This article examines the proposition that conservatives will be less willing than liberals to apologize and less likely to forgive after receiving an apology. In Study 1, we found evidence for both relationships in a nine-nation survey. In Study 2, participants wrote an open-ended response to a victim of a hypothetical transgression they had committed. More conservative participants were less likely to include apologetic elements in their response. We also tested two underlying mechanisms for the associations: social dominance orientation (SDO) and entity beliefs about human nature. SDO emerged as a stronger and more consistent mediator than entity beliefs. Apologies are theorized to be a rhetorical vehicle for removing power inequities in relationships posttransgression. Consistent with this theorizing, it was those who are relatively high in commitment to equality (i.e., those high in liberal ideology and low in SDO) who are most likely to provide and reward apologies.

During the 2016 race for Presidency of the United States, the Democrat nominee (Hillary Clinton) apologized 5 times in 8 months: for using her private e-mail server while secretary of state (September 2015), for praising Nancy Reagan’s record on promoting HIV awareness (March 2016), for supporting her husband’s 1994 crime bill (April 2016), for using the phrase “off the reservation” (April 2016), and for saying that she’d put the coal industry “out of business” (May 2016). This contrasts with the Republican nominee who, when pressed on when he had last apologized, said “It was too many years ago to remember. I have one of the great memories of all time, but it was too long ago” (Donald Trump; August 2015).

Other Republicans elevated the reluctance to be apologetic to the status of a moral virtue or rallying cry. Henry Kissinger—Secretary of State for successive Republican administrations—once said, “You are you and that is the beginning and the end—no apologies, no regrets.” In a similar vein, Ronald Reagan wrote, “I hope that when you’re my age, you’ll be able to say as I have been able to say: We lived in freedom, we lived lives that were a statement, not an apology.” The book Mitt Romney released in the lead-up to the 2012 election was titled No Apology.

We presented these examples merely to lend a human face to a broader research question: Do different political ideologies predict willingness to deliver apologies (and willingness to accept apologies when they are received)? The question is consequential because the presence of an apology is profoundly important for the forgiveness process. In a meta-analysis of 175 studies on interpersonal forgiveness (Fehr, Gelfand, & Nag, 2010), people were much more likely to forgive when the transgressor had apologized. Moreover, the effect size (r = .40) was comparable with other variables that are highly proximal to forgiveness, such as harm severity (r = −.26), trait forgiveness (r = .34), and whether the transgression was intentional (r = .50). In short, apologies matter: They help heal relationships that have been threatened by a breach of trust. It is perhaps not surprising, then, that proclivity to apologize—an individual difference variable—is positively associated with a range of well-being indices (Howell, Dopko, Turowski, & Buro, 2011). Neither is it surprising that victims have strong emotional needs to receive apologies (Hornsey, Okimoto, & Wenzel, in press; Leunissen, de Cremer, Folmer, & van Dijke, 2013).

Interestingly, research on proclivity to apologize has lagged well behind research on the effects of apologies. Existing research has focused on gender (Schumann & Ross, 2010) and the personality signatures of those with a proclivity to apologize (Dunlop, Lee, Ashton, Butcher, & Dykstra, 2015; Howell et al., 2011; Howell, Turowski, & Buro, 2012). However, there is no existing theorizing that has drawn a psychologically meaningful line between people’s political orientations and their willingness to apologize in interpersonal contexts.

In this article, we make a case that conservatives will be more reluctant to give apologies than liberals and less likely to reward apologies offered by others with forgiveness. We examine two variables that may help understand the links between political conservatism and willingness to give and receive apologies: implicit theories about the malleability of human nature (entity beliefs; e.g., Dweck, Chiu, & Hong, 1995) and ideological beliefs about power differentials (social dominance orientation [SDO]; e.g., Pratto, Sidanius, Stallworth, & Malle, 1994). We review the case for these pathways below.

