Journal of Research in Personality

Transcription

1 Journal of Research in Personality 44 (2010) Contents lists available at ScienceDirect Journal of Research in Personality journal homepage: Variance determines self-observer agreement on the Big Five personality traits Jüri Allik a, *, Anu Realo a, René Mõttus a, Tõnu Esko b, Janne Pullat b, Andres Metspalu b a Department of Psychology, Estonian Center of Behavioral and Health Sciences, University of Tartu, Estonia b Institute of Molecular and Cell Biology, Estonian Biocenter, University of Tartu, Estonia article info abstract Article history: Available online 3 June 2010 Keywords: Personality ratings NEO PI-R Self-observer agreement Visibility of traits Assumed similarity Restriction for range It is widely believed that, on those personality traits that are more visible to an external observer, two judges will reach a higher level of agreement than on those traits that are more difficult to judge. This view is challenged in the current paper, using a sample of 672 participants in the age range of years who described their own personality and were judged by an external observer who knew them well, using the NEO PI-3 questionnaire (McCrae, Costa, & Martin, 2005). The self-observer agreement on the 30 personality subscales varied from.38 (O3: Feelings) to.57 (E5: Excitement Seeking). Approximately one-half of the variance in the agreement level was explained by the standard deviation of the sum scores of these subscales: self-observer agreement was higher in the subscales on which individual differences were larger. After correction for the range of variance, differences in self-observer agreement substantially diminished. It is proposed that judges who know each other well reach an approximately equal level of agreement on all the Big Five personality traits. Crown Copyright Ó 2010 Published by Elsevier Inc. All rights reserved. 1. Introduction In order to function efficiently in the social environment, people need to regularly make personality judgments about themselves and other people. Typically, the accuracy of such judgments is estimated by the degree of agreement between ratings made by different judges. Although self-observer and between-observer agreement do not prove that personality ratings are accurate, it is a necessary prerequisite for their accuracy. When self-ratings are compared to those of well-informed observers, judges tend to achieve at least moderate cross-observer agreement for most personality traits (Funder & Colvin, 1997; Kenny, 1994). For example, the mean or median interobserver trait agreement among well-acquainted informants is almost invariably.40 or higher for all the Big Five personality dimensions (Connolly, Kavanagh, & Viswesvaran, 2007; Hall, Andrzejewski, Murphy, Schmid Mast, & Feinstein, 2008; McCrae et al., 2004). This level of agreement is not trivial considering the complicated chain of events required for an accurate personality judgment: the target of judgment must display behaviors and cues that are relevant to the trait being judged and the judge must detect these cues and correctly use them to make judgments (Funder, 1999; Funder & Colvin, 1997). Considering the intricacies of personality judgments, it is not surprising that researchers are inclined to believe that this agreement between judgments can be easily compromised and that * Corresponding author. Address: Department of Psychology, University of Tartu, Tiigi 78, Tartu 50410, Estonia. address: juri.allik@ut.ee (J. Allik). there are several moderators that can substantially reduce selfobserver agreement. One of these moderators which has received much attention is judgability. For instance, psychologically better adjusted individuals are easier to judge than less well adjusted people (Colvin, 1993; Furr, Dougherty, Marsh, & Mathias, 2007). Another equally powerful moderator is the ability to make correct personality judgments from available information: some individuals are believed to be better judges of personality than are others (Letzring, Wells, & Funder, 2006; Realo et al., 2003; Taft, 1955). However, judgability and good judges appear to be less influential than previously thought: as shown by Allik and colleagues (2010), self-observer agreement does not generalize easily from one personality trait to another since targets and raters from the same target rater pairs may occupy identical or nearly identical positions in their respective rankings on one personality trait but can have a considerable disparity in their rankings on another personality trait. In this paper, we focus on the question of whether some personality traits are easier to judge than the others and what might be the reason for this Are some traits easier to judge than others? Over the years, several different studies have shown that it is easier to reach self-observer agreement on some personality traits than on others. For example, numerous studies have shown that traits defining Extraversion are easier to judge than traits defining Neuroticism (Connolly et al., 2007; Funder & Dobroth, 1987; Hall et al., 2008; Park & Judd, 1989). Another recent meta-review /$ - see front matter Crown Copyright Ó 2010 Published by Elsevier Inc. All rights reserved. doi: /j.jrp

