Running head: POLITICAL PARTISANSHIP AND RESPONSES TO SEXUAL HARASSMENT ALLEGATIONS AGAINST POLITICIANS 1

Running head: POLITICAL PARTISANSHIP AND RESPONSES TO SEXUAL HARASSMENT ALLEGATIONS AGAINST POLITICIANS 1 Political Partisanship and Responses to Sexual Harassment Allegations against Politicians Edward J R Clarke 1, Anna Klas 2, Morgana Lizzio-Wilson 3, & Emily Kothe 2 1 Monash University 2 Deakin University 3 University of Queensland Author Note Please note that this pre-print has not undergone peer review. Correspondence concerning this article should be addressed to Edward J R Clarke, School of Psychology, Monash University. E-mail: eddie.clarke@monash.edu

ALLEGATIONS AGAINST POLITICIANS 2 Abstract This preliminary study aimed to examine whether participants political party affiliation would affect the perceived legitimacy of sexual misconduct allegations against partisan-aligned and non-aligned political candidates. Three hundred and ninety three participants were recruited for this study. After excluding participants based on study criteria (correctly responding to at least one out of three manipulation check items, and identifying as an independent), 341 participants remained. These participants (Democratic and Republican affiliates) were randomly allocated to one of three conditions (Democratic, Republican, and non-affiliated politician accused groups) where they read a fictitious news story describing a sexual misconduct allegation made by a female staffer against a male politician. They were then asked to rate the perceived legitimacy of the allegation in the story. Findings indicate that participants were no more likely to find the allegations to be legitimate depending on whether the allegation was directed at a party-congruent (participant affiliation matching politician affiliation) or party-incongruent politician (participant affiliation does not match politician affiliation). However, when the unaffiliated politician condition was removed from the analysis for exploratory purposes, Republican, but not Democratic affiliates were more likely to perceive a sexual misconduct allegation against a Democrat as more legitimate than against a Republican. Although no partisanship effect existed for Democratic affiliates, Democratic females perceived the allegation as more legitimate than Democratic males. These findings suggest that Republicans but not Democrats may engage in partisan-motivated reasoning at least in the domain of sexual misconduct allegations against Republican politicians. Keywords: partisanship, political psychology, gender, sexual misconduct, motivated reasoning

ALLEGATIONS AGAINST POLITICIANS 3 Political Partisanship and Responses to Sexual Harassment Allegations against Politicians Introduction Recently, a number of high profile politicians and political figures, particularly in the United States, have faced sexual misconduct allegations. However, many deny the legitimacy of the accusations despite the level of evidence, the number of allegations against many of the accused, and low rates of false allegations in sexual harassment (Pew Research Center, 2018).Furthermore, patterns of denial appear to suggest partisan bias. Anecdotally, this appeared to be the case in the 2017 Alabama special election, where Republican Senate candidate Roy Moore faced multiple sexual misconduct allegations. Even more recently, allegations were made against Judge Brett Kavanaugh, a Republican-backed Supreme Court nominee who was at the time undergoing the nomination review process. Support was split along party lines one week after Kavanaugh s testimony before a U.S. Senate committee where he denied allegations that he sexually assaulted Professor Christine Blasey Ford when they were both in high school. A Reuters/Ipsos poll showed that 71% of Democratic affiliates did not support his nomination, while 70% of Republican affiliates supported him (Hurley & Kahn, 2018). This is in contrast to a poll taken after misconduct allegations surfaced against Minnesotan Democratic Senator Al Frankin, which suggested that 70% of Minnesotan Republicans believed the allegations were true compared to just 45% of Democrats (Brooks & Rao, 2018). It appears as though the political affiliation of the alleged perpetrator may affect the evaluation of events among voters who are party identifiers, and may influence subsequent actions taking against the alleged. Past research suggests that voters will react to scandals against politicians in a biased manner based on their political affiliation (Bhatti, Hansen, & Olsen, 2013). Indeed, partisan-aligned individuals (those identifying with a particular political party) can engage in a type of politically-motivated reasoning (Taber & Lodge, 2006). That is, they appraise the evidence in such a way that reduces blame and even

