I use a path/confirmatory factor analytic framework to present an under-the-hood introduction to (a) the effects of measurement error and method variance on observed measures, (b) the effects of common latent Trait and Method factors on relationships between observed measures in the presence of measurement error and (c) examine implications of developments presented here for the possible reinterpretation of Campbell and Fiske's original criteria for convergent and discrimainat validity and presence of common method bias. I use results from one previously published and often-cited primary study and four meta-analyses of multitrait-multimethod studies to provide empirical examples to illustrate points made here. I show analytically that (a) some common interpretations of method effects are correct, (b) others are incorrect and (c) still others are more nuanced than is usually recognised. Finally, I offer some suggested directions for future research.

Background and Objectives We explored whether relationship quality, as measured by mutual communal behavior, would serve as a buffer against caregiver resentment and, consequently, depressive symptoms when care recipients (CRs) engage in problem behavior (i.e., controlling and manipulative behavior [CMB]). Using the common core model of caregiver distress, we hypothesized that caregiver resentment would explain why caregivers were depressed when their CRs engage in CMB. We predicted that the indirect effect of CRCMB on depressive symptoms via resentment would depend on the strength of the communal bond between the caregiver and CR. Resentment was expected to play a significant role in explaining the association between CRCMB and depressive symptoms for caregivers in less communal relationships, but a small or nonexistent role for those in highly communal relationships. We also investigated whether these effects were different for Black and White caregivers. Research Design and Methods Data were obtained from 187 Black and 247 White caregivers from the second Family Relationships in Late Life Project. Before testing our model, we confirmed the measurement equivalence/invariance of the four scales used in this study. Results Resentment mediated the association between CRCMB and depressive symptoms. However, the indirect effect was larger among highly communal caregivers. Caregiver race did not moderate the moderated mediation. Discussion and Implications The communal bond between the caregiver and CR does not entirely protect the caregiver from depressive symptoms, particularly among highly communal caregivers. Interventions aimed at improving caregiver outcomes should acknowledge the unique vulnerabilities of caregivers in close relationships.

There has been little theoretical and empirical attention to the role of time in studying levels of work–family conflict. Contrary to theoretical assumptions that work–family conflict is a highly dynamic construct, we borrow from established theories to describe a Stability and Change Model, which posits that work–family conflict levels are primarily stable with some meaningful change over time. The proposed model also outlines predictors of stability (and change) in work–family conflict levels: stability (or change) in‐person (i.e., neuroticism) and situational factors (i.e., role demands and, to a lesser extent, resources). Findings across two longitudinal studies with timeframes of 5 months (one‐month lags) and 3 years (one‐year lags) suggest that 75–80% of the observed variance in work–family conflict levels is stable over mid‐ to long‐term timeframes. In line with our Stability and Change Model, trait‐state‐occasion modelling results indicate that stability in work–family conflict is primarily due to stability in workload demands and to a lesser extent, personality factors (i.e., neuroticism), but, unexpectedly, not resources. There was some meaningful change in work–family conflict levels over time, primarily due to changes in demands and to a lesser extent, changes in person characteristics. Theoretical, methodological, and analytical considerations for future studies of work–family conflict levels are discussed.Practitioner points WFC levels may be very resistant to change. Our findings highlight the importance for organizations to make a sustained commitment to addressing factors that might create WFC rather than short‐term efforts. Organizations may design impactful interventions by aiming to significantly and durably reduce job demands, specifically: High involvement work practices that foster empowerment, information sharing, rewards, and training have been shown to reduce job demands including role overload (Kilroy et al., 2016, Human Resource Management Journal, 26, 408). Supervisors could reduce workload by eliminating unnecessary tasks or prioritizing most important tasks and/or developing employees’ time management skills so they can accomplish important job tasks in the time available. Reducing hindrance stressors such as role ambiguity and hassles tend not to be as relevant. Organizations may also want to consider person variables (e.g., personality, affect) by training for individual characteristics that may help employees manage WFC such as mindfulness training.