2 422 J. Allik et al. / Journal of Research in Personality 44 (2010) showed that the mean correlation in observer-ratings, corrected for coefficient alpha in self-ratings and interrater reliability, was.62 for Extraversion, which was higher than for all other traits (.51 for Neuroticism,.59 for Openness,.46 for Agreeableness, and.56 for Conscientiousness, respectively) (Connolly et al., 2007). In several other meta-reviews, Extraversion has invariably been shown to achieve higher self-observer agreement than Neuroticism (Hall et al., 2008). Despite such consistent findings, the between-trait differences in agreement are still relatively small. In fact, median or mean values of cross-observer correlations for all personality traits are basically within the same range. The median value of Extraversion (.47) for cross-observer agreement in studies using different measures of the Five-Factor Model with single raters was only slightly higher than for Neuroticism (.43), Openness (.43), Agreeableness (.40), and Conscientiousness (.41) (McCrae et al., 2004). Researchers have proposed several explanations for why it is easier to reach agreement on some personality traits than on others Visibility of traits The concept of visibility has been the most widely used as an explanation for why some personality traits are easier to agree upon than others. This term encompasses several related concepts such as judgability, confirmability, observability, and availability of traits (Funder, 1995; Funder & Dobroth, 1987; Gangestad, Simpson, DiGeronimo, & Biek, 1992; Paunonen, 1989; Tausch, Kenworthy, & Hewstone, 2007). For example, Funder and Dobroth (1987) asked participants to rate 100 California Q-sort items on nine subjective dimensions. One of these dimensions was how easy it is to imagine specific, observable behaviors that would confirm or disconfirm a trait and another one was how easy it is to judge the degree to which another person had the trait. Six of these dimensions grouped into one factor, which was interpreted as reflecting each trait s easy visibility to an outside observer. Most visible traits belonged to the Extraversion domain, while the least visible traits were from the Neuroticism domain. The correlation between agreement scores and visibility ratings was moderately positive: r =.42, p <.001 (Funder & Dobroth, 1987). Later studies confirmed that better agreement is reached on more visible traits (Funder & Colvin, 1988). Thus, variation in self-observer agreement may be attributed to systematic differences in items content Assumed similarity Assumed similarity refers to the tendency to perceive others as similar to oneself (Beer & Watson, 2008) or some generalized other (Cronbach, 1955). It seems logical to suppose, when trait information is not readily available, that judgments are made on the assumption that others are similar to oneself or some hypothetical, idealized person. The usual way to calculate assumed similarity is by correlating an individual s self-rating with the average of his or her ratings of each of the other group members (i.e., grouping within judges). As expected, Beer and Watson (2008) found that assumed similarity was statistically significant for Neuroticism (r =.32) and near zero for Extraversion (r =.07). This seems to confirm the principle that more visible traits are judged on the basis of veridical information, while less visible traits are described according to a self-based heuristic (Beer & Watson, 2008). The degree to which raters own personalities contributed to target ratings also correlated negatively with self-observer agreement (r =.60) (Beer & Watson, 2008; Ready, Clark, Watson, & Westerhouse, 2000) and exhibited a strong relationship with the visibility of traits (r =.73). Thus, when an observer is asked to rate a target on difficult-to-judge traits, he or she is more inclined to project his or her own personality on the target. Instead of relying on one s own personality, it is also possible to imagine someone who has socially desirable traits. It has been noticed that self-observer agreement is stronger on neutral rather than socially favorable personality traits (John & Robins, 1993). On socially more desirable personality traits, judges seem to be more guided by their expectations rather than actual information about personality traits which, in turn, could lead to lower self-observer agreement. Unfortunately, it is impossible to calculate assumed similarity when a person has assessed only one target in addition to her- or himself. In many cases, it is neither economic nor possible to obtain ratings of multiple targets from one rater. Another problem is that well-acquainted samples have actually shown higher assumed similarity correlations than stranger samples (Beer & Watson, 2008; Kenny, 1994). This casts doubt on the universality of the assumption that the absence of appropriate information is always compensated for with assumed information about oneself or a hypothetical person. Nevertheless, assumed similarity seems to be, so far, the strongest moderator of self-observer agreement Affectivity Item visibility and assumed similarity are not the only known moderators of self-observer