psychology

A Multi-Rater Framework for Studying Personality: The Trait-Reputation-Identity Model

Samuel T. McAbee Illinois Institute of Technology

Brian S. Connelly University of Toronto

Personality and social psychology have historically been divided between personality researchers who study the impact of traits and social– cognitive researchers who study errors in trait judgments. However, a broader view of personality incorporates not only individual differences in underlying traits but also individual differences in the distinct ways a person’s personality is construed by oneself and by others. Such unique insights are likely to appear in the idiosyncratic personality judgments that raters make and are likely to have etiologies and causal force independent of trait perceptions shared across raters. Drawing on the logic of the Johari window (Luft & Ingham, 1955), the Self–Other Knowledge Asymmetry Model (Vazire, 2010), and Socioanalytic Theory (Hogan, 1996; Hogan & Blickle, 2013), we present a new model that separates personality variance into consensus about underlying traits (Trait), unique self-perceptions (Identity), and impressions conveyed to others that are distinct from self-perceptions (Reputation). We provide three demonstrations of how this Trait-Reputation-Identity (TRI) Model can be used to understand (a) consensus and discrepancies across rating sources, (b) personality’s links with self-evaluation and self-presentation, and (c) gender differences in traits. We conclude by discussing how researchers can use the TRI Model to achieve a more sophisticated view of personality’s impact on life outcomes, developmental trajectories, genetic origins, person–situation interactions, and stereotyped judgments.

Keywords: accuracy and error, bifactor model, personality perception, reputation, self-enhancement

Supplemental materials: http://dx.doi.org/10.1037/rev0000035.supp

Personality is among the oldest and most pervasively studied topics throughout psychology, making its way into fields as di- verse as neuroscience (e.g., Pickering & Gray, 1999), culture (e.g., Benet-Martínez & Oishi, 2008), and education (e.g., De Raad & Schouwenburg, 1996). Throughout its history, personality has been dogged by a tension between the effects of “accuracy” and “error” in personality judgments (Funder, 1987). On the one hand, large-scale meta-analyses indicate that personality judgments are relatively stable over time (Roberts & DelVecchio, 2000) and converge across raters (Connelly & Ones, 2010). Researchers have identified an impressive litany of life outcomes that are tied to personality traits, ranging from relationship quality to occupational attainment and success to health and life expectancy (Ozer & Benet-Martínez, 2006; Roberts, Kuncel, Shiner, Caspi, & Gold- berg, 2007). Indeed, it is difficult to view this body of research

without concluding that not only must perceptions of personality traits be at least somewhat accurate but also that the personality traits are potent influences on people’s lives.

On the other hand, researchers studying topics in social cogni- tion like person perception and “the self” have identified a litany of errors people tend to make in forming personality judgments (Kunda, 1990; Malle, 2006; Ross, 1977). For example, we tend to have overly positive views of ourselves because we tend to take credit for our successes, make excuses for our failures, assume the best about ourselves, and neglect disconfirming feedback about our self-views (Dunning, Heath, & Suls, 2004; Robins & John, 1997; Sedikides, 1993). In the 1970s and 1980s, the variety of perceptual errors identified (which include the fundamental attri- bution error, the false consensus effect, the actor-observer differ- ence, and the above-average effect) was so numerous that the prospects for personality traits to not only be judged accurately but to even exist beyond a figment of raters’ imaginations seemed dismal. Moreover, even if it was possible to accurately perceive a trait, burgeoning research on impression management at the time suggested that people effortfully adjust their behaviors to mask underlying traits (e.g., Schlenker, 1980). Coupled with notions of situational specificity of behavior (Mischel, 1968), this extensive research on errors in trait judgments put personality research on ice for nearly two decades (Kenrick & Funder, 1988).

In contrast to the polarization that characterized the person- situation debate, contemporary person perception has used multi- rater designs to find that trait perceptions contain elements of both accuracy and error (e.g., Funder, 1995; West & Kenny, 2011; Zaki

This article was published Online First August 8, 2016. Samuel T. McAbee, Department of Psychology, Illinois Institute of

Technology; Brian S. Connelly, Department of Management, University of Toronto.

Both authors contributed equally to the manuscript; authorship order was determined by a gymnastics flip. We thank SiSi Tran, Erika Carlson, Jacob Hirsh, Fred Oswald, David Kenny, and David Funder for their insights on the manuscript and analytic approach.

Correspondence concerning this article should be addressed to Brian S. Connelly, Department of Management, University of Toronto, 1265 Mil- itary Trail, Toronto, Ontario M1C 1A4, Canada. E-mail: brian.connelly@ utoronto.ca

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Psychological Review © 2016 American Psychological Association 2016, Vol. 123, No. 5, 569 –591 0033-295X/16/$12.00 http://dx.doi.org/10.1037/rev0000035

569

& Ochsner, 2011). In general, accuracy in trait judgments pushes raters’ perceptions toward consensus, whereas perception errors tend to make raters’ judgments more discrepant and unique. Thus, the field has amassed an impressive body of knowledge about how strong consensus is across raters (both observer-observer correla- tions and self-observer correlations) and what factors moderate these relationships (e.g., Connelly & Ones, 2010; Funder, 1995).

Unfortunately, however, the general community of personality psychologists rarely implements multirater designs to disentangle shared and unique perceptions when studying traits’ etiologies and outcomes. Instead, most studies of personality only use self- reports. Whether personality effects from such studies are attrib- uted to “accuracy” or “error” substantially colors how findings are interpreted. For example, Roberts, Walton, and Viechtbauer (2006) found that people tend to self-report becoming more con- scientious as they age. However, what if these effects were not driven by the accurate trait variance in the measures but by self-perception errors (i.e., people mistakenly believe they become more conscientious as they age, though their patterns of conscien- tious behaviors, thoughts, and feelings remain unchanged)? Or, do errors that creep into self-perceptions attenuate the actual effects of conscientiousness (e.g., people become much more conscien- tious as they age but are biased to perceived themselves consis- tently, underestimating their true change in conscientiousness)? Personality psychologists have found some comfort in replicating self-report findings using observer-reports. However, even in such cases, results can and do differ (cf. Watson & Humrichouse, 2006). Such differences in findings raise a bevy of questions: How different is too different? Which effect is the “true” effect? Is the effect the product of a consensually viewed trait or a rater’s unique perception?

In the present paper, we present a model that will arm the broader community of personality psychologists with the tools of consensus and uniqueness from the field of person perception. Separating perceptions that are shared from those that are unique

is an invaluable (though imperfect) mechanism for distinguishing actual underlying personality traits from perceptual errors and idiosyncratic trait insights. Apart from traits themselves, individual differences in how people perceive themselves (self-definition) and are perceived by others (position within a social context) represent important components of a broader personality system (McAdams & Pals, 2006). Though we know a great deal about the conditions under which different raters agree on personality trait judgments, we know markedly less about the etiology and out- comes associated with disagreement. In the pages that follow, we describe this model and the theoretical meaning of its components. We then offer three examples of how this model can inform multiple avenues of personality research, and discuss its potential applications, limitations, and directions for future research.

The Trait-Reputation-Identity Model

Similar to the Self-Other Knowledge Asymmetry (SOKA) Model (Vazire, 2010), our Trait-Reputation-Identity (TRI) Model builds on the logic of the Johari window (Luft & Ingham, 1955). Popular as a counseling psychology tool, the Johari window asks both self-respondents and observers to provide trait information about a target person. Crossing information “known” and “un- known” to the self and “known” and “unknown” to observers produces a grid with four quadrants (see the left portion of Figure 1). The “Arena” reflects trait information that is shared between the self and others (i.e., known to the self and known to observers). The “Façade” represents a set of self-knowledge that is not shared with others (i.e., known to the self but unknown to observers). In contrast, the “Blind-Spot” reflects parts of targets’ personality that others see but about which the target is unaware or does not endorse (i.e., known to observers but unknown to the self). Finally, some trait information may remain “Unknown” to (or not per- ceived by) both the self and observers.

Figure 1. The Johari window (Left; Luft & Ingham, 1955) and the Trait-Reputation-Identity Model (Right). Residual correlations between like-indicators for TRI Model not depicted.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

570 MCABEE AND CONNELLY

Note that our TRI Model relabels the quadrants of the Johari window as Trait (Arena), Reputation (Blind-Spot), and Identity (Fa- çade; see Figure 1). The “Trait” label reflects personality psycholo- gy’s historical emphasis on corroborating traits through consensual validation. In contrast, “Reputation” and “Identity” are appropri- ated from long traditions of research that has explored the way people define themselves and are defined by their relationships with others (e.g., symbolic interactionism, Blumer, 1986; Socio- analytic Theory, Hogan, 1996; Hogan & Blickle, 2013; and even William James’s distinction between the “I” and the “me,” James, 1890). Beyond traits themselves, the reputation we convey to others serves as a means to achieving our goals, and the identity we form provides self-definition and sense-making functions. Al- though the Johari window was designed to sort multiple traits as being characteristic or uncharacteristic of a person, it also serves as a useful framework for characterizing components of a person’s standing on a single continuous trait dimension. For example, Bill may agree with Ted that he is below average on conscientiousness (Trait). However, Ted may see Bill as extremely unconscientious (Reputation), whereas Bill believes himself to be only slightly below average on conscientiousness (Identity).

From a mono-trait, multirater vantage point, the Johari window implies a confirmatory factor analytic measurement model for separating variance attributable to the Trait, Reputation, and Iden- tity. The right portion of Figure 1 depicts a hypothetical model in which both the self and several observers have completed multiple measures of a single trait for a target (e.g., multiple items, scales from different personality inventories, or repeated measurements). The Trait is a general factor that captures variance common across both self- and observer-reports. In contrast, Identity captures vari- ance in self-reports that is not shared with observers and that remains in self-reports after controlling for the Trait factor. Thus, the unique variance captured in the Identity comprises both errors in self-perception (e.g., mistakenly believing oneself to be espe- cially thoughtful and cultured) and trait-relevant information not available to the observers in question (e.g., harboring a hidden passion for art and nature). Similarly, Reputation captures residual variance unique to the observer-reports arising both from errors in observers’ perceptions (e.g., physical appearance stereotypes) and from trait-relevant information not available to the self (e.g., facial expressions in response to novel foods or thought-provoking ques- tions).

A major advantage of the TRI Model is the explicit separation of these three components of personality when aligning traits with external variables. In contrast, the SOKA Model similarly draws on concepts from the Johari window to generate predictions about whether observed trait-criterion relations will be stronger for self- or observer-reports. For instance, drawing on the SOKA Model, Carlson, Vazire, and Oltmanns (2013) predicted and found that observer-reports of conscientiousness were stronger (negative) predictors of externalizing personality disorders than were self- reports of conscientiousness. However, from a TRI Model per- spective, three potential mechanisms could explain this difference in correlations. First, it is possible that all of the prediction of externalizing personality disorders stems from the Trait factor (paths for the Identity and Reputation factors equal zero), and observer-reports are simply better indicators of the Trait factor (i.e., the Trait factor loadings are stronger for observer-reports than for self-reports for conscientiousness). Second, self- and observer-

reports could load equally on the Trait factor, but the unique insights that observers’ have of targets’ conscientiousness may supplement prediction from the general Trait factor (i.e., a negative path for the Reputation factor in predicting externalizing disor- ders). Finally, it is possible that the Identity factor also predicts externalizing disorders, but in the opposite direction of the Trait factor (i.e., a positive path for the Reputation factor but a negative path for the Trait factor).

We note that these three circumstances present substantially different implications and would lead to very different ap- proaches for assessing and treating at-risk patients. In the first instance (all prediction from the Trait factor, differentially assessed by self- and observer-reports), a general conscientious- ness trait alone is an important risk factor, though it may be somewhat better assessed by observers than by self-reports. In the second instance (supplemental prediction from the Reputa- tion factor), conscientiousness stands out as an important risk factor not only as a general trait but also as manifest in a social context, and therapies that focus on contextual triggers might be more effective. In the final instance (opposing prediction from the Identity factor), people’s misperception of their conscien- tiousness contributes to their externalizing problems, and ther- apies designed to bring self-perceptions more closely in line with reality may be helpful. We highlight these alternate inter- pretations not to argue that Carlson et al.’s (2013) conclusions are somehow inappropriate (after all, their methodology repre- sents a substantial advance in studying personality from a multirater perspective). Rather, our intention is to show the analytic advantages of examining the effects of latent factors rather than observed correlations. We next further discuss the Trait, Reputation, and Identity factors and align these with supporting literatures across personality and social psychology.

The Trait Factor: The Domain of Personality Trait Psychology

In capturing the shared variance across self- and observer- reports of a trait, the Trait factor represents what classic trait psychologists would conceptualize as “true trait variance.” Consensual validation (the correspondence across different rat- ers) has been a cornerstone of personality psychology since before even the appearance of the multitrait-multimethod ma- trix (Campbell & Fiske, 1959; e.g., Cleeton & Knight, 1924). The shared variance across raters captured in the Trait factor corresponds directly to how many models of perception have defined “accuracy” (e.g., Funder & West, 1993; Kenny, Al- bright, Malloy, & Kashy, 1994). For researchers interested in studying the effects of a personality trait on a criterion, these can best be captured by soliciting ratings from self- and observer-reports and modeling the shared variance across sources in the Trait factor.

Socioanalytic Theory (Hogan, 1996; Hogan & Blickle, 2013) offers an alternate perspective on the meaning of the Trait factor. Hogan and colleagues have argued that when people complete self-reports, they are conveying a particular impression in service of their goals. Rather than viewing these self-reports as “contam- inated” with impression management, Socioanalytic Theory argues that self-reports are meaningful because they model how people will present themselves to the world. Thus, from a socioanalytic

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

571TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

perspective, the Trait factors’ shared variance between self- and observer-reports would not reflect an underlying psychological “trait” entity but a description of how a person has conveyed their desired reputation both to others and in their self-description. On the other hand, people may conceal or distort some of their self-views (captured in our Identity factor) or may convey other unintended impressions (captured in our Reputation factor). Thus, even for Socioanalytic Theory, which has less focus on traits as psychological entities, our TRI Model is useful for separating perceptions that are shared in the Trait factor from those that are unique in the Reputation and Identity factors.

The Identity Factor: The Domain of Research on “The Self”

Philosophers have long emphasized the meaning and impor- tance of self-knowledge, and scientific study of how self- knowledge is structured and formed dates to the earliest psy- chologists (e.g., James, 1890). The Identity factor captures self-perceptions that are unique to self-reports. In some cases, these unique perceptions may reflect relevant trait variance that observers are simply never exposed to. For instance, some people may successfully hide their depressive thoughts and feelings from casual or even intimate acquaintances, with such unobservable trait variance appearing in the Identity factor. Such “hidden knowl- edge” is more likely to be present in the Identity factor when traits are lower in visibility or when observers are less intimately ac- quainted with the target.

In other cases, self-perceptions may deviate from observers’ perceptions because of distortions in self-perceptions and self- descriptions. Robins and John (1997) provide three metaphors to describe how self-reports can be swayed from seeking accuracy. First, self-perceptions often take the form of “The Egoist.” That is, people tend to maintain positive self-impressions because they assume the best about themselves and overlook their flaws (e.g., Dunning et al., 2004). Second, even when people have accurate self-perceptions, they may respond to self-reports as “The Politi- cian.” That is, people can be motivated to convey a particular impression in describing themselves (Leary & Kowalski, 1990), particularly when the assessment context provides incentives to be perceived favorably or unfavorably (e.g., Viswesvaran & Ones, 1999). Finally, apart from general positive or negative biases, people may respond as “Consistency-Seekers.” Indeed, the self- verification literature has provided strong support that people attend to, seek out, and recall information that is consistent with their self-views, even when this information may be negative (e.g., Swann, 1981; Swann, Rentfrow, & Guinn, 2003). In a similar vein, Funder (1999) highlights that people may maintain consistency by viewing themselves through the prism of their intentions, to the neglect of the effects of their behaviors.

