Construct Validity of Leader Effectiveness Criteria

摘要

Previous research on leadership effectiveness has often used criteria of one type and from one source. In this research, data on leadership effectiveness were collected with a variety of measures from different sources (leaders, peers, subordinates). Findings were analyzed for statistical convergence and by a multitrait-multimethod intercorrelation matrix. A satisfying degree of convergence but relatively little divergence appeared, and 47% of the variance was not accounted for by either one. Apparently, many measurement techniques or sources of data are essentially interchangeable. One cannot safely interpret results across different studies, however, because of the large amount of variance that remains unaccounted for by either convergent or divergent validity. The report is wirtten for psychologists. (Author)