|TITLE:||Development of a Biodata Measure of Leadership Skills|
Wayne State University
Doctoral Dissertation: May 1999
To develop a biodata (background characteristics) measure of leadership.
Participants are volunteers from the Southeast region of the super-center division of a large national discount retail chain. Fifty managers completed a life history essay questionnaire (16 of these managers had been identified previously by senior management as poor performers and the others as good performers). From this analysis a total of 157 biodata items were generated and grouped into seven dimensions. Fifty-six managers participated in the second phase (31% response rate) aimed at refining the instrument. Sixty items were subsequently retained and grouped into three general categories (educational experiences, problem solving, and mentor relationships). The third sample consisted of 734 participants (74% response rate), of which two-thirds were managers and one-third were hourly employees. The criterion measure consisted of supervisory ratings of performance on eight dimensions. Respondents also completed the Leadership Practices Inventory Individual Contributor, the Leader Behavior Questionnaire (Sahkin & Fulmer, 1985), the Multifactor Leadership Questionnaire (Bass, 1985), and measures of Self-Efficacy (Sherer, Maddux, Mercandante, Prentice-Dunn, Jacobs & Rogers, 1982) and Impression Management (Paulhus, 1988).
The internal reliability for the LPI-IC scales for this sample ranged from .73 to .86. The author reports that: "Overall, the pattern of correlation’s between the biodata scales and other measures of leadership suggests that each of the measures is distinct from the others, but that they are sufficiently related to suggest that they are measuring somewhat similar constructs.....None of the measures are so highly correlated as to suggest redundancy. That is, non of the measures could be substituted for another and be said to measure the same constructs. The pattern of correlation’s indicates that the leadership measures display evidence of construct validity" (p. 111).
Overall, the LBQ had the highest percentage of significant correlation’s with criterion dimensions (33%), followed by impression management (30%), and the LPI-IC (16%). The MLQ and self-efficacy scale did not correlate significantly with any of the criterion measures. There were no differences reported for gender, marital status, or age for the LPI-IC scales, although non-minorities scored significantly higher than minorities on all five of these leadership practices, as did managers versus non-managers.