Theorizing and empirical scrutiny has pitted these metaphors against one another in representing “the” model of self- knowledge (e.g., Swann, 1990; Tetlock & Manstead, 1985). However, there are almost certainly individual differences in self- descriptions’ links to consistency, impression management, and self-enhancement, and in the extent to which people’s self- perceptions resultantly become distorted (John & Robins, 1994). In the context of our TRI Model, Identity operationalizes and measures these individual differences in concealed trait manifes-

tations and distorted self-perceptions by capturing residual vari- ance in self-reports after partialing out variance shared with observer-reports. Thus, our operationalization of Identity corre- sponds closely to research paradigms for operationalizing self- distortion from Allport’s (1937) classic definition of self-insight to Paulhus and John’s (1998) self-criterion residual and Kwan, John, Kenny, Bond, and Robins’ (2004) interpersonal approach to self- enhancement. However, the Identity factor of the TRI Model provides an outlet for studying self-enhancement’s outcomes and causes in tandem with the Trait and Reputation factors.

Considerable debate has emerged regarding the mental health implications of self-enhancement. Although inaccuracy in self- perception has long been thought to indicate maladjustment and psychopathology, Taylor and Brown (1988) proposed that pos- itive self-illusions may buffer people from the negative impacts of harsh feedback. This proposition has been contentious (e.g., Colvin & Block, 1994; Taylor & Brown, 1994), and ensuing empirical research has been contradictory. Kwan, John, Robins, and Kuang (2008) distinguished between operationalizations of self-enhancement (based on contrasting self-reports with ratings from others) versus social comparison (based on contrasting self- reports with ratings of others). Whereas social comparison tends to be positively associated with psychological adjustment, self- enhancement is associated with reduced resiliency and heightened defensiveness, hypersensitivity, and narcissism. Such relationships between self-enhancement and psychopathology have major im- plications for research that has used self-reports to link personality traits with psychopathology. Indeed, multirater research has shown discrepancies when predicting both physical and mental health from self- versus observer-reports (Carlson et al., 2013; Smith et al., 2008). Accordingly, adopting the lens of the TRI Model to this research question can shed light on how much these linkages are a byproduct of personality traits as generally perceived (Trait) and as self-enhanced (Identity).

The Identity factor also has implications for research linking personality traits to success in pursuing goals across a variety of fields. Indeed, personality traits have been implicated as major antecedents to how goals are chosen, pursued, and monitored (DeYoung, 2014; Van Egeren, 2009). But how we perceive our personality may also play a pivotal role in this process. Specifi- cally, how we evaluate our personality (and other characteristics) provides a basis of information about which goals we might be successful in pursuing (Dweck, 1990). For example, people who believe themselves to be agreeable may be more likely to take on jobs in customer service or step into leadership roles. However, misperceptions of agreeableness could set some people up for failure. In addition, pursuing goals also necessitates that people monitor how they are progressing. In this respect, misperceptions associated with self-enhancement may indicate that people’s ba- rometer is faulty for how their behavior is perceived. In line with these notions, self-enhancement of personality traits has been linked to lower performance at work, at school, and on tasks in the laboratory (Connelly & Hülsheger, 2012; Kwan et al., 2004). Thus, wherever personality traits are linked to success through pursuing goals, it may be fruitful to separate the underlying personality traits captured in the Trait factor from self-enhancement captured in Identity.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

572 MCABEE AND CONNELLY

The Reputation Factor: The Domain of Interpersonal Perception

Beyond the identity we form, the impression we convey to others consumes much of our attention and determines many of our life outcomes (Goffman, 1959; Hogan & Blickle, 2013; Leary, 1995). Indeed, the reputations people create can readily become divorced from their self-perceptions either through actors’ willful manipulation (a calculated impression) or through the inadvertent stains their behaviors leave on others’ perceptions (a secondary impression; Schneider, 1981). Such distinctions harken back to early disagreements about how personality should be defined (Craik, 2008). In particular, some scholars (e.g., May, 1932; Ver- non, 1933) argued that personality exists beyond a collection of trait properties residing “within” a person; in addition, personality is also a socially emergent phenomenon that exists in the impres- sions conveyed to others. Similar premises can be found in the symbolic interactionism movement prominent in philosophy and sociology, which derives meaning from shared social definitions emerging from interactions among people (e.g., Blumer, 1986; Mead, 1913). Historically, however, a social definition of person- ality has presented conceptual and methodological difficulties for the field (Allport, 1937): What if observers disagree, or are prone to biases in judgments, or are swayed by rumors and gossip, or simply have inadequate opportunity to observe characteristics of a target? As a result, personality psychology has generally evolved independent from research studying how people’s reputations are crafted and perceived by others, with few exceptions (e.g., Hogan, 1996).

More recently, however, personality researchers have shown increasing interest in the reputations we create among others and, in particular, how others might be better able to detect certain aspects of our personality than we can ourselves (Vazire, 2010). From this vantage point, self-reports alone provide an incomplete picture of a person, and observers often provide valuable insights into personality that are not captured by self-ratings (Vazire & Carlson, 2011; Vazire & Mehl, 2008). Indeed, the knowledge asymmetries perceived by others but not by oneself may be par- ticularly meaningful in determining myriad social outcomes. The TRI Model offers a mechanism for contrasting the perceptions of others with self-perceptions within the Reputation factor (particu- larly when modeled as the shared perceptions across multiple observers that are distinct from self-perceptions).

From an integrative perspective, the personality, impression management, and person perception literatures have generally highlighted four reasons why observers’ perceptions may system- atically diverge from self-perceptions. First, targets may be suc- cessful in managing the impressions they convey in order to create a particular reputation for themselves. For example, some people may actively hide their feelings of anxiety and stress from those around them in order to be perceived as poised and calm. Second, observers may perceive targets in a particular set of situations that solicits contextually bound manifestations of traits (Fournier, Mos- kowitz, & Zuroff, 2009; Mischel & Shoda, 1995). For example, a person may be sociable and outgoing in general, but they may be reserved and introverted around their peers at work. Third, such communal errors among observers may stem from physical ap- pearance stereotypes (e.g., Dion, Berscheid, & Walster, 1972; Srivastava, Guglielmo, & Beer, 2010). In particular, observers

often report more favorable trait judgments of targets higher in physical attractiveness, including higher ratings of social compe- tence and intelligence. Notably, however, these positive percep- tions of physically attractive targets often differ from the targets’ true standing on such traits (e.g., Feingold, 1992) and from the targets’ self-perceptions (e.g., Furnham, Badmin, & Sneade, 2002). Though the effects of stereotypes on trait perceptions dis- sipate with increasing acquaintance, they substantially impact even moderately well acquainted observers (Kenny, 2004). Finally, observers may communicate both accurate and inaccurate percep- tions of targets with one another (Kenny, 1991; Malloy, Agatstein, Yarlas, & Albright, 1997). Indeed, a body of research has demon- strated that reputations exist in a social network, with much of what we know of people stemming from intermediaries rather than from direct observations of targets (Craik, 2008).

Thus, a person’s reputation contains much information about a person that extends beyond trait perceptions afforded by observing a target’s behavior (Anderson & Shirako, 2008). Within the TRI Model, much of what is captured in the Reputation factor might classically be considered “error” by trait psychologists, but also includes trait relevant information not accessible to the self (see also Vazire, 2010). Critically, how a person is known reflects an important and stable element of who they are and holds important consequences for many of their life outcomes. Socioanalytic The- ory (Hogan, 1996; Hogan & Blickle, 2013) has consistently high- lighted how the reputations people create determine how they pursue the goals of getting ahead and getting along in social groups throughout the course of their lives. The Reputation factor reflects this component that is uniquely perceived by observers (see also Vazire, 2006, 2010). For researchers studying the effects of per- sonality on social value (including job performance), power, social influence, and friendship, variance reflected in the Reputation factor is likely to be especially important.

The Trait-Reputation-Identity Model: Three Demonstrations

In the remainder of our manuscript, we present three dem- onstrations of ways that the TRI Model can be applied to use multirater data to enhance our understanding of personality, its measurement, and its outcomes. Specifically, we compare how strongly Trait, Reputation, and Identity influence each Big Five trait (Demonstration 1), examine self-enhancement and self- presentation correlates of the Trait, Reputation, and Identity factors (Demonstration 2), and explore the source of gender differences in personality ratings (Demonstration 3). These demonstrations provide examples of how the TRI Model can be used to construct hypotheses about personality traits, how to collect and statistically model data in the TRI Model, and how results from this model can be interpreted. Finally, we conclude by discussing variations in how models can be fit to address alternate domains of research and limitations associated with the TRI Model.

Demonstration 1: Modeling Self-Observer Perceptions

The purpose of Demonstration 1 was to establish the factor structure of the Trait-Reputation-Identity Model. Using archival data, we tested confirmatory factor models designed to examine (a)

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

573TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

the covariance (consensus) across self- and observer-reports of personality (i.e., Trait), (b) information unique to the self (i.e., Identity), and (c) information that is common across multiple observers but orthogonal to that perceived solely by the self (i.e., Reputation) for each of the Big Five traits.

According to Vazire’s (2010) SOKA Model, observer reports should have their greatest advantage over self-reports of person- ality for traits that are (a) high in visibility and (b) low in evalu- ativeness. Note, however, that these qualities of personality traits have different implications for the importance of the Trait, Repu- tation, and Identity factors within the TRI Model. In the context of the TRI Model, the Trait factors for traits that are higher in visibility should explain a greater proportion of the variance across self- and observer-reports of these traits, as such behaviors are available to both the self and to observers. Research suggests that extraversion and conscientiousness are among the more visible of the Big Five (e.g., Funder & Dobroth, 1987; John & Robins, 1993). Thus, we would expect the Trait factor to account for more variance for these traits. In contrast, differences in the evaluative qualities of the Big Five are likely to be manifest in the Identity and Reputation factors, and thus should account for more variance in self- and observer-ratings of those traits that are higher evalu- ativeness, such as neuroticism and are lower in visibility, such as openness. Conceptually, traits that are lower in visibility and higher in evaluativeness should create greater distinctions between self-views and those of observers, resulting in a smaller proportion of trait variance shared across sources (i.e., Trait) and, conversely, more variance accounted for within a single rating source (i.e., Identity and Reputation).

Method

Participants. Data for the present demonstration were a sub- set of the Eugene-Springfield Community Sample (ESCS), a lon- gitudinal sample of home owners who completed a variety of personality inventories and related psychological measures begin- ning in 1993 (Goldberg, 1999; Goldberg et al., 2006). Self-report data were available 478 participants (targets) who completed the Big Five Inventory (BFI; Benet-Martínez & John, 1998; John & Srivastava, 1999), collected during the fall of 1998. The target sample was age-diverse (M � 49.24, SD � 12.76) and composed of slightly more than half (64.9%) males, with nearly all partici- pants (97.5%) being Caucasian. Three observers (N � 1,434) rated each of these 478 targets using the observer-report version of the BFI. Observers had roughly comparable demographics to targets (for age, M � 48.06, SD � 18.04, and 62.2% were female). In total, 2.4% of the observers described themselves as a “significant other” of the target, 21.8% as a “spouse,” 28.3% as a “friend,” 11.5% as a “coworker,” 28.3% as a “relative,” 1.2% as an “ac- quaintance,” and 7.6% as “other” or did not report their relation- ship with the target. Observers were randomly assigned into one of three groupings (Observers A, B, and C).

Measures. The personality inventory used for the current demonstration was the BFI (Benet-Martínez & John, 1998; John & Srivastava, 1999). The BFI includes 44 items reflecting the Big Five personality traits, ranging from eight to 10 items per scale. Participants in the target sample rated the BFI for how accurate each item was descriptive of him/herself (i.e., “I see myself as someone who . . .” [emphasis added]) using a 5-point Likert scale

(1 � extremely inaccurate, 5 � extremely accurate). Participants in the observer sample rated these same items; however, the instructions were revised refer to the target (“I see this person as someone who . . .” [emphasis added]). Example items include: “Has an assertive personality” (extraversion) and “Makes plans and follows through with them” (conscientiousness). Alphas for the target sample ranged from .81 (agreeableness) to .87 (extra- version), and from .84 (extraversion) to .88 (agreeableness) for the full observer-report sample. Descriptive statistics and scale inter- correlations are available as a supplemental document.

Analysis. We constructed latent variable models delineating shared versus unique rater variance across item ratings from one self-report and three observer-reports for each Big Five trait, respectively. Specifically, we compared the fit of (a) a bifactor model (Holzinger & Swineford, 1937; Reise, 2012) with items loading on both a general Trait factor and specific rater factors (see the right portion of Figure 1) to (b) a higher-order model (e.g., Markon, 2009; Rindskopf & Rose, 1988; Yung, Thissen, & McLeod, 1999) in which the Trait factor is a higher-order factor defined by lower-order Identity and Reputation factors. Finding comparable fit for the higher-order model to that of the bifactor model would suggest that the meaning of the Trait factor is isomorphic with that of the Reputation and Identity factors; con- versely, finding improvements in fit of the bifactor model beyond the higher-order model would suggest that these different patterns of loadings are important to model (e.g., Markon, 2009). Models were identified by fixing the variances for all latent factors to 1.0. All models were assessed using common rules-of-thumb for close model fit (e.g., CFI � .95, RMSEA � .06; Browne & Cudeck, 1993; Hu & Bentler, 1999; Kline, 2011).

A number of a priori constraints were placed on the models to reflect the interchangeable assignment of observers to roles (Ob- servers A, B, and C). Specifically, we constrained to equivalence (a) factor loadings from the Trait factor to parallel items across observers, (b) factor loadings from the observer factors to parallel items across observers, (c) intercepts and residual variances for parallel observer-report items across observers, (d) loadings from the second-order Reputation factor to the first-order observer- report factors, (e) factor variances for the observer-report factors, (f) residual correlations between the self- and observer-reports for like-items, and (g) residual correlations between all like-items across observers (Olsen & Kenny, 2006). Notably, correlations between the residuals of like-items across raters reflect consensus in perceiving item characteristics above and beyond consensus for the broader trait factor. All models were similarly specified to include relevant constraints where implied.

Note that the model implied �2, degrees of freedom, and esti- mates of close fit reported by standard statistical programs for the TRI Model are generally inappropriate for SEM models with interchangeable raters (Olsen & Kenny, 2006). As such, several adjustments to model fit indices are needed to account for the random assignment of observers in the TRI Model. For each TRI Model, we estimated an “interchangeable saturated” (I-SAT) model, for which all means, variances, and covariances across parallel observed indicators are constrained to equivalence for our interchangeable observer groups (see Olsen & Kenny, 2006, for technical details). Model fit estimates for the TRI Model of interest are then adjusted by subtracting the �2 value and associated degrees of freedom from the respective I-SAT model. Indices of

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

574 MCABEE AND CONNELLY

close model fit are then recalculated using these adjusted values. Model fit estimates presented for all analyses were subject to these adjustments.

Results and Discussion

Table 1 presents the model fit statistics and average standard- ized factor loadings for the bifactor and higher-order models for each the Big Five personality traits.1 Across all models, the fit of the bifactor model was superior to that of the higher-order model. With respect to close fit, the RMSEA and CFI values for all models demonstrated moderate to strong fit, with estimates well within the typical range for analyses of item-level personality data (see Hopwood & Donnellan, 2010). As such, we retained the bifactor TRI Models of the Big Five personality traits for all subsequent analyses.

Figure 2 presents the proportion of the explained variance accounted for by the Trait, Reputation, and Identity factors for the TRI Models of the Big Five. The calculations for Figure 2 are an extension of the Explained Common Variance (ECV) statistic (see Reise, 2012; Ten Berge & Sočan, 2004). ECV reflects the relative strength of a general factor relative to group factors and is a useful statistic for judging whether it is feasible to collapse multidimen- sional data into a unidimensional model (Rodriguez, Reise, & Haviland, 2015). In the present application, the percentage of variance attributable to the Trait factor is a direct analogue to ECV, and, when low to moderate, suggests that much can be gained by parsing out Identity and Reputation factors. We supplement the ECV by also calculating variance attributable to Identity, Reputa- tion, and Observer-Idiosyncrasy factors, which speak to the rela- tive strength of these subgroup factors in the model. Recall, however, that our model included three observer-reports but only a single self-report, which by itself would cause our ECV approach to favor Reputation variance over Identity variance. To facilitate comparisons between the strength of the Identity and Reputation factors, we averaged the factor loadings across observers and represented these observer loadings only once in the numerator and denominators of our ECV calculations. (Given the constraints imposed by TRI Model with interchangeable raters [Olsen & Kenny, 2006], these average loadings are equivalent to those obtained for a single observer.)

