Journal of Management 1983, Voi 9, No 2, 113-126

A Multitrait-Munimethod Matrix Analysis of the PRF and M/VQ Need Sca/es
Bronston T. Mayes Daniel C. Ganster
University of Nebraska—Lincoln

The reliability and validity of the Manifest Needs Questionnaire (MNQ) and the Personality Research Form (PRF) nAch, nAff, nAut, and nPow scales were assessed in a heterogeneous sample (N = 136). Reliability was assessed by computing coefficient alpha for each scale. Validity assessments were based on criteria suggested by Campbell and Fiske (1959) for evaluating the multitrait-multimethod (MTMM) matrix. An additional analysis of variance suggested by Kavanagh, MacKinney, andWolins (1971) was applied to the MTMM matrix; thus allowing the quantitative identification of the effects of subjects, traits, and methods. The scales demonstrate convergent and discriminant validity according to these analyses. Where validity is in question, the cause seems to be due to the overall poor reliabilities of the MNQ scales. Only the MNQPOW scale was acceptably reliable, while all the PRF scales were acceptable. The unreliability of the MNQ seems to be a function of item heterogeneity and scale variance, while the PRF does not appear as sensitive to these factors.

Contingency theories of organizational behavior atwund and, as Korman (1973) suggested, all theories are contingency theories since boundary conditions should be specified to limit their applications. One such set of boundary conditions relevant to many theoretical domains is the strength of various needs assumed to characterize individuals. Of particular interest to students of organizational behavior are the needs for achievement, autonomy, affiliation, and power. Prior research has demonstrated the importance of these needs in moderating individual responses to goal characteristics (Steers, 1975), task difficulty (McClelland & Watson, 1973), job scope (Stone, Mowday, & Porter, 1977; Steers & Spencer, 1977), organizational