A Hero in the Name of Truth : Mortality Salience Increases Heroic Perceptions of Edward Snowden

Edward Snowden has made significant personal sacrifices by revealing the truth about secret surveillance activities by the National Security Agency. According to research on terror management theory, reminding people of their own death strengthens worldview-bolstering reactions, such as following cultural standards and supporting people who uphold those standards. Assuming Snowden’s disclosures were motivated by fighting for the truth and for the value of honesty, and further that cultural values become more important under mortality salience, we hypothesized that reminding people on their own death increases heroic perceptions of Snowden. As predicted, results of our study indicated that participants in the mortality salience condition reported stronger heroic perceptions compared to a control group. This effect was found to be independent of patriotism or political orientation.


Introduction
The former National Security Agency (NSA) contractor Edward Snowden has become famous worldwide for disclosing top-secret NSA documents containing, among other things, information about global surveillance activities run by the NSA, apparently also including intensive surveillance of Americans (e.g., Risen & Poitras, 2013).According to Snowden, his actions were based on the motivation "to inform the public as to that which is done in their name and that which is done against them" (Greenwald, MacAskill, & Poitras, 2013).As recent polls have indicated, U.S. citizens are divided as to whether Snowden is a hero or a villain-with Republicans showing less support for Snowden than Democrats (e.g., Newport, 2013).Despite being a subject of controversy, it appears that Snowden has made significant personal sacrifices (such as living in exile apart from his family, being accused of breaking the law, giving up his career) for the benefit of others; namely, revealing the truth about surveillance activities by the NSA.Thus, one can assume that he acted in favor of (one might even say fight for) important cultural values, such as honesty and truth.Given now that people face an existential threat like terroristic attacks, how will this affect perceptions of a person like Snowden?
According to terror management theory (TMT; e.g., Greenberg, Solomon, & Arndt, 2008), culture functions as an anxiety buffer against the ever-present potential terror of death by providing a meaningful, orderly conception of reality that contains a set of norms and values.By living up to cultural standards, people believe that they are valuable beings in a meaningful reality, what consequently increases self-esteem and reduces the fear of a final end (e.g., for a review see e.g., Pyszczynksi, Greenberg, Solomon, Arndt, & Schimel, 2004).Based on this idea, the Mortality Salience (MS) hypothesis states that reminding people of their mortality should lead them to increase their defenses and bolster their cultural worldview, resulting in derogating those who violate important cultural values (e.g., McGregor et al., 1998;Rosenblatt, Greenberg, Solomon, Pyszczynski, & Lyon, 1989) and supporting those who uphold them (e.g., Greenberg et al., 1990;Greenberg, Simon, Pyszczynski, Solomon, & Chatel, 1992;Mikulincer & Florian, 1997).Rosenblatt et al. (1989, Study 3), for example, found MS to lead to larger reward recommendations for a woman who helped the police apprehending a mugger, that is, for a "hero who upheld cultural values" (p.681).Moreover, Greenberg et al. (1990), for example, found American participants to show greater affection for a pro-American author after having been reminded on their own death, whereas an anti-American author was devaluated under MS.
Building on the assumptions that Edward Snowden's disclosures were motivated by fighting for the truth and for the value of honesty, and that cultural values become more important under MS, we hypothesized that MS should increase heroic perceptions of Snowden.As patriotic attitudes and political orientation were shown to affect worldview defending reactions (Pyszczynski et al., 2006), we included those factors in our investigation.However, assuming fighting for the truth to be a cultural value independent of patriotism and political orientation, we did not expect any interaction of those factors with MS on support of Snowden.
Although our idea seems to be straightforward, it nevertheless contributes to TMT research, especially regarding the general need of heroes in the face of threat.Based on the idea that mortality concerns can be solved by having faith in charismatic, prominent figures (like religious or political leaders; Becker, 1973), studies on TMT showed that MS increases the attraction of such "larger-than-life" persons (Cohen, Solomon, Maxfield, Pyszczynski, & Greenberg, 2004;Landau et al., 2004;Landau, Greenberg, & Sullivan, 2009).Landau and his colleagues (2004), for example, demonstrated that MS (as well as reminders on the 9/11 terroristic attacks) increased support for George W. Bush.According to the authors, MS increases the need for a powerful, charismatic authority who manages their (literal and symbolic) mortality concerns-in this case represented by the former President Bush.In contrast to the current study, in all the studies on MS and charismatic leaders, heroic perceptions were not directly measured.Furthermore, recent research identified features like bravery, moral integrity, honesty, selflessness, and helpfulness to be perceived as prototypical for a hero (Kinsella, Ritchie, & Igou, 2015).According to this approach (especially regarding selflessness and bravery), Edward Snowden can be assumed to fulfill the necessary (prototypical) criteria of being a hero.Taking this into account, our study aims to extend previous knowledge on MS effects by directly investigating the idea that MS increases not only the need for charismatic leaders but also for heroes.