Figure 2 reveals that the Trait factor accounted for the largest proportion of the explained variance in the models for conscien- tiousness, agreeableness, neuroticism, and, most notably, extraver- sion, suggesting a strong degree of consensus between self- and observer-reports of these traits. This finding is consistent with research which indicates that extraversion and conscientiousness, in particular, tend to be rated most consistently across self- and observer-reports in comparison to the remainder of the Big Five, even at low levels of acquaintance (e.g., Connelly & Ones, 2010; Connolly, Kavanagh, & Viswesvaran, 2007; Kenny, 1994). Con- versely, the Trait factor for openness contributed a substantially smaller proportion of the explained variance across self- and observer-reports for this trait.

Relative to the Trait factors, the Identity factors generally ac- counted for a smaller proportion of the explained variance in the self- and observer-reports of the Big Five personality traits. Note, however, that the Identity factor for openness accounted for a greater proportion of the explained variance in the responses for

this trait than did the Identity factors for the remaining Big Five. These findings suggest that self-perceptions may play a compar- atively larger role for openness, consistent with past findings suggesting that aspects of this trait is among the least observable of the Big Five (Funder & Dobroth, 1987; John & Robins, 1993). Conversely, the Identity factors for extraversion, in particular, accounted for a markedly smaller proportion of the explained variance in self- and observer-reports of this trait. In general, Identity was somewhat stronger in traits lower in visibility, without strong ties to the traits’ evaluativeness. Thus, much of the differ- ence across Big Five traits’ composition of Identity variance is likely tied more to the presence of “hidden” trait information in the Identity factor than to misperceptions.

Reputation variance was relatively more constant across traits; the one exception to this pattern was for extraversion, where Reputation variance (like Identity variance) was quite small rela- tive to the strong Trait factor variance. With this set of informants (who are fairly well acquainted with the targets), the strength of the Reputation factor appears less tied to specific trait characteristics aside from the strong consensus present for extraversion. In gen- eral, though, the strength of the Reputation factor was commen- surate with the strength of the Identity factor. Thus, the communal views of observers that are independent of self-other consensus and self-views represents an important factor in how observers judge the targets’ behavior.

Demonstration 1 elaborated the development of the Trait- Reputation-Identity Model for self-and observer-ratings of the Big Five personality traits. The TRI Models consistently demonstrated superior fit compared with higher-order models for self- and observer-reports. Moreover, a variance decomposition of the Trait, Reputation, and Identity factors across the Big Five suggested that the Trait factor generally accounts for the most substantial pro- portions of the explained variance in self- and observer-reports of these traits, with implications for when self-observer consensus versus consensus among multiple observers may play a larger role in contributing to assessments of behavior across different trait dimensions. Nonetheless, our findings also show that much of personality perception lies in the eyes of the beholder. As a result, failing to separate Trait variance from Reputation and Identity (as is commonly done when single-source personality measures are used or when multirater data is simply aggregated) has the poten- tial to obscure subtler personality processes that are at play.

Although the TRI Models themselves are interesting to consider, our goal in designing this model was to examine the independent relationships between the Trait, Reputation, and Identity factors and external variables of interest. Thus, we next elaborate a series of models examining how the TRI Models for the Big Five

1 In fitting some models, an initial model resulted in one or more negative factor loadings for the items on the latent factors. To facilitate interpretation, these factor loadings were constrained to be greater than or equal to zero and reestimated in a subsequent model. The imposition of these constraints results in a trivially small increase in the model implied �2. Such constraints are likely to be appropriate when, for example, personality researchers are interested in the Trait, Reputation, and Identity factors as an impression of a particular FFM dimension. However, in the context of person perception, where researchers are more interested in the forces that may cause idiosyncratic perceptions, omitting these constraints may be more appropriate.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

575TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

personality traits relate to a number of external correlates to demonstrate the applicability of the model.

Demonstration 2: Relationships With External Variables

A fundamental question in self-observer personality research is: “Who knows what about a person?” (Carlson et al., 2013; Vazire, 2010; Vazire & Carlson, 2011). Indeed, many researchers have suggested that the convergence between self- and observer-reports represents the standard by which accuracy should be judged (e.g., Connolly et al., 2007). Yet, another means of assessing accuracy across self- and observer-reports of personality is to identify cri- teria for which self- and observer-ratings agree in terms of predic- tion (Funder, 1995), and research has shown that self- and observer-reports can and do coincide in their prediction of external outcomes (e.g., Kolar, Funder, & Colvin, 1996).

Although finding that self- and observer-reports of personality traits predict criteria similarly is useful for demonstrating accu- racy, there are numerous instances when the self or observers might demonstrate unique insight for prediction (Vazire, 2010). This understanding is marked by a recent shift toward studying the unique information provided by self- and observer-reports (e.g., Connelly & Hülsheger, 2012; Connelly & Ones, 2010; Oh, Wang, & Mount, 2011; Vazire & Mehl, 2008). The Trait-Reputation- Identity Model is designed for just such purposes: to tease apart prediction from consensual views of a targets’ personality traits from those held solely by the self, and from the consensual views

of observers that are independent of self-perceptions. To this end, the TRI Model lends itself to research on self-presentation.

Teasing Apart Self-Presentation With the Trait- Reputation-Identity Model

The nature and impact of self-presentation for self-reports of personality traits has received a great deal of attention in the literature (e.g., Birkeland, Manson, Kisamore, Brannick, & Smith, 2006; Li & Bagger, 2006; Ones, Viswesvaran, & Reiss, 1996). According to Leary and Kowalski (1990), self-presentation is motivated by three distinct but related sources: (a) the need to enhance self-esteem, (b) the need to establish and maintain desired identities, and (c) the desire to obtain rewards from others in the social environment, including the desire to be accurately perceived by others (Baumeister, 1982; Schlenker & Weigold, 1992). Al- though conceptually distinct, these three motives substantially overlap in that attempts at self-presentation intended to “obtain rewards from others are often those that raise self-esteem and establish desired identities as well” (Leary & Kowalski, 1990, p. 38). In the context of the TRI Model, enhancing self-esteem and establishing and maintaining identities should be reflected primar- ily within the Identity factor (Hogan, 1996; Hogan & Blickle, 2013). In contrast, obtaining rewards from others should be re- flected within the Reputation and the Trait factors, as behaviors that are designed to elicit responses from the external environment (vs. internal acts in support of the self-concept) are more likely to be observed by others. Thus, in Demonstration 2 we examine

Table 1 Model Fit Statistics for I-SAT Adjusted Trait-Reputation-Identity Models

Model �ADJ2 (dfADJ) CFI RMSEA SABIC

Standardized factor loadings

Self Observera

��T � �

I � �

T � �

O � �

R � �

H

Extraversion TRI model 379.56 (107) .956 .073 [.064, .082] 1736.624 .61 .21 .39 .50 .52 — Higher-order 476.57 (122) .943 .078 [.068, .087] 1788.698 — .67 — .63 .75 .92

Agreeableness TRI model 429.87 (143) .959 .065 [.056, .074] 2104.483 .51 .30 .25 .63 .54 — Higher-order 547.87 (160) .945 .071 [.062, .080] 2171.556 — .58 — .68 .62 .77

Conscientiousness TRI model 243.60 (143) .983 .038 [.028, .047] 1918.209 .53 .34 .25 .58 .54 — Higher-order 367.36 (160) .966 .052 [.043, .061] 1991.044 — .63 — .63 .63 .74

Neuroticism TRI model 337.22 (107) .965 .067 [.058, .076] 1694.291 .57 .30 .29 .62 .56 — Higher-order 436.32 (122) .952 .073 [.064, .082] 1748.454 — .64 — .69 .66 .81

Openness TRI model 628.49 (184) .944 .071 [.062, .080] 2653.602 .29 .55 .24 .56 .60 — Higher-order 1050.59 (203) .893 .094 [.083, .103] 3018.785 — .60 — .60 .64 .85

Note. N � 478. I-SAT � interchangeable saturated model (Olsen & Kenny, 2006). CFI � comparative fit index; RMSEA � root mean square error of approximation; SABIC � sample adjusted Bayesian information criterion; TRI Model � Trait-Reputation-Identity Model. The �ADJ2 statistic and dfADJ and other fit statistics based on the �2 and df have been adjusted by subtracting out the corresponding values estimated in the I-SAT model for each trait. ��T � average loading on Trait factor; �

� I � average loading on Identity factor; �

� O � average

loading on Observer factors; ��R � average loading on Reputation factor; � �

H � average loading on higher-order Trait factor (constrained to equivalence across Reputation and Identity). Model fit statistics and averaged factor loadings are presented for these revised models. a Factor loadings are constrained to be equal across observers. Negative factor loadings were constrained to be greater than zero for all TRI models.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

576 MCABEE AND CONNELLY

relations between the TRI Models of the Big Five and targets’ responses on self-reports of three traits associated with these alternative self-presentation motives: self-esteem (Rosenberg, 1965), impression management, and self-deceptive enhancement (Paulhus, 1984, 1991).

Self-esteem reflects a person’s fundamental appraisal of their self-worth, and is thus an essential component of the self-concept (Campbell, 1990; Rogers, 1981; Rosenberg, 1979). As such, self- esteem should primarily relate to self-perceptions as manifest within the Trait and Identity factors. Because self-esteem reflects an internal appraisal that may not fully be communicated to others through overt behaviors, we expected self-esteem would correlate positively with the Identity factors for extraversion, conscientious- ness, agreeableness and openness, and negatively with that for neuroticism (John & Robins, 1993; Judge, Locke, Durham, & Kluger, 1998). Note, however, that personality expression related to self-esteem has clear implications for social interaction (see, e.g., Leary & Baumeister, 2000; Leary, Tambor, Terdal, & Downs, 1995), and research has consistently demonstrated relationships between self-esteem and a number of the Big Five personality traits (Judge et al., 1998; Judge, Van Vianen, & De Pater, 2004). As such, consistent with meta-analytic findings from the broader trait literature (e.g., Robins, Tracy, Trzesniewski, Potter, & Gos- ling, 2001), we expected self-esteem to negatively correlate with the Trait factor for neuroticism but to positively correlate with the Trait factors the remaining Big Five traits. We did not, however, expect self-esteem to correlate with the Reputation factors across the Big Five.

Self-deceptive enhancement reflects the targets’ implicit mis- perceptions regarding their own standing on a personality trait (Paulhus, 1984, 1991), typically in a socially favorable direction (Diener & Diener, 1995; Li & Bagger, 2006). Such misperceptions may arise from distinctions between current and desired identities

(Baumeister, 1982; Leary & Kowalski, 1990; Schlenker & Wei- gold, 1992). This is consistent with Hogan’s (1991, 2007) assertion that identity captures information regarding a person’s future mo- tives, behaviors, and intentions. Similar to self-esteem, we ex- pected self-deceptive enhancement to primarily relate to the TRI Models for the Big Five personality traits through its relations with Trait and Identity. Specifically, we expected self-deceptive en- hancement to positively correlate with the Identity factors for conscientiousness and openness, and to negatively correlate with that for neuroticism (John & Robins, 1993). In addition to relations with Identity, there are reasons to expect correlations between self-deceptive enhancement and the Trait factors for each of the Big Five. Self-deceptive enhancement reflects the implicit ten- dency to view the self in an positive light (Paulhus, 1984, 1991), which not only is manifest in the targets’ responses to personality items, but also influences observable patterns of behavior (Baumeister, 1982; Kwang & Swann, 2010; Seih, Buhrmester, Lin, Huang, & Swann, 2013). Thus, consistent with past meta-analytic findings (e.g., Li & Bagger, 2006), we expected self-deceptive enhancement to positively correlate with the Trait factors for conscientiousness, extraversion, and—to a lesser extent—agree- ableness and openness, and to negatively correlate with that for neuroticism. In line with our predictions for self-esteem, we did not expect self-deceptive enhancement to correlate with the Rep- utation factors across the Big Five.

Finally, impression management reflects a person’s attempts to present themselves to others in a socially desirable manner (Leary & Kowalski, 1990; Paulhus, 1984, 1991). From a traditional lens, the selective presentation of oneself to others captured in impres- sion management should relate to the Reputation factor of traits that are higher in evaluativeness (positively with conscientiousness and agreeableness, and negatively with neuroticism; John & Rob- ins, 1993). From a socioanalytic perspective, however, personality

Figure 2. Proportion of the explained variance accounted for by the Trait, Reputation, and Identity factors for the Trait-Reputation-Identity (TRI) Models of the Big Five. Observer uniqueness represents the combined proportion of the explained variance accounted for by the observer factors that is not attributable to the Reputation factor. The proportion of explained variance accounted for by each factors were calculated as the ratio of the sum of the squared loadings for an individual factor to the sum of the squared loadings across all factors (see, e.g., Reise, 2012; Reise, Moore, & Haviland, 2010), omitting correlated residuals and residual variances. The proportion of explained variance accounted for by the Reputation factor was similarly calculated using the product of the squared loadings on the first-order observer factors and the higher-order Reputation factor. To facilitate interpretation, parameter estimates were averaged across observer factors, and calculations were based on these averaged estimates. Calculations were performed using unstandardized factor loadings.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

577TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

self-reports model the impression that people are likely to convey to others through their behavior (Hogan, 1996; Hogan & Blickle, 2013). Thus, to the extent that impression management among others is mirrored in self-reports, impression management may also relate to the Trait factors for conscientiousness, agreeableness, and neuroticism. Finally, people might manage their impression successfully in a self-report personality measure but might meet less success in swaying the perceptions of others. If such a dis- parity exists, we expected targets’ impression management to be positively related to the Identity factors for conscientiousness and agreeableness and to be negatively related to that for neuroticism.

Method

Participants. Data for the present demonstration were again obtained from the ESCS (Goldberg, 1999; Goldberg et al., 2006) as reported in Demonstration 1. Data were unavailable on the measures of self-esteem, impression management, and self- deceptive enhancement for 26 targets from the initial sample. Missing values for these participants were estimated using full information maximum likelihood procedures in Mplus version 7.2 (Muthén & Muthén, 1998-2012).

Measures. In addition to the self- and observer-reports on the BFI (John & Srivastava, 1999; Benet-Martínez & John, 1998) reported in Demonstration 1, targets’ responses on a number of self-report measures available in the ESCS (Goldberg et al., 2006) were obtained for the purposes of the present demonstration. All measures were completed in a low stakes setting, where there are minimal anticipated effects of intentional faking.

Self-esteem. The 10-item Rosenberg Self-Esteem scale (Rosenberg, 1965) was implemented as the measure of targets’ self-esteem for the present demonstration. Targets rated each item on a 5-point Likert scale of agreement (1 � strongly disagree, 5 � strongly agree). Example items include: “On the whole, I am satisfied with myself” and “I feel that I have a number of good qualities.” Alpha for the target sample was .87.

Impression management and self-deceptive enhancement. The 40-item Balanced Inventory of Desirable Responding (BIDR: Paulhus, 1991) was used as the measure of targets’ impression management and self-deceptive enhancement for the present dem- onstration. The BIDR includes two subscales, one for each quality

of socially desirable responding, with 20 items per subscale. Tar- gets rated each item on a 5-point Likert scale of agreement (1 � strongly disagree, 5 � strongly agree). Example items include: “I never take things that don’t belong to me” (impression manage- ment) and “My first impressions of people usually turn out to be right” (self-deceptive enhancement). Alphas for the target sample were .82 and .69 for impression management and self-deceptive enhancement, respectively.

Analysis. Analyses for Demonstration 2 extend from the Trait-Reputation-Identity Models of the Big Five personality traits presented in Demonstration 1 (see Table 1). Specifically, we examined five additional CFA models— one for each dimension of the Big Five. To each of these models we incorporated the targets’ responses on the measures of self-esteem, self-deceptive enhance- ment and impression management, where we examined the corre- lations between these external variables and the Trait, Reputation, and Identity factors for each of the Big Five, respectively.