Study 1 Section: Choose Top of page Abstract Study 1 << Method Results and Discussion Study 2 Method Results and Discussion General Discussion References CITING ARTICLES We conducted a survey study across nine nations to provide the first test of whether people’s political ideologies are associated with their attitudes toward the giving (and receiving) of apologies. We predicted that participants identifying themselves as more conservative would be less willing to apologize. We also predicted that the link between political ideology and proclivity to apologize would be mediated through SDO and/or entity beliefs about human nature. In addition to measuring proclivity to apologize as a criterion variable, we measured the extent to which participants reported a greater proclivity to forgive upon receiving an apology. In line with the meta-analysis reported earlier (Fehr et al., 2010), we expected that forgiveness would be higher in the event of an apology. However, we predicted that this “apology bonus” would be greater among liberals than conservatives, an association that would be mediated through SDO and/or entity beliefs. The decision to sample participants from nine nations reflects a determination to draw conclusions that translate beyond the Western, educated, industrialized, rich, and democratic (WEIRD) samples that are already overrepresented in the literature (Henrich, Heine, & Norenzayan, 2010). Our goal was not to conduct a formal cross-cultural analysis, and we make no a priori predictions about whether results might be moderated by culture. However, for the sake of thoroughness, we also examined how the key variables vary as a function of the national sample.

Method Section: Choose Top of page Abstract Study 1 Method << Results and Discussion Study 2 Method Results and Discussion General Discussion References CITING ARTICLES Participants and Sampling Participants were community members recruited through online data collection companies (N = 2,130, ns, for each nation ranged from 211 to 287). Sample size was determined so that we would have sufficient power to test our model not just at the overall level but also within each nation. Participants (51.2% female, M age = 39.67 years, standard deviation [SD] = 13.75) responded to an invitation to take part in a survey titled “Personal ideals and views.” We sampled from developed and developing countries spanning five continents: Australia, Chile, China, Hong Kong, India, Japan, Peru, Russia, and the United States. Questionnaires were translated into the native language of non-English-speaking samples using translation/back-translation procedures. Measures Entity beliefs were measured using a 3-item scale devised by Chiu, Hong, and Dweck (1997). An example item is “The kind of person someone is, is something very basic about them and it can’t be changed very much” (1 = very strongly disagree, 6 = very strongly agree; α = .77). SDO was measured using a 4-item scale validated across an international sample (Pratto et al., 2013). Participants rated their support for four statements, including “Group equality should be our ideal” (reversed) and “Superior groups should dominate inferior groups” (1 = extremely oppose, 10 = extremely favor; α = .53).2 Reluctance to apologize was measured using 6 items from the proclivity to apologize measure (Howell et al., 2011). Example items include “If I think no-one will know what I have done, I am likely not to apologize” and “I don’t apologize very often because I don’t like to admit that I’m wrong” (1 = strongly disagree, 7 = strongly agree; α = .87). To measure the apology–forgiveness bonus, participants were presented with three scenarios in which they had been transgressed against. In each case, participants were asked “Do you think he or she should be forgiven?” which they answered using a sliding scale from 0 (never, regardless of the circumstances) to 100 (all the time, regardless of the circumstances). Within each scenario, the question was posed twice: First in the event that the transgressor does not apologize and then again on the understanding that the person “apologizes and seems genuinely sorry.” The scenarios included a workplace transgression (“A work colleague loses their temper at you in front of other work colleagues”), a social transgression (“You find out that an acquaintance has been criticizing you behind your back”), and a romantic transgression (“Your partner cheats on you by sleeping with someone else”). Levels of forgiveness measured before (α = .84) and after (α = .75) being told that the transgressor had apologized were highly correlated across the three scenarios and so were combined into separate scales of pre- and postapology forgiveness. Apology bonus was calculated by subtracting the mean pre-apology forgiveness score from the mean postapology forgiveness score, such that higher scores reflected a greater forgiveness bonus for receiving an apology. Political ideology was measured by asking participants where they would place themselves on a scale of “left” versus “right” (1 = left, 9 = right) and “liberal” versus “conservative” (1 = very liberal, 9 = very conservative). The 2 items were positively correlated (r = .33, p < .001) so were combined into a single scale of political conservatism.3 Covariates Five control variables were measured at the end: age, sex (1 = male, 2 = female), income (1 = much lower than the average national income, 5 = much higher than the average national income), education (1 = did not finish high score; 8 = professional degree), and religiosity (1 = not at all religious, 7 = extremely religious).

