The role of mortality awareness in hero identification

Abstract Three studies examine hypotheses derived from terror management theory to investigate the relationship between mortality concerns and hero identification. Study 1 found reminders of death, followed by a distraction task and a self-prime, led to greater inclusion of heroes in the self. Study 2 found that writing about a personal hero, but not other’s heroes or acquaintances, led to lower death-thought accessibility after being reminded of mortality. Finally, Study 3 found that after death reminders, participants led to identify with a hero exemplifying traits of legacy and/or sacrifice showed lower death thought accessibility. Findings are discussed as generative for heroism research, informing a previously overlooked motivation underlying hero identification and the existential function of such identification.


Terror management theory
Initially inspired by the works of ernest becker (e.g. becker, 1973), TMT posits conscious and non-conscious awareness of mortality engage different motivational orientations aimed at buffering the threat associated with mortality-related cognition (Pyszczynski, Greenberg, & Solomon, 1999). Conscious thoughts of death are argued to instigate responses focused on removing death thoughts from focal attention. In contrast, subtle or fleeting thoughts of mortality are thought to motivate people to strive for cultural value. Culture, in turn, confers existential security by allowing individuals to feel a part of something more enduring than the self. This, in turn, enables them to live with relative equanimity in the face of knowledge that life will, at least physically, end. Culture also provides standards of value that enable people to maintain self-esteem, or the sense of being a significant contributor to the cultural meaning system.
TMT has inspired considerable research suggesting, for example, that reminders of mortality motivate people to identify with their cultural beliefs and strive to enhance their self-worth, and that defense of these structures then reduces cognitions about mortality (see e.g. Greenberg, 2015, andHayes, Schimel, Arndt, & for conceptual reviews; burke, Steinman &Updegraff, 2015 for meta-analytic reviews). However, despite the many different directions of terror management research, very little work has examined heroism. This is surprising given that not only was heroism a central topic for becker (e.g. 1971), but the first empirical TMT paper included a study that assessed reactions to an ostensibly heroic individual. Specifically, in the context of examining how mortality salience (MS) affected reactions to those who support one's worldview, Rosenblatt, Greenberg, Solomon, Pyszczynski, and Lyon (1989) found participants reminded of mortality (vs. no reminder) recommended a greater monetary reward to a person who helped apprehend a criminal (and thus presumably upheld the participants' worldview). Rosenblatt et al. referred to the "worldview supporter" as a "hero" in their discussion of the study, but an explicit focus on the existential implications of reacting to heroes specifically was never developed.
Recent research has started to fill this gap by probing the relationship between mortality concerns and heroic enactment. McCabe, Carpenter, and Arndt (2015) theorized that if heroes are culturally valued and death reminders enhance striving for cultural values, then death reminders should enhance effort on an ostensibly heroic task. They further reasoned that feedback about whether or not one completed the heroic task successfully should predict the extent to which the topic of death continues to be cognitively active. Consistent with these hypotheses, after reminders of mortality (vs. a control topic) and linking pain endurance to heroism (vs. a positive personality), participants reported less pain during a cold pressor task. Further, participants who had their heroism or positive personality confirmed (vs. disconfirmed) by false feedback showed lower death thought accessibility (DTA; as assessed by a word completion task where some words could be completed with death or non-death words, e.g. COFF_ _ could be COFFIN or COFFee [see Hayes et al., 2010]). While this study suggests perceptions of being heroic can confer existential security, questions remain concerning the role of mortality awareness in motivating identification with heroes, and whether such identification also serves an existential function.

The present research: the existential underpinnings of hero identification
Following Rank (e.g. 1909), becker (1971,1973) argued that feelings of worth (self-esteem) reflect cultural significance and foster a sense that one will transcend mortality by gaining symbolic immortality. Heroism was considered to be the idealization of this process, in that heroes not only at times transcend death themselves, but also embody cultural values and, thus, achieve symbolic immortality. As an ideal, heroes could then provide others the opportunity to vicariously experience feelings of worth and immortality by fostering a sense of identification with said hero. Consequently, becker argued awareness of mortality not only spurs people to heroism, but also plays a central role in people's propensity to identify with heroes.
by identifying with heroes, becker (1973) suggested death concerns are quelled as the death transcending qualities of heroes are transferred to the self. Although this idea has not yet been tested, some research is consistent with aspects of this analysis. Sullivan and Venter (2005) presented participants with a Stroop-like task involving making judgments about whether certain traits were self-descriptive or not. Participants were faster to endorse traits shared with their heroes, but not their non-heroes, suggesting inclusion of heroes in the self-concept. As noted above, however, it is currently unknown whether certain motivational factors might exacerbate this kind of identification. The present work examines the possibility that awareness of mortality prompts greater identification with heroes, and thus, a greater incorporation of heroes into the self. This hypothesis is examined in Study 1.
Concerning hero identification, there are many ways such identification may be functional. For example, heroes have been argued to provide utility as leaders to guide others through difficult circumstances (Freud, 1922;Le bon, 1895;Fromm, 1941), vicarious experiences of success (bernhardt, Dabbs, Fielden, & Lutter, 1998), increases in positive affect and selfesteem (as with basking in reflected glory; Cialdini & Kenrick, 1976), ideal self-images that guide behavior toward attaining a possible self (Caughey, 1984;Higgins, 1987;Sullivan & Venter, 2005), and role models (yoon & Vargas, 2014). However, one function of hero identification yet to be explored is the protective utility of identified heroes against existential concerns. Indeed, the relationship between identification with heroes and mortality concerns may be bidirectional, such that not only may hero identification be motivated by mortality concerns, but such identification might quell the psychological threat of mortality. This second possibility is examined in Study 2, and with a more nuanced look at specific heroic traits, in Study 3.
Taken together, the present studies examine two central questions about motivations and functions related to hero identification. First, do death reminders motivate hero identification? Second, does hero identification reduce activation of death-related thoughts?

