Difficulty and Other Correlates of Criticalness Response Style At the Item Level

Abstract

LARGELY as a result of Cronbach's (1946; reviews of the early literature on response sets-or response styles-on tests, the view is widely held that tests and test items that are difficult or ambiguous are most affected by those response styles, such as acquiescence, evasiveness, and extremeness, that are linked to the response format. Despite the prevalence of this view, the results of relevant studies are not altogether convincing, and are largely limited to two response styles-acquiescence, and the position of the chosen alternative on multiple-choice tests.One major group of relevant studies examined acquiescence on tests or items of varying difficulty or ambiguity. Only one of the studies (Gage, Leavitt, and Stone, 1957) concerns the difficulty issue. In a comparison of difficult and easy general information items, the reliability of the number of true responses was .68 for the 50 difficult items, and .09 for the 40 easy ones. The two kinds of items, however, did not represent exactly the same content areas.The other studies of this kind concern various facets of item ambiguity. Two of these studies directly concern ambiguity. In

Secular Trends in an Adjective Checklist1,2

“…In a recent study (Stricker, 1964) that reviews most of the literature on secular trends (cf. Loevinger, 1957, p.…”