agreement. Watson and his colleagues (2000) noted that Neuroticism scales lead to considerably higher self-observer agreement than the PANAS negative affectivity scales (r =.46 versus.29), in spite of the fact that they both measure approximately the same content (Watson et al., 2000). The only obvious difference seems to be the format of items. While items in personality questionnaires, such as the NEO PI-R (Costa & McCrae, 1992), for instance, are formulated in short statement form (e.g., feel inferior to others or have a low opinion of myself ), in the PANAS Negative Affect Scale, people are asked to rate how they have felt during a certain period of elapsed time or at the present moment, using single words such as distressed, hostile, or angry. Thus, it could be wording, format, or instruction, not visibility of traits alone, which determines these differences in self-observer agreement Restriction of range However, there may be purely statistical reasons why the correlation between two variables, x and y, can vary in its magnitude, A common problem facing researchers is calculating correlation in some population of interest on the basis of a restricted sample. It is well known that a sample correlation can deviate from a population correlation for a variety of reasons, including sampling error, measurement error, and restriction of range (Sackett & Yang, 2000). Usually the sampling problem is understood in terms of individuals: instead of the entire sample, a small fraction of individuals is available for investigation. Another domain in which the restriction of range may have some relevance is the selection of items for personality questionnaires. In the study of inclination towards artistic experiences, for example, researchers can devise such items as I m not really interested in the arts (reversed item) or Certain kinds of music have an endless fascination for me (McCrae et al., 2005). For the study of altruistic dispositions, in turn, the following items have been successfully devised: Some people think I m selfish and egoistical (reversed item) or Most people I know like me (Costa & McCrae, 1992). Although impeccable specimens, these items are but a small fraction from a large pool of items which can potentially be used to measure Openness to Aesthetics or altruistic dispositions. For this very reason, it is possible that a given sample of items is only measuring a restricted range of Openness or Altruism, compared to what could conceivably be covered by the whole imaginable set of items. Any restriction in the range possibly resulting in depressed variance in traits

3 J. Allik et al. / Journal of Research in Personality 44 (2010) scores, however, could result in a reduction of correlation in which a given variable is involved Aim of the study Surprisingly, we know of no previous studies in which variance of personality scores has been taken into account when discussing the accuracy of personality judgments. It is obviously premature to claim that any particular factor, visibility of traits or assumed similarity, for example, has an impact on the magnitude of self-observer agreement without controlling for simpler and more fundamental reasons, such as restriction in the range of variance. This study is, to the best of our knowledge, the first attempt to see how interjudge agreement is influenced by the variance of trait scores on which two judges are supposed to agree. 2. Method 2.1. Participants Participants of this study came from the Estonian Genome Center, University of Tartu. The whole project is conducted in accordance with the Estonian Gene Research Act and all participants have given broad informed consent ( Subjects were randomly selected from individuals visiting general practitioners (GP) offices and hospitals and then recruited by the GPs and hospital physicians. In addition to donating blood samples and answering a medical questionnaire, participants were asked to complete the self-report version (Form S) of the NEO PI Questionnaire The NEO PI-3 (McCrae et al., 2005) is a slightly modified version of the NEO PI-R questionnaire (Costa & McCrae, 1992) that was adapted into Estonian by Kallasmaa and colleagues (2000). Like the original NEO PI-R, the NEO PI-3 has 240 items which measure 30 personality traits grouped into the five factors. The total number of participants in this sample was 672 (284 men and 388 women). The age range of participants was years (M = 46.0, SD = 16.4). All participants were asked to nominate someone who knew them well to fill in the observer report (Form R) version of the NEO PI-3. The age range of external observers or judges was (M = 43.3, SD = 16.4). Of these observers, 198 were men, 451 were women, and 23 did not report their sex. In order to develop a social desirability index for the NEO PI-R, questionnaire items were assessed by 88 judges (24 men and 64 women, mean age 37.6 years, SD = 12.7) who independently rated the social desirability of each of the 240 NEO PI-R items. The instruction stated, Descriptors of people often contain evaluative information. Some personality characteristics are considered more desirable, receiving approval from other people, whereas others are undesirable. If someone agrees strongly with this item, does this present that person in a favorable or unfavorable light, or is agreeing with this item neutral as regards to others approval? Ratings were made on a 7-point Likert scale, ranging from extremely undesirable ( 3) to extremely desirable (+3), with zero as a neutral point. Estonian desirability ratings were reported in a previous study (Konstabel, Aavik, & Allik, 2006). 