CONTROL OF “HALO” IN FACTOR ANALYSIS OF A SUPERVISORY BEHAVIOR INVENTORY

Publisher Website
Google Scholar
Abstract
Summary: This paper has described a factor analysis of supervisory behavior ratings using a procedure of standardizing both by variable and by subject, in order to eliminate a component of correlation due to differences in subjective means and standard deviation. The results show that:1. The procedure results in a large reduction in the relative size of the first principle component.2. The procedure results in a systematic reduction in the correlations between the primary vectors of the oblique factor solution.3. The procedure results in a more clearly defined simple structure resulting from the oblique rotation.It seems likely that these results are due primarily to the elimination of the component of a variance attributable to the differences between subjective means. It also seems likely that some of the difference between the two solutions has resulted by giving less weight to the extreme responding, nondiscriminating subjects. Both of these adjustments are judged to be desirable in helping the researcher understand the true organization of reports of individual perceptions of behavior and attitudes where the universe concerned is homogeneous with retudes where the universe concerned is homogeneous with respect to some valuative or other over‐all response tendency.