A FAIRLY COMMON AMBIGUITY - COMPARING RATING AND APPROVAL MEASURES IN PUBLIC-OPINION POLLING

Although public assessments of the performance of elected officials are often measured by job rating questions using easily understood language, answers to such questions can be difficult to interpret because of their use of generic words containing multiple meanings. The word 'fair,'' which is often presented as a response intended to signify mild disapproval, is particularly problematic in this regard. Because of this problem, we have used a statewide survey to compare results generated by job rating (excellent, good, fair, poor) and job approval (strongly approve, somewhat approve, somewhat disapprove, strongly disapprove) questions for elected officials. Magnitude estimation techniques are employed to enhance our understanding of the general value attached to responses and the stability of their meaning for respondents. Finally, a structural model of assessments developed. We conclude that, while job approval questions have some advantages over job rating questions, a number of potential problems must be considered before either strategy is chosen.