3. Results We started by computing product moment correlations between self- and observer-ratings on all 30 subscales. Coefficients of agreement were all highly significant (p <.0001) and varied from.38 (O3: Feelings) to.57 (E5: Excitement Seeking). The mean selfobserver correlations were.45,.53,.44,.41, and.45 for Neuroticism, Extraversion, Openness to Experience, Agreeableness, and Conscientiousness scales respectively, being in the same range of previous reports (Connolly et al., 2007; Hall et al., 2008). Next we found the standard deviation for the sum scores on the same 30 subscales (individual deviations from the mean score of the group) for self- and observer-ratings. For self-ratings, the standard deviations varied from 3.84 (A3: Altruism) to 6.35 (O2: Aesthetics), median = Observer-ratings had approximately the same range of standard deviations, from 3.57 (O6: Values) to 6.32 (E4: Activity), median = There was a significant, positive correlation between coefficients of agreement and the subscale standard deviation for both self- and observer-ratings. Fig. 1 shows the correlation between self-observer agreement and the mean standard deviation of selfand observer-ratings [r(28) =.67, p <.0001]. We combined selfand observer-ratings since their standard deviations were significantly correlated [r(28) =.82, p <.0001]. This means that personality traits with larger scatter of sum scores were the traits on which higher self-observer agreement was achieved. It is also important to note that the magnitude of correlation between self-observer agreement and the standard deviation of answers is equal to or even larger than the previously observed impact of visibility or assumed similarity of judgment. Fig. 2 shows that the difference in the level of agreement is not related to the Big Five content of the subscales (i.e., there is as much variation in agreement within the Big Five domains than between them). We chose to demonstrate this with two Openness subscales: that with the smallest and that with the largest self-observer agreement. It can be clearly seen, on the O2: Aesthetics subscale, that the sum scores of participants had a considerably larger scatter than on the O6: Values subscale, where all answers were concentrated in a restricted area of the plain. This restricted scatter may explain why self-observer agreement is smaller on O6 [r(670) =.38, p <.0001] than on O2 [r(670) =.55, p <.0001]. What determines the standard deviation of the sum scores? Since each subscale of the NEO PI-3 consists of eight items there are two potential sources of variance. First, the variance of sum Self-Observer Agreement O6 C3 O4 A3 O3 E1 A4 A6 C1 r =.67 A1 N5 N4 N6 C5 N3 C4 C6 O1 C2 N2 N1 E5 E6 O5 A5 A Standard Deviation (Self + Observer Ratings) Fig. 1. Self-observer agreement as a function of the standard deviation of subscale sum scores. E2 E3 E4 O2

4 424 J. Allik et al. / Journal of Research in Personality 44 (2010) A O2: Aesthetics (Observer) B O6: Values (Observer) r = O2: Aesthetics (Self) r = O6: Values (Self) Fig. 2. Two-dimensional plot of self- and observer-sum scores on two Openness subscales: O2: Aesthetics (A) and O6: Values (B). scores is influenced by the variation of answers on eight items within participants. Secondly, the sum score variances can be influenced by the variance across participants for each of the eight items. As it turned out, the mean standard deviation of answers on these eight items was significantly correlated with the standard deviation of the sum scores: r(28) =.90 and.80 for self- and observer-ratings, respectively (p <.0001). Thus, the distribution of answers on single items is the main reason why the sum scores of the various subscales have different standard deviations. It is entirely plausible that, with an appropriate selection of items, it would be possible to increase the standard deviations of the subscales that currently have low variance. It would theoretically be possible to predict the true magnitude of the self-observer correlation, given that the restriction on the sum scores variation can be compensated for. Let us assume that, with appropriate item selection, it would be possible to increase the standard deviation up to the largest observed variation, SD = 6.35 (O2: Aesthetics). Assuming that the unrestricted variance for all subscales is 6.35, it would be possible to calculate the correction for the observed self-observer correlation. The correction formula for unrestricted variance is known as the Thorndike Case 2 (Sackett & Yang, 2000; Thorndike, 1949): kr xy R xy ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; 1 þ r 2 xy ðk2 1Þ where r xy is the correlation between variables x and y obtained from the restricted sample and k = S x /s x where S x and s x are the standard deviations of the variable x for the restricted sample