|
Abstract:
|
Social scientists often rely on survey data to examine group differences in various outcomes. An inherent problem in survey data, however, is the potential misidentification of group membership due to poorly trained interviewers, inconsistent responses, or simple errors in marking questions. In data that contain drastically unequal sub-sample sizes, the detrimental consequences of misidentification can be considerable, especially concerning groups with a small sample size. In this study, we develop a new mixture model that allows researchers to address the problem using the data that they have in hand. By supplying additional information from the data, this two-stage model is estimated using a Bayesian method. As anticipated, the more information we supply to adjust for group membership, the better the model performs. We conduct sensitivity analysis on choices of priors.
|
