Controlling for Acquiescence Response Set in scale development.

Abstract

Acquiescence Response Set (ARS), the tendency to agree with questionnaire statements regardless of content, is a source of bias in attitude measurement. ARS is especially problematic in scale development when attitude structure is not well known, because it heightens the correlations among items that are worded similarly, even when they are not conceptually related. This paper describes a partial correlation technique for measuring and controlling for ARS using the method of matched pairs. Respondents (N = 1,351) earned an ARS score from the frequency with which they agreed with pairs of items that are logically opposite. Principal-components analysis was then performed on the firstorder interitem partial correlation matrix, controlling for ARS score. Evidence is presented that this procedure reduces the average interitem correlation among like-worded items, increases the average interitem correlation among un-likeworded items measuring the same concept, and produces a principal components solution that is more interpretable. These conclusions emerge from comparisons with analyses of untransformed attitude scores and attitude scores excluding respondents who demonstrated the greatest acquiescence.Acquiescence Response Set (ARS), the tendency to agree with attitude statements regardless of content, is a potential threat to validity in attitude measurement (Bradburn & Sudman, 1979;Cronbach, 1960). ARS threatens the validity of attitude ratings because it is a source of "correlated errors" that can bias scale scores and produce misleading conclusions about group differences in subsequent analyses (Go've & Geerken, 1977).ARS also poses problems during scale development, when attitude structure is less well known. ARS disguises true relationships between attitude items by falsely heightening the correlations among items that are worded in the same direction at the expense of items

Extent and impact of response biases in cross-national survey research

“…However, the mean itself may not capture all of yea-saying and may treat yea-saying and nay-saying as polar opposites. A second approach is to balance positively and negatively valenced items in the overall scale (Baumgartner & Steenkamp, 2001Johnson et al, 2005;Leavitt & Walton, 1975;Martin, 1964;Winkler, Kanouse, & Ware, 1982). Here, the agreement with contradictory statements is presumed to reflect acquiescence (Winkler et al, 1982).…”

Acquiescence in personality questionnaires: Relevance, domain specificity, and stability

“…During the last decades, acquiescence has been investigated from different perspectives. Several studies investigated how acquiescence can be measured (e.g., Billiet & McClendon, 2000;Paulhus, 1991;Winkler, Kanouse, & Ware, 1982), which situational or item characteristics affect acquiescence (e.g., Elliott, 1961;Krosnick & Presser, 2010;McBride & Moran, 1967;Trott & Jackson, 1967), and how acquiescence is related to personality or demographic variables (e.g., DiStefano, Morgan, & Motl, 2012;Knowles & Nathan, 1997;Weems, Onwuegbuzie, Schreiber, & Eggers, 2003). However, few studies have examined the structure of the construct acquiescence itself.…”

The Relation Between Culture and Response Styles

“…A second pair of questions included, "I am optimistic about the future of my company," and "I am frequently worried about the future of my company." Using the coding scheme developed by Winkler, Kanouse, and Ware (1982), values of 1 were assigned to each question pair for which respondents agreed or tended to agree with both items; otherwise, a value of 0 was assigned. These items were then summed to create an index that ranged from 0 to 9, with higher values representing greater levels of acquiescence.…”