Men and Women Defending Themselves from Political Scandals: Gender Stereotypes and Proneness to Forgive Scandalous Politicians

Many studies have highlighted the role of gender stereotypes in impressions of politicians. In general, a politician’s image benefits from behaving consistently within gender stereotypes. The present study tested whether this also applied to different image restoration tactics employed by male versus female politicians after a scandal. We assessed participants’ evaluation of a fictitious male or female politician soon after a scandal and then after his or her defence. We used a 3 × 2 experimental design to compare the effectiveness of three defensive tactics involving different degrees of assertiveness vs. submissiveness as a function of a politician’s gender. All tactics were overall effective in improving the damaged reputation of the fictitious politician but, as expected, using the excuse of mitigating circumstances (submissive) was more beneficial to female politicians while diffusing responsibility by accusing another person (assertive) was more beneficial to male politicians, both in terms of global attitude and stereotypical evaluations (communality and agency, respectively). The tactic of mortification and requests for forgiveness did not have different effects by gender. Theoretical and practical implications are discussed.

Another crucial factor influencing how and to what extent a scandal will end a political career is the politician's ability to communicate a plausible and justifying interpretation of the event to the public (Bull & Fetzer 2010). Politicians can employ different strategies to rebuild their reputations after a scandal (Coombs 2006;McGraw 2001). While most empirical studies to date have focused on the comparative efficacy of these strategies, less attention has been paid to their relative efficacy in relation to the gender of a politician. This possibility merits consideration, because the effects of gender stereotypes in politics are well documented. Politics is still considered a masculine activity where stereotypically male personality traits such as assertiveness, confidence, independence, decisionmaking ability, are judged among the foremost requisites (e.g., Schneider & Bos 2014). The salience of stereotypical female features, such as sympathy, sociability, interdependence, and warmth make women appear less worthy of a vote (Cavazza & Pacilli 2021). Therefore, the present study aims to verify whether gender stereotypes influence voters' evaluation of female versus male politicians as a function of the communication strategy s/he employs to limit the reputational damage of being involved in a scandal.

Scandals and Political Communication Strategies
In addition to the role of digital media, the personalisation of politics also fosters the spread of scandals: in modern political campaigns, candidates more than parties are the focus of communication and public attention, as they represent the actual 'product' to be sold (Barisione 2009;Caprara, Schwartz, Capanna, Vecchione & Barbaranelli 2006;Kepplinger et al. 2012;Schoen & Schumann 2007). For political actors and the institutions they represent, it is therefore essential for a politician to manage their own reputation and preserve the moral credit they have obtained through positive behaviour (e.g. Harris Lock Davis & Mian 2010). The need to manage reputation is even greater when politicians face a scandal and the negative consequences it may entail. In this case, politicians can employ different image restoration strategies to limit their own responsibility for a scandal and minimise the gravity of the offence in the court of public opinion.
