Data from surveys often include errors, and such errors can have a serious effect on inferences about behavior or perceptions. In this paper a model is developed for making inferences based on dichotomous survey data with possible errors. A likelihood analysis reveals an identification problem,...