Multistage selection strategies: Simulating the effects on adverse impact and expected performance for various predictor combinations.

Examination of the trade-off between mean performance and adverse impact has received empirical attention for single-stage selection strategies; however, research for multistage selection strategies is almost nonexistent. The authors used Monte Carlo simulation to explore the trade-off between expected mean performance and minority hiring in multistage selection strategies and to identify those strategies most effective in balancing the trade-off. In total, 43 different multistage selection strategies were modeled; they reflected combinations of predictors with a wide range of validity, subgroup differences, and predictor intercorrelations. These selection models were examined across a variety of net and stage-specific selection ratios. Though it was still the case that an increase in minority hiring was associated with a decrease in predicted performance for many scenarios, the current results revealed that certain multistage strategies are much more effective than others for managing the performance and adverse impact trade-offs. The current study identified several multistage strategies that are clearly more desirable than those strategies previously suggested in the literature for practitioners who seek a practical selection system that will yield a high-performing and highly representative workforce. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Testing for adverse impact when sample size is small.

Adverse impact evaluations often call for evidence that the disparity between groups in selection rates is statistically significant, and practitioners must choose which test statistic to apply in this situation. To identify the most effective testing procedure, the authors compared several alternate test statistics in terms of Type I error rates and power, focusing on situations with small samples. Significance testing was found to be of limited value because of low power for all tests. Among the alternate test statistics, the widely-used Z-test on the difference between two proportions performed reasonably well, except when sample size was extremely small. A test suggested by G. J. G. Upton (1982) provided slightly better control of Type I error under some conditions but generally produced results similar to the Z-test. Use of the Fisher Exact Test and Yates's continuity-corrected chi-square test are not recommended because of overly conservative Type I error rates and substantially lower power than the Z-test. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Modeling the behavior of the 4/5ths rule for determining adverse impact: Reasons for caution.

The Equal Employment Opportunity Commission's 4/5ths rule has been used for over 20 years in applied psychology and employment law. The rule signals that there is adverse impact when the protected group selection ratio is less than 80% of the highest scoring group's selection ratio. We conducted several simulations and found, consistent with some previous management science literature, that the 4/5ths rule often resulted in false-positive readings of adverse impact even when there were no underlying (population) standardized group differences between subgroups. We then incorporated tests of statistical significance and found that adding such tests to the 4/5ths rule eliminated many false-positive indications of adverse impact. We also examined simulated selection systems based on meta-analytic values from the selection literature. The frequency of adverse impact signals from the 4/5ths rule increased markedly relative to simulations with no subgroup population differences. Adding statistical tests mitigated the number of indications of adverse impact to some extent. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Bootstrapping to test for nonzero population correlation coefficients using univariate sampling.

This article proposes 2 new approaches to test a nonzero population correlation (ρ): the hypothesis-imposed univariate sampling bootstrap (HI) and the observed-imposed univariate sampling bootstrap (OI). The authors simulated correlated populations with various combinations of normal and skewed variates. With αset=.05, N≥10, and ρ≤0.4, empirical Type I error rates of the parametric r and the conventional bivariate sampling bootstrap reached .168 and .081, respectively, whereas the largest error rates of the HI and the OI were .079 and .062. On the basis of these results, the authors suggest that the OI is preferable in alpha control to parametric approaches if the researcher believes the population is nonnormal and wishes to test for nonzero ρs of moderate size. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Recruitment efforts to reduce adverse impact: Targeted recruiting for personality, cognitive ability, and diversity.

Noting the presumed tradeoff between diversity and performance goals in contemporary selection practice, the authors elaborate on recruiting-based methods for avoiding adverse impact while maintaining aggregate individual productivity. To extend earlier work on the primacy of applicant pool characteristics for resolving adverse impact, they illustrate the advantages of simultaneous cognitive ability– and personality-based recruiting. Results of an algebraic recruiting model support general recruiting for cognitive ability, combined with recruiting for conscientiousness within the underrepresented group. For realistic recruiting effect sizes, this type of recruiting strategy greatly increases average performance of hires and percentage of hires from the underrepresented group. Further results from a policy-capturing study provide initial guidance on how features of organizational image can attract applicants with particular job-related personalities and abilities, in addition to attracting applicants on the basis of demographic background. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Educational attainment as a proxy for cognitive ability in selection: Effects on levels of cognitive ability and adverse impact.