The communication literature offers various descriptive typologies of such strategies, each including specific and concrete tactics, used by transgressors to deal with organisational or individual reputation crises (Benoit 1997;Coombs 1995Coombs , 2006Coombs , 2007McGraw, 2001). All of these typologies stipulate that, when possible, initial crisis response strategy is the denial of involvement in the scandal (Benoit 1997;McGraw 2001). However, this is not always possible, as sometimes evidence can be uncontroversial. Thus, when perpetrators cannot deny their involvement in a scandal, they can activate the following four strategies (Benoit 1997;McGraw 2001): a) Excuses. The perpetrator can use different types of excuses to mitigate the intentionality or foreseeability of the event, such as attributing the causes of the scandal to other actors (i.e., diffusion of responsibility to other actors) or underlining the crucial role of external circumstances (i.e., claiming mitigating circumstances). In these cases, the involved actor admits that the offensive behaviour was performed but denies being the only one responsible. b) Justification. This strategy implies a reframing of the situation in which the actor admits responsibility for the scandalous behaviour but strives to reduce the perceived severity of the act (McGraw, 2001). In particular, perpetrators can use a bolstering tactic: they strengthen an audience's positive feelings toward them by encouraging the audience to remember the perpetrator's previous good actions and results in order to offset the negative feelings connected with the wrongful act. They can also employ a transcendence tactic by highlighting the benefits brought about by the wrongful act (Benoit 1997). Finally, through a minimisation tactic, they can try to convince the public that the act is less serious than it appears (Benoit 1997). c) Concession. This category encompasses all tactics that aim to restore a positive reputation through full acceptance of responsibility for the wrongful act and its consequences (McGraw 2001). In particular, this strategy category includes mortification -that is, confessing and begging forgiveness for one's transgression (Benoit 1997;Coombs 2006); implementation, or the promise to implement corrective action in the future to prevent the wrongful act's reoccurrence (Benoit 1997); a reward offer to the victims (Benoit 1997;Coombs 2006); and the expression of feelings of sorrow and compassion (Coombs 2007). d) Attack on the accusers. Perpetrators may even defend themselves by questioning the accusers' reliability and credibility, in an attempt to cast doubt and shift the audience's attention away from the target to the new victim (Benoit 1997;Coombs 1995Coombs , 2006. In general, mortification and the promise of corrective action have been found to be more effective and appropriate than other strategies (Benoit & Drew 1997), even more than denial. These strategies and their specific tactics have been primarily studied in the organisational or interpersonal (face-threatening) domains. However, they are also manifestly used in the political domain, and in fact some studies have verified their appropriateness and efficacy in restoring a damaged political reputation (Craig & Rippere 2016;Smith et al. 2005). Most of these studies have analysed real cases of scandals followed by some attempt at self-defence on the part of the protagonist. Strategy efficacy was captured through surveys measuring liking for a political actor or the outcome of their career after their self-defence (e.g., Benoit & Brinson 1999). In addition, these investigations have mostly focused on male politicians. There is a scarcity of empirical studies in the political context verifying the effectiveness of different image reparation tactics in relation to a politician's gender. However, based on studies about the influence of gender stereotypes in politics, we propose that such differences may be observable.

Gender Stereotypes and Communication in the Political Field
The influence of gender stereotypes on voters' evaluation of politicians is well documented. Voters use a politician's gender to infer information about their characteristics and their positions on political issues and to formulate expectations regarding their behaviour (Alexander & Andersen 1993;Huddy & Terkildsen 1993;Kahn 1994;King & Matland 2003;Koch 1999;Leeper 1991;Matland 1994;Rosenwasser & Seale 1988;Sapiro 1981). In general, people tend to attribute different personality traits to male and female politicians: women are seen as more compassionate and honest while men are considered stronger and more assertive (e.g., Alexander & Andresen 1993;Dolan 2010;Huddy & Capelos 2002;Sapiro 1981). This difference is rooted in more general gender stereotypes according to which women are characterised by communal characteristics (i.e., relationship skills, care and hospitality) and men are characterised by agentic characteristics (i.e., assertiveness and self-assertion; e.g., Eagly 1987;Eagly & Steffen 1984).
Behaving consistently with gender stereotypes seems to be an effective strategy to communicate positive impressions in politics (Bauer 2017;Bauer, Harbridge & Krupnikov 2017;Bauer & Carpinella 2018;Brooks 2013;Cassese & Holman 2018;Herrnson, Lay & Stokes 2003;Hitchon & Chang 1995;Hitchon, Chang & Harris 1997;Krupnikov & Bauer 2014). For example, a female candidate who focuses her election campaign on typically feminine issues (i.e., compassion issues such as abortion rights, education/school finance and health care) increases her chances of success compared to male colleagues who adopt the same strategy or female colleagues who do not address such issues (Herrnson et al. 2003).