Over the next two decades, many countries, including the USA, will experience an unprecedented and seismic demographic transition as their older adult populations grow, both in number and as a proportion of the overall population. During this time, the number of working adults with eldercare responsibilities will grow substantially, posing a threat to the well-being and economic stability of both workers and their organizations. This phenomenon is a long anticipated symptom of the greater “eldercare crisis,” yet the response of organizations to this impending change has been relatively slow, and the experiences of working caregivers remain an almost entirely neglected area of research in the organizational sciences. In this special issue, we begin a conversation with the IO/OB community regarding the importance of studying the experiences of caregivers and their employing organizations. We introduce six original research papers, which highlight various issues facing eldercarers and the implications of these findings for employees and their employers. We also summarize themes from across these and other published papers on eldercare, providing an overview and directions for future research on eldercare with the intention of springboarding more research on this critical and timely topic. We end with next steps for researchers and practitioners who wish to collaborate to bring about meaningful innovative solutions to this problem.

Spielberger's state and trait anxiety and anger scales are widely used and documented, but there is little or no direct evidence that they actually measure their respective state and trait aspects as was intended. We conducted latent state-trait analyses on data collected from 310 community-dwelling caregivers of older adult care recipients and found that (a) both state and trait scales reflected a mixture of state and trait aspects of their latent constructs, (b) state scales reflected more state-like variance than did corresponding trait scales, but (c) both state and trait scales were dominated by stable trait-like variance. Follow-up bivariate latent state-trait analyses indicated that correlations between trait components of anger and anxiety correlated more strongly with trait components of caregiver-care recipient mutually communal behavior and care recipient problem behavior than did state-state component correlations. Implications for the measurement of state and trait components of psychological constructs are discussed.

Furloughs refer to placing employees on a temporary leave with no pay for the period of the leave. The current study draws from conservation of resources (COR) theory to examine how furloughs affect employees’ experiences of burnout, work–family conflict, and life satisfaction. Results gathered from 212 individuals show that being furloughed during the 2013 U.S. federal government shutdown was associated with perceived personal resource loss, which was related to decreased life satisfaction and increased work–family conflict and physical, cognitive, and emotional burnout 5 weeks after the shutdown ended. The relationships between furlough status and all outcomes were fully mediated by perceived resource loss. Our findings show that furloughs can and do negatively affect employees and that these effects last long after the furlough has ended.

Objectives: Two studies and data sources were used to empirically compare and cross-validate three bidimensional acculturation measurement models. Method: Following a cross-sectional design and confirmatory factor analysis (CFA), Study 1 examined the models in a sample of Central Florida Puerto Ricans (n = 484). In Study 2, the measurement model was retested on a separate sample of island Puerto Ricans (n = 399). To ensure adequate comparisons across the two samples, measurement invariance tests were performed. Results: A bidimensional six-factor model (B6F) resulted in the best model fit for both samples. Results indicated full configural, metric, uniqueness, factor variance, covariance, and partial scalar invariance between the 2 groups. An examination of mean differences between the samples revealed that Central Florida Puerto Ricans had a significantly stronger endorsement of Puerto Rican and European American values than those living in the island. Conclusions: This examination revealed that both groups share the same acculturation structure. That is, acculturation among Central Florida and island Puerto Ricans can be conceptualized as a bidimensional construct of simultaneous adherence to Puerto Rican and European American cultural aspects across behavioral, values, and ethnic identity domains. The B6F model of island Puerto Rican acculturation suggests that even in the absence of a migration experience, island Puerto Ricans may experience bidimensional and multifaceted acculturation.

Trust propensity is typically conceptualized as a stable, trait-like, exogenous variable. Drawing on the social investment principle of personality change, we argue that trust propensity has situationally specific components and is likely to be less stable during periods of career transition. Using a latent curve-latent state-trait model, we present evidence that suggests that trust propensity has stable (trait) and unstable (state) components during career transition periods and that it has the potential to change over time. Our results are replicated across two, transitional workplace populations during a process of (re)socialization into an organization. In our second study, we also expand our focus to examine correlates of trust propensity and demonstrate the relationship between state and trait trust propensity and cognitive depletion. Our paper significantly extends knowledge of the nature of trust propensity and raises questions about the stability of this construct, one of the core tenets of trust theory.

