A Comparison of Statistical Techniques for Assessing the Effects of Moderator Variables in the Job Enrichment Process.
Abstract
Research efforts have used a variety of statistical analysis techniques--moderated regression analysis, subgroup analysis, analysis of variance (ANOVA), analysis of covariance (ANCOVA), and the Ghiselli technique--to assess the effect of moderator variables in the job enrichment process. Since analysis of the same set of data by various techniques has tended to produce different results, this research effort was designed to investigate the power of these five techniques to identify the effects of moderator variables. Monte-Carlo simulation was employed to generate data sets which either exhibited a moderator effect at a prespecified level or were devoid of such an effect. The simulated data were subjected to analysis with each of the techniques. Comparative results evidenced that the Ghiselli technique is not appropriate when the measurement of primary variables is based on a common scale; moderated regression analysis is always superior to ANOVA, ANCOVA, and subgroup analysis when the moderator variable is continuous; and a change in explained variation due to interaction of a moderator variable as small as two percent may be a good indicator of the presence of a moderator effect. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1978
- Accession Number
- ADA061358
Entities
People
- Homer L. Tackett
Organizations
- Air Force Institute of Technology