The TRI portion of the model was specified using the model parameterization reported in Demonstration 1. Note, however, that we differed in our specification of the external correlates for these models. Specifically, self-esteem, impression management, and self-deceptive enhancement were specified as single-indicator la- tent variables with the targets’ averaged scale scores serving as the indicator for each trait. To identify these variables, factor variances for the latent traits were fixed to 1.0, and we fixed the indicator uniqueness to adjust for measurement error (see Brown, 2006). All external correlates were allowed to intercorrelate in each model.

Results and Discussion

Table 2 presents the model fit statistics and fully standardized latent correlations for the TRI Models of the Big Five personality traits with the targets’ self-reports of self-esteem, self-deceptive enhancement, and impression management. With respect to the latent correlations with external variables, a number of interesting patterns were observed. Largely consistent with past meta-analytic findings (e.g., Robins et al., 2001), self-esteem demonstrated sta- tistically significant and meaningful correlations with the majority of the Trait factors for the Big Five, the exception being openness. Specifically, self-esteem was positively correlated with the Trait factors for conscientiousness, extraversion, and agreeableness, and

Table 2 Model Fit Statistics and for TRI Models of the Big Five Personality Traits and Correlations With External Variables

Model �ADJ2 (dfADJ) CFI RMSEA

Standardized latent correlations with external variables

Self-esteem Self-deceptive enhancement

Impression management

T I R T I R T I R

Extraversion 490.16 (146) .947 .070 [.061, .079] .15 .32 .07 .11 .33 .03 �.03 �.01 .05 Agreeableness 480.65 (188) .960 .057 [.048, .066] .20 .21 �.12 .04 .10 �.10 .48 �.01 .02 Conscientiousness 321.25 (188) .979 .039 [.028, .048] .38 .29 .03 .42 .28 �.08 .30 .06 .06 Neuroticism 452.54 (146) .956 .066 [.057, .075] �.42 �.31 .02 �.47 �.26 .11 �.20 �.34 .00 Openness 699.86 (235) .943 .064 [.055, .073] �.04 .25 .04 �.17 .29 �.08 .03 .01 �.08

Note. N � 478. CFI � comparative fit index; RMSEA � root mean square error of approximation; T � Trait; I � Identity; R � Reputation. The �ADJ2

statistic and dfADJ and other fit statistics based on the � 2 and df have been adjusted by subtracting out the corresponding values estimated in the I-SAT model

for each trait. Correlations in bold are statistically significant, p � .05. Average correlations across models were r� � .64 for self-deceptive enhancement with self-esteem; .49 for self-deceptive enhancement with impression management; and .27 for self-esteem with impression management. Negative factor loadings were constrained to be greater than zero for all TRI models.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

578 MCABEE AND CONNELLY

negatively correlated with that for neuroticism. This finding sug- gests that the behaviors associated with self-esteem within the Big Five personality traits are communally observed (and reported) by both the self and by observers. Moreover, targets’ reports of self-esteem correlated with the Identity factors for each of the Big Five in the expected direction. Consistent with our expectations, self-esteem was weakly correlated with the Reputation factors across the Big Five.

Relations between self-deceptive enhancement and the TRI Models of the Big Five personality traits were generally consistent with our expectations. In particular, self-deceptive enhancement was positively related to the conscientiousness Trait factor and was negatively correlated with the Trait factor for neuroticism. Note, however, that self-deceptive enhancement was not meaningfully associated with the agreeableness Trait or Identity factors, or the extraversion Trait factor. Interestingly, self-deceptive enhance- ment was slightly negatively related to the openness Trait factor, contrary to our initial expectations. One potential explanation for this finding is that people who are higher in openness are likely more thoughtful and reflective than those lower in this trait, and may therefore be less prone to self-deception (see, e.g., Christian- sen, Wolcott-Burnam, Janovics, Burns, & Quirk, 2005). Turning to relations with the Identity factor, self-deceptive enhancement was positively correlated with the Identity factors for conscientiousness and openness, yet negatively correlated with that for neuroticism, consistent with our predictions. Unexpectedly, self-deceptive en- hancement was positively correlated with the extraversion Identity factor, but was not related to the extraversion Trait factor, as noted previously. Consistent with our expectations, self-deceptive en- hancement was not meaningfully associated with the Reputation factors across the Big Five.

Finally, impression management correlated with the Trait fac- tors for conscientiousness, agreeableness, and neuroticism in the expected direction (see Table 2). Beyond the Trait factors, impres- sion management was negatively associated with the Identity factor neuroticism in the expected direction; however, impression management was not associated with the Identity factors for agree- ableness and conscientiousness, contrary to our initial expecta- tions. Interestingly, the correlation observed for the Identity factor for neuroticism was stronger in magnitude than that observed for the Trait factor for this trait. This finding suggests that attempts at impression management with respect to neuroticism may have less of an impact on others’ perceptions of the target than on the targets’ own self-presentations. Notably, impression management was not meaningfully associated with the Reputation factors across the Big Five. On the one hand, these findings may indicate that Reputation reflects largely individual differences in “secondary impressions” that observers form rather than positive “calculated impressions” that conceptually align with our impression manage- ment scale (cf. Schneider, 1981). On the other hand, social desir- ability scales (like the impression management scale on the BIDR) have been strenuously critiqued for being poor indicators of pre- senting a false impression (e.g., McCrae & Costa, 1983; Ones et al., 1996). Thus, an alternate interpretation may be that impression management scales are ineffective in capturing any unique ways people present themselves to others.

Taken together, findings for Demonstration 2 were generally consistent with our initial expectations. First, self-esteem and self-deceptive enhancement motives reflect both an internal and

external focus, as evidenced by the substantial correlations ob- served for these motives and the Identity and Trait factors for the Big Five. Second, impression management motives are primarily externally focused, as evidenced by the moderate to strong corre- lations observed with the Trait factors for more socially desirable traits than those correlations observed with the Identity factors for these same traits—although the inverse pattern was observed for the TRI Model for neuroticism. Importantly, findings indicate that each of the self-presentation motives examined (i.e., self-esteem, self-deceptive enhancement, impression management) simultane- ously serve both internal and external functions depending on the trait under examination, underscoring the vital interplay between the self-presentation motives manifest within self-reports of the Big Five and the behavioral manifestations of self-presentation that are observable by others. Having explored the nature of self-presentation within the TRI Models of the Big Five, we next turn to a final example in which the TRI Model might usefully lend itself to understanding origins of self- and observer-ratings of personality.

Demonstration 3: Effects of Gender on Self- and Observer-Ratings of Personality

Aside from its utility for examining relations with external correlates, the Trait-Reputation-Identity Model is useful more gen- erally when separating self-perceptions and communal observer- perceptions from general trait perceptions is desirable. One area in which such distinctions might be of interest is in research on gender differences in personality (e.g., Feingold, 1994; Schmitt, Realo, Voracek, & Allik, 2008). The existence and magnitude of gender differences in personality has drawn considerable attention over the past several decades (e.g., Costa, Terracciano, & McCrae, 2001; Schmitt et al., 2008). Gender differences are most consis- tently observed and largest in magnitude for neuroticism and agreeableness, such that women generally score higher on agree- ableness and neuroticism for both self- (e.g., Costa et al., 2001) and observer-reports (e.g., McCrae, Terracciano, & the Personality Profiles of Cultures Project, 2005).

But whence do gender differences in personality ratings emerge? One potential explanation for differences in personality trait expression between genders relates to evolutionary distinc- tions in behavior and gender roles (e.g., Buss, 1995; MacDonald, 1995; Wood & Eagly, 2002). Specifically, differences in the ex- pression of certain traits (primarily those related to kin altruism; Ashton & Lee, 2007) might result from the societal roles assumed by men and women and differences in reproductive strategies between genders. Although such evolutionary distinctions between genders are viable and reasonable contenders underlying differ- ences in personality trait expression, there are other potential explanations for why men and women might report differences in their personality traits that are related to their gender. For example, women might feel that they need to overrepresent their standing on measures of warmth and nurturance to avoid disapproval from others (Heilman & Chen, 2005; Heilman & Okimoto, 2007). Of course, men and women may or may not be aware of such attempts at stereotype-consistent self-presentation.

Beyond differences in self-presentation of personality traits, observers might in fact perceive behaviors differently depending on the gender of the target (Löckenhoff et al., 2014). Such differ-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

579TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

ences in trait perceptions can stem from within the target and/or from within the observer. For example, women tend to be more emotionally expressive than men (Kring & Gordon, 1998), engage in more social interaction, and demonstrate less social anxiety (Feingold, 1994). Thus, it may be easier to judge women on certain traits than it is to judge men on those same traits (e.g., neuroticism; see Funder, 1995; Human & Biesanz, 2013). Moreover, the same behaviors might be interpreted differently depending if the actor is male or female. For example, a study by Abbey (1982) found that males were much more likely to judge acts of friendliness by female targets as signifying sexual interest than were females judging acts of friendliness by male targets.

Given these alternative explanations, it is useful to distinguish when gender differences might arise from mechanisms unique to the self, the stereotypical and idiosyncratic attributions of others, or differences that are congruent across self- and observer-reports of these traits. Thus, as an illustration of how the Trait, Reputation, and Identity factors can be separated as outcomes, Demonstration 3 examines relationships between the TRI Model for the Big Five personality traits and the gender of the target. Consistent with past findings (e.g., McCrae et al., 2005; Schmitt et al., 2008), we expected to observe differences between genders for judgments of neuroticism and agreeableness; however, we examine each of the Big Five personality traits, as the distinction between Trait, Rep- utation, and Identity factors might illuminate differences in per- sonality ratings contingent on the targets’ gender that have yet to be discovered.

Method

Participants. Data for Demonstration 3 were once again taken from the ESCS (Goldberg, 1999; Goldberg et al., 2006) sample reported in Demonstration 1. As previously noted, slightly more than half (64.9%) of the target sample was male.

Measures. Once again, the self- and observer-reports from the BFI (John & Srivastava, 1999; Benet-Martínez & John, 1998) were used for the present demonstration. In addition, the targets’ gender was included in the analysis.

Analysis. Analyses for Demonstration 3 once again utilize the TRI Models of the Big Five personality traits presented in Dem- onstration 1 (see Table 1). Specifically, we regressed the Trait, Reputation, and Identity factors for each of the Big Five traits on the targets’ gender, respectively. Gender was treated as a dichot- omous observed variable, and path coefficients between gender and the personality factors were standardized for personality fac- tors only to the reflect the number of standard deviations in the TRI Model factors between men and women (reported as “STDY” in Mplus; see Muthén & Muthén, 1998-2012).

Results and Discussion

Table 3 presents the model fit statistics and standardized regres- sion coefficients for the Trait-Reputation-Identity Models of the Big Five personality traits regressed on target gender. Turning to the path coefficients, note that a positive coefficient reflects higher mean-levels of trait expression on a given factor for women, whereas a negative coefficient reflects higher mean-levels of trait expression on a given factor for men. In general, differences between genders on the Trait, Reputation, and Identity factors were not large across the Big Five. However, several interesting findings did emerge. First, and unexpectedly, men had higher Reputation factors for neuroticism than did women. This finding is particularly interesting given the findings of higher levels of neu- roticism reported in the literature for women using both self- and observer-reports of this trait (e.g., Costa et al., 2001; McCrae et al., 2005). Why then might observers rate men as higher in neuroti- cism apart from the consensus between self- and observer-reports of these traits? One potential explanation is the differences ob- served between men and women with respect to social interaction: Specifically, women typically score higher on measures of social competence and lower on measures of social anxiety (Feingold, 1994). Because Reputation taps aspects of social interaction of which the target is unaware, observers might perceive men as less socially competent and more socially anxious, thereby influencing the unique variance captured by observer reports in ways that are not captured across rating sources.

Table 3 Model Fit Statistics and Standardized Path Coefficients for TRI MIMIC Models of the Big Five Regressed on Gender

Model �ADJ2 (dfADJ) CFI RMSEA

Standardized path coefficientsa

Trait Identity Reputation

Extraversion 403.91 (120) .954 .070 [.061, .079] .05 �.12 .18 Agreeableness 458.17 (158) .957 .063 [.054, .072] �.08 �.04 .30 Conscientiousness 261.89 (158) .983 .037 [.026, .046] .02 �.24† �.04 Neuroticism 355.71 (120) .964 .064 [.055, .073] �.02 .10 �.38 Openness 646.82 (201) .944 .068 [.059, .077] �.11 �.09 .17

Note. N � 478. TRI � Trait-Reputation-Identity; MIMIC � multiple indicator multiple causes; CFI � comparative fit index; RMSEA � root mean square error of approximation. The �ADJ2 statistic and dfADJ and other fit statistics based on the �2 and df have been adjusted by subtracting out the corresponding values estimated in the I-SAT model for each trait. Gender coded 0 � male, 1 � female. Correlations in bold are statistically significant, p � .05; †p � .06. Negative factor loadings were constrained to be greater than zero for all TRI models. a Model estimates are standardized for endogenous variables only (i.e., the Trait, Reputation, and Identity factors) as reported in MPlus STDY output (Muthén & Muthén, 1998 –2012). Thus, these parameter estimates can be interpreted analogously to standardized d-values on Trait, Reputation, and Identity factors between men and women.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

580 MCABEE AND CONNELLY

Beyond Neuroticism, two other gender differences are worth mentioning. First, there was a statistically significant relationship between the target’s gender and the agreeableness Reputation factor, such that observers viewed the target as more agreeable if the target was female, independent of the consensus with self- ratings of agreeableness. This finding is consistent with the broader literature on gender differences in personality, where women generally rate themselves and are rated higher than men on this trait (e.g., Costa et al., 2001; Feingold, 1994; McCrae et al., 2005). Second, we observed a modest, though not statistically significant, relationship between the targets’ gender and the Iden- tity factor for conscientiousness, such that men view themselves as more conscientious than do observers. This finding is largely consistent with past research indicating gender differences in the conscientiousness facet of competence (e.g., Costa et al., 2001), though this gender difference may be an artifact of self-perception rather than a consensually substantiated gender difference in the trait. Taken together, findings for gender differences in the Trait, Reputation, and Identity factors for the Big Five personality traits indicate the importance of understanding social effects in self- and observer-reports of personality.

General Discussion

In the present paper we described the development of the Trait-Reputation-Identity Model of self- and observer-reports of personality traits. Building on the framework of the Johari window (Luft & Ingham, 1955), and extending from Vazire’s (2010) SOKA Model and Socioanalytic Theory (e.g., Hogan, 1996; Ho- gan & Blickle, 2013), the TRI Model decomposes self- and observer-ratings of personality into three components: Trait, Rep- utation, and Identity.

In Demonstration 1, we illustrated how TRI Models can be estimated and we assessed the proportion of the variance explained in self- and observer-reports of the Big Five personality traits that was accounted for by the Trait, Reputation, and Identity factors using a large archival dataset (Goldberg et al., 2006). In general, the (bifactor) TRI Models demonstrated superior fit to the data relative to higher-order models of the Big Five personality traits for multirater data. Furthermore, the Trait factor accounted for the greatest proportion of the explained variance for the majority of the Big Five personality traits. This is particularly notable because some of the variance captured in the Identity and Reputation factors likely reflects substantive “trait” characteristics that are not available to be consensually viewed across raters (e.g., unex- pressed thoughts and feelings assessed in the Identity factor). The strength of the Trait factor was more pronounced for Big Five dimensions that are higher in observability (e.g., extraversion; Funder & Dobroth, 1987; John & Robins, 1993), whereas the Identity factor accounted for greater proportions of the explained variance in self- and observer-reports for traits that are lower in observability (notably, openness). Though personality psycholo- gists can take comfort in the finding that most variance in mea- sures can be ascribed to trait factors (with some additional sub- stantive trait variance additionally lurking within Reputation and Identity), it is also clear that the Reputation and Identity factors account for a substantial portion of variance in personality mea- sures.