Subjects and Design
We calculated the sample size to obtain sufficient power (80% to detect an effect if one exists ;Cohen, 1988).Power analysis assuming an effect between the two conditions of r = .25(cf.non-student sample in Burke et al., 2010) revealed an N of 120.Given the possibility of easily obtaining large samples through Internet experiments (Reips, 2002), 202 U.S. citizens (118 male, 76 female; ages 18-71 years, M = 33.94,SD = 11.95) participated in our Internet study via Amazon Mechanical Turk (cf.Buhrmester, Kwang, & Gosling, 2011).Regarding ethnicity, 156 participants indicated being Caucasian, 19 African American, 9 Asian, 3 Asian American, 7 Hispanic, and 6 indicated "other".Regarding the current occupational status, 102 participants reported being employed, 53 self-employed, 21 work seeking, 18 studying, and 8 reported "other".Of them, we excluded three participants because they guessed our hypothesis and five participants because they did not know whom Edward Snowden was, leaving a total of 194 participants.All participants were randomly assigned to the experimental conditions in a 1 × 2 (MS: mortality condition vs. dentist-visit control condition) between-subjects factorial design.

Procedure and Measures
First, participants filled out the demographic measures, including one item assessing patriotism and one item assessing political orientation (conservative vs. liberal; e.g., Nail & McGregor, 2009).Next, they were randomly assigned to a mortality condition or a dentist-visit control condition: participants in the MS (dentist-visit) condition answered two open-ended questions about death (a visit to the dentist; c.f. Jonas et al., 2008).When using such explicit death primes, a distractor is necessary to elicit effects of worldview defense and bolstering (Arndt, Greenberg, & Cook, 2002).Therefore, as in many studies on TMT (Burke et al., 2010), participants filled out 60 items of the Expanded Form of the Positive and Negative Affect Schedule (PANAS-X; Watson & Clark, 1992).Afterwards, we assessed whether participants perceived the person Edward Snowden as a hero, using nine items (α = .96;e.g., "Edward Snowden is a hero"; "Edward Snowden tells the truth"; "Edward Snowden is a betrayer").Participants responded to all nine items on a 9-point Likert-type scale ranging from 1 (fully disagree) to 9 (absolutely agree).

Results
In a first step, two one-way ANOVAs were calculated with MS as factor and positive (α = .92)and negative (α = .90)affect scores of the PANAS-X as dependent variables, respectively, showing that, consistent with prior research on TMT, MS had no effect on participants' mood, both ps > .248.To test our hypothesis, we used a multiple linear regression analysis with MS, patriotism, and political orientation in the first step, to predict heroic perceptions of Snowden.In a second step, we included the two-way interaction terms of MS × Patriotism and MS × Political Orientation, and the three-way interaction term of MS × Patriotism × Political Orientation.MS was centered (coded-1 for the dentist condition and 1 for the MS condition), and patriotism and political orientation was z-standardized.
The first model explained 11.8% (10.4% adjusted) of heroic perceptions of Snowden, F(3, 190) = 8.46, p < .001.In line with our predictions, a significant main effect of MS occurred, b = .31,SE b = .13,t(190) = 2.30, p = .023,indicating that participants in the mortality condition reported stronger heroic perceptions compared to participants in the control group.Additionally, there was a significant main effect of patriotism, b = -.52,SE b = .14,t(190) = .370,p < .001,indicating that heroic perceptions decreased with increasing patriotism.Political orientation was not a significant predictor, p > .183.Including the four interaction terms in the second model did not significantly improve the overall model fit, R 2 change = .03,p = .189.In this model, MS, b = .34,SE b = .14,t(186) = 2.44, p = .016,and patriotism were again significant predictors, b = -.46,SE b = .15,t(186) = -3.02,p = .003.No significant effects of political orientation or the interaction terms occurred, all ps > .147.Controlling for sex and age did not affect the results significantly.

