Double agreement with reversed items: the plausibility of an alternative explanation to response bias

Evidence was presented which bears on the plausibility of an hypothesis proposed by Rokeach (8) as an alternative explanation to response bias for double agreement on authoritarian attitude scales. The results of a series of analyses of responses to attitude scale items indicated that (a) the average amount of double agreement with pairs of opposite items, in a reversed item modification of the Dogmatism Scale, was inconsistent by about one category on the response continuum, (b) inconsistent double rejection occurred infrequently among respondents who disagreed most with original, non reversed items, (c) contrary to the proposed alternative hypothesis, the Modified Dogmatism Scale is balanced for item social desirability, (d) agreement set appears to be a better explanation for double agreement than responsiveness to item social desirability, and (f) inconsistent responding in the “acquiescent” direction generalized somewhat to attitude items unrelated to the Modified Dogmatism Scale. It was suggested that confidence in the generality of these findings depends upon similar analyses of responses to other attitude scales, as well as the use of other samples of respondents. Also, additional exploration of double agreement might be relevant to another alternative explanation (“a weak need for logical consistency”) ta “acquiescence” suggested by Rokeach (8).