Improving the Measurement of Hostile Sexism

Abstract

AbstractIn recent years, sexism has played an increasingly pivotal role in American politics, and scholarship examining the importance of gender attitudes for political behavior has surged. Researchers have largely relied on the hostile sexism scale to measure prejudice against women, and this scale seems particularly relevant to political science research. However, this scale measures attitudes with an agree-disagree response format, which has long been recognized as a source of substantial measurement error. In this paper, we introduce a revised version of the hostile sexism scale that instead relies on an item-specific question format. Across three studies, we show that the item-specific scale is strongly related to the agree-disagree scale, but that the item-specific version reduces problems with truncation and tends to improve discriminant and predictive validity. Given these advantages, we conclude by recommending that researchers adopt the item-specific hostile sexism scale.