Does this particular item give the impression of someone who is

624 Sylvain Caruana et al.

likely to succeed in his/her professional life (+3) or someone who is liable to fail (�3)?” All items were taken from the French versions of the NEO PI-R (sample 1) and the Alter Ego (sample 2). The participants in both samples only had to rate the items for one of the two rating contexts (Facebook vs. Viadeo). For the NEO PI-R, only 80 of the 240 items were presented to each participant, to minimize the fatigue effect. Thus, each NEO PI-R item was evaluated by approximately 25 participants. Because of its lower number of items (132), the Alter Ego was integrally presented to each participant.

Result Analyses and Questionnaire Elaboration

The aim of this pilot study was to build a questionnaire that measures the five factors via the same amount of agentic, com- munal, and neutral items. The items were selected following two principles. Using one-sample t-tests, we first compared each item score with the central point of the scale (0). Two thirds of these comparisons were statistically significant for both the agentic and communal dimensions, indicating a posi- tive bias for the major part of the items. Second, we performed independent group t-tests to compare the agency and commu- nion ratings. By crossing these two criteria, we obtained four categories of items: neutral, agentic, communal, and both agentic and communal items. We should first note that for both samples, these four categories of items were not equally represented among the Big-Five factors (sample 1: χ2(12) =49.27, p< .001; sample 2: χ2(12) = 22.92, p < .05). For both questionnaires, conscientiousness was primarily an agentic factor, and agreeableness was mainly a communal factor. To construct a more balanced inventory, we proceeded as follows. For agentic versus communal items, we selected the items with a high rating on only one dimension (agency vs. communion) and which were the most neutral on the other. We also verified that the agentic or communal ratings were significantly (or at least marginally) higher on the dimension of interest. Finally, we selected only the items with the same valences on the personality dimension and the social value. For neutral items, we selected the more neutral items on both agency and communion with no significant difference between agentic and communal ratings. A 30-item questionnaire was developed, which measures each personality dimension with six items (two agentic, two communal, and two neutral; see Appendix 1). The mean scores of agency and communion as a function of the item’s type are reported in Table 1.

This questionnaire allowed us to test the following hypotheses:

(i) At equal scores on the five factors, agentic-oriented profiles should lead to higher performance ratings compared with communion-oriented or neutral profiles (Experiment 1).

(ii) Descriptive information variation should have less impact on performance ratings compared with agentic informa- tion variation (Experiment 2).

(iii) Information about job performance should have a higher im- pact on agentic items compared with communal or neutral ones, independent of the Big-Five factors (Experiment 3).

EXPERIMENT 1

This study was designed to show that the item’s social value is perceived before the Big-Five traits and is sufficient to make job performance inferences via personality items. We elabo- rated three strictly similar personality profiles that varied only on the social value orientation (agentic, communal, or neutral). In line with the social judgment models, we first postulated that the task performance attribution would be higher for the agentic profile compared with the communal or neutral ones. Second, we hypothesized that the contextual performance pre- diction (i.e., OCBs) would be influenced by both the agentic and communal values, leading to a higher contextual perfor- mance for the agentic and communal profiles compared with the neutral one.

Participants

The participants included 74 third-year psychology undergraduates (64 women, 10 men, Mage = 21.96 years, SD = 3.83), who took part in the experiment during a tutorial. This experiment was presented as part of a study on how individuals form impressions about professional abilities. The participants were randomly assigned to one of three different experimental conditions (agency profile, communion profile, or neutral). They were not familiar with the theoretical background and did not receive any remuneration for participating in the study.

Materials and Procedure

The participants had to put themselves in the place of a recruiter and read through a fictitious job applicant’s responses to a personality inventory. They were given 5 minutes to con- sult the response profile and form an overall impression of the applicant. Then, the participants had to predict the applicant’s task performance and contextual performance and indicate how likely the applicant was to possess different agentic or communal personality traits (manipulation check). The order in which the manipulation check and the dependent variables were performed was counterbalanced.

Independent Variable: Profile Type

The fictitious applicant was introduced to the participants via his responses to the personality inventory, which had been constructed following the pilot study. This inventory allowed us to construct three response profiles (agency-oriented, commu- nion-oriented, and neutral; see Appendix 2 for sample items).

The item ratings were calculated so that all the personality factors were scored equally (8/18 for each factor) and were held constant across the three profiles. In contrast, the agency,

Table 1. Agency and communal mean scores as a function of the item’s value orientation (pilot study)

Item’s type

Agency-oriented Neutral Communion-oriented

Agency 1.47 0.16 0.43 Communion 0.36 0.20 1.38

Note: The mean scores can vary from 0 to 3.

Personality, evaluative information, and performance inference 625

communion, and neutral values of the personality profiles sys- tematically varied. In the agency-oriented profile, the agency items were given the most positive or negative ratings (+3 or �3, depending on the positive or negative formulation of the items), whereas the communion and neutral items were given middle ratings (0, 1, �1). For the communion-oriented profile, the communion items were given the most positive or negative ratings (+3, �3), whereas the agency and neutral items were given middle ratings (0, 1, �1). Finally, for the neutral profile, the neutral items were given the highest positive ratings, and the agency and communion items were given moderate ratings. Accordingly, none of the profiles contained a particularly salient personality characteristic, and there was no descriptive difference between the three profiles. Each participant saw only one of the three profiles.

Manipulation Check: Agency and Communion Ratings

To check that the experimental manipulations had the expected effect, we asked the participants to assess the profiles on the basis of 15 agency traits (nine positive: ambitious, competent, motivated, successful, self-confident, hard working, active, inventive, and ingenious; six negative: indecisive, inefficient, slow, negligent, lazy, and modest) and 15 commu- nion traits (nine positive: altruistic, friendly, warm, honest, sensitive, sincere, considerate, indulgent, and temperate; six negative: distant, selfish, uncommunicative, hypocritical, heartless, and arrogant). All traits were also applicable to the Big Five. The trait choice was based on the items contained in the French version of the Big-Five Inventory (Plaisant, Courtois, Réveillère, Mendelsohn, & John, 2010). The participants rated each trait on a 7-point scale ranging from 1 (Don’t agree) to 7 (Completely agree). We computed a communion score (α = .70) and an agency score (α = .82).

Dependent Variables: Task Performance and Organizational Citizenship

Two main types of indicators of job performance are reported and have been linked with personality in the literature: task and contextual performance.

Task performance (objective or other-assessed) is directly linked with the job description. In our study, task performance predictions took the form of ratings on six items taken from the task performance scale developed by Tsui et al. (1997; α = .89). This scale is commonly used to probe aspects, such as the quality (e.g., “Employee’s standards of quality are higher than formal standards on this job”), quantity (e.g., “Employee’s quantity of work is higher than average”), and efficiency (e.g., “The employee’s efficiency is much higher than average”) of the work completed by a co-worker from the point of view of his or her superior.