The authors examined the differences in mean level of cognitive ability and adverse impact that can be expected when selecting employees solely on educational attainment as a proxy for cognitive ability versus selecting employees directly on cognitive ability. Selection using cognitive ability worked as a more efficient cognitive screen. Imposing an educational attainment standard of at least 1 year of college, though, did result in noticeably higher levels of cognitive ability in potential applicant pools than did random selection, meaning that educational attainment does work as a cognitive screen. These results held not only in a nationally representative sample but also within and across 6 different occupational groups. Finally, adverse impact is examined for selection using educational attainment, compared with selection on the basis of cognitive ability. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Adjustments to the correction for attenuation.

With respect to the often-present covariance between error terms of correlated variables, D. W. Zimmerman and R. H. Williams's (1977) adjusted correction for attenuation estimates the strength of the pairwise correlation between true scores without assuming independence of error scores. This article focuses on the derivation and analysis of formulas that perform the same function for partial and part correlation coefficients. Values produced by these formulas lie closer to the actual true-score coefficient than do the observed-score coefficients or those obtained by using C. Spearman's (1904) correction for attenuation. The new versions of the formulas thus allow analysts to use hypothetical values for error-score correlations to estimate values for the partial and part correlations between true scores while disregarding the independence-of-errors assumption. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

An examination of factors contributing to a reduction in subgroup differences on a constructed-response paper-and-pencil test of scholastic achievement.

The authors investigated subgroup differences on a multiple-choice and constructed-response test of scholastic achievement in a sample of 197 African American and 258 White test takers. Although both groups had lower mean scores on the constructed-response test, the results showed a 39% reduction in subgroup differences compared with the multiple-choice test. The results demonstrate that the lower subgroup differences were explained by more favorable test perceptions for African Americans on the constructed-response test. In addition, the two test formats displayed comparable levels of criterion-related validity. The results suggest that the constructed-response test format may be a viable alternative to the traditional multiple-choice test format in efforts to simultaneously use valid predictors of performance and minimize subgroup differences in high-stakes testing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

"On the question of professional standards for computer-based test interpretation": Response to Kramer on Matarazzo.

Responds to the comments by J. J. Kramer (see record 1990-58925-001) on the current author's original article (see record 1986-19878-001), in which Matarazzo contends that computerized clinical psychological test interpretations offer considerable potential for improving the work of health service providers. However, cost additions associated with this technology have increased the probabilities that legislative, judicial, and administrative restrictions and sanctions will be imposed. Here, Matarazzo notes that he finds nothing in Kramer's comments that is inconsistent with the opinions that Matarazzo himself offered. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Using the Delay Discounting Task to test for failures in ego control in substance abusers: A meta-analysis.

A meta-analysis was conducted to test a core feature of the self-medication hypothesis of substance use disorders (SUDs) that posits that SUDs are associated with failures of ego control. This hypothesis was tested by quantitatively synthesizing studies that used the delay discounting task in individuals with SUDs compared with nonclinical controls. An additional aim of the study was to identify variables that might moderate the relationship between SUDs and delay discounting task performance. The results were based on 10 studies that consisted of 925 participants who used a variety of psychoactive substances, and all studies used matched-groups designs. Participants with SUDs showed significant failures in ego control compared with nonclinical controls (d = 0.587). Variance across studies was nil, and therefore no moderators were sought. The results of this study support the self-medication hypothesis and suggest that SUD treatment should focus on improving ego control. The results also show that the delay discounting task is an experimental method that can be used to operationalize a number of psychoanalytic constructs and thus opens the door to further psychoanalytic experimental research on failures of ego control. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Using the five-factor model to identify a new personality disorder domain: The case for experiential permeability.

Two studies were conducted to demonstrate that maladaptive aspects of high and low Openness to Experience were related to characterological impairment and that this aspect of personality may define a new domain of personality dysfunction. The 55-item Experiential Permeability Inventory (EPI; containing 4 scales) was developed and demonstrated to have acceptable psychometric properties. Evidence of convergent, discriminant, and incremental validity was provided. These studies provide a methodological framework for identifying and developing aspects of personality dysfunction that can expand the comprehensiveness of the current set of Axis II disorders. Theoretical implications of the EPI are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)