and the unrestricted population, respectively. Fig. 3 demonstrates how the selfobserver correlation would change once the restriction for standard deviation is eliminated. Except for E1: Warmth, the subscales of Extraversion changed relatively little after correction for the restriction of standard deviation. On all other dimensions, the predicted increase in self-observer correlation was substantial. The largest increase occurred on C3: Dutifulness, where the correlation rose from.43 to.61. After the correction, the mean self-observer correlations became.51,.56,.52,.51, and.56 for Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness subscales respectively. An ANOVA revealed that, before the correction, the mean level of standard deviations for the various personality factors was significantly different, F(4, 25) = 6.32, p =.001, mainly due to higher standard deviation on the Extraversion subscales. After correction, however, there was no significant difference in the mean level of standard deviations of the Big Five factors; F(4, 25) = 1.77, p =.167. Not only corrected correlations were more similar to one another, they were slightly above previously reported meta-analytic correlations (Connolly et al., 2007; Hall et al., 2008). Although the profile of corrected self-observer correlations across the 30 subscales maintained similarity to the observed profile of correlations [r(28) =.42, p =.021], there was no difference across the various personality factors in mean self-observer agreement. We were also interested how social desirability ratings were related to self-observer agreement. The data provided no support for previous findings (John & Robins, 1993) that socially desirable traits lead to lower interjudge agreement. Social desirability ratings were not correlated either with uncorrected [r(28) =.02, p =.916] or corrected [(r(28) =.26, p =.162) self-observer agreement. Finally, it is possible to assume, given a lack of veridical information, that raters are more inclined to report their own or an average personality. Indeed, we found that observed self-observer agreement coefficients were negatively correlated with mean scores of the facet scales: r(28) =.45, p =.030. This may indicate a tendency to (incorrectly) perceive others as similar to an average person. However, this correlation becomes insignificant when the correction for the restriction of range is taken into account [r(28) =.03, p =.874]. 4. Discussion Nobody seems to doubt that some personality traits, such as Extraversion, are easier to judge than other traits, such as Neuroticism. For this very reason, a higher self-observer agreement is usually reached on easy-to-judge traits than on traits which are not as visible to an external observer. This study provides a serious challenge to this intuitively appealing picture by demonstrating that a considerable proportion of differences in personality trait agreement can be explained by the standard deviation of the sum scores on these traits. Traits which have a larger standard deviation are more likely to be among those for which higher interjudge agreement is reached. The reverse is also true: on traits with

5 J. Allik et al. / Journal of Research in Personality 44 (2010) Observed (r xy ) Corrected (R xy ) Self-Observer Agreement N1:Anxiety N2:Angry Hostility N3:Depression N4:Self-Consciousness N5:Impulsiveness N6:Vulnerability E1:Warmth E2:Gregariousness E3:Assertiveness E4:Activity E5:Excitement Seeking E6:Positive Emotions O1:Fantasy O2:Aesthetics O3:Feelings O4:Actions O5:Ideas O6:Values A1:Trust A2:Straightforwardness A3:Altruism A4:Compliance A5:Modesty A6:Tender-Mindedness C1:Competence C2:Order C3:Dutifulness C4:Achievement Striving C5:Self-Discipline C6:Deliberation Fig. 3. Self-observer agreement before (r xy ) and after (R xy ) correction for the restricted range of standard deviation. a smaller variance of answers around the mean score, a lower level of interjudge agreement is more likely to be achieved. Since the effect of restricted variance was very powerful, explaining approximately 50% of the interjudge agreement, it imposes limitations on other explanations as well. The effects of visibility and assumed similarity, for instance, cannot be proved unless it is demonstrated that they are not mediated through restrictions in the standard deviation of sum scores. Our own analysis demonstrated, when the effect of restricted range was corrected, that interjudge agreement on the subscales of Extraversion was not higher than any some subscales of Openness, Agreeableness, Conscientiousness, or even Neuroticism. If the restriction of range is compensated for, the predicted agreement on the subscales N2: Angry Hostility and N6: Vulnerability is on the same level as interjudge agreement on E4: Activity and E6: Positive Emotions. Thus, it seems that the often observed advantage of Extraversion in terms of interjudge agreement is at least partly caused by the fact that the standard deviation of Extraversion subscales happens to be one of the highest. The most intriguing question is, of course, what determines the standard deviation of personality traits. There are two major theoretical options, substantial and methodological. According