That an experimenter can very easily influence his subjects to give him the response he wants is a problem that every investigator recognizes and takes precautions to avoid. But how does one cope with the problem of unconscious influence? It is possible that a good many contradictory or unexpected findings are due to the fact that the experimenter unknowingly communicated his desires or expectations to his subjects. Though this problem has been generally recognized and much discussed, there has heretofore been no systematic test of the hypothesis that an experimenter can obtain from his subjects the data he expects or wants to obtain. This paper reports just such a test, and discusses what can be done to avoid experimenter influence, both conscious and unconscious.