In Demonstration 2 we examined how the Trait, Reputation, and Identity factors can be distilled in relating personality to other constructs. Our analysis focused on three distinct but related mo- tives for self-presentation, namely the need to enhance self-esteem, the need to establish and maintain desired identities, and the desire to obtain material and social rewards from others (Baumeister, 1982; Leary & Kowalski, 1990; Schlenker & Weigold, 1992). These three motives were operationalized using measures of self- esteem (Rosenberg, 1965), self-deceptive enhancement and im- pression management (Paulhus, 1984, 1991), respectively. In most cases, self-esteem and self-deceptive enhancement correlated with the Identity and Trait factors for the Big Five, suggesting that behaviors associated with these self-presentation motives are not entirely internal but are indeed manifest in the social context (Leary & Kowalski, 1990). Impression management was related to the Trait factors for conscientiousness, neuroticism, and agreeable- ness. Moreover, impression management was related to the Iden- tity factor for neuroticism, suggesting the vital interplay between internal and external self-presentation acts for this trait (see also Kwang & Swann, 2010; Schlenker & Weigold, 1992). Interest- ingly, self-esteem, self-deceptive enhancement, and impression management did not meaningfully relate to the Reputation factors for any of the Big Five personality traits, suggesting that these external correlates might not meaningfully impact how observers communally perceive the target independent of the target’s at- tempts at self-presentation.

Finally, in Demonstration 3 we illustrated how the TRI Model could be used to study the impact of gender on self- and observer- ratings of the Big Five. Although most of these effects were small, a number of interesting suggestive effects were observed. Consis- tent with past findings for self- (e.g., Costa et al., 2001; Schmitt et al., 2008) and observer-reports (e.g., Löckenhoff et al., 2014; McCrae et al., 2005) of the Big Five, women were generally perceived as higher than men on the Reputation factor for agree- ableness. Interestingly, and contrary to the typical pattern observed in the literature on gender differences in personality, men were generally perceived as higher in neuroticism then were women in terms of the Reputation factor for neuroticism, suggesting that observers might be more sensitive to differences in social anxiety between genders than are the targets themselves (e.g., Feingold, 1994). Though smaller in magnitude, there was an observed trend for men to perceive themselves as higher in conscientiousness, reflected in the Identity factor for this trait. These findings help direct researchers to attribute gender differences to differences in self-perception, self-presentation, and/or social influences (e.g., gender stereotypes). These different attributions tell markedly dif- ferent stories about how and where gender differences are likely to emerge. Without disaggregating the Reputation and Trait factors in observer reports, such distinctions are likely to be missed.

The purpose of the present manuscript was to demonstrate some of the many different ways in which the TRI Model can be applied to personality research. The examples provided are by no means intended to be exhaustive; rather they are designed to establish the basis for the TRI Model and to demonstrate its applicability for a wide range of topics in personality research. In addition, for person perception researchers, the TRI Model could readily be applied to study consensus and discrepancies in perceiving nearly any char- acteristic within or beyond personality (abilities, attractiveness,

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

581TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

values, etc.). We next discuss a number of potential extensions to the TRI Model presented here.

Extending the Trait-Reputation-Identity Model

Alternate designs. Thus far, our discussion of the TRI Model has focused on designs in which observer-raters are nested within targets (e.g., Peter rates himself and is rated by his three friends Ray, Winston, and Egon, who rate no one else in the study). Such designs are common when personality psychologists use a multi- rater design to study how personality aligns with other constructs because it generally allows researchers to obtain observer-reports from closely acquainted informants. However, researchers (partic- ularly those studying person perception) also frequently adopt other designs for soliciting observer-reports. For example, in a block design, a set of targets is rated by a constant set of observers (e.g., a team of 6 trained undergraduates observes a videotape of each target behaving in the lab and rates the personality of every- one in the sample); in a round-robin design, groups of subjects both rate everyone else in the group and are rated by everyone else in the group (e.g., Peter, Ray, Winston, and Egon all rate them- selves and one another; Kenny, 1996). Such designs are common among other models of interpersonal perception, such as the Weighted Accuracy Model (WAM; Kenny, 1991), the Social Relations Model (SRM; e.g., Back & Kenny, 2010; Kenny, 1994; Malloy & Kenny, 1986), the PERSON Model (Kenny, 2004), and the Truth and Bias Model (West & Kenny, 2011). Although potentially less powerful (Kenny, 1996), nested rater designs are often more practical and are less burdensome to implement. Spe- cifically, the use of round-robin and similar designs is often not feasible when also collecting criterion measures or in applied research contexts, such as research on clinical assessment or per- sonnel selection. Thus, researchers should select the most appro- priate design (e.g., nested, block, round robin) for the purposes of their research (Funder, 1999; Kenny, 1994).

The TRI Model is flexible in its ability to accommodate block and round-robin designs. For a block design, raters would not necessarily be treated as interchangeable in the model. Thus, researchers could contrast parameters estimates across raters as a mechanism for studying person perception processes (e.g., per- ceiver effects could be studied by comparing the latent means of each rater method factor; “good judge” effects could be repre- sented in contrasting raters’ factor loadings on the Trait factor). Round-robin designs would necessitate somewhat more complex modeling using multilevel SEM, wherein participants at level 1 would be modeled as in the block design but nested within their round-robin group (e.g., round-robin tetrads) at level 2. Such a design would be powerful not only for isolating person perception processes as in the block design but also for studying how group- level variables impact TRI parameters (e.g., do work teams who are more cohesive have higher reputations for members’ agree- ableness?). However, the increased complexity of estimating ad- ditional parameters for block and round-robin designs would likely necessitate commensurately larger sample sizes. Nonetheless, the TRI Model can be a useful tool to separate shared versus rater- specific variance in these alternate designs that are more common in person perception research.

As a related issue, researchers should also be mindful of how informants are selected for targets. When left to nominate their

own informants, targets typically choose close friends and family members, who not only are likely to know targets well but also to have more positive views of targets’ personality traits. However, research suggests that there is a trade-off in selecting raters be- tween knowing targets well (which promotes accuracy; Connelly & Ones, 2010) and liking targets (which inhibits informants’ ability to make distinctions; Leising, Erbs, & Fritz, 2010). Relat- edly, we affiliate more closely with those who affirm our self- perceptions (e.g., Swann, 1984, 2012). Accordingly, the strength of the Trait, Reputation, and Identity factors may depend in part on the freedom afforded targets in nominating their informants.

We have generally presumed that personality assessments will be derived from targets and observers completing questionnaires. However, personality is manifest in many subtle ways (e.g., Gos- ling, 2009), and with the “big data” movement researchers have begun developing actuarial mechanisms to distill personality as- sessments from more objectively assessed cues (e.g., Kosinski, Stillwell, & Graepel, 2013; Youyou, Kosinski, & Stillwell, 2015). Our model could be applied in such cases to separate out Traits, Reputation, and Identity components even when personality as- sessments come from life-history data rather than raters’ percep- tions. Such an approach can be useful to separate out a person’s traits and identity from, for example, the digital persona they create.

Multiwave and longitudinal assessments. Our three exam- ples treated the Trait, Reputation, and Identity factors as static entities measured at a single point in time. However, the TRI Model could also be applied to multiwave or longitudinal data to great effect in several ways. Lurking within each rater factor is likely some degree of transient measurement error. In line with discussions of common method variance (see Kammeyer-Mueller, Steel, & Rubenstein, 2010; Lance, Dawson, Birkelbach, & Hoff- man, 2010; Spector, 2006), such transient error is most problem- atic when predictors or outcomes of the Trait, Reputation, and Identity factors are measured at the same time by a single rater. These concerns can be abated more easily for observers by spec- ifying a Reputation factor defined by multiple raters. Because there is only a single self-rater for a given target, however, multiwave assessments of personality offer a useful mechanism for disentan- gling transient error from truly unique self-perceptions in terms of Identity (e.g., LaGrange & Cole, 2008; see Figure 3 for one potential model). Such a model would likely be most appropriate when the interval between assessments is relatively short (e.g., two weeks) and when substantive change in traits, reputation, and identity is unlikely to occur.

In other cases, however, researchers might adopt multiwave assessments to specifically study substantive change in personality scores over time. Infusing multirater measurement within this multiwave assessment can clarify whether changes in personality scores reflect changes in the Trait, Reputation, and/or Identity factors (and whether correlates of personality change are tied to these factors). Marrying the TRI Model with latent growth models (e.g., Cole, Martin, & Steiger, 2005; Courvoisier, Nussbeck, Eid, Geiser, & Cole, 2008) or multilevel SEM (e.g., Mehta & Neale, 2005) offers a useful mechanism for studying such stability and change in personality from a multirater perspective. For example, researchers could specify both latent intercepts and latent slopes for Trait, Reputation, and Identity components of the TRI Model to

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

582 MCABEE AND CONNELLY

study, for example, whether men and women show different tra- jectories for these components within the neuroticism domain.

Acquaintanceship contexts. Until now, we have generally neglected the importance of the context through which raters are acquainted for with the target. That is, Reputation is defined by the commonality across observer ratings independent of the specific context(s) in which the observer and the target typically interact. However, the reputations that targets create among family, friends, coworkers, and complete strangers are related but still distinct (Funder, 1995). These context effects appear to be most strong among less intimately acquainted observers (Geeza, Connelly, & Chang, 2010). Thus, when finding effects associated with the Reputation factor, researchers may be particularly interested to know whether the effects stem from context-specific manifesta- tions of personality or from a more general way that targets present themselves. Indeed, a substantial body of personality theory about person-situation interactions suggests that these context-specific manifestations are an important signature of personality itself (Fleeson & Jayawickreme, 2015; Fournier et al., 2009; Mischel & Shoda, 1995). In this way, collecting data from observers across different contexts offers a particularly useful research method for representing situation-specific personality profiles. In addition, obtaining observers from multiple contexts may offer a way to mitigate possible communication between observers (e.g., gossip) that could artificially inflate Reputation variance (Kenny, 2004). Similarly, including raters of differing levels of acquaintance may provide further insights into the effects of acquaintanceship on observer reports (e.g., Funder & Colvin, 1988; Funder et al., 1995).

Our TRI Model provides a framework with which to formally separate context-specific standing on traits from the general trait (Trait), self-perceptions (Identity), and the distinct reputation an individual creates (Reputation). Figure 4 depicts how context effects could be separated in a design that has collected observer- ratings across multiple contexts (e.g., friendship, school, and work- place contexts). Note that such a design would allow researchers to

ask questions like, “How distinct are trait manifestations in the workplace (strength of factor loadings on the workplace context factor)?” and “Is job performance more determined by an employ- ee’s general standing on conscientiousness (path for Trait) or by adaptive responses of conscientiousness to the demands of the workplace (path for the workplace context)?” Questions such as these are central to understanding how much personality states shift in response to general situations defined by an acquaintance- ship context and how important these contextual versus general manifestations are in determining life outcomes. Even if the ac- quaintanceship context is not an explicit focus of substantive inquiry, researchers should be sensitive to how context could affect whether perceptions are shared or unique across raters. For exam- ple, a design soliciting personality ratings from self-reports, moth- ers, fathers, and a roommate should account for the shared context between mothers and fathers by correlating these raters’ method factor disturbances.

A related stream of research has examined self-reports’ ability to distinguish context-specific personality manifestations. Specif- ically, research using “frame-of-reference” approaches to person- ality adds contextual delimitations to typical self-report items (e.g., “I am outgoing at school”; Schmit, Ryan, Stierwalt, & Powell, 1995; Woo, Jin, & LeBreton, 2015). Frame-of-reference ap- proaches appear to generally improve personality measures’ reli- ability and predictive validity for in-context criteria (Lievens, De Corte, & Schollaert, 2008), though they do not necessarily improve consensus with context-consistent observers (Kurtz & Palfrey, 2016; Small & Diefendorff, 2006). Similarly, research on meta- perceptions asks self-reports to predict how observers from a particular context will describe their personality (Kenny & De- Paulo, 1993). Though metaperceptions correlate strongly with general self-reports, research suggests that people are sensitive to the distinct impressions they make on others (Carlson, Vazire, & Furr, 2011). Researchers in these areas could fruitfully supple- ment the model in Figure 4 with frame-of-reference measures or metaperception measures. Such a model would allow research- ers to estimate the sensitivity of frame-of-reference/metaper- ceptions to context effects while controlling for the effects of

Figure 4. Incorporating multiple rater contexts in the Trait-Reputation- Identity Model.

Figure 3. Incorporating multiwave self-report data in the Trait- Reputation-Identity Model.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

583TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

general traits (Trait) and self-perception biases (Identity). In addition, the TRI Model provides a mechanism for examining outcomes associated with systematic tendencies to over- or underestimate one’s standing on frame-of-reference and meta- perception measures.

Stereotypes. Undoubtedly, one component of Reputation as we have thus far modeled it is stereotypes commonly held across observers. We would expect that stereotypes would be less influ- ential in the Reputation factor as intimacy and opportunity to observe increase (Kenny, 2004). However, beyond the kernel of truth that may lie in stereotypes (Penton-Voak, Pound, Little, & Perrett, 2006), many social and personality psychologists alike would find it useful to distinguish between reputation created through observing a target’s behaviors and that derived simply from physical appearance cues.

One methodological approach to achieve such separation would be to supplement our base TRI Model with observer-reports from strangers viewing only a limited set of cues (e.g., for stereotypes based on physical appearance, observer-ratings of personality after viewing a neutral photograph). Specifically, researchers could additionally specify a “trait stereotype” factor defined by both acquainted observers and multiple stereotype-based strangers’ rat- ings (see Figure 5). Such a factor would have the effect of separating out trait stereotypes that are distinct from the Trait and Reputation factors. Stereotype factors could thus be used to study (a) what affects observers’ perceptions and (b) what impact trait- based stereotypes may have on life outcomes. Parallel approaches could be adopted for other sources of bias potentially affecting observer-reports.

Potential Applications of the Trait-Reputation-Identity Model

Work and education. Personality measures have long been used to predict success in the workplace among job applicants (Zickar & Kostek, 2013) and are increasingly being expanded for use in educational admissions (e.g., Kyllonen, 2008). However, recent findings suggesting that observer-reports are better predic- tors of job and academic performance (Connelly & Ones, 2010; Oh et al., 2011) has raised many questions about the shared and unique insights of self- and observer-reports (Connelly & Hülsheger, 2012). The TRI Model offers a viable research mechanism for understanding the behavioral outcomes of these insights. Opera- tionally, this model could be extended to provide Trait, Reputation, and Identity scores for job or university applicants.

Additionally, the past two decades have produced a surge of research on the workplace consequences of misperceptions, whether reflected in performance appraisals (e.g., Atwater, Os- troff, Yammarino, & Fleenor, 1998), task simulations (Atkins & Wood, 2002), or leadership (Fleenor, Smither, Atwater, Braddy, & Sturm, 2010). Thus, providing multisource feedback (e.g., 360- degree performance feedback) has become a major business en- terprise aimed at bringing employees’ self-perceptions more closely in line with others (e.g., Woo, Sims, Rupp, & Gibbons, 2008). However, this movement has not generally been extended to personality measurement aimed at employee development, which remains dominated by reliance on only self-report measures. The distinct perspectives of rater groups likely has substantial ecological value (e.g., Hoffman, Lance, Bynum, & Gentry, 2010; Hoffman & Woehr, 2009). Thus, the TRI Model offers a mecha- nism for bringing counseling psychology’s historic multirater ap- proach of the Johari window in line with contemporary organiza- tional personality feedback practices.

Health, wellness, and adjustment. Multirater assessment has made recent strides in diagnosing clinical symptoms in general and in assessing personality disorders in particular (for a discussion, see Oltmanns & Turkheimer, 2009). For example, diagnostic in- formation gleaned from self-reports may be combined with clini- cians’ ratings and with informant-reports. Although the informa- tion in these rating sources overlap, it is also clear that each source contains unique perspectives on clients. How to best make use of these shared and unique perspectives remains an important out- standing question for both researchers and clinicians. The TRI Model offers a valuable tool for assessing such effects directly. Drawing from Oltmanns and Turkheimer (2009), Identity variance may incrementally predict internalizing problems, whereas Trait and/or Reputation variance may provide useful prediction for externalizing problems.