Similarly, counter-stereotypical gender behaviours lead to negative impressions. The vast literature on the backlash effects (for a review, see Rudman, Moss-Racusin, Glick & Phelan 2012) suggests that women showing agentic traits induce a negative impression in observers more than similarly agentic man (for a review, see Phelan & Rudman 2010). In the political domain, for example, Cassese and Holman (2018) found that, when a female candidate was accused of being unwilling to work hard and uncooperative (i.e., she violated stereotypical expectations for women), she was evaluated more harshly (i.e., respondents considered her less warm, feminine, sensitive and expressed lower intention to vote for her) compared to a male politician attacked for the same reasons. A male politician can also be penalised for adopting counterstereotypical gender behaviour: when a fictitious male politician was described as a caring, compassionate and loving person, he received lower evaluations on leadership and competence dimensions compared to the control condition where he was not described in terms of gender-stereotypical traits (Bauer 2017). This is in line with the Backlash and Stereotype Maintenance Model (BSMM; Rudman & Fairchild, 2004) which considers how perceivers and actors work in concert to perpetuate stereotypes as a means of preserving the social status quo. Focusing on the perceiver side of the coin, that relevant for the present study, the BSMM proposes that observers compare targets with a normative standard provided by stereotypes. When targets fail to meet the standard, they get a backlash effect (i.e., some kind of penalty) from observers. In particular, the status incongruity hypothesis poses that perceivers sanction targets who violate those stereotypic expectancies that legitimize social hierarchies (Rudman, Moss-Racusin, Phelan & Nauts 2012). Gender stereotypes are prescriptive or proscriptive rules that support the status hierarchy.
In terms of politicians' communicative behaviour, women are penalised more heavily than men when attacking a rival-a stereotypically male behaviour (Krupnikov & Bauer 2014)-but also when responding to an attack (Craig & Rippere 2016). This seems to suggest that gendered expectations may also play a role in how politicians defend themselves from accusations, as in the case of a scandal. Indeed, through their reaction to a scandal, the perpetrators can choose a typically 'masculine' (assertive) or 'feminine' (submissive) strategy, thereby strengthening a stereotypical or counter-stereotypical gender image. We found only one study directly comparing the efficacy of different communication strategies for a woman vs. a man after a political scandal. Smith et al. (2005) manipulated three defence strategies through a fictitious newspaper article: justifications, aimed to weaken the perception of negativity of the act; excuses, aimed to reduce the perception of personal responsibility for the act; and denial. This study failed to find a difference in defence efficacy based on the interaction between the gender of the politician and the type of defence tactic used. Instead, the results showed that, irrespective of the candidate's gender, participants preferred politicians who justified or denied their personal involvement in a scandal compared to offering excuses. However, in Smith et al.'s (2005) study, the perpetrator was evaluated only after using the defence strategy. As such, the researchers could not verify the actual ability of each strategy to restore the politician's previous reputation. Furthermore, the tactics compared were very different from one another and the possibility of employing them depends strongly on the disputed facts, as it is not always possible for a politician to deny any personal involvement in a scandal or contest its negativity (Kim, Ferrin, Cooper & Dirks 2004).