The purposes of this study were to (a) provide additional competitive tests between three models of relationships between overall and life facet satisfaction [Bottom-Up (BU), Top-Down (TD), and Bidirectional (BD) models], and (b) explore whether culture moderates these relationships. Models were tested using data collected as part of Michalos' (1991) global study of student well-being from samples of college students in 32 different countries. The BD model received strongest support in 29 of the 32 samples (countries); the TD model was most strongly supported in the remaining 3 samples. Cluster analyses of samples in which the BD model was supported indicated a 7-cluster solution of reasonably homogeneous sets of OLS-LFS relationships. However, clusters were not easily interpretable in terms of evident cultural commonalities. We conclude that (a) a BD model describes OLS-LFS relationships globally, (b) and although culture appears to moderate OLS-LFS relationships, (c) additional research is needed to explain why.

The correlated trait-correlated method (CTCM) model for the analysis of multitrait-multimethod (MTMM) data is known to suffer convergence and admissibility (C&A) problems. We describe a little known and seldom applied reparameterized version of this model (CTCM-R) based on Rindskopf's reparameterization of the simpler confirmatory factor analysis model. In a Monte Carlo study, we compare the CTCM, CTCM-R, and the correlated trait-correlated uniqueness (CTCU) models in terms of C&A, model fit, and parameter estimation bias. The CTCM-R model largely avoided C&A problems associated with the more traditional CTCM model, producing C&A solutions nearly as often as the CTCU model, but also avoiding parameter estimation biases known to plague the CTCU model. As such, the CTCM-R model is an attractive alternative for the analysis of MTMM data.

Almost all goodness-of-fit indexes (GFIs) for latent variable structural equation models are global GFIs that simultaneously assess the fits of the measurement and structural portions of the model. In one sense, this is an elegant feature of overall model GFIs, but in another sense, it is unfortunate as the fits of the 2 different portions of the model cannot be assessed independently. We (a) review the developing literature on this issue, (b) propose 6 new GFIs that are designed to evaluate the structural portion of latent variable models independently of the measurement model, (c) that are couched within a general taxonomy of James, Mulaik, and Brett's (1982) Conditions 9 and 10 for causal inference from nonexperimental data, (d) conduct a Monte Carlo simulation of the usefulness of these 6 new GFIs for model selection, and (e) on the basis of simulation results provide recommended criteria for 4 of them. Supplemental analyses also compare 2 of the new GFIs to 2 other structural model selection strategies currently in use. (PsycINFO Database Record

We compared six different analytic models for multitrait-multimethod (MTMM) data in terms of convergence, admissibility, and model fit to 258 samples of previously reported data. Two well-known models, the correlated trait-correlated method (CTCM) and the correlated trait-correlated uniqueness (CTCU) models, were fit for reference purposes in comparison to four other under- or unstudied models, including (a) Rindskopf's reparameterization of the CTCM (CTCM-R) model, (b) a correlated trait-constrained uncorrelated method model and two of its more general cases, (c) a correlated trait-constrained correlated method model, and (d) a correlated trait-uncorrelated method model. Results show that (a) the CTCM-R model often solved convergence and admissibility problems with the CTCM model at rates equivalent to the CTCU model and (b) constrained models often provided convergent and admissible solutions but significantly worse model fit, indicating that they are often not plausible when analyzing real data. A follow-up simulation study showed that the CTCM-R model also provided the most accurate estimates of the full range of parameters relevant to a confirmatory factor analytic model of MTMM data.

This study uses meta-analysis and a qualitative review of exercise descriptions to evaluate the content, criterion-related, construct, and incremental validity of 5 commonly used types of assessment center (AC) exercises. First, we present a meta-analysis of the relationship between 5 types of AC exercises with (a) the other exercise types, (b) the 5-factor model of personality, (c) general mental ability (GMA), and (d) relevant criterion variables. All 5 types of exercises were significantly related to criterion variables (ρ = .16-.19). The nomological network analyses suggested that the exercises tend to be modestly associated with GMA, Extraversion and, to a lesser extent, Openness to Experience but largely unrelated to Agreeableness, Conscientiousness, and Emotional Stability. Finally, despite sparse reporting in primary studies, a content analysis of exercise descriptions yielded some evidence of complexity, ambiguity, interpersonal interaction, and fidelity but not necessarily interdependence. (PsycINFO Database Record