Contextual performance (e.g., OCBs) measures the employee’s broader behaviors toward the working team, management, and organization. The predictions of the OCB took the form of ratings on eight items (e.g., “The employee expresses opinions honestly when others think differently,” “The employee makes suggestions to improve work procedures,” and “The employee informs management of potentially unproductive

policies and practices”), which were taken from the same authors’ citizenship behavior scale (Tsui et al., 1997; α = .91). For both scales, the participants rated items with the same 7-point scale.

Results

Manipulation Check: Agency or Communion Profile

We subjected the data to an analysis of variance (ANOVA) with the profile type (agency, communion, or neutral) as the independent variable and the agency and communion ratings as the repeated-measures factors. The ANOVA revealed an interaction between the profile type and the perceived value, F(2, 71) = 64.96, p < .001, η2p = .65. Helmert-type contrasts validated the experimental manipulations, showing that the agentic ratings were higher for the agency-oriented profile (M = 6.00, SD = 0.49) than they were for the other two profiles (communion-oriented: M = 4.43, SD = 0.79; neutral: M = 4.37, SD = 0.82), F(2, 71) = 48.89, p < .001, t(71) = 9.26, p < .001, which did not differ from each other (t(71) = .34, ns). The communal ratings also varied in the expected direction and were significantly higher for the communion-oriented profile (M = 5.16, SD = 0.74) compared with the neutral (M = 4.65, SD = 0.92) and agency-oriented (M = 4.04, SD = 0.61) profiles, F(2, 71) = 13.85, p < .001, t(71) = 4.35, p < .001. Finally, the neutral profile had higher communal ratings compared with the agency-oriented profile (t(71) = 2.86, p < .001).

Task Performance and Organizational Citizenship Behavior Predictions

The task performance and OCB predictions were analyzed separately (r = .42, p < .001). We ran two ANOVAs on the data, with the profile type as the independent variable and the task performance and organizational citizenship as the dependent variables. Helmert-type contrasts confirmed our hy- pothesis on task performance predictions and showed that the agency-oriented profile was perceived as being more success- ful (M = 5.27, SD = 0.76) compared with both the communion- oriented (M = 3.65, SD = 1.12) and neutral (M = 3.70, SD = 0.93, t(71) = 6.88, p < .001) profiles, F(2, 71) = 23.752, p < .001, η2p = .40, which did not differ from each other (t(71) = �0.181, ns).

Concerning organizational citizenship, as expected, the agency-oriented profile elicited more favorable predictions (M = 5.34, SD = 0.81) compared with both the communion-ori- ented (M = 4.46, SD = 0.92) and neutral (M = 3.70, SD = 0.99, t(71) = 5.70, p < .001) profiles, F(2, 70) = 19.725, p < .001, η2p = .36. Furthermore, the communion-oriented profile attracted more favorable OCB predictions compared with the neutral profile (t(71) = 2.86, p < .001).

Discussion

The aim of this experiment was to show that the variation in social value on a personality inventory is sufficient to develop inferences about professional performance. The results con- firmed this hypothesis. All personality factors being equal, an agency-oriented profile elicited higher task performance

626 Sylvain Caruana et al.

and OCB predictions compared with a communion-oriented or a neutral profile. If reasoning about the performance was based on the descriptive information (i.e., personality factors), the participants would not have been able to distinguish between the profiles. This suggests that the agency information is linked to the performance inferences and is well perceived by people. Communion-oriented profile also attracted higher OCB predictions compared with the neutral profile. This suggests that in addition to agency, information about an individual’s degree of communion also helps to shape impressions based on the responses to personality inventories.

A limitation of the experiment concerns the over-representa- tion of the evaluative information. While the descriptive informa- tion is provided by six items per factor (two high-scored items per factor), the evaluative information comes from 10 high-scored items. The imbalance between the two types of information may explain the prevalence of the evaluative part of items. Experiment 2 was designed to overcome this shortcoming.

EXPERIMENT 2

Experiment 2 was designed to test the hypothesis that evalua- tive information takes primacy, even when the profiles differ on personality factors that are theoretically more or less sought after in an occupational setting: Agreeableness versus Consci- entiousness (Dunn et al., 1995; see also Pauls & Crost, 2005). Conscientiousness almost always predicts performance (Barrick et al., 2001; Kuncel et al., 2010), whereas agreeable- ness has far less predictive power (or only in specific cases; Mount et al., 1998). The results of the previous experiment and the unequal repartition of agency and communion among the Big-Five factors showed by the pilot study led us to expect this divergence to be due to the agentic part of conscientious- ness. If evaluative information takes primacy, we can postulate that the impact of agreeableness and conscientiousness on performance inferences should be the same when both factors are viewed as strongly agentic. Moreover, when the agentic part of Conscientiousness is reduced, this factor should no longer be linked to performance. To test this hypothesis, iden- tical procedure as in Experiment 1 was used, except that the descriptive information (agreeableness vs. conscientiousness) and the evaluative information (agency vs. communion) were simultaneously manipulated.

Participants

The participants included 133 students who were either in their third year of a management science course or their first year of a Master’s degree in auditing or financial management at Reims University (62 women, 66 men, 5 n/a, Mage =22.11, SD= 2.54). They completed the experiment during a tutorial. The participants were randomly divided between six experimental conditions. They were not familiar with the theoretical background and did not receive any remuneration for participating in the study.

Materials and Procedure

As in Experiment 1, the participants had to put themselves in the position of a recruiter and form an impression of a fictitious

job applicant on the basis of the individual’s responses to a per- sonality inventory. This time, however, the fictitious question- naire only concerned two personality factors (agreeableness and conscientiousness). The personality profiles operationalized the crossing of the personality (agreeableness, conscientious- ness) and social value (agency, communion) dimensions.

The personality questionnaire presenting the target was constructed along the same principle as the one in Experiment 1, except that it only contained the items that measured the agreeableness and conscientiousness factors. The applicants were therefore introduced via their responses to six agreeable- ness items and six conscientiousness items (i.e., two agentic, two communal, and two neutral items per personality dimen- sion). As in the previous experiment, we used the responses the applicant had supposedly provided to these items to vary both the social value of their profiles (agency-oriented vs. com- munion-oriented) and their personality scores (Agreeableness vs. Conscientiousness vs. Standard). The agency and communal mean scores by the type of items are shown in Table 2.

We began by constructing two standard profiles that had equivalent scores (8/18) on the agreeableness and conscien- tiousness factors, but conveyed different values: one profile was agency-oriented and the other profile was communion-ori- ented. With the exception that they were focalized on consci- entiousness and agreeableness items, these two standard profiles were the same as the agentic and communal profiles of Experiment 1. To construct the agreeableness and conscien- tiousness profiles without modifying the degree of agency or com- munion, we started from these standard profiles and gave the neutral items (measuring either agreeableness or cons- cientiousness) the most positive ratings. Thus, we increased the agreeableness/conscientiousness score to 13/18. The participants encountered one of six profiles crossing descriptive (agreeable, conscientious, or standard) and evaluative (agency-oriented or communion-oriented) information (see Appendix 3 for more details about the profiles).