ALLEGATIONS AGAINST POLITICIANS 4 justifies the actions of those aligned with their party and political ideology. This tendency to engage in biased reasoning also appears to be bipartisan (Ditto et al., 2018). If voters are less likely to engage in objective, rational reasoning processes when determining whether a candidate might have engaged in acts of misconduct, this then could have implications on the likelihood of victims to come forward with such allegations against politicians. Additionally, if there is a political dimension to public judgement of such allegations, this may further influence of the likelihood of victim disclosure when the perpetrator is a politician or member of the judiciary. Aims and Hypotheses This preliminary study aimed to examine whether partisanship affects perceptions of sexual misconduct allegations directed toward a political candidate. We predicted that: 1. Self-identified Republicans would be less likely than self-identified Democrats to perceive an allegation against any politician as legitimate; 2. Partisanship (identification as a Democrat or Republican) would influence reactions to sexual misconduct allegations against politicians contained in the vignette, such that Republican identifiers would be less likely to perceive the allegation as legitimate if it is directed at a Republican candidate, compared to a Democratic candidate and a non-affiliated candidate. A similar effect was predicted for Democrat identifiers in that they would be less likely to perceive allegations against a Democratic politician as legitimate when compared to a Republican or non-affiliated candidate. Methods We report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures in the study.

ALLEGATIONS AGAINST POLITICIANS 5 Participants Participants were recruited via Amazon MTurk and the TurkPrime platform. Workers who noted that they were affiliates of either the Democratic or Republican parties in their TurkPrime worker profiles were eligible for the study. Participants received $0.70USD to compensate them for their time. 393 participants completed the study. Since some people may have changed party affiliation since that information was included in their worker profile, a screening item was included in the survey to identify participants who did not identify as Democrats or Republicans; 46 were excluded on the basis of party affiliation. A further 6 were excluded on the basis of preregistered manipulation checks. Table 1 describes the demographic characteristics of the final sample. Materials Demographic and political information. We asked participants to report their age, gender, sexual orientation, highest level of education completed, who they voted for in the 2016 presidential election, political orientation, voting frequency, party affiliation and strength of party identification. Aggression-Submission-Conventionalism (ASC). This multifactorial 18-item ASC scale is a three-factor scale (Dunwoody & Funke, 2016). It designates six items to each of the three factors (Aggression, Submission and Conventionalism), measured on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Aggression is defined as support for intentional harm to a person or group if the proper authority approves of it. Submission is an acceptance of actions of the authority and a willingness to comply with their instructions. Finally, conventionalism is an acceptance and commitment to societal norms and traditions.

ALLEGATIONS AGAINST POLITICIANS 6 Social Dominance Orientation (SDO). This multifactorial 8-item short form of the SDO7 scale developed by Ho et al. (2015) measures two SDO subtypes, both measured on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree): SDO-Dominance (SDO-D) which reflects a preference for group-based dominance and oppression of low status groups by high status groups, and SDO-Anti-egalitarianism (SDO-E) which refers to a preference for hierarchy and group-based inequality. Fictitious news articles. Participants were presented with one of three fictitious news stories describing a complaint made against a political representative. The basic format of the story was based on a real news article and presented in a similar manner. We have not included the stories in this pre-print as they may be used in future studies pertaining to this broad project. In each story, the representative is male and the victim is a female junior staffer working for the representative. The story outlines an allegation that Rep. Williams touched the female staffer inappropriately. The severity of the allegation and level of detail is relatively low compared to many of the recent allegations made against politicians and political figures, however we wished to reduce the relative severity of the allegation and therefore the impact on participants. The only variation across the three news stories is the political affiliation of the accused (Democrat, Republican, No affiliation). Perceived Legitimacy. This 6-item scale measures the perceived legitimacy of the sexual misconduct allegation on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Higher scores indicate that the complaint is perceived as more legitimate. The scale had high internal consistency (α = 0.83; Ω = 0.89). All items loaded on to a single factor. Voting Intention. Which consists of likelihood of: voting for the politician, voting for the candidate from the opposing party, voting for an independent or third party candidate, not vote at all.