Study 2 Section: Choose Top of page Abstract Study 1 Method Results and Discussion Study 2 << Method Results and Discussion General Discussion References CITING ARTICLES Although relatively consistent, the size of the relationship between political conservatism and proclivity to apologize in Study 1 ranged from small to moderate. Furthermore, it is unclear how much of the relationship could be explained by common method variance and pattern responding; methodological artifacts associated with the approach of correlating two self-report scales. To address this, we sought to replicate the main relationship between political orientation and proclivity to apologize, but this time operationalizing proclivity to apologize by coding open-ended responses. In Study 2, participants imagined they had committed a transgression toward a neighbor and were asked to record what they would say to the neighbor in response. Consistent with Study 1, it was expected that more conservative participants would be less likely to use apologetic statements in their response than would relatively liberal participants. Given the inconsistent mediational evidence for the entity belief measure in Study 1, we again measured entity beliefs but this time with a different scale.

Method Section: Choose Top of page Abstract Study 1 Method Results and Discussion Study 2 Method << Results and Discussion General Discussion References CITING ARTICLES Participants Study 2 was conducted using Amazon’s Mechanical Turk. Eighty-three participants initially completed the study, but 5 left their open-ended response blank, and 13 did not follow instructions (i.e., they either imagined themselves from the victim’s perspective or wrote something that was clearly not a response to the neighbor). This left 65 usable participants (56.9% male; M age = 32.75), comprising 38 from India and 27 from the United States. Materials and Procedure Participants first completed a number of individual differences scales, including a 3-item measure of entity beliefs adapted from Erdley and Dweck (1993; e.g., “Someone’s personality is a part of them that they can’t change very much;” 1 = strongly disagree, 7 = strongly agree; α = .96).5 Participants were then asked to imagine that they had committed a transgression against a neighbor (by not watering their plants as frequently as promised, leaving the neighbor’s plants distressed) and to write down what they would say to the neighbor.6 Two assistants blind to hypothesis coded the responses for the presence (coded 1) or absence (coded 0) of apologetic statements (e.g., I’m sorry or I apologize). Interrater reliability was high (κ = .93); discrepancies were resolved through discussion. At the end of the study, participants recorded their demographic details, including political conservatism (1 = extremely liberal; 7 = extremely conservative) and the control variables of age, sex, nationality (measured as country of birth), and religiosity (measured as in Study 1).

Results and Discussion Section: Choose Top of page Abstract Study 1 Method Results and Discussion Study 2 Method Results and Discussion << General Discussion References CITING ARTICLES As predicted, the more conservative participants were, the less likely they used apologetic statements in their open-ended response (r = −.38, p = .002). This relationship remained significant after controlling for nationality, age, religiosity, and sex (r partial = −.37, p = .004). Although the main association documented in Study 1 proved robust in Study 2, the role of entity beliefs in explaining the association was not: Entity beliefs did not significantly correlate with either political orientation (r = .21, p = .093) or apologetic statements (r = −.08, p = .51). This is despite the fact that the two cultures examined in this study—India and the United States—were the cultures in Study 1 for which there was robust evidence that entity beliefs played a role.