The participants had 2 minutes to familiarize themselves with the profile that they were given. Then, they had to predict the applicant’s task performance and organizational citizen- ship with the same material as Experiment 1. Finally, to check that the personality characteristics had been correctly encoded, the participants were instructed to recall the responses the applicant had made.

Manipulation Check: Perception of Descriptive Information

After collecting the task performance and OCB predictions, we checked the encoding and the perception of the descriptive information. The participants had to recall the applicant’s responses on three conscientiousness and three agreeableness

Table 2. Agency and communal mean scores as a function of the item’s value orientation (Experiment 2)

Items’ type

Agency-oriented Neutral Communion-oriented

Agency 1.73 0.20 0.62 Communion 0.44 0.35 1.51

Note: The scores can vary from 0 to 3.

Personality, evaluative information, and performance inference 627

items. For both dimensions, one of the items was agency- oriented, one item was communion-oriented, and one item was neutral. Thus, the possible effects of the item’s value on memorization were neutralized. The participants had to recall the applicant’s responses on the same 7-point scales as those presented in the profiles.

Results

Manipulation Check: Perception of Personality Information

We conducted two ANOVAs with the personality profile (standard, agreeable, or conscientious) as the independent variable and the conscientiousness and agreeableness scores as the dependent variables. As expected, these ANOVAs revealed a main effect of the personality profile, F(4, 254) = 3.51, p < .01, η2p = .05. In line with our expectations, the conscientious profiles were perceived as more conscientious (M = 5.42, SD= 1.10) compared with the other two personality profiles (Mstandard =4.86, SD= 0.97; MAgreeable =4.98, SD= 0.79; F(2, 132) = 4.10, p < .05). Similarly, the agreeable profiles were perceived as more agreeable (M = 4.02, SD = 0.86) com- pared with the other two personality profiles (Mstandard = 3.56, SD = 0.85; MConscientious = 3.82, SD = 1.15; F(2, 132) = 2.69, p = .07). The perceptions of the profiles therefore matched the experimental manipulations.

Task Performance and Organizational Citizenship Behavior Predictions

The task performance and the OCB predictions were analyzed separately (r = .29, p < .001). Gender effects were tested for both dependent variables. No significant interaction with the profile value or the descriptive information on the task performance was observed, F(2, 133) = .08, ns, or on the OCB prediction, F(2, 133) = .43, ns. Gender was not taken into account in the final analyses.

Regarding the task performance (cf. Figure 1), the ANOVA revealed a main effect of social value, F(1, 133) = 10.37, p < .01, η2p = .08. As expected, the agency-oriented profiles elicited higher task performance predictions compared with the commu- nion-oriented profiles. The analyses failed to reveal either a main effect of descriptive personality information, F(2, 133) = .73, ns, or an interaction effect, F(2, 133) = .77, ns. Moreover, in line with our expectations, the performance predictions were higher for the agentic–agreeable profile compared with the commu- nal–conscientious profile, t(43) = 2.14, p < .05, when these two experimental conditions were compared separately.

The ANOVAs also revealed a main effect of social value on the OCB predictions, F(1, 133) = 7.44, p < .01, η2p = .06 (see Figure 1). The agency-oriented profiles attracted higher OCB predictions compared with the communion-oriented profiles, independent of the variations in the descriptive personality scores. There was no main effect of personality information, F(2, 133) = .07, ns, and no interaction effect, F(2, 133) = .42, ns. Moreover, as expected, the performance predictions were higher for the agentic–agreeable profile compared with the communal–conscientious profile, t(43) = 1.97, p = .05, when these two experimental conditions were specifically compared.

Discussion

The aim of this experiment was to demonstrate that descriptive information has less impact on performance predictions than evaluative information, even though individuals are fully capa- ble of perceiving differences in the personality between the profiles. The results confirmed that impressions formed from the responses to a personality questionnaire are predominantly based on the agentic versus communal value of those responses (see Wojciszke, Abele, & Baryla, 2009). Moreover, although the participants correctly perceived and memorized the descriptive personality information, an increased conscien- tiousness score had no impact on the task performance predictions if the responses were not agency-oriented. Further- more, agreeableness personality profile is judged to be just as effective as a conscientious (or standard) profile, provided that it is agency-oriented. The third experiment further explored the reasoning of the respondents (cf. Kuncel, Borneman, & Kiger, 2011). We reversed the procedure and tested our hypothesis from the respondent’s point of view.

EXPERIMENT 3

Hogan (1996; Hogan et al., 1998; Hogan & Shelton, 1998; Johnson & Hogan, 2006) described the completion of an in- ventory as a form of social situation in which an individual

Figure 1. The mean task performance and organizational citizenship behavior (OCB) predictions according to the personality profile and the social value (Experiment 2)

628 Sylvain Caruana et al.

(actor) communicates something to someone (observer). The author argued that personality measures elicit self-presenta- tions in which individuals manage their reputation according to the context. Therefore, our third experiment hypothesize that when respondents seek to present themselves as effective employees, they base their reasoning more on the value con- veyed by the responses than on the personality information. This time, the participants had to put themselves in the place of a more or less successful fictitious employee. They were given information about this employee’s task performance and asked to complete a personality inventory in which each personality factor was equilibrated in terms of its social value. We expected that the information on the task performance would lead to inferences in terms of social value but not personality. We expected this pattern to be observed for each of the personality factors.

Participants

The participants included 74 students (39 women, 35 men, Mage = 21.29, SD = 2.23) recruited from a range of courses (e.g., management, economics, finance and accounting, law, and languages). They were approached either during tutorials or in the university library to take part in a study on personality and professional abilities. The participants were randomly assigned to three different experimental conditions and did not receive any remuneration for participating in the study.

Procedure

The procedure was based on the identification paradigm (Gilibert & Cambon, 2003). The participants had to identify themselves with a low-performance, moderate-performance, or high-performance employee (between-participants design) and then respond to a personality questionnaire. They were provided with one of three performance profiles created from the responses that the employee’s manager had supposedly given to the task performance questionnaire used in the previous two experiments. After completing the personality questionnaire, the participants rated the employee on two single-item agency and communion scales.

Independent Variable: Profile Type

The applicants’ profiles were built from six items taken from the task performance scale developed by Tsui et al. (1997; see Experiments 1 and 2). The managers’ responses about their employees were rated on 7-point scales, which provided an opportunity to vary the performances according to the three modalities (high, moderate, and low performance). For the high-performance profile (ideal employee), the manager’s ratings on the task performance items were extremely favorable (6 or 7). The moderate-performance profile (standard employee) was operationalized by central ratings (3, 4, or 5), and the low-performance profile (unproductive employee) consisted of negative ratings (1 or 2). Deviations in task performance scores were standardized between the three experimental conditions.