ALLEGATIONS AGAINST POLITICIANS 7 Recommended Punishment. Recommended punishment for the accused was measured by asking participants their level of agreement with specific types of punishment to be put in place, such as firing Representative Williams. Procedure All participants completed demographic measures and baseline measures of strength of party identification, and the ASC and SDO scales. Participants were randomized to one of three conditions. In each condition participants read a vignette which described a scenario where a political candidate (Representative Williams) has been accused of sexual misconduct against a staffer in their office. The only content that varied across the three conditions was party affiliation of the alleged perpetrator (Democrat, Republican, No party affiliation). After reading the vignette, participants completed measures of voting intention, perceived legitimacy of the allegation, and recommended punishment for the accused. Data analysis We used R (Version 3.5.1; R Core Team, 2018) and the R-packages apatables (Version 2.0.4; Stanley, 2018), bindrcpp (Version 0.2.2; Müller, 2018), dplyr (Version 0.7.6; Wickham, François, Henry, & Müller, 2018), labelled (Version 1.1.0; Larmarange, 2018), papaja (Version 0.1.0.9842; Aust & Barth, 2018), psych (Version 1.8.4; Revelle, 2018), readr (Version 1.1.1; Wickham, Hester, & Francois, 2017), and tableone (Version 0.9.3; Yoshida & Bohn., 2018) for all our analyses. Multiple regression was conducted to test the main effect of partisanship on perceived legitimacy and the hypotheses that partisanship would moderate the effect of the experimental conditions on perceived legitimacy. We controlled for participant gender in this analysis.

ALLEGATIONS AGAINST POLITICIANS 8 Results The mean response to the allegation stratified by condition is shown in Table 2. Pre-registered analyses H1: Self-identified Republicans would be less likely than self-identified Democrats to perceive an allegation against any politician as legitimate and H2. Participants would be less likely to perceive a complaint as legitimate when it was directed at a politician affiliated with the party that they identify with (relative to an opposing party or non-affiliated politician). A linear regression was conducted to test the effects of participant party affiliation, congruence between personal and politician affiliation, and gender on perceived legitimacy. As shown in Table 3, there were main effects of gender (t(336) = 2.60, p =.010) and participant party affiliation (t(336) = 2.72, p =.007). However, the perceived legitimacy of the complaint did not differ when the politician was affiliated with the same party as the participant rather than being holding an ambiguous or opposing affiliation (t(336) = 0.70, p =.486). All other things being equal, women rated the complaint as more legitimate than men and Democrats rated the complaint as more legitimate than Republicans. Exploratory analyses A series of exploratory analyses were conducted to examine the effect of partisanship under conditions where affiliation congruence was less ambiguous (namely the conditions were the party affiliation of the alleged perpetrator was stated). For these analysis, only data from participants randomized to the conditions where the politician party affiliation was stated as Republican or Democrat were retained. First, a linear regression was conducted to test the effects of participant party affiliation, the party affiliation of the alleged perpetrator, and gender on perceived

ALLEGATIONS AGAINST POLITICIANS 9 legitimacy. As shown in Table 4, there was an interaction between politician party affiliation and participant party affiliation, the influence of the party affiliation of the alleged perpetrator on individuals perception of the legitimacy of the complaint was moderated by participant party affiliation b = 0.62, 95% CI [ 1.13, 0.11], t(226) = 2.41, p =.017. Follow-up tests were conducted to examine these relationships for self-identified Democrats and Republicans separately. As shown in Table 5, among Democrats, the party affiliation of the alleged perpetrator was not a predictor of perceived legitimacy b = 0.27, 95% CI [ 0.09, 0.64], t(121) = 1.50, p =.136. Among Republicans (see Table 6), the party affiliation of the alleged perpetrator was a predictor of perceived legitimacy b = 0.36, 95% CI [ 0.71, 0.01], t(104) = 2.03, p =.045. Republican participants were less likely to perceive the complaint as legitimate when the alleged perpetrator was a Republican. Discussion Overall, Democratic affiliates are more likely than their Republican counterparts to perceive sexual misconduct allegations of any politician, regardless of party affiliation, as legitimate. When controlling for gender, the interaction of party affiliation and politician affiliation did not produce an effect on perceived legitimacy. However, As part of an exploratory analysis we investigated the same interaction excluding the non-affiliated politician condition, to probe a simple Republican versus Democratic politician effect. Here we found a significant interaction effect for participant party affiliation and party of the accused politician (Democratic v. Republican). Further post-hoc analyses indicated that Republicans, but not Democrats, perceived the allegation as more legitimate when the politician is from the opposing party, when controlling for gender. This may indicate that a partisan effect occurs in the domain of sexual misconduct allegations against politicians,

