From critical-incident analysis and judgments by subject-matter experts, a low-fidelity simulation was developed for selecting entry-level managers in the telecommunications industry. The simulation presents applicants with descriptions of work situations and five alternative responses for each situation. It asks them to select one response they would most likely make and one they would least likely make in each situation. In a sample of approximately 120 management incumbents, simulation scores correlated from .28 (p < .01) to .37 (p < .01) with supervisory ratings of performance. These results show that samples of even hypothetical work behavior can predict performance, without the props, equipment, or role players often required by high-fidelity simulations, such as work-sample tests or assessment centers.Simulations used for employee selection typically present applicants with a task stimulus that mimics an actual job situation and elicit responses that are interpreted as direct indicators of how applicants would handle the task situation if it were actually to occur on the job. Thus, job simulations are designed more to sample job behaviors than to provide signs of underlying ability, temperament, or other traits presumed necessary for job performance. Arguing the basic tenet of behavioral consistency-that past performance is the best predictor of future performance-Wernimont and Campbell (1968) claimed that behavioral samples like those elicited by simulations can be very useful for predicting job performance, probably more useful than predispositional signs, such as results of standard ability, personality, or interest tests. Some forms of simulation, such as work samples and assessment centers, are by now familiar selection tools, and results of meta-analytic reviews show impressive support for their validity (Hunter & Hunter, 1984;Schmitt, Gooding, Noe, & Kirsch, 1984).Simulations vary in the fidelity with which they present a task stimulus and elicit a response. The highest fidelity simulations use very realistic materials and equipment to represent a task situation and provide applicants with an opportunity to respond almost exactly as if they were actually in the job situation. Fidelity decreases as stimulus materials and responses become less and less exact approximations of actual job stimuli and responses. At the lower end of the fidelity continuum are simulations that simply present a verbal description of a hypothetical work situation, instead of a concrete representation, and that ask applicants to describe how they would deal with the situation, instead of having them actually carry out some action to deal with it. In this article, simulations that present a

Recruiters from 8 telecommunications companies interviewed applicants or incumbents in four studies of the psychometric properties of structured behavioral interviews for management and marketing positions. Results yielded an interrater reliability estimate of .64 (n = 37), a mean criterion-related validity estimate of .22 (n ~ 500), evidence of convergent and discriminant validity, and small race and sex differences. In a fifth study, 3 doctoral students rated audiotapes and written summaries of 146 interviews conducted by recruiters. Results show that valid judgments are possible from information about interviewees' past behavior even without access to nonverbal cues in the interview itself and that judgmental accuracy is related to amount of relevant behavioral information, which, in turn, is related to interviewers' questioning skills.Early reviews of the selection interview are tinged with pessimism about its reliability and validity (e.g.,

A new method is presented for conducting differential prediction analyses that makes it possible to test differential prediction hypotheses with adequate statistical power even when the sample size within a job or a job family is very small. This method, called synthetic differential prediction analysis, represents an application of the logic of synthetic validation to differential prediction analyses. The authors explain this new method and describe its application in a selection-system validation study conducted in a large organization.

In this study, we demonstrated the usefulness of Occupational Information Network (O*NET) job analysis data for identifying applied adult literacy requirements across occupations. This was accomplished using a job component validity approach relating O*NET knowledge, skill, ability, and generalized work activity descriptor data to literacy test scores on the national adult literacy survey (NALS). Mean scores on prose, document, and quantitative literacy were estimated for 902 O*NET‐SOC occupations using descriptor data at the occupation level. Multiple correlation coefficients ranging from .79 to .81 (corrected for shrinkage) were found when predicting literacy scores from O*NET descriptors. Results of this study showed that 3 different types of applied adult literacy were highly predictable from the O*NET descriptors.

No abstract