ERIC Number: EJ1115195
Record Type: Journal
Publication Date: 2016-Sep
Pages: 20
Abstractor: As Provided
ISBN: N/A
ISSN: EISSN-1531-7714
EISSN: N/A
Accuracy of Range Restriction Correction with Multiple Imputation in Small and Moderate Samples: A Simulation Study
Pfaffel, Andreas; Spiel, Christiane
Practical Assessment, Research & Evaluation, v21 n10 Sep 2016
Approaches to correcting correlation coefficients for range restriction have been developed under the framework of large sample theory. The accuracy of missing data techniques for correcting correlation coefficients for range restriction has thus far only been investigated with relatively large samples. However, researchers and evaluators are often faced with a small or moderate number of applicants but must still attempt to estimate the population correlation between predictor and criterion. Therefore, in the present study we investigated the accuracy of population correlation estimates and their associated standard error in terms of small and moderate sample sizes. We applied multiple imputation by chained equations for continuous and naturally dichotomous criterion variables. The results show that multiple imputation by chained equations is accurate for a continuous criterion variable, even for a small number of applicants when the selection ratio is not too small. In the case of a naturally dichotomous criterion variable, a small or moderate number of applicants leads to biased estimates when the selection ratio is small. In contrast, the standard error of the population correlation estimate is accurate over a wide range of conditions of sample size, selection ratio, true population correlation, for continuous and naturally dichotomous criterion variables, and for direct and indirect range restriction scenarios. The findings of this study provide empirical evidence about the accuracy of the correction, and support researchers and evaluators in their assessment of conditions under which correlation coefficients corrected for range restriction can be trusted.
Descriptors: Correlation, Sample Size, Error of Measurement, Accuracy, Statistical Analysis, Statistical Bias, Predictive Validity, Monte Carlo Methods, Bayesian Statistics, Maximum Likelihood Statistics, Multivariate Analysis
Center for Educational Assessment. 813 North Pleasant Street, Amherst, MA 01002. e-mail: pare@umass.edu; Tel: 413-577-2180; Web site: https://scholarworks.umass.edu/pare
Publication Type: Journal Articles; Reports - Research
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Grant or Contract Numbers: N/A