Role conflict and role ambiguity: Do the scales measure these two constructs?

Abstract

Confirmatory factor analysis was applied to the conflict and ambiguity scales developed by Rizzo, House, and Lirtzman (1970). Alternative models were contrasted to evaluate the possibilities that (a) the 14 items comprising the scales do measure the two purported constructs, (b) the 14 items measure only one construct, or (c) the 14 items load complexly on a second-order factor model. The second-order factor model was superior across three independent subject samples (total n = 913), indicating that these measures do not establish role conflict and role ambiguity as two factorially independent constructs. The authors conclude that alternative scales are needed; suggestions for scale development are discussed.

Item content versus item wording: Disentangling role conflict and role ambiguity.

“…We posited a two-factor model comprising a general role-stress factor and a second, orthogonal, factor comprising all seven negatively worded items (i.e., the six role ambiguity items and the one negatively worded role overload item). Though similar to the method-factor model examined by McGee et al (1989), the two-factor model examined in the current study goes further by positing a method factor composed of items from two scales, thus disentangling the potential effects of item wording from those attributable to item content. Finally, the fourth model posits the existence of two method factors: one factor defined by all the positively worded items, and a second defined by all the negatively worded items.…”

“…A second issue arising from McGee et al's (1989) analysis is addressed in this study. Although their method-factor model provided the best fit to the data, this is to be expected because the fit of a model improves with the increase in the number of parameters estimated (Mulaik et al, 1989).…”

“…This possibility emerges because the six items comprising the ambiguity scale reflect the absence of ambiguity, In contrast, the eight conflict items reflect the presence of role conflict. Thus, the substantive interpretation of these two scales is perfectly confounded with the direction of item wording, rendering any interpretation open to plausible, rival hypotheses.In a recent confirmatory factor analysis of the role conflict and role ambiguity scales, McGee et al (1989) contrasted the goodness of fit of a one-factor (role stress) model, a two-factor (role conflict and role ambiguity) model, and another two-factor model (general role stress together with a method factor on which all the negatively worded items loaded). ~ McGee et al suggested that the last model provided the best fit to the data and concluded that the confound of item wording with item content compromises the construct validity of the scales, thereby precluding any substantive interpretations.…”

“…In a recent confirmatory factor analysis of the role conflict and role ambiguity scales, McGee et al (1989) contrasted the goodness of fit of a one-factor (role stress) model, a two-factor (role conflict and role ambiguity) model, and another two-factor model (general role stress together with a method factor on which all the negatively worded items loaded). ~ McGee et al suggested that the last model provided the best fit to the data and concluded that the confound of item wording with item content compromises the construct validity of the scales, thereby precluding any substantive interpretations.…”

“…To do so, we empirically contrasted the goodness of fit of four plausible models by using confirmatory factor analysis. The first model represents a general role-stress i McGee, Ferguson, and Steers (1989) referred to this as a secondorder model, comprising a single role-stress factor and a second-order method factor. We believe that their use of the term second-order factor, although correct, is potentially misleading.…”