The Value of Outside Support for Male and Female Politicians Involved in a Political Sex Scandal

ABSTRACT

This research examined how third party statements impact the evaluation of male and female politicians caught in a scandal (i.e., extramarital affair). Governor's sex was crossed with three types of support statements: third party supportive (TPS), third party non-supportive (TPNS), and governor self-supportive (GSS). In Experiment 1, a female politician was evaluated more positively than a male politician. The TPS and the GSS conditions were both evaluated more positively than the TPNS condition. Experiment 2's design was similar to Experiment 1's, except it involved multiple affairs. In Experiment 2, participants used the third party's statements as an information source and thus reduced their use of gender stereotypes in the TPS and TPNS conditions compared to the GSS condition. We also found that male respondents gave more negative evaluations of the female governor than female respondents. Implications for the gender stereotype and social influence literatures are discussed.

Keywords:

• accounts
• gender stereotypes
• political scandal

Acknowledgments

Part of this article was presented at the 22^nd Annual Convention of the Association for Psychological Science, in Boston, MA; and at the 84^th Midwestern Psychological Association Meeting in Chicago, IL. We thank the anonymous reviewers for their comments on an earlier version of this article. The authors gratefully acknowledge funding from the Psychology Endowment Fund at the University of Minnesota, Morris.

Notes

1. In terms of the temporal sequence of the experiments, Experiment 2 was run first and then Experiment 1. However, in hindsight, we felt that presenting them in reverse order was the more logical presentation.

2. The manipulation check was only in the supportive and non-supportive conditions. The participants were asked to circle a statement asking them whether the third party had either supported or not supported the governor's decision to stay in office.

3. The study included two types of third party support. In one condition, the statements came from the spouse, while in another they came from a former male Republican governor. Our interest was in seeing if the results applied across third party sources. However, a preliminary 2 (politician sex) × 2 (support: present vs. absent) × 2 (third party source: spouse vs. former governor) between groups factorial revealed neither a main effect nor any interactions with third party source so we collapsed across this variable.

4. For the manipulation check, all participants wrote down the number of affairs mentioned in the article. Participants in the supportive and non-supportive conditions also circled which of two statements (i.e., a supportive statement vs. a non-supportive statement) best represented the spouse's position about the governor staying to finish his/her term. Participants were dropped for failing either manipulation check. Also for the manipulation check, the first 10 participants had the female governor referred to with a male pronoun. This should not be a problem because the manipulation check occurs only after the main dependent variables.

5. We also ran a preliminary analysis with these three variables and order ( filibuster-extramarital vs. extramarital-filibuster), but we found no effect and removed it from later analysis.

6. Although the instrumental measures were primarily designed to look at H4 and H5, we also examined H3a in which we found parallel results to our earlier global evaluation measure. Namely, respondents above the masculine median gave evaluation to female politicians (M = 3.00, SD = 1.39) that were lower than the average of the other three conditions (M = 3.69, SD = 1.12) for the instrumental measure, t(113) = 2.63 p = .01.

7. Finally, one concern about study 2 was that there was a possible confound in the support versus the non-support condition. In the support condition, the spouse both supports the governor and also mentioned he/she is doing a good job—i.e., expressing forgiveness and also making a statement about his/her competency. In the non-support condition, the spouse does not mention his/her competency, but states their non-support. We think there are two reasons why this confound is not too great an issue. First, Experiment 2 did not show many effects involving a significant difference between support versus non-support; so the confound seems like less of a concern. Second, when we did find an effect for support level in the second study (i.e., the instrumental measure), it generally mirrored the finding from the first study, which did not have the problem with the confound.

8. We would like to thank our anonymous reviewers for several specific suggestions that shaped our article and particularly our discussion section.