The Present Study
In the present study, we aimed to compare the efficacy of different communicative tactics as a function of the gender of the perpetrator in restoring (i.e., significantly improving post-scandal) the reputation of a politician involved in a scandal which could not be denied. Since behaviours which do not conform to gender-stereotypical expectations tend to be sanctioned and those which fall in line with gender stereotypes tend to induce positive evaluations (Bauer 2017;Bauer & Carpinella 2018;Bauer et al. 2017;Brooks 2013;Herrnson et al. 2003;Hitchon & Chang 1995;Krupnikov & Bauer 2014), different defensive strategies should induce different effects in relation to their degree of assertiveness or submissiveness and the politician's gender. From the typologies of defence strategies outlined above, we selected three tactics based on their level of assertiveness or submissiveness (i.e., their stereotypical feminine vs. masculine connotations): diffusion of responsibility by accusing another person; claiming mitigating circumstances; and mortification and requests for forgiveness. Though these strategies do not deny the misdeed itself, they involve a decreasing degree of assertiveness and aggressiveness and as such can be considered respectively more or less in line with what is expected from a man vs. a woman (Chanley, Sullivan, Gonzales & Kovera 1994;Gonzales, Kovera, Sullivan & Chanley 1995;McGraw 2001;Smith et al. 2005). On this basis, we formulated the following hypotheses: H1: A male politician should be successful in improving his post-scandal evaluation by using an assertive-aggressive defensive tactic, such as trying to accuse another person, whereas a woman should benefit mainly from defending herself by claiming mitigating circumstances and through mortification. In operative terms, we expected a two-way interaction between the type of defensive tactic and a politician's gender. H2: In particular, we expected that excuses which involved accusing another person would improve post-scandal evaluation in terms of the agentic traits of a male politician (H2a), whereas excuses involving claims of mitigating circumstances (H2b) and mortification with request for forgiveness (H2c) would improve post-scandal evaluation in terms of the communal traits of a female politician. In operative terms, we expected a three-way interaction among type of post-scandal defensive tactic, politician gender and trait dimension.

Method Participants
We issued a link to our online questionnaire via a mailing list of students at the University of Parma, asking them to also send the survey to their family members, and enrolled all individuals who volunteered to participate over the course of five weeks. A total of 192 participants (49.5% women) ranging from 22-68 years old (mean = 33.18 years; SD = 10.61) took part in the study. They were predominantly university students (40.6%) or employees (40.6%) from the north of Italy (64.1%). Their average political self-placement on the 10-point left-right continuum was 4.58 (SD = 2.38). Informed consent was obtained from participants at the start of the questionnaire.

Procedure
First, to control for stereotypical expectations about women's and men's honesty, we asked participants to rate comparative honesty by gender on a scale from 1 (women are generally more honest than men) to 5 (men are generally more honest than women). For the experimental paradigm, we adapted previously used procedures and material ( Cucchi & Cavazza 2020) to design a two-step fictional scenario concerning a corruption scandal. In the first step, participants were invited to read a false newspaper article about a politician of their own preferred party. 1 The news article was specifically created for the purposes of this study and reported that an investigation by the public prosecutor discovered a €10,000 bribe given to a (male vs. female) councillor to support the building of a waste treatment plant. The news article was accompanied by the politician's alleged image (the same as previously used by Cucchi & Cavazza 2020). After participants read the article, we asked them to complete a first questionnaire, including their judgment of the politician's physical attractiveness (1 item), the seriousness of his or her behaviour (1 item), an evaluation of the perpetrator of the scandal (7 items; see below) and three other items not used for the present study because of their excessive skewness (intention to vote for the candidate, evaluation of his or her career continuation and support for his or her resignation).
In the second step, a second false news article described the perpetrator's reaction to the scandal. We manipulated three tactics to restore the politician's reputation. In the first condition-' diffusion of responsibility by accusing another person'-the perpetrator admitted that he or she had taken the bribe but without knowing it, accusing a secretary of accepting the money and lodging it in the politician's account (with the politician promising to take legal action against the secretary). In the second condition-' excuses with claims of mitigating circumstances'the perpetrator admitted to taking the money to alleviate a personal dramatic economic situation induced by his or her spouse's dismissal from a bankrupt company. In the third condition-'mortification and request for forgiveness'-the perpetrator admitted that he or she had taken the money but stated his/her extreme remorse for the enormous error.
After this second article, participants completed a second brief questionnaire which included the same items as the first for the evaluation of the perpetrator. A final section concerned participants' level of political interest, political self-placement on the 10-point left-right continuum, sex, age, area of residence and education.
Participants were randomly assigned to one of six experimental conditions derived from the 2 (perpetrator's gender) × 3 (restoration tactic) between participants factorial design. The experimental material and dataset are available at https://osf.io/3pe48/?view_only=6b803cf93 b534db88fd6cd3287ef32f9.
