The Greater Effects of Sexual Harassment: A Conjoint Analysis Assessing Variation in Political Scandal Effects on Candidate Evaluations in the UK

The rich literature on valence finds that involvement in scandals results in potential sanctions by voters on the basis of low assessments of valence attributes. However, the literature largely does not differentiate between whether some scandals are more politically costly to elites than others, particularly with respect to the (recently) heightened salience of sexual harassment scandals.

To assess the presented theoretical framework, we recruit 241 respondents and employ a conjoint experimental design to estimate and compare the causal effect of various types of political scandals on evaluating electoral candidates. Because each of our respondents evaluates a total of 6 pairs of candidates, the total sample size for analysis is 2892 under our design.

Candidates who are involved in any type of scandal are statistically less preferred by the experimental respondents. Between the selected scandal types, a candidate’s preference drops the most while he is involved in sexual harassment. Finally, except when the voters’ and candidate’s gender are in congruence, the negative effect of sexual harassment is conditional.

We find strong evidence that involvement in sexual harassment scandals lowers candidate support to a greater degree than other types of scandals. We also find that male voters tend to ease the negative effect of sexual harassment when the violators are male candidates.

With sexual harassment’s large and various (i.e., on victims, citizens, and parties) impacts, our findings emphasize the need for better finding responsible parties that may pressure candidates involved in sexual harassment scandals to either resign if they are currently serving in office or withdrawal from electoral competition.