Dependent Variable: Personality Measure

The personality questionnaire was exactly the same as the one used in Experiment 1, but it served as a dependent variable. This inventory permitted us to compute three social value scores (agency, communion, and neutral items scores), which were crossed with the Big-Five factor scores (conscientious- ness, agreeableness, neuroticism, extraversion, and openness scores).

Manipulation Check

The participants had to rate the employee’s agency and com- munion on a 7-point scale, which ranged from 1 (Don’t agree at all) to 7 (Agree completely). Agency was defined as the employee’s propensity to have “all the qualities required in order to perform well in the company and have a successful ca- reer.” Communion was defined as the propensity to have “the qualities required to be on good terms with other people.” The agency ratings allowed us to confirm that agency was induced by our experimental manipulation of the task performance. The communion ratings allowed us to identify halo effects generated by the positive task performance assessments.

Results

Manipulation Check: Perceptions of Agency and Communion

We ran an ANOVA on the agency and communion ratings. In line with expectations, the manipulation of the employee’s task performance had a significant effect on the perceived agency, F(2, 71) = 32.15, p < .001, η2p = .49. The high-performance pro- file scored significantly higher on agency (M = 5.71, SD = 1.00) compared with the moderate-performance profile (M = 4.91, SD= 0.95), which, in turn, scored significantly higher than the low-performance profile (M = 3.33, SD = 1.17). The per- ceived communion scores did not vary with the perfor- mance level, F(2, 71) = 1.34, ns.

Impact of Performance Information on the Responses to the Personality Inventory

Gender did not interact with the Profile Type, the Social Value, or the Big-Five scores, F(20, 540) = 1.15, ns. Analyses were then conducted independently of gender. We first ran a multivariate ANOVA (MANOVA) with the performance profile as the independent variable and the Big-Five and social value scores as the measures factors. As expected, the MANOVA revealed a main effect of performance on the social value scores, F(4, 114) = 4.94, p < .01, η2p = .15. There was also a main effect of performance on the Big-Five scores, F(8, 110) = 2.103, p < .05, η2p = .13, as well as a Perfor- mance Profile × Big Five × Social Value interaction, F(16, 102) = 1.78, p < .05, η2p = .22. We broke these effects down into their component parts.

Analysis of personality scores. Looking more closely at the effect of performance profiles on the Big-Five scores, we

Personality, evaluative information, and performance inference 629

found that the participants in the ideal employee condition described themselves as more conscientious, F(2, 71)=16.40, p < .001, η2p = .32, extraverted, F(2, 71) = 4.33, p < .05, η2p = .11, and emotionally stable, F(2, 71) = 6.95, p < .01, η2p = .16 (see Table 3 for detailed means) compared with the participants in the other two experimental conditions (stan- dard, unproductive). No effect of performance profile was found on agreeableness, F(2, 71) = 1.43, ns, or openness, F(2, 71) = .06, ns.

Analysis of social value scores. In line with our hypothe- sis on the item’s social value, the ideal employee profiles scored higher on the agentic items compared with the standard employee profiles, which, in turn, scored higher on agency than the unproductive employee profiles, F(2, 71) = 13.97, p < .01 (see Figure 2). The communion scores did not vary with the experimental condition, F(2, 71) = 1.41, ns. Contrary to expectations, the profiles’ performance levels also had an impact on the scores for the neutral items, F(2, 71) = 4.00, p < .05. Bonferroni post-hoc comparisons showed that the unproductive profile scored lower on the neutral items com- pared with the standard and ideal profiles, which did not differ from each other. Generally speaking, however, the difference between the agentic and neutral items was greater for the ideal profile than it was for the unproductive and standard profiles, F(4, 142) = 5.76, p < .001.

We broke the interaction effect down into separate persona- lity factors, crossing the performance profile and the item’s value (see Table 4 for detailed means and results). The results showed that the agentic items were more sensitive compared with the communal or neutral items to information about per- formance. Table 4 shows that with the exception of openness, this expected increase in the scores was observed for the agentic items of every personality factor (extraversion, conscientiousness, agreeableness, and emotional stability). A detailed analysis showed that the agreeableness–agency item scores were higher in the ideal condition than in the standard and unproductive conditions (communion scores followed the opposite tendency). With regard to the emotional stability and extraversion, the information about the performance affected the agentic items but appeared to have no impact on the communal and neutral ones. For conscientiousness, the communal and neutral items followed the same trend as the agentic items, with significantly higher scores for the standard and ideal profiles compared with the unproductive profile. However, this effect seemed stronger for the agentic items compared with the communal or neutral ones. Finally, comple- mentary analyses showed that the agentic openness items attracted significantly higher scores compared with the other

Table 3. Mean Big-Five scores (SD) as a function of the profile type (Experiment 3)

Profile

Unproductive Standard Ideal

Conscientiousness 3.93 0.98 4.98 0.71 5.20 0.76 Extraversion 3.91 0.89 4.13 0.85 4.59 0.71 Emotional stability 3.41 0.75 4.08 0.72 4.20 0.89 Agreeableness 4.20 0.98 4.47 0.80 4.06 0.79 Openness 3.84 1.06 3.91 0.99 3.92 0.69

Figure 2. The mean scores as a function of the profile type and the item value (Experiment 3)

Table 4. The mean scores as a function of the item’s type (Social Value × Big Five) and the performance profile type (Experiment 3)

Big-Five factor Item type

Task performance profile

F(4, 142) p < ƞ2Unproductive Standard Ideal

Agreeableness Agency 4.21a 4.54a 5.02b 4.03 .005 .10 Communion 4.54a 4.79a 3.56b

Neutral 3.96 4.10 3.67 Conscientiousness Agency 4.71a 5.23a 6.15b 2.47 .05 .07

Communion 3.63a 4.96b 5.06b

Neutral 3.38a 4.73b 4.33b

Emotional stability Agency 3.58a 5.02b 5.15b 3.36 .05 .09 Communion 3.52 3.63 4.06 Neutral 3.06 3.58 3.42

Extraversion Agency 4.06a 4.62b 5.75c† 4.50 .005 .11 Communion 4.42 4.19 4.15 Neutral 3.21 3.54 3.88

Openness Agency 3.92 4.04 4.33 0.46 ns Communion 3.71 3.63 3.52 Neutral 3.85 4.06 3.92

Note: Two different letters denote a significant difference (p < .05; Helmert-type contrast). †p = .05.

630 Sylvain Caruana et al.

two types of items for the ideal profile (t(73) = 2.21, p < .05), while this was not the case for the standard and unproductive profiles.

Therefore, the Profile Type × Big Five × Social Value inter- action mainly appeared to be due to the different ways in which the scores changed according to the value of the items and highlighted the expected pattern of the results for four of the five personality factors.