All procedures performed in this study were in accordance with the ethical standards of the national research committee and with the 1964 Helsinki declaration and its later amendments.

Evaluation of the political actor
Participants reported their global attitude towards the fictitious politician on a 10-point scale (1 = completely negative to 10 = completely positive) twice, once at T1 (post-scandal) and once at T2 (post-defensive reaction).
The efficacy of the defensive tactic in restoring reputation was captured through the change in evaluation of the protagonist.
Evaluation of the politician's communality and agency before and after his or her justification of the scandal The respondents evaluated the scandal perpetrator on a 5-point scale ranging from 1 (not at all) to 5 (very much) twice, once after the scandal (T1) and once after the perpetrator's defensive reaction (T2). Participants rated the perpetrator on three adjectives each for the communality dimension (honest, sensible and empathetic) and agency dimension (determined, competent and strong; e.g., Bauer 2017). The indices built on the mean items achieved satisfactory reliability: α communality post-scandal = 0.80; α communality post-justification = 0.74; α agency postscandal = 0.63; α agency post-justification = 0.67.
Because the answers were given on different scales, we normalised all scores to range from 0 to 1.

Results
Preliminary regression analyses including the politician's attractiveness, respondents' gender, respondents' political orientation, respondents' degrees of interest in politics and respondents' expectations about the comparative honesty of women and men did not yield any main nor interaction effect on the dependent variables. As such, they will not be considered in subsequent analyses. Table 1 reports descriptive statistics and correlations.

Effects of Defensive Tactics
A preliminary 3 (type of defensive tactic) × 2 (global attitude toward the politician before and after defensive reaction) mixed-design analysis of variance (ANOVA) with the latter as the repeated factor showed that all the tactics were globally effective in improving the T2 evaluation (M = 0.28; SD = 0.17) in comparison to T1 (M = 0.23; SD = 0.14), F (1, 189) = 12.39; p = 0.001; η 2 p = .06. We tested H1 using a 3 (type of defensive tactic) × 2 (perpetrator's gender) ANOVA on the global attitude toward the politician after his/her defensive reaction including the previous attitude (T1) as a covariate. The hypothesised two-way interaction between the independent variables proved to be the only significant effect, F (2, 185) = 6.06; p = 0.003; η 2 p = 0.06 (Figure 1). Further inspection of this interaction revealed that the ' diffusion of responsibility with accusation of another person' was effective only when used by the man, simple slope = 0.11, SE = 0.04, 95% CI [0.03, 0.19], whereas the ' excuses with claims of mitigating circumstances' was effective only when used by the woman, simple slope = 0.09, SE = 0.04, 95% CI [0.00, 0.17]. Finally, contrary to our expectations, 'mortification with request for forgiveness' did not vary its efficacy as a function of the politician's gender, simple slope = 0.00, SE = 0.04, 95% CI [-0.08, 0.08].
