The problem of model selection uncertainty in structural equation modeling

Model selection in structural equation modeling (SEM) involves using selection criteria to declare one model superior and treating it as a best working hypothesis until a better model is proposed. A limitation of this approach is that sampling variability in selection criteria usually is not considered, leading to assertions of model superiority that may not withstand replication. We illustrate that selection decisions using information criteria can be highly unstable over repeated sampling and that this uncertainty does not necessarily decrease with increases in sample size. Methods for addressing model selection uncertainty in SEM are evaluated, and implications for practice are discussed.     

CAML—Maximum likelihood consensus analysis

Consensus analysis enables estimation of individual differences in competencies and response tendencies when answer keys to dichotomous forced-choice questions are unknown. CAML, a set of functions written in R, implements maximum likelihood estimation for the general Condorcet model that underlies consensus analysis. CAML avoids problems of alternative approaches that have often rendered consensus analysis impractical or unfeasible in the past. It provides (1) measures of model fit, (2) a measure of consensus, (3) point and interval estimates of competencies and response tendencies, and (4) an estimate of the unknown answer key. The present article describes the general Condorcet model, the CAML algorithms, and the handling of the software. In addition, the validity of CAML results is tested in a recognition memory study using selective experimental manipulations of the parameters. The results show that CAML works very well in practice and provides valid estimates of competencies, response tendencies, and answer keys.     

Using General Mental Ability and Personality Traits to Predict Job Performance in Three Chilean Organizations

Although the field of personnel selection has amounted around 100 years of research, there has been an overrepresentation of American and Western European samples in these studies. In particular, samples from Latin America have been almost entirely absent from industrial and organizational psychology journals. Thus, it is unknown whether well‐documented findings, such as the prediction of job performance based on general mental ability and conscientiousness, replicate in this region. This research intended to address this gap in the literature with three studies conducted in Chilean organizations, using different research designs, and different operationalizations of predictors and criteria. Results are generally consistent with previous studies, showing that conscientiousness and general mental ability significantly predict job performance in these Chilean samples.