For decades researchers, analysts, and organizational professionals have utilized correction equations to adjust for the effects of various statistical artifacts. However, every correction method has certain assumptions that must be satisfied to work properly. These assumptions are likely rarely satisfied for range restriction corrections. As a result, these correction methods are used in a manner that can lead to incorrect results.
The current study employed a Monte Carlo design to examine the direct range restriction correction. Analyses were conducted to examine the accuracy of adjustments made with the direct range restriction correction when its assumption of perfect top-down selection was violated to varying degrees. Analyses were conducted on two datasets, each representing a population of 1,000,000 cases. The following variables were manipulated: the population correlation, the selection ratio, and the probability that the hypothetical applicant would accept the job if offered. Results of the accuracy of the direct range restriction correction equation for the optimal (all job offers accepted) versus realistic (job offers refused at various rates) conditions demonstrated small differences in bias for all conditions. In addition, small differences in squared bias were observed for most of these conditions, with the exception of conditions with both low selection ratios and low probabilities of job offer acceptance. In a surprising finding, the direct range restriction correction equation exhibited greater accuracy for realistic job offer acceptance (some job offers refused) than for optimal job offer acceptance (all offers accepted). It is recommended that researchers further explore the violations of assumptions for correction methods of indirect range restriction as well.
| Identifer | oai:union.ndltd.org:WKU/oai:digitalcommons.wku.edu:theses-2586 |
| Date | 01 April 2016 |
| Creators | Hall, Austin J |
| Publisher | TopSCHOLAR® |
| Source Sets | Western Kentucky University Theses |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Masters Theses & Specialist Projects |
Page generated in 0.0023 seconds