Regarding H2, the ANOVA including the agentic and communal dimensions measured at T2 as a repeated factor (and those measured at T1 as covariates) revealed the main effect of the repeated factor, F (1, 184) = 9.13, p = 0.003, η 2 p = 0.05, signalling that the defensive reactions were globally more effective in improving the perpetrator agentic traits (M = 0.30, SD = 0.17) than the communal ones (M = 0.27, SD = 0.18). This analysis also elicited the twoway interaction between the type of defensive tactic and the repeated factor, F (1, 184) = 9.13, p < 0.001, η 2 p = 0.14, indicating that the ' diffusion of responsibility' improved the evaluation of the perpetrator on agentic traits more than the other defensive tactics, that did not differ from each other at the LSD post-hoc test, whereas none of the tactics differed from each other as for the efficacy in improving the perceived perpetrator communality. In addition, the two-way interaction between the gender of the politician and the repeated factor, Finally, and more importantly, the hypothesised threeway interaction among type of defensive tactic, politician  gender and trait dimension, F (2, 184) = 5.56; p = 0.005; η 2 p = 0.06 emerged. In line with H2a, the ' diffusion of responsibility by accusing another person' improved the agentic traits of the male politician more than those of the female one, simple slope = 0.19, SE = 0.03, 95% CI [0.12, 0.26], whereas neither ' excuses with claims of mitigating circumstances', simple slope = -0.01, SE = 0.03, 95% CI [-0.08, 0.06], nor 'mortification and request for forgiveness', simple slope = 0.05, SE = 0.04, 95% CI [-0.01, 0.13] varied their efficacy in terms of agentic perception as a function of politician's gender (see Figure 2). In line with H2b, ' excuses with claims of mitigating circumstances' was effective in improving the communal traits of the female politician more than those of the male one, simple slope = 0.13, SE = 0.03, 95% CI [0.06, 0.20], whereas ' diffusion of responsibility by accusing another person', simple slope = 0.00, SE = 0.03, 95% CI [-0.07, 0.07], and unexpectedly even 'mortification and request for forgiveness', simple slope = 0.06, SE = 0.03, 95% CI [-0.01, 0.13] did not vary as a function of politician's gender (see Figure 3).

Discussion
People tend to judge the same behaviour differently when it is performed by a male vs. a female politician as a function of the congruency between the behaviour and gender-stereotypical expectations. In general, a politician is evaluated more positively when his or her behaviour is in line with gender stereotypes than when he or she performs a counter-stereotypical gender behaviour (Bauer 2017; Bauer & Carpinella 2018;Bauer et al. 2017;Brooks 2013;Dolan 2010, Herrnson et al. 2003Hitchon & Chang 1995;Krupnikov & Bauer 2014). The present study aimed to expand the knowledge in this field by verifying whether gender stereotypes impacted the effectiveness of certain image restoration tactics employed by a male vs. female politician after a scandal. To achieve this goal, we selected three defensive tactics involving different degree of assertiveness (a stereotypically masculine trait) and submissiveness (a stereotypically feminine trait). Globally, all tactics were effective in improving the damaged reputation of the fictitious politician. However, their efficacy varied as a function of the politician's gender in the expected direction, confirming that gender stereotypes play a role in this domain. On the one hand, the fictitious female politician experienced greater benefits than her male counterpart when using a submissive defensive tactic, such as appealing to mitigating circumstances justifying her misconduct. Indeed, the claims of mitigating circumstances improved participants' global attitude toward and evaluation of the communal traits of the female politician. On the other hand, the male politician had an advantage over the female politician when he defended himself using an assertive tactic, such as diffusing responsibility by accusing another person. This tactic improved participants' global attitudes towards the male politician and their evaluation of his agentic traits. Therefore, when a politician faces the consequences of a scandal, the best defensive tactic seems to be to behave in accordance with voters' gender expectations. Following the BSMM (Rudman & Fairchild 2004), this could imply that political actors and their communication consultants, made aware of the potential consequences, may tend to use stereotypical defensive tactics to avoid backlash, but thus ends up reinforcing the status quo.
Unexpectedly, mortification-the most submissive of the three tactics-was ineffective in improving respondents' attitudes toward the politician after a scandal, regardless of gender. This result contrasts that of Benoit and Drew (1997), who found that mortification was the most effective tactic to repair a face-threatening interpersonal situation. However, it is consistent with other studies comparing the effectiveness of mortification in relation to the type of the wrongful act committed (morality violation vs. competence violation; Bertolotti et al. 2013;Brambilla et al. 2011;Cislak & Wojciszke 2008). Since negative moral aspects are particularly salient and diagnostic for the audience (negativity effect) and a scandalous event directly calls into question the morality of the politician involved, it is plausible that the mere admission of guilt and request for forgiveness are insufficient to restore a politician's reputation. To regain an audience's trust, the mortification strategy should ensure that the benefits, in terms of the redemption of the protagonist, outweigh the costs associated with the admission of responsibility for the transgression committed (Kim et al. 2004). In our study, reading a newspaper article about the involvement of a politician in a scandal could have led to an association between the politician and the morally wrongful act. On the basis of the associative-propositional evaluation model (Gawronski & Bodenhausen 2011), we speculate that the admission of guilt-even if followed by a request for forgiveness-strengthens the association between the politician and the scandal in the audience's memory, making their evaluation of that politician even more negative. When the politician emphasises the role of mitigating circumstances, he or she modifies this link, transforming it into a scandal-external circumstances association. Emphasising the responsibility of other actors likewise alters this connection by transforming it into a scandalother person association. Future investigations should test this interpretive hypothesis.