Discussion

The aim of this third experiment was to show that information about the task performance is coded in terms of the agency and communion, rather than in terms of the personality dimen- sions. The results mostly supported our hypothesis. Individuals seeking to convey the image of an ideal employee primarily use agentic items to describe themselves. Although the increase in the scores for the conscientiousness, extraver- sion, and emotional stability factors was predicted by the Big-Five theory (Pauls & Crost, 2005), the application of the two-dimensional model of social judgment allowed us to add an important nuance: manipulating performance leads to a po- tential increase in the scores on all factors—provided that they are measured on agentic items.

GENERAL DISCUSSION

All three experiments explored the use of the descriptive ver- sus evaluative properties of the personality inventory items. The first experiment demonstrated that when participants are exposed to strictly similar psychological profiles in terms of the personality dimensions, they attribute higher performances to the agency-oriented profiles than to the communion-ori- ented or neutral ones. The second experiment replicated this result with personality profiles that also varied on a descriptive characteristic that is more (i.e., conscientiousness) or less (i.e., agreeableness) associated with job success. The results then showed that conscientiousness and agreeableness led to the same performance inferences when they were measured with agency items. In a complementary fashion, the third experiment showed that participants seeking to convey the image of an ideal employee based their responses more on the evaluative compo- nent (social value) than on the descriptive component (personality dimensions). These results can be due to the evaluative stakes linked to the recruitment settings. The agency items were globally more sensitive to performance than the communion or neutral items. Finally, the neutral items (and thus, the purely descriptive ones) seemed barely sensitive, if at all, to the infor- mation about task performance. With the exception of the con- scientiousness factor, the neutral items did not undergo any significant difference linked to the profile type.

Our results showed that personality descriptions have at least an evaluative and a descriptive meaning that can be relatively independent. They also indicate that the former takes primacy on the latter in recruitment settings (Beauvois & Dubois, 2009; Dubois & Aubert, 2010). Moreover, compared with communal and descriptive information, agentic behavioral information is the most linked to job success perception

(Abele, 2003; Abele et al., 2008). Indeed, agentic profiles are perceived as more efficient and productive and constitute self- description and other-judgment norms in recruitment settings.

Finally, and more broadly, our observations also support the hypothesis that the two social judgment dimensions allow inferences to be drawn about what individuals can bring in a given context (see Beauvois & Dubois, 2009; Dubois & Beauvois, 2005). Agency discriminates competent individuals in a given system, and communion provides additional infor- mation on the relational value. In this sense, they pursue the growing interest for the recent study of egoistic (i.e., agentic) and moralistic (i.e., communion) biases in personality literature. To date, only one study has brought together research on social judgment and research on faking on personality inventories in an occupational context. In this study, Honkaniemi and Feldt (2008) tested the hypothesis that respondents adopt an agentic (vs. communal) response mode in recruitment situations. Real- life job applicants had to complete a personality inventory. Half of the participants were simply given the standard instructions. The other half of the participants were warned of the presence of a “socially desirable responding sub-scale” designed to mea- sure faking. The results showed that the scores in the warning condition were lower not only on some of the agentic subscales but also on some of the communal subscales. As in our experi- ments, Honkaniemi and Feldt (2008)’s results confirm the use- fulness of pairing the classic approach to personality with a social judgment approach.

However, they also highlighted a limitation arising from the assumption that respondents reason directly in terms of per- sonality traits (see Kuncel, Goldberg, & Kiger, 2011). These results underline the interest of an item-level analysis (Kuncel, Goldberg, et al., 2011; see also, Pace & Brannik, 2010; Zickar & Robie, 1999) when studying the effects of the item’s social value on performance or faking behavior (Backström et al., 2009, 2011). In fact, we suggest that the higher Big-Five scores recurrently observed in recruitment settings can be the consequence of distinct agentic or communal faking modes on different types of items (Paulhus & Trapnell, 2008). Recruitment settings could drive higher scores on agentic items but not necessarily on communal or neutral items. This assumption would have to be tested using the same type of procedure and material we used.

Limitations and Future Directions

Our research presents a limit because we did not investigate the real performance. It would be interesting to demonstrate that taking notice of the agentic and communal responses to personality inventories enhances their predictive validity in real-life settings. Dompnier, Darnon, and Butera (2009) found that the predictive strength of a mastery-goal scale in an academic setting varied according to the respondents’ motives. For students who associated their responses with communal factors (winning the teacher’s approval), the mastery goals were only weakly predictive of academic achievement, whereas for students who associated them with agentic factors (succeeding at university), there was a positive relationship between the endorsement of mastery goals and final grades. The results reported by Hogan and colleagues (1998) support this approach to personality. In occupational settings where

Personality, evaluative information, and performance inference 631

there is an explicit chance of promotion, assessments of the employees’ performance appear to be more closely correlated with personality dimensions linked to the need for status (agency) compared with those linked to acceptance (commu- nion); the opposite finding is observed in settings where there is no chance of promotion. Congruent with our results, the criteria for inferring performance have more to do with the motives surrounding the personality assessment situation (e.g., job promotion) than with the personality factors per se. Hence, integrating Hogan’s socioanalytic theory and the agency–communion framework could enhance research on self-presentation in personality inventories and performance prediction. Taken with the results of our experiments, it sug- gests that any Big-Five factor could be predictive of task per- formance, provided that its component items are associated with agentic motives in the reference environment.

Beyond job performance criteria, it would also be interesting to replicate our findings with outcomes dependent on com- munion stakes. The literature has considered the two evaluative dimensions to be supra-ordinate factors of the Big Five (Blackburn et al., 2004; Digman, 1997). While this research field considers that the Big Five can be resumed with the two evaluative dimensions, our research proposed an alternative model, which considered that each Big-Five factor can be mea- sured in terms of agency and communion (and descriptive) and that different types of items could have a differential impact on performance prediction. In other words, conscientiousness and agency do not appear to be confounded, and agreeableness could gain predictive strength when distinguished from com- munion. Proposing an agency and communion standardized measure of each of the Big-Five factors, the interest of our research was to initiate this alternative conception. Our results showed that it could be interesting to use personality inventories that would be standardized on social value (agency, communion, and neutral behavior descriptions) to predict agentic (or communal) outcomes from the FFM dominant model of personality. We could then compute indices of satisfaction of the social demand and, as a result, optimize the prognostic job or relational performance. Some authors have worked on a similar hypothesis on integrity tests and showed that the individual’s ability to identify the selection criteria during the evaluation procedures contrib- utes to the validity of the measures (e.g., König, Melchers, Kleinmann, Richter, & Klehe, 2006).

As evaluation is part of the self and other judging pro- cesses, it should also be interesting to precisely determine the apparition condition of descriptive versus evaluative rea- soning when responding to personality inventories and to rep- licate our findings in real-life settings. The social judgment approach appears to be a good way of overcoming the limita- tions of the descriptive personality approach, in regard to studying the processes behind responding to personality in- ventories or optimizing the latter’s predictive power.

