Researchers often consider how personal attributes affect the data gathered, and how that effect can be eliminated, or at least minimalized. This question takes on added significance in the realm of politics. Numerous studies have examined the issue of interviewer/interviewee race. The results of these studies indicate that members of one race avoid answers they think will alienate an interviewer of another race, or that would be considered socially unacceptable. According to these studies, the results of surveys are most skewed for "racial questions" (questions about civil disobedience, racial intermarriage, and tolerance of racists). It would seem that the results of surveys conducted concerning candidates of a different race would have the same effects. Yet no results of such studies have been published. Carrying the analysis one step further, it can be argued that even if the interviewer and interviewee are of the same race, but the questions involve a candidate of a different race, the poll results could be misleading. This is a factor that has not been accounted for in the survey data to date. The following research hypothesis should be explored: the race of a political candidate in a campaign involving at least one white and one black candidate will serve as an uncontrolled independent variable in surveys regarding at least one black and one white candidate. Once the validity or invalidity of such a hypothesis is determined, researchers can work to educate the media and public about the potential invalidity of the results of such polls. (MM)