This article appraises the construct validity of role conflict and role ambiguity in organizationa l psychology research by tracing the process of validation from initial theory introduction to the present. Although the theoretical bases for the constructs were relatively well explicated, some conceptual shortcomings have carried over into empirical work. There are problems with the content validity of measures, a lack of convergent and discriminant evidence, and inattention to the full nomological network of hypothesized relationships. Recommendations for future role-conflict and role-ambiguity research are offered. Since their incorporation into the organizational psychology literature in the late 1950s and early 1960s, the role-theory constructs of role conflict and role ambiguity have assumed a prominent position in empirical work. In recent years they have been the focus of three major literature reviews (Fisher & Gitelson, 1983; S. E. Jackson & Schuler, 1985; Van Sell, Brief, & Schuler, 1981), two of which used a meta-analytic approach (Hunter, Schmidt, & Jackson, 1982; Schmidt & Hunter, 1977; Schmidt, Hunter, Pearlman, & Shane, 1979). Although documenting significant correlations between these two role constructs and a variety of other work-related variables, all three reviews lamented, to a greater or lesser degree, inconsistencies in reported study findings. Extrapolating from the three reviews, a primary explanation offered for inconsistent results seems to be that research on role conflict and role ambiguity has suffered from a deficient theoretical framework, leading to a lack of coherent research direction, conceptual omissions, failure to consider important moderator variables, and inadequate linkages to other, related theoretical frameworks. This concern for deficient theory, however, tends to deemphasize the importance of measurement in evaluating the explanatory viability of such hypothetical abstractions as role conflict and role ambiguity. The role-conflict and role-ambiguity literature has been overwhelmingly dominated by studies using the self-report measures introduced by Rizzo, House, and Lirtzman (1970). A small stream of studies (Breaugh, 1980; House, Schuler, & Levanoni, 1983; Schuler, Aldag, & Brief, 1977; Tracy & Johnson, 1981) has been directed toward the psychometric qualities of these instruments, with a generally positive appraisal to date that the scales are well suited and appropriate to research in organizations. Indeed, all three of the major reviews
Read full abstract