Researchers have also been long-interested in the effects of personality and self-enhancement on well-being (DeNeve & Coo- per, 1998; Taylor & Brown, 1994), with these effects often being difficult to disentangle. Kwan et al.’s (2004, 2008) componential model (separating self-enhancement, social comparisons, and un- derlying traits) has brought resolution to seemingly incompatible views of the effects of self-enhancement on well-being. To this model, the TRI Model adds avenues for studying the effects of reputation on well-being through the Reputation factor. Particu- larly in highly collectivist cultures, targets’ happiness may be tied less to underlying personality traits (Trait) or to self-enhancement

Figure 5. Disentangling stereotype effects using the Trait-Reputation- Identity Model.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

584 MCABEE AND CONNELLY

(Identity), yet more to how peers construe targets’ personalities (Reputation).

Genetic and developmental antecedents. In many ways, be- havioral genetics and developmental research has led personality in using multirater assessments. For instance, the German Obser- vational Study of Adult Twins collected self-reports, multiple informant-reports, experimenter-ratings, confederate-ratings, and zero-acquaintance ratings on a sample of 300 adult twin pairs (Borkenau et al., 2004). In short, personality research projects of such scope often begin by ensuring that personality has been measured well. Research within this literature (Kandler, Bleidorn, et al., 2010; Kandler, Riemann, Spinath, & Angleitner, 2010) has already shown that (a) the unique information in both self- and peer-reports is genetically influenced and (b) unique self-report variance is somewhat stable over time but unique peer-report variance is not.

Such analyses map closely to the framework of the TRI Model and yet raise important questions. For example, what biological mechanisms might mediate genetic effects in propagating pheno- typic Identity and Reputation variance? Similarly, what life events might change one’s Identity or Reputation for a trait, beyond changes in the latent trait itself? Multirater assessments provide not only more reliable estimates of latent traits but also the op- portunities to study unique perspectives of raters.

Considerations

Perhaps the largest constraint on the application of the Trait- Reputation-Identity Model is the data collection requirements. The TRI Model as described requires researchers to not only obtain self-ratings of a target’s personality traits, but also ratings from multiple observers. Note that it is possible to estimate the TRI Model with just one observer in instances where collecting data from multiple observers is difficult or even impossible (e.g., rat- ings from spouses). However, when a single observer-report is used, the Reputation factor will be confounded with the idiosyn- cratic views of the rater. Although this is the case with self-reports in the Identity factor, these idiosyncratic self-views likely have much more relevance for a broader understanding of the target’s personality than do idiosyncratic observer-reports. Thus, when possible, we encourage researchers to gain a much cleaner measure of the Reputation factor by using multiple observer-reports and modeling the Reputation factor as the shared variance across them (see Vazire, 2006 for valuable suggestions on soliciting such multirater data).

Aside from the logistics involved in collecting data from mul- tiple raters per target, the model itself requires larger sample sizes to achieve parameter stability and appropriate estimation, a com- monly recognized feature of structural equation modeling (see, e.g., Chen, West, & Sousa, 2006). In the context of the TRI Model, sample size requirements grow with the number of additional observers included in the model (see, e.g., Marsh & Bailey, 1991, for recommendations in multitrait-multimethod research). Thus, future simulation research is needed to determine how well the TRI Model recovers model parameter estimates and subsequent model fit, and to detect meaningful associations with external correlates.

Another consideration for the TRI Model is its focus on one trait at a time, rather than general patterns that exist across traits. Thus,

although the extraction of factors within traits is meaningful by itself, there may be overlap in the external correlates one consid- ers. As observed in Demonstration 2, self-esteem and self- deceptive enhancement correlated with the Identity factors for extraversion, conscientiousness, openness, and neuroticism. Ex- amining these relations in isolation potentially biases the magni- tude of the correlations observed within each model. For instance, the correlation observed between the conscientiousness Identity factor and self-esteem might diminish after accounting for any overlap with neuroticism. Indeed, other models of rater consensus have recently explored how to marry multitrait approaches to accuracy (typical of profile correlations; see, e.g., Biesanz, 2010) with multitarget approaches to accuracy (typical for self-informant correlations; see, e.g., Human & Biesanz, 2013). Future large- sample research should consider the impact of overlap between personality traits to better clarify the source(s) of trait variance that constitute the personality Trait, Reputation, and Identity factors across the Big Five.

In a similar vein, multidimensionality in the indicators used within a Big Five domain can potentially complicate interpretation of the Trait, Reputation, and Identity factors. Indeed, it is widely accepted that each Big Five domain is a multifaceted construct comprising a number of narrower tendencies (e.g., extraversion comprises facets including sociability, assertiveness, positive emo- tionality, and energy; Costa & McCrae, 1995). Such multidimen- sionality could potentially bend the interpretation of Trait, Repu- tation, and Identity factors if consensus differs across the facets. For example, it is feasible that the Reputation factor for extraver- sion could come to largely reflect items from a highly observable facet like energy (which produces stronger rater consensus), whereas the Identity factor might be largely defined by a less- visible facet like positive emotions (which produces weaker rater consensus; Connelly, 2008). Such multidimensionality might sub- stantially change the interpretation of the TRI Model factors (e.g., in the above example, the presumed effect of Identity on self- esteem could be produced by the unique variance associated with the positive emotions facet rather than a person’s Identity for extraversion). When encountering such multidimensionality, we recommend researchers follow several steps. First, we recommend conducting an exploratory factor analysis of a trait’s indicators separately for self- and informant ratings to provide an initial appraisal of potential multidimensional structure. Second, we rec- ommend that authors inspect the pattern of factor loadings when fitting the model to ensure that the Identity and Reputation factors are not dominated by any particular narrow (facet) trait. Finally, if Identity or Reputation factors are dominated by one particular facet but the effects of consensus and discrepancies across raters remain of interest, the researchers could supplement the factors of the TRI Model with additional factors to model multidimension- ality of the item content. In the present three demonstrations, the loadings for Identity and Reputation factors did not reflect any discernible pattern suggestive of a predominance from particular narrower personality facets. Note, however, that items reflecting less observable content tended to have stronger loadings on both the Identity and Reputation factors than did items that with higher degrees of observability, underscoring the findings of Demonstra- tion 1. Moreover, this pattern of loadings is consistent with find- ings regarding trait observability from the broader personality literature (e.g., Funder & Dobroth, 1987; John & Robins, 1993;

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

585TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

Vazire, 2010). Nonetheless, greater exploration of potential mul- tidimensionality within the TRI Model is certainly warranted, particularly as it related to narrower facets of personality.

A final consideration for the TRI Model is the need for repli- cation. Findings for the present demonstrations are likely influ- enced by the characteristics of the sample and the specific mea- sures used. As such, future research should seek to replicate and extend the findings of the present demonstrations using diverse samples and alternative measures of the Big Five. In addition, our goal for the development of the TRI Model was to examine the contribution of the Trait, Reputation, and Identity factors for predicting external criteria. Future research should strive to include more objective, behaviorally based outcomes to inform our under- standing for how the components of the model relate to important life outcomes.

Conclusion

As an unfortunate side effect of the person-situation debate and the cognitive revolution, research on traits and the errors made in evaluating them was fractured into separate fields of personality, the self, impression management, and person perception. Indeed, the judgment errors, distinct perceptions, and separate personas that social– cognitive psychologists have meticulously documented affecting trait perceptions do not undermine personality research but merit research concomitant with traits. The Trait-Reputation- Identity Model offers a unified approach to studying individual differences in our underlying personality traits, our unique self- perceptions, and the reputations we create.

References

Abbey, A. (1982). Sex differences in attributions of friendly behavior: Do males misperceive females’ friendliness? Journal of Personality and Social Psychology, 42, 830 – 838. http://dx.doi.org/10.1037/0022-3514 .42.5.830

Allport, G. W. (1937). Personality: A psychological interpretation. New York, NY: Holt.

Anderson, C., & Shirako, A. (2008). Are individuals’ reputations related to their history of behavior? Journal of Personality and Social Psychology, 94, 320 –333. http://dx.doi.org/10.1037/0022-3514.94.2.320

Ashton, M. C., & Lee, K. (2007). Empirical, theoretical, and practical advantages of the HEXACO model of personality structure. Personality and Social Psychology Review, 11, 150 –166. http://dx.doi.org/10.1177/ 1088868306294907

Atkins, P. W. B., & Wood, R. E. (2002). Self- versus others’ ratings as predictors of assessment center ratings: Validation evidence for 360- degree feedback programs. Personnel Psychology, 55, 871–904. http:// dx.doi.org/10.1111/j.1744-6570.2002.tb00133.x

Atwater, L. E., Ostroff, C., Yammarino, F. J., & Fleenor, J. W. (1998). Self-other agreement: Does it really matter? Personnel Psychology, 51, 577–598. http://dx.doi.org/10.1111/j.1744-6570.1998.tb00252.x

Back, M. D., & Kenny, D. A. (2010). The social relations model: How to understand dyadic processes. Social and Personality Psychology Com- pass, 4, 855– 870. http://dx.doi.org/10.1111/j.1751-9004.2010.00303.x

Baumeister, R. F. (1982). A self-presentational view of social phenomena. Psychological Bulletin, 91, 3–26. http://dx.doi.org/10.1037/0033-2909 .91.1.3

Benet-Martínez, V., & John, O. P. (1998). Los Cinco Grandes across cultures and ethnic groups: Multitrait multimethod analyses of the Big Five in Spanish and English. Journal of Personality and Social Psychol- ogy, 75, 729 –750. http://dx.doi.org/10.1037/0022-3514.75.3.729

Benet-Martínez, V., & Oishi, S. (2008). Culture and personality. In O. P. John, R. W. Robins, & L. A. Pervin (Eds.), Handbook of personality: Theory and research (pp. 542–567). New York, NY: Gilford Press.

Biesanz, J. C. (2010). The social accuracy model of interpersonal percep- tion: Assessing individual differences in perceptive and expressive ac- curacy. Multivariate Behavioral Research, 45, 853– 885. http://dx.doi .org/10.1080/00273171.2010.519262

Birkeland, S. A., Manson, T. M., Kisamore, J. L., Brannick, M. T., & Smith, M. A. (2006). A meta-analytic investigation of job applicant faking on personality measures. International Journal of Selection and Assessment, 14, 317–335. http://dx.doi.org/10.1111/j.1468-2389.2006 .00354.x

Blumer, H. (1986). Symbolic interactionism: Perspective and method. Berkeley, CA: University of California Press.

Borkenau, P., Mauer, N., Riemann, R., Spinath, F. M., & Angleitner, A. (2004). Thin slices of behavior as cues of personality and intelligence. Journal of Personality and Social Psychology, 86, 599 – 614. http://dx .doi.org/10.1037/0022-3514.86.4.599

Brown, T. A. (2006). Confirmatory factor analysis for applied research. New York, NY: Guilford Press.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136 –162). Newbury Park, CA: Sage.

Buss, D. M. (1995). Evolutionary psychology: A new paradigm for psy- chological science. Psychological Inquiry, 6, 1–30. http://dx.doi.org/10 .1207/s15327965pli0601_1

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105. http://dx.doi.org/10.1037/h0046016

Campbell, J. D. (1990). Self-esteem and clarity of the self-concept. Journal of Personality and Social Psychology, 59, 538 –549. http://dx.doi.org/10 .1037/0022-3514.59.3.538

Carlson, E. N., Vazire, S., & Furr, R. M. (2011). Meta-insight: Do people really know how others see them? Journal of Personality and Social Psychology, 101, 831– 846. http://dx.doi.org/10.1037/a0024297

Carlson, E. N., Vazire, S., & Oltmanns, T. F. (2013). Self-other knowledge asymmetries in personality pathology. Journal of Personality, 81, 155– 170. http://dx.doi.org/10.1111/j.1467-6494.2012.00794.x

Chen, F. F., West, S. G., & Sousa, K. H. (2006). A comparison of bifactor and second-order models of quality of life. Multivariate Behavioral Research, 41, 189 –225. http://dx.doi.org/10.1207/s15327906mbr 4102_5

Christiansen, N. D., Wolcott-Burnam, S., Janovics, J. E., Burns, G. N., & Quirk, S. W. (2005). The good judge revisited: Individual differences in the accuracy of personality judgements. Human Performance, 18, 123– 149. http://dx.doi.org/10.1207/s15327043hup1802_2

Cleeton, G., & Knight, F. (1924). Validity of character judgments based on external criteria. Journal of Applied Psychology, 8, 215–231. http://dx .doi.org/10.1037/h0072525

Cole, D. A., Martin, N. C., & Steiger, J. H. (2005). Empirical and conceptual problems with longitudinal trait-state models: Introducing a trait-state-occasion model. Psychological Methods, 10, 3–20. http://dx .doi.org/10.1037/1082-989X.10.1.3

Colvin, C. R., & Block, J. (1994). Do positive illusions foster mental health? An examination of the Taylor and Brown formulation. Psycho- logical Bulletin, 116, 3–20. http://dx.doi.org/10.1037/0033-2909.116.1.3

Connelly, B. S. (2008). The reliability, convergence, and predictive valid- ity of personality ratings: An other perspective (Unpublished doctoral dissertation). University of Minnesota, Minneapolis, MN.

Connelly, B. S., & Hülsheger, U. R. (2012). A narrower scope or a clearer lens for personality? Examining sources of observers’ advantages over self-reports for predicting performance. Journal of Personality, 80, 603– 631. http://dx.doi.org/10.1111/j.1467-6494.2011.00744.x

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

586 MCABEE AND CONNELLY

Connelly, B. S., & Ones, D. S. (2010). An other perspective on personality: Meta-analytic integration of observers’ accuracy and predictive validity. Psychological Bulletin, 136, 1092–1122. http://dx.doi.org/10.1037/ a0021212

Connolly, J. J., Kavanagh, E. J., & Viswesvaran, C. (2007). The convergent validity between self and observer ratings of personality: A meta- analytic review. International Journal of Selection and Assessment, 15, 110 –117. http://dx.doi.org/10.1111/j.1468-2389.2007.00371.x

Costa, P. T., Jr., & McCrae, R. R. (1995). Domains and facets: Hierarchical personality assessment using the revised NEO Personality Inventory. Journal of Personality Assessment, 64, 21–50. http://dx.doi.org/10.1207/ s15327752jpa6401_2

Costa, P. T., Jr., Terracciano, A., & McCrae, R. R. (2001). Gender differences in personality traits across cultures: Robust and surprising findings. Journal of Personality and Social Psychology, 81, 322–331. http://dx.doi.org/10.1037/0022-3514.81.2.322

Courvoisier, D. S., Nussbeck, F. W., Eid, M., Geiser, C., & Cole, D. A. (2008). Analyzing the convergent and discriminant validity of states and traits: Development and applications of multimethod latent state-trait models. Psychological Assessment, 20, 270 –280. http://dx.doi.org/10 .1037/a0012812

Craik, K. H. (2008). Reputation: A network interpretation. New York, NY: Oxford University Press. http://dx.doi.org/10.1093/acprof:oso/ 9780195330922.001.0001

DeNeve, K. M., & Cooper, H. (1998). The happy personality: A meta- analysis of 137 personality traits and subjective well-being. Psycholog- ical Bulletin, 124, 197–229. http://dx.doi.org/10.1037/0033-2909.124.2 .197

De Raad, B., & Schouwenburg, H. C. (1996). Personality in learning and education: A review. European Journal of Personality, 10, 303–336. http://dx.doi.org/10.1002/(sici)1099-0984(199612)10:5%3C303::aid- per262%3E3.0.co;2-2

DeYoung, C. G. (2014). Cybernetic Big Five Theory. Journal of Research in Personality. Advance online publication. http://dx.doi.org/10.1016/j .jrp.2014.07.004

Diener, E., & Diener, M. (1995). Cross-cultural correlates of life satisfac- tion and self-esteem. Journal of Personality and Social Psychology, 68, 653– 663. http://dx.doi.org/10.1037/0022-3514.68.4.653

Dion, K., Berscheid, E., & Walster, E. (1972). What is beautiful is good. Journal of Personality and Social Psychology, 24, 285–290. http://dx .doi.org/10.1037/h0033731

Dunning, D., Heath, C., & Suls, J. M. (2004). Flawed self-assessment: Implications for health, education, and the workplace. Psychological Science in the Public Interest, 5, 69 –106. http://dx.doi.org/10.1111/j .1529-1006.2004.00018.x

Dweck, C. S. (1990). Self-theories and goals: Their role in motivation, personality, and development. Nebraska Symposium on Motivation, 38, 199 –235.