ALLEGATIONS AGAINST POLITICIANS 10 but only for Republican affiliates. Furthermore, it may suggest that in the face of sexual misconduct allegations against their own political representatives, Democrats may be less likely to hold support for them. Future Research Directions This preliminary study established that a partisan effect may exist for Republicans, but not for Democrats. However, this study does not explain why the effect occurs, and also why it may only occur for Republicans. A follow-up study aims to further investigate the role of party identification and partisanship to examine the robustness of the first finding, and also to understand why Republicans are potentially more willing to side with their accused politician and/or view an allegation against a non-aligned candidate as more legitimate.

ALLEGATIONS AGAINST POLITICIANS 11 References Aust, F., & Barth, M. (2018). papaja: Create APA manuscripts with R Markdown. Retrieved from https://github.com/crsh/papaja Bhatti, Y., Hansen, K. M., & Olsen, A. L. (2013). Political hypocrisy: The effect of political scandals on candidate evaluations. Acta Politica, 48 (4), 408 428. Brooks, J., & Rao, M. (2018). Minn. Poll: Franken accusers believed; resignation not needed. Star Tribune. Retrieved from http://www.startribune.com/ majority-of-minnesotans-believe-franken-accusations-but-fewer-think-he-should-have-quit/ 469451383/ Ditto, P. H., Liu, B. S., Clark, C. J., Wojcik, S. P., Chen, E. E., Grady, R. H.,... Zinger, J. F. (2018). At least bias is bipartisan: A meta-analytic comparison of partisan bias in liberals and conservatives. Perspectives on Psychological Science, 1745691617746796. Dunwoody, P. T., & Funke, F. (2016). The aggression-submission-conventionalism scale: Testing a new three factor measure of authoritarianism. Journal of Social and Political Psychology, 4 (2), 571 600. Ho, A. K., Sidanius, J., Kteily, N., Sheehy-Skeffington, J., Pratto, F., Henkel, K. E.,... Stewart, A. L. (2015). The nature of social dominance orientation: Theorizing and measuring preferences for intergroup inequality using the new sdo7 scale. Journal of Personality and Social Psychology, 109 (6), 1003. Hurley, L., & Kahn, C. (2018). Opposition to Kavanaugh grows after Senate hearing: Reuters/Ipsos poll. Reuters. Retrieved from https://www.reuters.com/article/us-usa-court-kavanaugh-poll/ opposition-to-kavanaugh-grows-after-senate-hearing-reuters-ipsos-poll-iduskcn1md27v Larmarange, J. (2018). Labelled: Manipulating labelled data. Retrieved from

ALLEGATIONS AGAINST POLITICIANS 12 https://cran.r-project.org/package=labelled Müller, K. (2018). Bindrcpp: An rcpp interface to active bindings. Retrieved from https://cran.r-project.org/package=bindrcpp Pew Research Center. (2018, April). Sexual Harassment at Work in the Era of #MeToo. Retrieved from http://www.pewsocialtrends.org/2018/04/04/ sexual-harassment-at-work-in-the-era-of-metoo/ R Core Team. (2018). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.r-project.org/ Revelle, W. (2018). Psych: Procedures for psychological, psychometric, and personality research. Evanston, Illinois: Northwestern University. Retrieved from https://cran.r-project.org/package=psych Stanley, D. (2018). ApaTables: Create american psychological association (apa) style tables. Retrieved from https://cran.r-project.org/package=apatables Taber, C. S., & Lodge, M. (2006). Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50 (3), 755 769. Wickham, H., François, R., Henry, L., & Müller, K. (2018). Dplyr: A grammar of data manipulation. Retrieved from https://cran.r-project.org/package=dplyr Wickham, H., Hester, J., & Francois, R. (2017). Readr: Read rectangular text data. Retrieved from https://cran.r-project.org/package=readr Yoshida, K., & Bohn., J. (2018). Tableone: Create table 1 to describe baseline characteristics. Retrieved from https://cran.r-project.org/package=tableone

