Date of Award


Document Type


Degree Name

Master of Arts




In a recent article, Hobert and Dunnette (1967) maintained that compared to the Absolute-difference and Algebraic-difference techniques Quadrant Analysis should yield the more effective moderators. Their assertion was based on the reasoning that more homogenous subgrouping should yield moderators with increased sensitivity to error. The present study was carried out as an empirical investigation of their assertions. Further, the design of the study permitted an investigation of moderator function.

The total sample (n=333) of male college students was randomly divided into a developmental sample and a cross-validation sample. Employing the aforementioned techniques a total of five moderator scales were developed and applied to the cross-validation sample. Based on the obtained findings, the following conclusions were drawn:

1. The tenability of the moderator model is questionable in the face of apparently fallacious reasoning concerning moderator characteristics and function.

2. Moderator function apparently can be more parsimoniously accounted for in terms of the suppression concept.

3. Previously suggested tests for the presence of moderator effects are inadequate in that they do not necessarily distinguish moderator effects from effects produced by suppressor variables

4. The use of scales developed by those techniques employing more homogeneous subgrouping results in the attainment of larger R's.