General Discussion Section: Choose Top of page Abstract Study 1 Method Results and Discussion Study 2 Method Results and Discussion General Discussion << References CITING ARTICLES This article provides the first evidence that people are more reluctant to apologize if they are more politically conservative. Because there is a temptation to simplify and magnify differences between ideological categories, we want to put on the record that this does not mean that conservatives are anti-apology. In Study 1, scores on our reluctance to apologize measure were generally below the midpoint, so the majority of our participants showed a proclivity to apologize, regardless of political orientation. But the data across both studies showed that this proclivity to apologize waned for participants who were more politically conservative. Interestingly, in Study 1, politically conservative participants were also less influenced by apologies when determining whether a transgressor deserved forgiveness. Again, this is not to say that conservatives are unimpressed by apologies: Forgiveness levels were much higher in the presence of an apology than in the absence of an apology, and this was the case for the vast majority of participants in our sample. But the data show that the apology bonus in terms of forgiveness is less pronounced for conservatives than for liberals. So not only are conservatives less likely to want to provide an apology than liberals, but they are also less likely to weight an apology when determining whether to forgive. We do not wish to imply that there is an intrinsic moral virtue or psychological advantage in giving or accepting apologies. Elsewhere it has been argued that there is a downside to the normative expression of apologies as a default response to hurt (Okimoto, Wenzel, & Hornsey, 2015), and there is also a case to be made that it might be premature to forgive in the face of thoughtless or qualified apologies (e.g., Barlow et al., 2015). However, it is undeniable that in general the proclivity to apologize is associated with positive well-being (Howell et al., 2011) and that receiving apologies helps enable victims to move on (Fehr et al., 2010; Tavuchis, 1991). Forgiving transgressors is also associated with a range of positive consequences in terms of physical and psychological health (e.g., Exline & Baumeister, 2000; Wade, Hoyt, Kidwell, & Worthington, 2014). Given this, the current data suggest that conservatives’ relative unwillingness to give and accept apologies might present an obstacle to intrapsychic health and interpersonal reconciliation. We examined two mechanisms through which political conservatism and attitudes toward apologies might be connected: SDO (Study 1) and entity beliefs about human nature (Studies 1 and 2). The positive link between conservatism and SDO reflects a long-established effect in the literature (Ho et al., 2015). What is new is the finding that SDO was positively associated with a reluctance to apologize and negatively associated with the tendency to reward an apology with forgiveness. We argue that this reflects people’s underlying orientations to the importance of reducing or maintaining power inequalities. Apologies are widely construed to be a rhetorical vehicle for removing power inequities in relationships posttransgression. As such, it makes sense that low-SDO people—who are committed to equality—are more likely to provide and reward apologies than higher SDO people, who are comfortable with maintaining power differences. We also theorized that people may be more reluctant to provide apologies the more they held entity beliefs about human nature. Given that apologies are often seen as markers of character reform or transformation, it seemed plausible that people would be less congenial to apologies the more they believe that the people cannot fundamentally change who they are. But despite the evidence for an indirect effect displayed in Figure 1, we are cautious in concluding that entity beliefs are a major reason for why conservatives are more likely to give apologies than liberals. First, the mediating role of entity beliefs emerged reliably in only two of nine national samples in Study 1 and was not replicated in Study 2. Second, even in Study 1, there was no evidence that entity beliefs mediated the relationship between political conservatism and an apology bonus in terms of forgiveness. Together, the results offer only limited evidence for the hypothesis that the association between political ideology and the willingness to provide and reward apologies is tied to broader beliefs about the capacity for transformation. Strengths, Limitations, and Future Directions Nine national samples allowed us to test the generalizability of the observed effects outside WEIRD samples, which is a strength of the study. However, one challenge of examining this question outside WEIRD samples is that the notion of left and right ideology has different psychological resonances in different countries. Five of the nations in the current samples have communist, postcommunist, or socialist governments. In China and Russia, for example, “left-ist” values may be less tied to “progressive” or “liberal” values, as they are in long-standing democracies such as the United States or Australia (Huber & Inglehart, 1995; Piurko, Schwartz, & Davidov, 2011). Indeed, in China, Hong Kong, and Russia, the correlations between the two indices of political ideology were relatively weak. However, we are reassured that the noise attached to our operationalization of political ideology in these countries would have worked against our hypotheses rather than for them, meaning that the observed effects are unlikely to be artifacts of psychometric issues. Of course, the cross-sectional nature of our designs raises the usual questions about causality. We can only state with confidence that political ideology and attitudes to apology are associated (rather than being linked causally). However, it should be noted that the reverse causal path to that shown in Figures 1 and 2 seems less theoretically plausible: It is hard to envisage how attitudes specifically about apology could influence broader political ideologies or the other variables of interest (SDO and implicit theories about human nature). Finally, it remains to be seen whether the results would generalize to more public apology contexts (e.g., many-to-many apologies). Previously, theorists have cautioned against taking conclusions drawn from research on interpersonal contexts and uncritically extrapolating it to the collective context (Hornsey & Wohl, 2013; Philpot & Hornsey, 2008). So although there are theoretical reasons to expect that the link between political ideology and apology beliefs would also emerge at the collective level, this link should not be presumed. This is especially the case given that public/collective apologies are often wrapped up in political debates about where the boundaries of morality and responsibility lie (Wohl, Hornsey, & Philpot, 2011), so whether conservatives or liberals are more likely to want to apologize might well depend on the sociopolitical context surrounding a transgression. Summary This research uncovered a previously undocumented relationship: Conservatives are less likely than liberals to want to provide apologies and are less likely to reward received apologies with forgiveness. These data—and the mediating role of SDO—reinforce theorizing that construes apologies as rhetorical tools for diminishing power differences posttransgression and suggest that willingness to give and receive apologies may be related to individual differences about the desirability of equality as a principle. More broadly, the data reinforce how political ideologies infuse how people conduct themselves interpersonally, in this case in ways that have implications for interpersonal reconciliation and forgiveness.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data were collected using funds from an Australian Research Council grant awarded to Hornsey and Wohl. Translations were partly supported by funds from FONDECYT (1161371), the Centre for Social Conflict and Cohesion Studies (FONDAP 15130009), and the Center for Intercultural and Indigenous Research (FONDAP 15110006) awarded to González.