Being familiar with their earlier work investigating the factor structures of the Armed Services Vocational Aptitude Battery and the Air Force Officer Qualifying Test, we read with interest Ree, Carretta, and Teachout's (2015) proposal to extend the idea of a dominant general factor (DGF) beyond the realm of cognitive abilities to other areas of research and practice in industrial–organizational (I-O) psychology. We found their ideas intriguing and their arguments compelling, but we stumbled on a reference to an article of one of the present authors (Lance, Teachout, & Donnelly, 1992) and Ree et al.’s claim that Lance et al. (1992) had found a DGF that accounted for 59% of the variance in job performance ratings in a military job. They did not. Rather, Lance et al. reported a hierarchical confirmatory factor analysis (CFA) model that explained parsimoniously the correlations among 15 job performance first-order factors in terms of four Job Proficiency and four other Measurement Source second-order factors. We reasoned that Ree et al. must have conducted some secondary analysis on the results presented by Lance et al., and indeed we replicated their claim by finding that the first unrotated principal component accounted for 59% of the variance in correlations among the four Proficiency second-order factors reported in Lance et al.’s Table 6.

Being familiar with their earlier work investigating the factor structures of the Armed Services Vocational Aptitude Battery and the Air Force Officer Qualifying Test, we read with interest Ree, Carretta, and Teachout's (2015) proposal to extend the idea of a dominant general factor (DGF) beyond the realm of cognitive abilities to other areas of research and practice in industrial–organizational (I-O) psychology. We found their ideas intriguing and their arguments compelling, but we stumbled on a reference to an article of one of the present authors (Lance, Teachout, & Donnelly, 1992) and Ree et al.’s claim that Lance et al. (1992) had found a DGF that accounted for 59% of the variance in job performance ratings in a military job. They did not. Rather, Lance et al. reported a hierarchical confirmatory factor analysis (CFA) model that explained parsimoniously the correlations among 15 job performance first-order factors in terms of four Job Proficiency and four other Measurement Source second-order factors. We reasoned that Ree et al. must have conducted some secondary analysis on the results presented by Lance et al., and indeed we replicated their claim by finding that the first unrotated principal component accounted for 59% of the variance in correlations among the four Proficiency second-order factors reported in Lance et al.’s Table 6.

This special issue is focused on how null results can meaningfully advance science and practice. This editorial describes some of the unique aspects of creating a special issue devoted to null results, offers opinions regarding why and under what conditions null results should be valued, and offers recommendations for key stakeholders (i.e., editors, reviewers, and authors) in the publication enterprise with respect to submitting and evaluating research with null findings. This editorial also provides summaries of the papers included in this special issue with emphasis on how they serve as example for what constitutes meaningful null results.

Impulsivity has been consistently associated with pathological gambling (PG), but the diversity of definitions and measures of impulsivity has led to ambiguity with regard to which indices are independently relevant. Toward clarifying this relationship, the current study examined indices from an array of commonly used impulsivity measures in relation to PG severity in an adult community sample of frequent gamblers (N = 353). These included both survey assessments and behavioral tasks. Using a factor analytic approach, 4 latent factors were identified among 19 indices and were designated reward sensitivity, punishment sensitivity, delay discounting, and cognitive impulsivity. All 4 latent variables were positively and independently related to PG severity, albeit at a trend level for cognitive impulsivity in a combined model. These findings reveal 4 generally independent domains of impulsivity that are related to PG severity, clarify which assessment measures aggregate in each domain, and illustrate the importance of measurement specificity in studying impulsivity in relation to PG and other psychiatric disorders.

AbstractMallows Cp statistic is used in multiple regression analysis to select models that contain smaller numbers of predictors from a larger number that is available for inclusion. Low values of Cp indicate regression models that are parsimonious yet still have good predictive ability.

The separate sections of this book build the case for the traditional dimension-based design of assessment centers, for designing assessment centers with a focus on exercises, rather than dimensions, and for a dimension-exercise hybrid approach. Each section contains a useful blend of theory, research, and practice, and serve as an outstanding example of Industrial/Organizational psychology's scientist-practitioner model. (PsycINFO Database Record (c) 2012 APA, all rights reserved). (foreword)