The study has some limitations. First, we did not include in the questionnaire a manipulation check assuring us that the three restoration strategies actually differed mainly on the agency dimension. To overcome this limitation, we performed a post-test asking 60 university students (47 women, age mean = 24.78, SD = 3.11) to carefully read one out of the three defensive passage used in the main experiment and rate it on eight items, four tapping the agency dimension (i.e. assertive, determined, submissive and yielding) and four tapping the message efficacy (i.e. efficacious, opportune, compelling, irritating). A multivariate analysis of variance showed significant statistical differences on determined, F (2, 55) = 12.17, p < 0.001, η 2 p = 0.31; submissive, F (2, 55) = 8.88, p < 0.001, η 2 p = 0.24; and yielding, F (2, 55) = 24.09, p < 0.001, η 2 p = 0.47, whereas no effect was detected on the other items. The post-hoc LSD test confirmed that the ' diffusion of responsibility by accusing another person' was indeed perceived as the most agentic defensive strategy, since the message tone was evaluated as more determined and less submissive and yielding (all ps < 0.001) in respect to the other two (not differing from each other).
A second limitation concerns the fact that we built a very simple experimental paradigm in which the politician's affiliation was not manipulated and s/he employed only one defensive tactic. In real-life situations, the political affiliation of the perpetrator may be more salient and determinant, and citizens may consider it a base for their judgment (e.g. Russo 2017). Indeed, although the role of political leaders has increased in the last decades, the role of party identification seems still central in explaining Italian voters' choice in parliamentary elections (e.g., Garzia & Viotti 2012). Therefore, knowing the political affiliation of the candidate might have activated an inparty/outparty double standard of evaluation (Abrams et al. 2013). Future research should address the effects of this double standard evaluations in interaction with the gender of the scandalous politician.
In addition, a politician involved in a scandal, and his/ her consultant team, likely will not employ only one image restoration defence strategy, but instead will implement multiple tactics simultaneously through different communicative channels or change them depending on the way that the scandal and public opinion evolve. With regard to real-life situations, though we did not find any proper statistic about the relative percentage of women and men involved in political scandals, it would be interesting to relate these data to women's actual presence in politics (OECD 2021), also comparing across countries and parties, and assessing whether they affect the electorate's perceptions of male and female politicians and their justification strategies in real contexts. Although this is beyond the scope of the present work and thus is not a limitation, it could be the object of further studies.
Notwithstanding these limitations, our study contributes to the literature by expanding the knowledge of the effectiveness of different communication strategies in restoring politicians' reputation after a scandal, directly considering the evaluation change before and after the message incorporating the defensive tactic. Our study also confirms and extends the results of previous research on the role of gender stereotypes in politics (Bauer 2018;Dolan 2010;Sanbonmatsu 2003) by showing that men and women have different likelihoods of repairing the reputational damage of a scandal using the same tactics. These chances depend on gender-stereotypical expectations: a politician is more likely to regain his or her credibility if he or she is able to take advantage of gender-stereotypical expectations and adapt his or her communication tactics to be congruent with those expectations.

Note
1 The instruction read as follows: "Imagine to read this piece of news about a politician of your preferred party."