ALLEGATIONS AGAINST POLITICIANS 13 Table 1 Demographic characteristics of participants by the party affiliation of the alleged perpetrator (condition) Democrat Republican Unspecified pol n 114 117 110 Age in years (mean (sd)) 33.92 (10.40) 35.51 (11.07) 35.28 (11.42) Gender = Male (%) 67 (58.8) 70 (59.8) 57 (51.8) Party affiliation = Republican Voter (%) 51 (44.7) 56 (47.9) 52 (47.3) Ideological Orientation (mean (sd)) 5.60 (3.42) 6.11 (3.52) 6.09 (3.50)

ALLEGATIONS AGAINST POLITICIANS 14 Table 2 Response to allegation by the party affiliation of the alleged perpetrator (condition) Democrat Republican Unspecified pol n 114 117 110 Perceived legitimacy of complaint 5.30 (0.94) 5.28 (1.10) 5.12 (1.14) Intention to vote for Rep. Williams 3.18 (2.06) 2.91 (2.05) 3.10 (1.94) Intention to vote for opposing candidate 3.93 (1.93) 4.69 (2.14) 4.65 (1.65) Intention to vote for independent candidate 4.30 (1.86) 4.05 (1.87) 4.34 (1.75) Intention to not vote 3.26 (2.12) 3.22 (2.35) 3.49 (2.14) Note. Vales are: Mean (SD) Table 3 The effect of party affiliation, congruence between personal and politician affilation, and gender on perceived legitimacy Predictor b 95% CI t(336) p Intercept 5.55 [5.26, 5.84] 37.73 <.001 Congruence 0.11 [ 0.20, 0.43] 0.70.486 Affilitation -0.51 [ 0.88, 0.14] -2.72.007 Gender -0.29 [ 0.52, 0.07] -2.60.010 Congruence x Affiliation 0.05 [ 0.41, 0.51] 0.23.815 Note. Reference categories: Congruence - Congruent. Affiliation - Democrat. Gender - Female.

ALLEGATIONS AGAINST POLITICIANS 15 Table 4 The effect of party affiliation, condition, and gender on perceived legitimacy Predictor b 95% CI t(226) p Intercept 5.54 [5.25, 5.83] 37.65 <.001 Condition 0.29 [ 0.05, 0.64] 1.66.098 Affiliation -0.18 [ 0.54, 0.18] -0.99.325 Gender -0.28 [ 0.53, 0.02] -2.10.037 Condition x Party affiliation -0.62 [ 1.13, 0.11] -2.41.017 Note. Reference categories: Condition - Democrat. Affiliation - Democrat. Gender - Female Table 5 The role of perpertrator party and participant gender on perceived legitimacy among self-identifed Democrats Predictor b 95% CI t(121) p Intercept 5.67 [5.33, 6.01] 33.03 <.001 Condition 0.27 [ 0.09, 0.64] 1.50.136 Gender -0.49 [ 0.86, 0.13] -2.66.009 Note. Reference categories: Condition - Democrat. Gender - Female

ALLEGATIONS AGAINST POLITICIANS 16 Table 6 The role of perpertrator party and participant gender on perceived legitimacy among self-identifed Republicans Predictor b 95% CI t(104) p Intercept 5.22 [4.89, 5.54] 32.14 <.001 Condition -0.36 [ 0.71, 0.01] -2.03.045 Gender -0.02 [ 0.38, 0.34] -0.09.926 Note. Reference categories: Condition - Democrat. Gender - Female Table 7 Correlations between gender and dependent variable Variable 1 1. gender_num 2. legit_tot -.16**