to the substantial interpretation interindividual variance of personality traits is determined by their content. On some traits (for example more visible traits) personality judges have a higher variability of opinions than on some other traits. For some intrinsic reasons there are many different ways how people, for example, can express their appreciation for art and beauty but their readiness to reexamine social, political, and religious values demonstrates no such broad interindividual differences. In other words, different personality traits are perceived to vary in different extent and estimated accordingly. Although not entirely unlikely we are not aware of any solid evidences in favor of this theoretical interpretation. Another theoretical interpretation presumes that differences in the standard deviation of personality traits are caused by some methodological reasons. It is not the content of traits that poses restrictions on the interindividual variance but the way how personality items are constructed and selected for questionnaires. For example, the rank order of standard deviations of the 30 subscales in this study was highly correlated [r(28) =.82, p <.0001] with the rank order of the standard deviations in the North American normative adult sample (Costa & McCrae, 1992). This is not very surprising given that the Estonian translation generally preserved the wording of the original questionnaire. More revealing is the fact that the rank order of standard deviations in this study replicated the rank order of standard deviations in the Estonian version of the International Personality Item Pool NEO (Goldberg, 1999), EE.PIP-NEO (Mõttus, Pullmann, & Allik, 2006), which has different, shorter, and grammatically simpler items. Identical subscales of these two questionnaires also tended to have similar standard deviations: r(28) =.69, p < However, this does not prove that it would not be possible to create another set of items which would generate a wider range of responses. Indeed, it would be premature to think that it is only possible to measure, for example, Openness to Values (O6) with eight items, the summary scores of which cover just 75% of the whole range (see Fig. 2B). The replicability of the pattern of standard deviations from one NEO family questionnaire to another may speak first and foremost about a persistent item construction tradition, not about standard deviations that are intrinsic to some personality traits. One novel conclusion of this paper is a need to consider standard deviations of subscales during their construction. Test constructors are primarily interested in the loadings of a given item on the intended and other factors rather than equalizing standard deviations of sum scores on various subscales. But as soon as we want to say anything definite about differences between personality traits in their cross-observer agreement or predictive validity in general we need to take into account the scatter of responses around the mean values. Perhaps a requirement to have an equal standard deviation on all subscales is unrealistic. A more downto-earth recommendation would be to use standard deviation as a covariate whenever potential determinants of self-observer agreement are considered. The same is true when differences

6 426 J. Allik et al. / Journal of Research in Personality 44 (2010) between subscales in their ability to predict whatever criteria is investigated. Whatever is the reason of the differences in variance between the scales, the subscales with higher variance have higher chance of predicting any criteria. In other words, we can study individual differences in those traits only that vary across people. If the proposal that the range of sum scores can, in principle, be corrected is accurate, then the problem of determinants of self-observer agreement on the Big Five personality traits can be seen from a new perspective. Given that, after correcting for range of sum scores, the advantage of more visible traits, such as Extraversion, over less visible traits, such as Neuroticism, disappears, it seems more likely that there are no substantial differences between the Big Five personality traits and approximately the same level of interjudge agreement can be reached, at least when judges know their targets well. Although the notion that the visibility or judgability of traits may not be the main factor that moderates interjudge agreements is counterintuitive, it is still the finding which best agrees with the reported data (Connolly et al., 2007; Hall et al., 2008; McCrae et al., 2004). It also provides a new agenda for future studies: why is it that when people know each other well, they attain approximately equal accuracy in judgment on all Big Five personality traits? Acknowledgments This project was supported by a grant from the Estonian Ministry of Science and Education (SF s08). Andres Metspalu and Tõnu Esko received support from FP7 grants [ ENGAGE, BBMRI, ECOGENE (#205419, EBC)]. They also received targeted financing from Estonian Government SF s08 and by EU via the European Regional Development Fund, in the frame of Centre of Excellence in Genomics. We are very thankful to Delaney Michael Skerrett for his helpful