Feingold, A. (1992). Good-looking people are not what we think. Psycho- logical Bulletin, 111, 304 –341. http://dx.doi.org/10.1037/0033-2909 .111.2.304

Feingold, A. (1994). Gender differences in personality: A meta-analysis. Psychological Bulletin, 116, 429 – 456. http://dx.doi.org/10.1037/0033- 2909.116.3.429

Fleenor, J. W., Smither, J. W., Atwater, L. E., Braddy, P. W., & Sturm, R. E. (2010). Self-other rating agreement in leadership: A review. The Leadership Quarterly, 21, 1005–1034. http://dx.doi.org/10.1016/j .leaqua.2010.10.006

Fleeson, W., & Jayawickreme, E. (2015). Whole trait theory. Journal of Research in Personality, 56, 82–92.

Fournier, M. A., Moskowitz, D. S., & Zuroff, D. C. (2009). The interper- sonal signature. Journal of Research in Personality, 43, 155–162. http:// dx.doi.org/10.1016/j.jrp.2009.01.023

Funder, D. C. (1987). Errors and mistakes: Evaluating the accuracy of social judgment. Psychological Bulletin, 101, 75–90. http://dx.doi.org/ 10.1037/0033-2909.101.1.75

Funder, D. C. (1995). On the accuracy of personality judgment: A realistic approach. Psychological Review, 102, 652– 670. http://dx.doi.org/10 .1037/0033-295X.102.4.652

Funder, D. C. (1999). Personality judgment: A realistic approach to person perception. San Diego, CA: Academic Press.

Funder, D. C., & Colvin, C. R. (1988). Friends and strangers: Acquain- tanceship, agreement, and the accuracy of personality judgment. Journal of Personality and Social Psychology, 55, 149 –158. http://dx.doi.org/10 .1037/0022-3514.55.1.149

Funder, D. C., & Dobroth, K. M. (1987). Differences between traits: Properties associated with interjudge agreement. Journal of Personality and Social Psychology, 52, 409 – 418. http://dx.doi.org/10.1037/0022- 3514.52.2.409

Funder, D. C., Kolar, D. C., & Blackman, M. C. (1995). Agreement among judges of personality: Interpersonal relations, similarity, and acquain- tanceship. Journal of Personality and Social Psychology, 69, 656 – 672. http://dx.doi.org/10.1037/0022-3514.69.4.656

Funder, D. C., & West, S. G. (1993). Consensus, self-other agreement, and accuracy in personality judgment: An introduction. Journal of Person- ality, 61, 457– 476. http://dx.doi.org/10.1111/j.1467-6494.1993 .tb00778.x

Furnham, A., Badmin, N., & Sneade, I. (2002). Body image dissatisfaction: Gender differences in eating attitudes, self-esteem, and reasons for exercise. The Journal of Psychology: Interdisciplinary and Applied, 136, 581–596. http://dx.doi.org/10.1080/00223980209604820

Geeza, A. A., Connelly, B. S., & Chang, L. (2010, April). A meta-analytic examination of consistency in observers’ perspectives across contexts. Paper presented at the annual meeting for the Society for Industrial and Organizational Psychologists, Atlanta, GA.

Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday.

Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe (Vol. 7, pp. 7–28). Tilburg, the Netherlands: Tilburg University Press.

Goldberg, L. R., Johnson, J. A., Eber, H. W., Hogan, R., Ashton, M. C., Cloninger, C. R., & Gough, H. C. (2006). The International Personality Item Pool and the future of public-domain personality measures. Journal of Research in Personality, 40, 84 –96. http://dx.doi.org/10.1016/j.jrp .2005.08.007

Gosling, S. (2009). Snoop: What your stuff says about you. New York, NY: Basic Books.

Heilman, M. E., & Chen, J. J. (2005). Same behavior, different conse- quences: Reactions to men’s and women’s altruistic citizenship behav- ior. Journal of Applied Psychology, 90, 431– 441. http://dx.doi.org/10 .1037/0021-9010.90.3.431

Heilman, M. E., & Okimoto, T. G. (2007). Why are women penalized for success at male tasks?: The implied communality deficit. Journal of Applied Psychology, 92, 81–92. http://dx.doi.org/10.1037/0021-9010.92 .1.81

Hoffman, B. J., Lance, C. E., Bynum, B., & Gentry, W. A. (2010). Rater source effects are alive and well after all. Personnel Psychology, 63, 119 –151. http://dx.doi.org/10.1111/j.1744-6570.2009.01164.x

Hoffman, B. J., & Woehr, D. J. (2009). Disentangling the meaning of multisource performance rating source and dimension factors. Personnel Psychology, 62, 735–765. http://dx.doi.org/10.1111/j.1744-6570.2009 .01156.x

Hogan, R. T. (1991). Personality and personality measurement. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organiza-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

587TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

tional psychology (2nd ed., pp. 873–919). Palo Alto, CA: Consulting Psychologists Press.

Hogan, R. (1996). A socioanalytic interpretation of the Five-Factor Model. In J. S. Wiggins (Ed.), The Five-Factor model of personality (pp. 163–179). New York, NY: Guilford Press.

Hogan, R. T. (2007). Personality and the fate of organizations. Mahwah, NJ: Erlbaum.

Hogan, R., & Blickle, G. (2013). Socioanalytic theory. In N. D. Christian- sen & R. P. Tett (Eds.), Handbook of personality at work (p. 53). New York, NY: Routledge.

Holzinger, K. J., & Swineford, F. (1937). The bi-factor method. Psy- chometrika, 2, 41–54. http://dx.doi.org/10.1007/BF02287965

Hopwood, C. J., & Donnellan, M. B. (2010). How should the internal structure of personality inventories be evaluated? Personality and Social Psychology Review, 14, 332–346. http://dx.doi.org/10.1177/ 1088868310361240

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alterna- tives. Structural Equation Modeling, 6, 1–55. http://dx.doi.org/10.1080/ 10705519909540118

Human, L. J., & Biesanz, J. C. (2013). Targeting the good target: An integrative review of the characteristics and consequences of being accurately perceived. Personality and Social Psychology Review, 17, 248 –272. http://dx.doi.org/10.1177/1088868313495593

James, W. (1890). The principles of psychology (Vol. 1). London, UK: Macmillan. http://dx.doi.org/10.1037/11059-000

John, O. P., & Robins, R. W. (1993). Determinants of interjudge agreement on personality traits: The big five domains, observability, evaluative- ness, and the unique perspective of the self. Journal of Personality, 61, 521–551. http://dx.doi.org/10.1111/j.1467-6494.1993.tb00781.x

John, O. P., & Robins, R. W. (1994). Accuracy and bias in self-perception: Individual differences in self-enhancement and the role of narcissism. Journal of Personality and Social Psychology, 66, 206 –219. http://dx .doi.org/10.1037/0022-3514.66.1.206

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (pp. 102–138). New York, NY: Guilford Press.

Judge, T. A., Locke, E. A., Durham, C. C., & Kluger, A. N. (1998). Dispositional effects on job and life satisfaction: The role of core evaluations. Journal of Applied Psychology, 83, 17–34. http://dx.doi.org/ 10.1037/0021-9010.83.1.17

Judge, T. A., Van Vianen, A. E. M., & De Pater, I. E. (2004). Emotional stability, core self-evaluations, and job outcomes: A review of the evidence and an agenda for future research. Human Performance, 17, 325–346. http://dx.doi.org/10.1207/s15327043hup1703_4

Kammeyer-Mueller, J., Steel, P. D. G., & Rubenstein, A. (2010). The other side of method bias: The perils of distinct source research designs. Multivariate Behavioral Research, 45, 294 –321. http://dx.doi.org/10 .1080/00273171003680278

Kandler, C., Bleidorn, W., Riemann, R., Spinath, F. M., Thiel, W., & Angleitner, A. (2010). Sources of cumulative continuity in personality: A longitudinal multiple-rater twin study. Journal of Personality and Social Psychology, 98, 995–1008. http://dx.doi.org/10.1037/a0019558

Kandler, C., Riemann, R., Spinath, F. M., & Angleitner, A. (2010). Sources of variance in personality facets: A multiple-rater twin study of self-peer, peer-peer, and self-self (dis)agreement. Journal of Personality, 78, 1565–1594. http://dx.doi.org/10.1111/j.1467-6494.2010.00661.x

Kenny, D. A. (1991). A general model of consensus and accuracy in interpersonal perception. Psychological Review, 98, 155–163. http://dx .doi.org/10.1037/0033-295X.98.2.155

Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York, NY: Guilford Press.

Kenny, D. A. (1996). The design and analysis of social-interaction re- search. Annual Review of Psychology, 47, 59 – 86. http://dx.doi.org/10 .1146/annurev.psych.47.1.59

Kenny, D. A. (2004). PERSON: A general model of interpersonal percep- tion. Personality and Social Psychology Review, 8, 265–280. http://dx .doi.org/10.1207/s15327957pspr0803_3

Kenny, D. A., Albright, L., Malloy, T. E., & Kashy, D. A. (1994). Consensus in interpersonal perception: Acquaintance and the big five. Psychological Bulletin, 116, 245–258. http://dx.doi.org/10.1037/0033- 2909.116.2.245

Kenny, D. A., & DePaulo, B. M. (1993). Do people know how others view them? An empirical and theoretical account. Psychological Bulletin, 114, 145–161. http://dx.doi.org/10.1037/0033-2909.114.1.145

Kenrick, D. T., & Funder, D. C. (1988). Profiting from controversy. Lessons from the person-situation debate. American Psychologist, 43, 23–34. http://dx.doi.org/10.1037/0003-066X.43.1.23

Kline, R. B. (2011). Principles and practice of structural equation mod- eling (3rd ed.). New York, NY: Guilford Press.

Kolar, D. W., Funder, D. C., & Colvin, C. R. (1996). Comparing the accuracy of personality judgements by the self and knowledgeable others. Journal of Personality, 64, 311–337. http://dx.doi.org/10.1111/j .1467-6494.1996.tb00513.x

Kosinski, M., Stillwell, D., & Graepel, T. (2013). Private traits and attri- butes are predictable from digital records of human behavior. Proceed- ings of the National Academy of Sciences of the United States of America, 110, 5802–5805. http://dx.doi.org/10.1073/pnas.1218772110

Kring, A. M., & Gordon, A. H. (1998). Sex differences in emotion: Expression, experience, and physiology. Journal of Personality and Social Psychology, 74, 686 –703. http://dx.doi.org/10.1037/0022-3514 .74.3.686

Kunda, Z. (1990). The case for motivated reasoning. Psychological Bul- letin, 108, 480 – 498. http://dx.doi.org/10.1037/0033-2909.108.3.480

Kurtz, J. E., & Palfrey, A. P. (2016). On being the same in different places: Evaluating Frame-of-Reference effects across two social contexts. Jour- nal of Personality, 84, 113–120.

Kwan, V. S. Y., John, O. P., Kenny, D. A., Bond, M. H., & Robins, R. W. (2004). Reconceptualizing individual differences in self-enhancement bias: An interpersonal approach. Psychological Review, 111, 94 –110.

Kwan, V. S. Y., John, O. P., Robins, R. W., & Kuang, L. L. (2008). Conceptualizing and assessing self-enhancement bias: A componential approach. Journal of Personality and Social Psychology, 94, 1062– 1077. http://dx.doi.org/10.1037/0022-3514.94.6.1062

Kwang, T., & Swann, W. B., Jr. (2010). Do people embrace praise even when they feel unworthy? A review of critical tests of self-enhancement versus self-verification. Personality and Social Psychology Review, 14, 263–280. http://dx.doi.org/10.1177/1088868310365876

Kyllonen, P. C. (2008). The research behind the ETS® Personal Potential Index (PPI). Princeton, NJ: ETS.

LaGrange, B., & Cole, D. A. (2008). An expansion of the trait-state-occasion model: Accounting for shared method variance. Structural Equation Mod- eling, 15, 241–271. http://dx.doi.org/10.1080/10705510801922381

Lance, C. E., Dawson, B., Birkelbach, D., & Hoffman, B. J. (2010). Method effects, measurement error, and substantive conclusions. Orga- nizational Research Methods, 13, 435– 455. http://dx.doi.org/10.1177/ 1094428109352528

Leary, M. R. (1995). Self-presentation: Impression management and in- terpersonal behavior. Madison, WI: Brown & Benchmark.

Leary, M. R., & Baumeister, R. F. (2000). The nature and function of self-esteem: Sociometer theory. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 32, pp. 1– 62). San Diego, CA: Academic Press.

Leary, M. R., & Kowalski, R. M. (1990). Impression management: A literature review and two-component model. Psychological Bulletin, 107, 34 – 47. http://dx.doi.org/10.1037/0033-2909.107.1.34

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

588 MCABEE AND CONNELLY

Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as an interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68, 518 –530. http://dx .doi.org/10.1037/0022-3514.68.3.518

Leising, D., Erbs, J., & Fritz, U. (2010). The letter of recommendation effect in informant ratings of personality. Journal of Personality and Social Psychology, 98, 668 – 682. http://dx.doi.org/10.1037/a0018771

Li, A., & Bagger, J. (2006). Using the BIDR to distinguish the effects of impression management and self-deception on the criterion validity of personality measures: A meta-analysis. International Journal of Selec- tion and Assessment, 14, 131–141. http://dx.doi.org/10.1111/j.1468- 2389.2006.00339.x

Lievens, F., De Corte, W., & Schollaert, E. (2008). A closer look at the frame-of-reference effect in personality scale scores and validity. Jour- nal of Applied Psychology, 93, 268 –279. http://dx.doi.org/10.1037/ 0021-9010.93.2.268

Löckenhoff, C. E., Chan, W., McCrae, R. R., De Fruyt, F., Jussim, L., De Bolle, M., . . . Terracciano, A. (2014). Gender stereotypes of personality: Universal and Accurate? Journal of Cross-Cultural Psychology, 45, 675– 694. http://dx.doi.org/10.1177/0022022113520075

Luft, J., & Ingham, H. (1955). The Johari window: A graphic model of interpersonal awareness. Paper presented at the Proceedings of the Western Training Laboratory in Group Development, Los Angeles, CA: UCLA Extension Office.

MacDonald, K. (1995). Evolution, the Five-Factor Model, and levels of personality. Journal of Personality, 63, 525–567. http://dx.doi.org/10 .1111/j.1467-6494.1995.tb00505.x

Malle, B. F. (2006). The actor-observer asymmetry in attribution: A (surprising) meta-analysis. Psychological Bulletin, 132, 895–919. http:// dx.doi.org/10.1037/0033-2909.132.6.895

Malloy, T. E., Agatstein, F., Yarlas, A., & Albright, L. (1997). Effects of communication, information overlap, and behavioral consistency on consensus in social perception. Journal of Personality and Social Psy- chology, 73, 270 –280. http://dx.doi.org/10.1037/0022-3514.73.2.270

Malloy, T. E., & Kenny, D. A. (1986). The social relations model: An integrative method for personality research. Journal of Personality, 54, 199 –225. http://dx.doi.org/10.1111/j.1467-6494.1986.tb00393.x

Markon, K. E. (2009). Hierarchies in the structure of personality traits. Social and Personality Psychology Compass, 3, 812– 826. http://dx.doi .org/10.1111/j.1751-9004.2009.00213.x

Marsh, H. W., & Bailey, M. (1991). Confirmatory factory analyses of multitrait-multimethod data: A comparison of alternative models. Ap- plied Psychological Measurement, 15, 47–70. http://dx.doi.org/10.1177/ 014662169101500106

May, M. A. (1932). Psychology at work. New York, NY: McGraw-Hill. McAdams, D. P., & Pals, J. L. (2006). A new Big Five: Fundamental

principles for an integrative science of personality. American Psychol- ogist, 61, 204 –217. http://dx.doi.org/10.1037/0003-066X.61.3.204

McCrae, R. R., & Costa, P. T. (1983). Joint factors in self-reports and ratings: Neuroticism, extraversion and openness to experience. Person- ality and Individual Differences, 4, 245–255. http://dx.doi.org/10.1016/ 0191-8869(83)90146-0

McCrae, R. R., Terracciano, A., & the Personality Profiles of Cultures Project. (2005). Universal features of personality traits from the observ- er’s perspective: Data from 50 cultures. Journal of Personality and Social Psychology, 88, 547–561. http://dx.doi.org/10.1037/0022-3514 .88.3.547

Mead, G. H. (1913). The social self. Journal of Philosophy, Psychology & Scientific Methods, 10, 374 –380. http://dx.doi.org/10.2307/2012910

Mehta, P. D., & Neale, M. C. (2005). People are variables too: Multilevel structural equations modeling. Psychological Methods, 10, 259 –284.