REFERENCES

Abele, A. E. (2003). The dynamics of masculine-agentic and feminine- communal traits: Findings from a prospective study. Journal of Personality and Social Psychology, 85(4), 768–776. DOI:10.1037/0022-3514.85.4.768

Abele, A. E., & Brack, S. (2013). Preference for other persons’ traits is depen- dent on the kind of social relationship. Social Psychology, 44(2), 84–94. DOI:10.1027/1864-9335/a000138

Abele, A. E., & Bruckmüller, S. (2011). The bigger one of the “Big Two”? Preferential processing of communal information. Journal of Experimental Social Psychology, 47(5), 935–948. DOI:10.1016/j.jesp.2011.03.028

Abele, A. E., Rupprecht, T., & Wojciszke, B. (2008). The influence of success and failure experiences on agency. European Journal of Social Psychology, 38(3), 436–448. DOI:10.1002/ejsp.454

Abele, A. E., & Wojciszke, B. (2007). Agency and communion from the perspec- tive of self versus others. Journal of Personality and Social Psychology, 93(5), 751–763.

Abele, A. E., & Wojciszke, B. (2013). The Big Two in social judgment and be- havior. Social Psychology, 44(2), 61–62. DOI:10.1027/1864-9335/a000137

Backström, M. (2007). Higher-order factors in a five-factor personality inven- tory and its relation to social desirability. European Journal of Psycholog- ical Assessment, 23(2), 63–70.

Backström, M., Björklund, F., & Larsson, M. R. (2009). Five-factor invento- ries have a major general factor related to social desirability which can be reduced by framing items neutrally. Journal of Research in Personality, 43, 335–344.

Backström, M., Björklund, F., & Larsson, M. R. (2011). Social desirability in personality assessment: Outline of a model to explain individual differ- ences. In M. Ziegler, C. MacCann, & R. D. Roberts (Eds.), New perspec- tives on faking in personality assessment (pp. 201–213). New York, NY: Oxford University Press.

Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44, 1–26.

Barrick, M. R., & Mount, M. K. (1993). Autonomy as a moderator of the re- lationships between the Big Five personality dimensions and job perfor- mance. Journal of Applied Psychology, 78(1), 111–118. DOI:10.1037/ 0021-9010.78.1.111

Barrick, M. R., & Mount, M. K. (2005). Yes, Personality Matters: Moving on to More Important Matters Human Performance, 18, 359–372.

Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and perfor- mance at the beginning of the new millennium: What do we know and where do we go next? International Journal of Selection and Assessment, 9, 9–30.

Beauvois, J.-L., & Dubois, N. (2009). Lay psychology and the social value of persons. Social and Personality Psychology Compass, 6, 1082–1095.

Bi, C., Ybarra, O., & Zhao, Y. (2013). Accentuating your masculine side: Agentic traits generally dominate self-evaluation, even in China. Social Psychology, 44(2), 103–108. DOI:10.1027/1864-9335/a000144

Blackburn, R., Renwick, S. J. D., Donnelly, J. P., & Logan, C. (2004). Big Five or Big Two? Superordinate factors in the NEO Five Factor Inventory and the Antisocial Personality Questionnaire. Personality and Individual Differences, 37(5), 957–970. DOI:10.1016/j.paid.2003.10.017

Borman, W. C., & Motowidlo, S. J. (1997). Task performance and contextual performance: The meaning for personnel selection research. Human Perfor- mance, 10(2), 99.

Caprara, G. V., Barbaranelli, C., & Borgogni, L. (1997). Alter Ego: Les 5 facteurs fondamentaux de la personnalité. Paris: E.A.P.

Caprara, G. V., Barbaranelli, C., Borgogni, L., & Perugini, M. (1993). The “Big Five questionnaire”: A new questionnaire to assess the five factor model. Personality and Individual Differences, 15(3), 281–288. DOI:10.1016/0191-8869(93)90218-R

Chiaburu, D. S., Oh, I.-S., Berry, C. M., Li, N., & Gardner, R. G. (2011). The five-factor model of personality traits and organizational citizenship behav- iors: A meta-analysis. Journal of Applied Psychology, 96(6), 1140–1166.

Cislak, A. (2013). Effects of power on social perception: All your boss can see is agency. Social Psychology, 44(2), 138–146. DOI:10.1027/1864-9335/a000139

Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI). Professional manual. Odessa, Florida: Psychological Assessment Resources.

Costa, P. T., McCrae, R. R., & Rolland, J.-P. (1998). Manuel de l’inventaire de personalité NEO PI-R. Paris: ECPA.

Cottrell, C. A., Neuberg, S. L., & Li, N. P. (2007). What do people desire in others? A sociofunctional perspective on the importance of different valued characteristics. Journal of Personality and Social Psychology, 92(2), 208– 231. DOI:10.1037/0022-3514.92.2.208

Digman, J. M. (1997). Higher-order factors of the Big Five. Journal of Person- ality and Social Psychology, 73(6), 1246-1256.

Dompnier, B., Darnon, C., & Butera, F. (2009). Faking the desire to learn: A clarification of the link between mastery goals and academic achievement. Psychological Science, 20, 939–943.

Dubois, N. (2010). Theory of the social value of persons applied to organiza- tions: Typologies of good leaders and recruitment. European Review of Applied Psychology, 60(4), 255–266.

Dubois, N., & Aubert, E. (2010). Valeur sociale des personnes: deux informa- tions valent-elles mieux qu’une ? International Review of Social Psychology, 23(1), 57–92.

632 Sylvain Caruana et al.

Dubois, N., & Beauvois, J.-L. (2005). Normativeness and individualism. European Journal of Social Psychology, 35, 123–146.

Dubois, N., & Beauvois, J.-L. (2011). Are some rabbits more competent and warm than others? Swiss Journal of Psychology/Schweizerische Zeitschrift für Psychologie/Revue Suisse de Psychologie, 70(2), 63–73.

Dunn, W. S., Mount, M. K., Barrick, M. R., & Ones, D. S. (1995). Relative importance of personality and general mental ability in managers’ judg- ments of applicant qualifications. Journal of Applied Psychology, 80(4), 500–509. DOI:10.1037/0021-9010.80.4.500

Fiske, S. T., Cuddy, A. J., & Glick, P. (2007). Universal dimensions of social cog- nition: Warmth and competence. Trends in Cognitive Science, 11(2), 77–83.

Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from per- ceived status and competition. Journal of Personality and Social Psychol- ogy, 82(6), 878–902.

Gilibert, D., & Cambon, L. (2003). Paradigms of the sociocognitive approach. In N. Dubois (Ed.), A sociocognitive approach to social norms (pp. 38–69). London: Routledge.

Hall, J. A., Pennington, N., & Lueders, A. (in press). Impression management and formation on Facebook: A lens model approach. New Media & Society. DOI:10.1177/1461444813495166

Hogan, R. (1996). A socioanalytic perspective on the Five-Factor Model. In J. S. Wiggins (Ed.), The Five Factor Model of personality. Theoretical per- spectives. New York London: Guilford Press.

