Construct proliferation—the accumulation of ostensibly different but potentially identical constructs representing organizational phenomena—is a salient problem in contemporary research. While a number of construct validation procedures exist, relatively few validation studies conduct comprehensive assessments of the discriminant validity of theoretically distinct constructs. In this article, we outline the key considerations a researcher must take into account when attempting to establish the empirical distinctness of new or existing constructs and provide a step-by-step guide on how to assess the discriminant validity of constructs while accounting for three major sources of measurement error: random error, specific factor error, and transient error. Using a number of popular measures from the leadership literature, we provide an illustrative example of how to conduct a study of discriminant validity. We include several analytic strategies in our study and discuss the similarities and differences between the results they yield. We also discuss several additional issues related to this type of research and make recommendations for conducting discriminant validity analyses.

Drawing on the group-norms theory of organizational citizenship behaviors and person-environment fit theory, we introduce and test a multilevel model of the effects of additive and dispersion composition models of team members' personality characteristics on group norms and individual helping behaviors. Our model was tested using regression and random coefficients modeling on 102 research and development teams. Results indicated that high mean levels of extraversion are positively related to individual helping behaviors through the mediating effect of cooperative group norms. Further, low variance on agreeableness (supplementary fit) and high variance on extraversion (complementary fit) promote the enactment of individual helping behaviors, but only the effects of extraversion were mediated by cooperative group norms. Implications of these findings for theories of helping behaviors in teams are discussed.

This study describes a multilevel examination of person-group (PG) fit perceptions in a sample of 1023 individuals working in 92 teams at a private sector R&D firm. Using confirmatory factor analysis and multilevel random coefficient modeling, we provide evidence that perceptions of team-level collective fit are unique from aggregated individual-level PG fit perceptions at the individual and team levels. We demonstrate that collective values-based and abilities-based fit perceptions showed unique and positive relationships with team cohesion, team efficacy, and team performance, after accounting for aggregated individual perceptions of PG fit. Results also demonstrate that cohesion partially mediates the relationship between collective fit and team performance. Cross-level effects were also supported, indicating that collective fit explains additional variance in individual-level outcomes, beyond individual-level PG fit perceptions. The usefulness of employing a multilevel approach to studying PG fit is discussed. have begun to establish that such a construct exists, but research has not yet demonstrated that collective fit is differentiable from aggregated individual-level fit perceptions. These studies have also exclusively examined collective fit and team-level outcomes, disregarding possible cross-level effects that these emergent perceptions may have on individual-level outcomes (Bliese, 2000). By ignoring cross-level effects, current estimates of individual-level fit relationships may be biased because they fail to account for variance attributable to fit at higher levels. This study adds to the nascent literature on collective fit by using multilevel analysis to empirically demonstrate its difference from individual-level fit perceptions, and explore its relationships with work-relevant outcomes at both the team and individual levels.By doing so, this study contributes to the fit literature in multiple ways. First, building on early conceptual work distinguish between collective fit perceptions that form through referent-shift consensus (Chan, 1998) as derived from but conceptually distinct from individuals' experiences of fit within the team. We use multilevel confirmatory factor analysis (CFA) to provide evidence of this distinctiveness. Second, we use multilevel random coefficient modeling (MRCM) to test the cross-level effects of collective fit perceptions. The cross-level effects are particularly noteworthy, as they shed new light on possible bias in past estimates of PG fit relationships that did not account for the context of collective fit in the team. Finally, we extend prior work on outcomes of collective fit at the team-level, providing a holistic picture of the influence of collective fit on team-level and individual-level outcomes. Through these analyses, we validate collective fit perceptions as a construct that is unique from individual-level perceptions of PG fit, which has unique and incremental relationships with both team-level and individual-level outcomes.

This paper presents an account of the impact that research synthesis methods, in the form of psychometric meta-analysis, has had on industrial/organizational (I/O) psychology. This paper outlines the central contributions of psychometric meta-analysis in providing a method for developing cumulative knowledge. First, this paper describes the concerns and the state of the field before the development of meta-analytic methods. Second, the paper explains how meta-analysis addressed these concerns. Third, the paper details the development of psychometric meta-analysis through VG research and describes how the use of psychometric meta-analysis spread to other topic areas in the field. Finally, the paper presents illustrative example literatures, such as training and leadership, where meta-analysis had crucial impacts. Copyright © 2011 John Wiley & Sons, Ltd.

PurposeThe purpose of this study is to examine gender differences in personality predictors of a specific form of workplace aggression: counterproductive work behaviors directed at individuals (CWB‐I).Design/methodology/approachStudents (n=212) who were part‐time employees working at least 15 hours per week completed a measure of the five‐factor model (FFM) personality traits and two circumplex personality traits (Calmnesss and Pleasantness), as well as a measure of CWB‐I. Hierarchical regressions and tests of mean differences were used to examine hypotheses pertaining to gender differences in personality predictors of interpersonal aggression.FindingsResults generally supported the hypotheses as shown by the significant interactions between gender and personality traits in predicting CWB‐I. Agreeableness and Pleasantness significantly (negatively) predicted CWB‐I among males, but not females. Emotional Stability significantly (negatively) predicted CWB‐I among females, but not males.Research limitations/implicationsThe use of self‐report surveys may impact the results of this study. However, as this is the first study to explore the complex interactions between gender and personality in predicting workplace aggression, it is hoped that future research tests these relationships with alternate samples and methodologies.Practical implicationsThe results show that personality traits predict interpersonal workplace aggression differentially for males and females. Results also show that circumplex intersection traits are a useful supplement to the FFM traits in explaining interpersonal aggression in the workplace.Originality/valueTo the authors' knowledge, this is the first study to show that personality traits differentially predict interpersonal aggression for males and females; and to demonstrate the incremental validity of circumplex traits over FFM traits in predicting interpersonal aggression.