comments and suggestions. References Allik, J., Realo, A., Mõttus, R., & Kuppens, P. (2010). Generalizability of self-other agreement from one personality trait to another. Personality and Individual Differences, 48, Beer, A., & Watson, D. (2008). Personality judgment at zero acquaintance: Agreement, assumed similarity, and implicit simplicity. Journal of Personality Assessment, 90(3), Colvin, C. R. (1993). Judgable people: Personality, behavior, and competing explanations. Journal of Personality and Social Psychology, 64(5), 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(1), Costa, P. T., & McCrae, R. R. (1992). Revised NEO personality inventory (NEO PI-R) and NEO five-factor inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources. Cronbach, L. J. (1955). Process affecting scores on understanding others and assumed similarity. Psychological Bulletin, 52, Funder, D. C. (1995). On the accuracy of personality judgment: A realistic approach. Psychological Review, 102(4), Funder, D. C. (1999). Personality judgement: A realistic approach to person perception. San Diego: Academic Press. Funder, D. C., & Colvin, C. R. (1988). Friends and strangers: Acquaintanceship, agreement, and the accuracy of personality judgment. Journal of Personality and Social Psychology, 55(1), Funder, D. C., & Colvin, C. R. (1997). Congruence of others and self-judgements on personality. In R. Hogan, J. Johnson, & S. Briggs (Eds.), Handbook of personality psychology (pp ). San Diego: Academic Press. Funder, D. C., & Dobroth, K. M. (1987). Differences between traits: Properties associated with interjudge agreement. Journal of Personality and Social Psychology, 52(2), Furr, R. M., Dougherty, D. M., Marsh, D. M., & Mathias, C. W. (2007). Personality judgment and personality pathology: Self-other agreement in adolescents with conduct disorder. Journal of Personality, 75(3), Gangestad, S. W., Simpson, J. A., DiGeronimo, K., & Biek, M. (1992). Differential accuracy in person perception across traits: Examination of a functional hypothesis. Journal of Personality and Social Psychology, 62, 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.: Tilburg University Press. Hall, J. A., Andrzejewski, S. A., Murphy, N. A., Schmid Mast, M., & Feinstein, B. A. (2008). Accuracy of judging others traits and states: Comparing mean levels across tests. Journal of Research in Personality, 42(6), John, O. P., & Robins, R. W. (1993). Determinants of interjudge agreement on personality traits: The Big Five domains, observability, evaluativeness, and the unique perspective of the self. Journal of Personality, 61(4), Kallasmaa, T., Allik, J., Realo, A., & McCrae, R. R. (2000). The Estonian version of the NEO-PI-R: An examination of universal and culture-specific aspects of the fivefactor model. European Journal of Personality, 14, Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York: Guilford Press. Konstabel, K., Aavik, T., & Allik, J. (2006). Social desirability and consensual validity of personality traits. European Journal of Personality, 20(7), Letzring, T. D., Wells, S. A., & Funder, D. C. (2006). Information quantity and quality affect the realistic accuracy of personality judgment. Journal of Personality and Social Psychology, 91(1), McCrae, R. R., Costa, P. T., & Martin, T. A. (2005). The NEO-PI-3: A more readable revised NEO personality inventory. Journal of Personality Assessment, 84(3), McCrae, R. R., Costa, P. T., Martin, T. A., Oryol, V. E., Rukavishnikov, A. A., Senin, I. G., et al. (2004). Consensual validation of personality traits across cultures. Journal of Research in Personality, 38(2), Mõttus, R., Pullmann, H., & Allik, J. (2006). Toward more readable Big Five personality inventories. European Journal of Psychological Assessment, 22(3), Park, B., & Judd, C. M. (1989). Agreement on initial impressions: Differences due to perceivers, trait dimensions, and target behaviors. Journal of Personality and Social Psychology, 56(4), Paunonen, S. V. (1989). Consensus in personality judgements: Moderating effects of target rater acquaintanceship and behavior observability. Journal of Personality and Social Psychology, 56, Ready, R. E., Clark, L. A., Watson, D., & Westerhouse, K. (2000). Self- and peerreported personality: Agreement, trait ratability, and the self-based heuristic. Journal of Research in Personality, 34(2), Realo, A., Allik, J., Nolvak, A., Valk, R., Ruus, T., Schmidt, M., et al. (2003). Mindreading ability: Beliefs and performance. Journal of Research in Personality, 37(5), Sackett, P. R., & Yang, H. (2000). Correction for range restriction: An expanded typology. Journal of Applied Psychology, 85(1), Taft, R. (1955). The ability to judge people. Psychological Bulletin, 52, Tausch, N., Kenworthy, J. B., & Hewstone, M. (2007). The confirmability and disconfirmability of trait concepts revisited: Does content matter? Journal of Personality and Social Psychology, 92(3), Thorndike, R. L. (1949). Personnel selection: Test and measurement techniques. New York: Wiley. Watson, D., Hubbard, B., & Wiese, D. (2000). Self-other agreement in personality and affectivity: The role of acquaintanceship, trait visibility, and assumed similarity. Journal of Personality and Social Psychology, 78(3),