Hogan, R., Rybicki, S. L., Motowidlo, S. J., & Borman, W. C. (1998). Rela- tions between contextual performance, personality, and occupational ad- vancement. Human Performance, 11(2/3), 189–207.

Hogan, R., & Shelton, D. (1998). A socioanalytic perspective on job perfor- mance. Human Performance, 11(2/3), 129–144.

Honkaniemi, L., & Feldt, T. (2008). Egoistic and moralistic bias in real-life in- ventory responses. Personality and Individual Differences, 45(4), 307–311. DOI:10.1016/j.paid.2008.04.018

Hough, L. M.., Eaton, N. K., Dunnette, M. D., Kamp, J. D., & McCloy, R. A. (1990). Criterion-related validities of personality constructs and the effect of response distortion on those validities. Journal of Applied Psychology, 75(5), 581–595. DOI:10.1037/0021-9010.75.5.581

Johnson, J. A., & Hogan, R. (2006). A socioanalytic view of faking. In R. L. Griffith, & M. H. Peterson (Eds.), A closer examination of applicant faking behavior (pp. 209–231). Greenwich: Information Age Publishing.

Judd, C. M., James-Hawkins, L., Yzerbyt, V. Y., & Kashima, Y. (2005). Fundamental dimensions of social judgment: Understanding the relations between judgments of competence and warmth. Journal of Personality and Social Psychology, 89(6), 899–913.

Judge, T. A., Higgins, C. A., Thoresen, C. J., & Barrick, M. R. (1999). The Big Five personality traits, general mental ability, and career success across the life span. Personnel Psychology, 52, 621–652.

König, C. J., Melchers, K. G., Kleinmann, M., Richter, G. M., & Klehe, U.-C. (2006). The relationship between the ability to identify evaluation criteria and integrity test scores. Psychology Science, 48(3), 369–377.

Kuncel, N. R., Borneman, M. J., & Kiger, T. (2011). Innovative item response process and Bayesian faking detection methods: More questions than an- swers. In M. Ziegler, C. MacCann, & R. D. Roberts (Eds.), New perspec- tives on faking in personality assessment. New York, NY: Oxford University Press.

Kuncel, N. R., Goldberg, L. R., & Kiger, T. (2011). A plea for process in personality prevarication. Human Performance, 24(4), 373–378. DOI:10.1080/08959285.2011.597476

Kuncel, N. R., Ones, D. S., & Sackett, P. R. (2010). Individual differences as predictors of work, educational, and broad life outcomes. Personality and Individual Differences, 49, 331–336.

Marcus, B., Machilek, F., & Schütz, A. (2006). Personality in cyberspace: Personal web sites as media for personality expressions and impressions. Journal of Personality and Social Psychology, 90(6), 1014–1031. DOI:10.1037/0022-3514.90.6.1014

McAdams, D. P., Hoffman, B. J., Mansfield, E. D., & Day, R. (1996). Themes of agency and communion in significant autobiographical scenes. Journal of Personality, 64(2), 339–377. DOI:10.1111/1467-6494.ep9606164115

McCrae, R. R., & Costa, P. T. (2006). Perspectives de la théorie des cinq facteurs (TCF): Traits et culture. Psychologie Française, 51, 227–244.

McCrae, R. R., Yamagata, S., Jang, K. L., Riemann, R., Ando, J., Ono, Y.,… Spinath, F. M. (2008). Substance and artifact in the higher-order factors of the Big Five. Journal of Personality and Social Psychology, 95(2), 442–455. DOI:10.1037/0022-3514.95.2.442

Moskowitz, D. S., Suh, E. J., & Desaulniers, J. (1994). Situational influences on gender differences in agency and communion. Journal of Personality and Social Psychology, 66(4), 753–761. DOI:10.1037/0022-3514.66.4.753

Mount, M. K., Barrick, M. R., & Stewart, G. L. (1998). Five-Factor Model of personality and performance in jobs involving interpersonal interactions. Human Performance, 11(2/3), 145.

Oldmeadow, J. A., & Fiske, S. T. (2007). System-justifying ideologies moder- ate status = competence stereotypes: Roles for belief in a just world and so- cial dominance orientation. European Journal of Social Psychology, 37(6), 1135–1148. DOI:10.1002/ejsp.428

Oldmeadow, J. A., & Fiske, S. T. (2010). Social status and the pursuit of pos- itive social identity: Systematic domains of intergroup differentiation and discrimination for high- and low-status groups. Group Processes & Intergroup Relations, 13(4), 425–444. DOI:10.1177/1368430209355650

Pace, V. L., & Brannik, M. T. (2010). How similar are personality scales of the “same” construct? A meta-analytic investigation. Personality and Individ- ual Differences, 49, 669–676.

Paulhus, D. L., & Trapnell, P. D. (2008). Self-presentation: An agency–com- munion framework. In O. P. John, R. W. Robins, & L. A. Pervin (Eds.), Handbook of personality psychology (pp. 492–517). New York: Guilford.

Pauls, C. A., & Crost, N. W. (2005). Effects of different instructional sets on the construct validity of the NEO-PI-R. Personality and Individual Differ- ences, 39, 297–308.

Peeters, G. (1992). Evaluative meanings of adjectives in vitro and in context: Some theoretical implications and practical consequences of positive–negative asymmetry and behavioral-adaptive concepts of evaluation. Psychologica Belgica, 32(2), 211–231.

Plaisant, O., Courtois, R., Réveillère, C., Mendelsohn, G. A., & John, O. P. (2010). Validation par analyse factorielle du Big Five Inventory français (BFI-Fr). Analyse convergente avec le NEO-PI-R. Annales Médico- Psychologiques, Revue Psychiatrique, 168(2), 97–106. DOI:10.1016/j. amp.2009.09.003

Rothstein, M. G., & Goffin, R. D. (2006). The use of personality measures in personnel selection: What does current research support? Human Resource Management Review, 16(2), 155–180. DOI:10.1016/j.hrmr.2006.03.004

Salgado, J. F. (1997). The Five Factor Model of personality and job perfor- mance in the European community. Journal of Applied Psychology, 82 (1), 30–43.

Tsui, A. S., Pearce, J. L., Porter, L. W., & Tripoli, A. M. (1997). Alternative approaches to the employee–organization relationship: Does investment in employees pay off? Academy of Management Journal, 40(5), 1089–1121. DOI:10.2307/256928

Utz, S. (2010). Show me your friends and I will tell you what type of person you are: How one’s profile, number of friends, and type of friends influence im- pression formation on social network sites. Journal of Computer-Mediated Communication, 15(2), 314–335. DOI:10.1111/j.1083-6101.2010.01522.x

Woike, B., Gershkovich, I., Piorkowski, R., & Polo, M. (1999). The role of motives in the content and structure of autobiographical memory. Journal of Personality and Social Psychology, 76(4), 600–612. DOI:10.1037/ 0022-3514.76.4.600

Woike, B., Lavezzary, E., & Barsky, J. (2001). The influence of implicit mo- tives on memory processes. Journal of Personality and Social Psychology, 81(5), 935–945. DOI:10.1037/0022-3514.81.5.935

Woike, B., & Polo, M. (2001). Motive-related memories: Content, structure, and affect. Journal of Personality, 69(3), 391–415. DOI:10.1111/1467- 6494.00150

Wojciszke, B. (1994). Multiple meanings of behavior: Construing actions in terms of competence or morality. Journal of Personality and Social Psy- chology, 67(2), 222–232.

