Preparation and Analyses of Implicit Attitude Measures: Challenges, Pitfalls, and Recommendations

Abstract

Implicit measures such as the Implicit Association Test (IAT) and the Personalized-IAT can be useful tools for studying automatic processes and socially sensitive topics. But their reliance on reaction time data comes with issues for data preparation and analysis. Dealing with reaction time data can be complex – exacerbated by many steps and available alternatives. Greenwald et al. (2003) offer guidelines for handling IAT data. However, these guidelines are often cited with little information as to which steps and alternatives were chosen. This provides latitudes of freedom for researchers to choose the version that is most likely to give desired results, not necessarily the one that best reflects the data or matches other work. This manuscript reports what happens when steps in data cleaning and analysis are omitted or changed, finding variations in relationships between variables and the potential for significance tests to change depending on the version used.