Notes 1

Evidence for this latter notion can be extrapolated from the finding that prosocial individuals (who typically value equality) behaved more cooperatively in a negotiation with a hypothetical colleague who had apologized for an angry outburst than with a colleague who had not apologized. In contrast, those low in prosociality were more likely to exploit the apologetic counterpart (Van Kleef & De Dreu, 2010). 2

Given the modest internal consistency of the Social Dominance Orientation (SDO) Scale, we examined each item separately as a potential mediator of the link between political conservatism and the criterion variables. Indirect effects described in Figures 1 and 2 remained significant when the SDO Scale was replaced with any one of its individual items. 3

The measures were included after a set of questions on people’s folk theories about what constituted perfect societies, measures that were designed to answer an unrelated research question (the full set of measures is available upon request). Three hundred and five participants—who started the questionnaire but dropped out before completing any of the measures in the current manuscript—are not counted in the reported N for this study. 4

The model predicting apology bonus was then repeated but with two different methods of operationalizing the change score reflecting the difference between pre-apology and postapology forgiveness. First, we regressed the postapology forgiveness scores on the pre-apology forgiveness scores and used the associated standardized residuals as the criterion variable. We then ran a separate model in which we used only postapology forgiveness scores as the criterion variable but with pre-apology forgiveness scores included as a covariate. Whichever way the criterion variable was analyzed, the effect remained the same: SDO was a significant mediator, but entity beliefs were not. 5

Other measures—for example, of self-consciousness, self-esteem, need for cognition, and need for closure—were peripheral to hypotheses. Data can be obtained upon request from the first author. 6

Half the participants imagined that the neighbor’s plants were wilting and the foliage brown; the other half imagined that the plants were almost all dead and the foliage burned. Results were the same regardless of the severity of the consequences described in the scenario, so data are reported after collapsing across these conditions.