Wojciszke, B., Abele, A. E., & Baryla, W. (2009). Two dimensions of inter- personal attitudes: Liking depends on communion, respect depends on agency. European Journal of Social Psychology, 39, 973–990.

Wojciszke, B., Baryla, W., Parzuchowski, M., Szymkow, A., & Abele, A. E. (2011). Self-esteem is dominated by agentic over communal information. Eu- ropean Journal of Social Psychology, 41(5), 617–627. DOI:10.1002/ejsp.791

Wojciszke, B., & Sobiczewska, P. (2013). Memory and self-esteem: The role of agentic and communal content. Social Psychology, 44(2), 95–102. DOI:10.1027/1864-9335/a000149

Ybarra, O., Chan, E., Park, H., Burnstein, E., Monin, B., & Stanik, C. (2008). Life’s recurring challenges and the fundamental dimensions: An integration and its implications for cultural differences and similarities. European Jour- nal of Social Psychology, 38, 1083–1092.

Ybarra, O., Park, H., Stanik, C., & Lee, D. S. (2012). Self-judgment and rep- utation monitoring as a function of the fundamental dimensions, temporal perspective, and culture. European Journal of Social Psychology, 42(2), 200–209. DOI:10.1002/ejsp.854

Zickar, M. J., & Robie, C. (1999). Modeling faking good on personality items: An item-level analysis. Journal of Applied Psychology, 84(4), 551–563. DOI:10.1037/0021-9010.84.4.551

Personality, evaluative information, and performance inference 633

APPENDIX

Appendix A1: Agentic, communal, and neutral items for each of the Big-Five traits (pilot study)

Big-Five trait Orientation Item Agency Communion t p

Conscientiousness Agentic I strive for excellence in everything I do 2.24 0.72 5.18 .01 I am efficient and productive at work 2.08 0.04 3.43 .01

Communal I don’t think you ever need to tax your strength to the limit, even if there’s a deadline to meet

�0.24 �0.96 1.69 .10

I’m known for my prudence and common sense 1.37 1.10 0.60 ns Neutral I’m not compulsive about cleaning* �0.36 0.11 1.11 ns

I often do thing on the spur of the moment 0.15 0.57 1.75 ns Agreeableness Agentic If necessary, I do not hesitate to tell people not

to interfere 1.00 0.04 2.33 .05

I usually adopt a cordial attitude to people, even if I feel antipathy towards them

1.59 0.96 1.68 .10

Communal I’m better than most people, and I know it* 0.28 �2.09 �5.29 .01 Some people think of me as cold and calculating* �0.59 �1.90 �2.84 .01

Neutral I’d rather not talk about my achievements and myself 0.04 0.23 �0.58 ns With some people, you mustn’t be too easygoing* 0.24 0.47 1.11 ns

Emotional stability Agentic I feel comfortable in the presence of my bosses or other authorities

1.48 0.57 �2.15 .05

I often feel inferior to others* �1.42 �0.43 �2.44 .05 Communal I rarely feel lonely or blue 0.56 1.67 2.28 .05

It doesn’t embarrass me too much if people ridicule and tease me

0.06 �0.90 2.14 .05

Neutral When people I know do foolish things, I get embarrassed for them*

0.16 0.32 �0.38 ns

At times I have felt bitter and resentful* 0.20 0.00 �0.42 ns Extraversion Agentic I have often been a leader of groups I have belonged to* 1.46 0.05 2.72 .01

I am a very active person 1.20 0.36 1.86 .07 Communal I really feel the need for other people if I am

by myself for long 0.16 1.10 �2.87 .01

Sometimes I bubble with happiness 0.56 1.67 �2.52 .01 Neutral I’m attracted to bright colors and flashy styles 0.00 0.04 �0.09 ns

I rarely use words like “fantastic!” or “sensational!” to describe my experiences*

0.04 �0.19 �0.57 ns

Openness Agentic Aesthetic and artistic concerns aren’t very important to me

�1.04 0.23 1.87 .07

I have a very active imagination 1.15 0.23 1.01 .08 Communal Odd things—like certain scents or the names

of distant places—can evoke strong moods in me 0.32 1.40 �3.17 .01

Sometimes when I am reading poetry or at a work of art, I feel a chill or wave of excitement

0.12 1.00 �1.95 .06

Neutral Watching ballet or modern dance bores me* �0.30 0.04 0.70 ns I believe that the “new morality” of permissiveness is no morality at all*

0.12 �0.08 �0.45 ns

Note: The agentic items have average communion scores, and the communal items have average agency scores. The neutral items have average scores on both judgment dimensions. *Reversed items. Items are part of the authorized French versions of the NEO PI-R and the Alter Ego. The meanings could have been altered because of the French translation.

Appendix A2. Pattern of the target’s responses for Conscientiousness (Experiment 1)

Agentic profile Communal profile Neutral profile

Agentic items I strive for excellence in everything I do +3 1 1 I am efficient and productive at work +3 0 0

Communal items I don’t think you ever need to tax your strength to the limit, even if there’s a deadline to meet

�1 �3 �1

I’m known for my prudence and common sense 0 +3 0 Neutral items I’m not compulsive about cleaning* �1 �1 �3

I often do thing on the spur of the moment* 0 0 �3 Score on conscientiousness factor 8/18 8/18 8/18

*Reversed items.

634 Sylvain Caruana et al.

Appendix A3. Pattern of the target’s responses on Agreeableness for standard-personality and agreeable-personality profiles (Experiment 2)

Standard-personality profile Agreeable-personality profiles

Agentic profile Communal profile Agentic profile Communal profile

Agentic items If necessary, I do not hesitate to tell people not to interfere

+3 1 +3 1

I usually adopt a cordial attitude to people, even if I feel antipathy towards them

+3 0 +3 0

Communal items I’m better than most people, and I know it* �1 �3 �1 �3 Some people think of me as cold and calculating* 0 �3 0 �3

Neutral items I’d rather not talk about my achievements and myself +1 +1 +3 +3 With some people, you mustn’t be too easygoing* 0 0 �3 �3 Score on agreeableness factor 8/18 8/18 13/18 13/18

Note: For both profiles, the Conscientiousness score was maintained at 8/18 (cf. Appendix 2). The same logic was applied to build the conscientiousness profile. *Reversed items.

Personality, evaluative information, and performance inference 635

Copyright of European Journal of Social Psychology is the property of John Wiley & Sons, Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.