Mischel, W. (1968). Personality and assessment. New York, NY: Wiley. 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. http://dx.doi.org/10.1037/0033-295X.102.2.246

Muthén, L. K., & Muthén, B. O. (1998 –2012). Mplus user’s guide (7th ed.). Los Angeles, CA: Author.

Oh, I. S., Wang, G., & Mount, M. K. (2011). Validity of observer ratings of the five-factor model of personality traits: A meta-analysis. Journal of Applied Psychology, 96, 762–773. http://dx.doi.org/10.1037/a0021832

Olsen, J. A., & Kenny, D. A. (2006). Structural equation modeling with interchangeable dyads. Psychological Methods, 11, 127. http://dx.doi .org/10.1037/1082-989X.11.2.127

Oltmanns, T. F., & Turkheimer, E. (2009). Person perception and person- ality pathology. Current Directions in Psychological Science, 18, 32–36. http://dx.doi.org/10.1111/j.1467-8721.2009.01601.x

Ones, D. S., Viswesvaran, C., & Reiss, A. D. (1996). Role of social desirability in personality testing for personnel selection: The red her- ring. Journal of Applied Psychology, 81, 660 – 679. http://dx.doi.org/10 .1037/0021-9010.81.6.660

Ozer, D. J., & Benet-Martínez, V. (2006). Personality and the prediction of consequential outcomes. Annual Review of Psychology, 57, 401– 421. http://dx.doi.org/10.1146/annurev.psych.57.102904.190127

Paulhus, D. L. (1984). Two-component models of socially desirable re- sponding. Journal of Personality and Social Psychology, 46, 598 – 609. http://dx.doi.org/10.1037/0022-3514.46.3.598

Paulhus, D. L. (1991). Measurement and control of response bias. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 17–59). San Diego, CA: Academic Press. http://dx.doi.org/10.1016/B978-0-12-590241-0 .50006-X

Paulhus, D. L., & John, O. P. (1998). Egoistic and moralistic biases in self-perception: The interplay of self-deceptive styles with basic traits and motives. Journal of Personality, 66, 1025–1060. http://dx.doi.org/ 10.1111/1467-6494.00041

Penton-Voak, I. S., Pound, N., Little, A. C., & Perrett, D. I. (2006). Personality judgments from natural and composite facial images: More evidence for a “kernel of truth” in social perception. Social Cognition, 24, 607– 640. http://dx.doi.org/10.1521/soco.2006.24.5.607

Pickering, A. D., & Gray, J. A. (1999). The neuroscience of personality. Handbook of personality: Theory and research (Vol. 2, pp. 277–299). New York, NY: Guilford Press.

Reise, S. P. (2012). The rediscovery of bifactor measurement models. Multivariate Behavioral Research, 47, 667– 696. http://dx.doi.org/10 .1080/00273171.2012.715555

Reise, S. P., Moore, T. M., & Haviland, M. G. (2010). Bifactor models and rotations: Exploring the extent to which multidimensional data yield univocal scale scores. Journal of Personality Assessment, 92, 544 –559. http://dx.doi.org/10.1080/00223891.2010.496477

Rindskopf, D., & Rose, T. (1988). Some theory and applications of confirmatory second-order factor analysis. Multivariate Behavioral Re- search, 23, 51– 67. http://dx.doi.org/10.1207/s15327906mbr2301_3

Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126, 3–25. http://dx.doi .org/10.1037/0033-2909.126.1.3

Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Goldberg, L. R. (2007). The power of personality: The comparative validity of person- ality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science, 2, 313–345. http://dx.doi.org/10.1111/j.1745-6916.2007.00047.x

Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A meta- analysis of longitudinal studies. Psychological Bulletin, 132, 1–25. http://dx.doi.org/10.1037/0033-2909.132.1.1

Robins, R. W., & John, O. P. (1997). The quest for self-insight: Theory and

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

589TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY

research on accuracy and bias in self-perception. In R. Hogan, J. A. Johnson, & S. R. Briggs (Eds.), Handbook of personality psychology (pp. 649 – 679). San Diego, CA: Academic Press. http://dx.doi.org/10 .1016/B978-012134645-4/50026-3

Robins, R. W., Tracy, J. L., Trzesniewski, K., Potter, J., & Gosling, S. D. (2001). Personality correlates of self-esteem. Journal of Research in Personality, 35, 463– 482. http://dx.doi.org/10.1006/jrpe.2001.2324

Rodriguez, A., Reise, S. P., & Haviland, M. G. (2015). Evaluating bifactor models: Calculating and interpreting statistical indices. Psychological Methods, 21, 137–150. http://dx.doi.org/10.1037/met0000045

Rogers, T. B. (1981). A model of the self as an aspect of the human information-processing system. In N. Cantor & X. F. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp. 193–214). Hillsdale, NJ: Erlbaum.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

Rosenberg, M. (1979). Conceiving the self. New York, NY: Basic Books. Ross, L. (1977). The intuitive psychologist and his shortcomings: Distor-

tions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 173–220). London, UK: Academic Press.

Schlenker, B. R. (1980). Impression management: The self-concept, social identity, and interpersonal relations. Monterey, CA: Brooks/Cole.

Schlenker, B. R., & Weigold, M. F. (1992). Interpersonal processes in- volving impression regulation and management. Annual Review of Psy- chology, 43, 133–168. http://dx.doi.org/10.1146/annurev.ps.43.020192 .001025

Schmit, M. J., Ryan, A. M., Stierwalt, S. L., & Powell, A. B. (1995). Frame of reference effects on personality scale scores and criterion-related validity. Journal of Applied Psychology, 80, 607– 620. http://dx.doi.org/ 10.1037/0021-9010.80.5.607

Schmitt, D. P., Realo, A., Voracek, M., & Allik, J. (2008). Why can’t a man be more like a woman? Sex differences in Big Five personality traits across 55 cultures. Journal of Personality and Social Psychology, 94, 168 –182. http://dx.doi.org/10.1037/0022-3514.94.1.168

Schneider, D. J. (1981). Tactical self-presentations: Toward a broader conception. In J. T. Tedeschi (Ed.), Impression Management Theory and Social Psychological Research (pp. 23– 40). London, UK: Academic Press. http://dx.doi.org/10.1016/B978-0-12-685180-9.50007-5

Sedikides, C. (1993). Assessment, enhancement, and verification determi- nants of the self-evaluation process. Journal of Personality and Social Psychology, 65, 317–338. http://dx.doi.org/10.1037/0022-3514.65.2.317

Seih, Y.-T., Buhrmester, M. D., Lin, Y.-C., Huang, C.-L., & Swann, W. B., Jr. (2013). Do people want to be flattered or understood? The cross- cultural universality of self-verification. Journal of Experimental Social Psychology, 49, 169 –172. http://dx.doi.org/10.1016/j.jesp.2012.09.004

Small, E. E., & Diefendorff, J. M. (2006). The impact of contextual self-ratings and observer ratings of personality on the personality- performance relationship. Journal of Applied Social Psychology, 36, 297–320. http://dx.doi.org/10.1111/j.0021-9029.2006.00009.x

Smith, T. W., Uchino, B. N., Berg, C. A., Florsheim, P., Pearce, G., Hawkins, M., . . . Yoon, H.-C. (2008). Associations of self-reports versus spouse ratings of negative affectivity, dominance, and affiliation with coronary artery disease: Where should we look and who should we ask when studying personality and health? Health Psychology, 27, 676 – 684. http://dx.doi.org/10.1037/0278-6133.27.6.676

Spector, P. E. (2006). Method variance in organizational research: Truth or urban legend? Organizational Research Methods, 9, 221–232. http://dx .doi.org/10.1177/1094428105284955

Srivastava, S., Guglielmo, S., & Beer, J. S. (2010). Perceiving others’ personalities: Examining the dimensionality, assumed similarity to the self, and stability of perceiver effects. Journal of Personality and Social Psychology, 98, 520 –534. http://dx.doi.org/10.1037/a0017057

Swann, W. B., Jr. (1981). Self-verification processes: How we sustain our self-conceptions. Journal of Experimental Social Psychology, 17, 351– 372. http://dx.doi.org/10.1016/0022-1031(81)90043-3

Swann, W. B., Jr. (1983). Self-verification: Bringing social reality into harmony with the self. In J. Suls & A. G. Greenwald (Eds.), Social psychological perspectives on the self (Vol. 2, pp. 33– 66). Hillsdale, NJ: Erlbaum.

Swann, W. B., Jr. (1990). To be known or to be adored: The interplay of self-enhancement and self-verification. In E. T. Higgins & R. M. Sor- rentino (Eds.), Handbook of motivation and cognition (Vol. 2, pp. 408 – 448). New York, NY: Guilford Press.

Swann, W. B., Jr. (2012). Self-verification theory. In P. Van Lang, A. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (pp. 23– 42). London, UK: Sage. http://dx.doi.org/10.4135/ 9781446249222.n27

Swann, W. B., Rentfrow, P. J., & Guinn, J. S. (2003). Self-verification: The search for coherence. In M. R. Leary & J. P. Tangney (Eds.), Handbook of self and identity (pp. 367–383). New York, NY: Guilford Press.

Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193–210. http://dx.doi.org/10.1037/0033-2909.103.2.193

Taylor, S. E., & Brown, J. D. (1994). Positive illusions and well-being revisited: Separating fact from fiction. Psychological Bulletin, 116, 21–27. http://dx.doi.org/10.1037/0033-2909.116.1.21

Ten Berge, J. M. F., & Sočan, G. (2004). The greatest lower bound to the reliability of a test and the hypothesis of unidimensionality. Psy- chometrika, 69, 613– 625. http://dx.doi.org/10.1007/BF02289858

Tetlock, P. E., & Manstead, A. S. (1985). Impression management versus intrapsychic explanations in social psychology: A useful dichotomy? Psychological Review, 92, 59 –77. http://dx.doi.org/10.1037/0033-295X .92.1.59

Van Egeren, L. F. (2009). A cybernetic model of global personality traits. Personality and Social Psychology Review, 13, 92–108. http://dx.doi .org/10.1177/1088868309334860

Vazire, S. (2006). Informant reports: A cheap, fast, and easy method for personality assessment. Journal of Research in Personality, 40, 472– 481. http://dx.doi.org/10.1016/j.jrp.2005.03.003

Vazire, S. (2010). Who knows what about a person? The self-other knowl- edge asymmetry (SOKA) model. Journal of Personality and Social Psychology, 98, 281–300. http://dx.doi.org/10.1037/a0017908

Vazire, S., & Carlson, E. N. (2011). Others sometimes know us better than we know ourselves. Current Directions in Psychological Science, 20, 104 –108. http://dx.doi.org/10.1177/0963721411402478

Vazire, S., & Mehl, M. R. (2008). Knowing me, knowing you: The accuracy and unique predictive validity of self-ratings and other-ratings of daily behavior. Journal of Personality and Social Psychology, 95, 1202–1216. http://dx.doi.org/10.1037/a0013314

Vernon, P. E. (1933). The biosocial nature of the personality trait. Psy- chological Review, 40, 533–548. http://dx.doi.org/10.1037/h0072622

Viswesvaran, C., & Ones, D. S. (1999). Meta-analyses of fakability esti- mates: Implications for personality measurement. Educational and Psy- chological Measurement, 59, 197–210. http://dx.doi.org/10.1177/ 00131649921969802

Watson, D., & Humrichouse, J. (2006). Personality development in emerg- ing adulthood: Integrating evidence from self-ratings and spouse ratings. Journal of Personality and Social Psychology, 91, 959 –974. http://dx .doi.org/10.1037/0022-3514.91.5.959

West, T. V., & Kenny, D. A. (2011). The truth and bias model of judgment. Psychological Review, 118, 357–378. http://dx.doi.org/10.1037/ a0022936

Woo, S. E., Jin, J., & LeBreton, J. M. (2015). Specificity matters: Criterion-related validity of contextualized and facet measures of con- scientiousness in predicting college student performance. Journal of

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

590 MCABEE AND CONNELLY

Personality Assessment, 97, 301–309. http://dx.doi.org/10.1080/ 00223891.2014.1002134

Woo, S. E., Sims, C. S., Rupp, D. E., & Gibbons, A. M. (2008). Devel- opment engagement within and following developmental assessment centers: Considering feedback favorability and self-assessor agreement. Personnel Psychology, 61, 727–759. http://dx.doi.org/10.1111/j.1744- 6570.2008.00129.x

Wood, W., & Eagly, A. H. (2002). A cross-cultural analysis of the behavior of women and men: Implications for the origins of sex differences. Psychological Bulletin, 128, 699 –727.

Youyou, W., Kosinski, M., & Stillwell, D. (2015). Computer-based per- sonality judgments are more accurate than those made by humans. Proceedings of the National Academy of Sciences of the United States of America, 112, 1036 –1040. http://dx.doi.org/10.1073/pnas.1418680112

Yung, Y.-F., Thissen, D., & McLeod, L. D. (1999). On the relationship between the higher-order factor model and the hierarchical factor model. Psychometrika, 64, 113–128. http://dx.doi.org/10.1007/BF02294531

Zaki, J., & Ochsner, K. (2011). Reintegrating the study of accuracy into social cognition research. Psychological Inquiry, 22, 159 –182. http://dx .doi.org/10.1080/1047840X.2011.551743

Zickar, M. J., & Kostek, J. A. (2013). History of personality testing within organizations. In N. D. Christiansen & R. P. Tett (Eds.), Handbook of personality at work (pp. 173–190). New York, NY: Taylor & Francis.

Received April 1, 2015 Revision received May 4, 2016

Accepted May 21, 2016 �

Members of Underrepresented Groups: Reviewers for Journal Manuscripts Wanted

If you are interested in reviewing manuscripts for APA journals, the APA Publications and Communications Board would like to invite your participation. Manuscript reviewers are vital to the publications process. As a reviewer, you would gain valuable experience in publishing. The P&C Board is particularly interested in encouraging members of underrepresented groups to participate more in this process.

If you are interested in reviewing manuscripts, please write APA Journals at [email protected]. Please note the following important points:

• To be selected as a reviewer, you must have published articles in peer-reviewed journals. The experience of publishing provides a reviewer with the basis for preparing a thorough, objective review.

• To be selected, it is critical to be a regular reader of the five to six empirical journals that are most central to the area or journal for which you would like to review. Current knowledge of recently published research provides a reviewer with the knowledge base to evaluate a new submission within the context of existing research.

• To select the appropriate reviewers for each manuscript, the editor needs detailed information. Please include with your letter your vita. In the letter, please identify which APA journal(s) you are interested in, and describe your area of expertise. Be as specific as possible. For example, “social psychology” is not sufficient—you would need to specify “social cognition” or “attitude change” as well.

• Reviewing a manuscript takes time (1– 4 hours per manuscript reviewed). If you are selected to review a manuscript, be prepared to invest the necessary time to evaluate the manuscript thoroughly.

APA now has an online video course that provides guidance in reviewing manuscripts. To learn more about the course and to access the video, visit http://www.apa.org/pubs/authors/review- manuscript-ce-video.aspx.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

591TRAIT-REPUTATION-IDENTITY MODEL OF PERSONALITY