An experiment tested the hypotheses that the more important a prospective choice, (a) the greater the difference in private evaluations of the alternatives and (b) the less the difference in public evaluations of the alternatives. Female subjects expected to choose between two interviewers for a lengthy, personal interview (important choice) or a brief; superficial interview (unimportant choice). They rated the interviewers with the experimenter absent while "practicing" with a computer mouse apparently not registering responses (private evaluations), or did that and told the experimenter their ratings (public evaluations). The difference in private evaluations was significantly greater when the choice was important, supporting the first hypothesis. Results for public evaluations were consistent with the second hypothesis but non-significant. In a second experiment, subjects rated the interviewers on a questionnaire (public evaluations). The difference in public evaluations was significantly less when the choice was important, supporting the second hypothesis.