Discussion
Although the case of Edward Snowden is a subject of high controversy, there is less doubt that he has made significant personal sacrifices for revealing the truth about surveillance activities by the NSA.Thus, one can assume that he acted in favor of honesty and truth, which reflect important cultural values (Bishin et al., 2006;Weber & Ruch, 2012).According to research on TMT, indicating that individuals under MS show greater support for people who uphold cultural values (e.g., Greenberg et al., 1990;Rosenblatt et al., 1989), we predicted heroic perceptions of Edward Snowden to be enhanced after reminding people of their own death.Results of our study yielded support for our hypothesis.There was also a main effect of patriotism, indicating that heroic perceptions decreased with increasing patriotism.However, and in line with the assumption that fighting for the truth constitutes a cultural value independent of patriotism and political orientation, we did not find any interaction of those factors with MS in support of Snowden.Interestingly, there was no main effect of political orientation on heroic perceptions.This was only the case when patriotism was not included in the model, showing that political orientation does not account for variance beyond patriotism.
Given that we directly assessed heroic perceptions, the current findings are also interesting concerning the general need of heroes in the face of threat.Studies showed that MS increases support for charismatic leaders (Cohen et al., 2004;Landau et al., 2004Landau et al., , 2009)).However, in contrast to the current study, heroic perceptions were not directly measured in all those studies.Furthermore, according to the prototypical perception of heroes (Kinsella et al., 2015), we assumed Edward Snowden (rather than the previous investigated charismatic leaders) to fulfill the prototypical criteria of being a hero.Accordingly, our study contributes to TMT literature by providing first evidence for the idea that MS increases not only the need for charismatic leaders but also for heroes, namely by showing that MS leads people to perceive the actions of Edward Snowden as more heroic.This finding can be applied to existential threatening events, such as the terror attacks of 9/11: such an event is likely to increase heroic perceptions of people who uphold important cultural values.Nevertheless, given that we did not assess whether Snowden was categorized as an ingroup member (citizen of the USA), the role of perceived group membership as a necessary condition for increasing heroic perceptions after MS remains still an open question.
It should be emphasized that there are further circumstances that might lead to different outcomes than increasing support of Snowden.Research on TMT suggests salience of cultural norms and values to play a crucial role for MS guided reactions (e.g., Gailliot et al., 2008;Jonas et al., 2008;Schindler et al., 2013): norms and values only influence behavior to the extent that an individual's attention is focused on them.From this perspective, when bringing cultural values like loyalty towards the employer into focal attention, MS might rather lead to a devaluation of Snowden's actions.One could further speculate that the same effect would occur when pronouncing increased insecurity through Snowden's disclosures.Notably, regarding our online sample, one might speculate that people who register on a platform, such as Amazon Mechanical Turk, are more sensitive, in general, toward topics such as data security (or data security is at least a highly salient topic while participating), and show therefore more support for Snowden after MS.
Considering literature on social comparison (e.g., Lockwood & Kunda, 1997), it seems plausible that under certain conditions, MS decreases heroic perceptions even of a person who acts in line with their own worldview, namely when one's own achievements and contributions are devalued in comparison to those of the heroic person (which potentially further threatens one's self-esteem).Thus, future research should address the generalizability and bounding conditions of our findings.