An attempt was made to examine authoritarian and non-authoritarian scales of social attitudes and their reversals as a function of: (1) content consistency; (2) acquiescence; and (3) a tendency to use extreme categories of response. The study questioned whether Adorno's fascism (F), dogmatism (D), ethnocentrism (E), and anti-Semitism (A-S) scales measure authoritarian attitudes or whether the responses obtained are systematically biased in some way unrelated to overt content variables. A pilot study, which involved 75 male and 75 female college students, did not elicit the expected factor structure. Instead, eight factors emerged which were later replicated in a sample of 250 male college students. The eight factors showed responses to the F and D scales to be highly determined by an agreement response bias, whereas E and A-S are largely content-determined. Distinct sets to employ extreme responses and to be equivocal about item responses also emerged. A separate factor representing a strongly negativistic response tendency was found. In addition, the expected liberal conservative factor appeared as did a factor indicating that logical reversal of items is ineffective. (Author/CTM)