Study 1
To inform whether mortality reminders motivate hero identification, Study 1 examines self-representations. People's self-representations are multifaceted, and can include other people (Aron, Aron, & Smollan, 1992). Inclusion of others into the self has been argued to be a core facet of social identification (Coats, Smith, Claypool, & banner, 2000;Tropp & Wright, 2001), and includes such aspects as feelings of shared meaning (Schutz, 1970) and of being in communion (bakan, 1966;Merleau-Ponty, 1945). Further, as noted, previous research suggests heroes are included in the self (Sullivan & Venter, 2005). What are the motivational catalysts of this identification? Study 1 tested whether mortality reminders are one such trigger that spurs people to include their heroes in the self.
In addition, Study 1 sought to inform two other issues. The first concerns the role of the self in existentially motivated hero identification. To the extent a person's heroes form part of the self-identity network (e.g. Sullivan & Venter, 2005), self-salience should temporarily provoke activation not just of the self, but also of one's heroes, and the link between those constructs. In contrast, a focus on others might disrupt this link. Integrating this work with TMT gives rise to a novel hypothesis. If hero identification is motivated in part by awareness of mortality, the capacity of MS to motivate inclusion of the hero in the self may be moderated by the salience of the self. However, a manipulation of the salience of self is not typical of terror management studies on identification, and thus it is not clear whether experimentally induced self-salience is necessary. Study 1 included such a manipulation to examine this possibility.
Additionally, Study 1 examined the effects of conscious vs. non conscious mortality concerns and implications for hero identification. Research finds conscious thoughts of death tend to motivate people to make rationally-oriented efforts to remove death-related thought from focal awareness (e.g. through suppression, distraction, or vulnerability reduction; Greenberg, Pyszczynski, Solomon, Simon, & breus, 1994;Goldenberg & Arndt, 2008). In comparison, death thoughts outside of focal awareness motivate people to rely on cultural values that provide a sense of existential security (e.g. Greenberg, Arndt, Simon, Pyszczynski, & Solomon, 2000;Pyszczynski et al., 1999). Consequently, following prior research, the presence of a distraction task following the MS induction was manipulated to test if conscious thoughts of mortality have differential effects on hero identification (inclusion in the self ), relative to non-conscious thoughts of death. Conceptually, this tests the idea that if heroes are culturally valued and thoughts of death outside of focal attention spur cultural identification processes, then heroes should be more strongly included in the self when death thoughts are outside of focal attention (i.e. when cultural buffers are engaged).

Participants
One hundred sixty MTurk workers (age: M = 32.14, SD = 10.18; male = 100, female = 60) participated for financial compensation ($1.88). Participants were randomly assigned to conditions in a 2 (Salience: MS vs. uncertainty) X2 (no distraction vs. distraction) X2 (self vs. other prime) experimental design. Across studies, sample sizes were based on recommendations at the time the studies were run (Simmons, Nelson, & Simonsohn, 2011), and reflect conventions in the terror management literature supported by meta-analyses on MS effect sizes when longer delay tasks are included between the MS manipulation and assessment of the dependent variable (e.g. Martens, burke, Schimel, & Faucher, 2011;Steinman & Updegraff, 2015). We return to this issue in the general discussion.

Materials and procedure
Mortality salience. To manipulate awareness of mortality, a method extensively employed in TMT studies was used (burke et al., 2010). In the MS condition, participants completed two open-ended questions, "briefly describe the emotions that the thought of your own death arouses in you" and "Jot down, as specifically as you can, what you think happens to you as you physically die and once you are physically dead. " The control condition asked similar questions about another potential existential threat -uncertainty ("briefly describe the emotions that thought of being uncertain arouses in you" and "Jot down, as specifically as you can, what you think happens to you physically when you experience uncertainty"). This comparison condition helps to inform whether heroic identification is spurred by reminders of mortality relative to other existential concerns.

Distraction.
To manipulate the level of consciousness of existential thoughts, participants were next either given or not given a set of distracter tasks. The distraction tasks, as used in previous studies, were included to move thoughts of mortality out of direct conscious attention (see Martens et al., 2011;Steinman & Updegraff, 2015; for relevant meta-analyses). Those in the distraction condition completed the 20-item Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988; note this also allowed for an assessment of initial effects on affect) and read an extract from a short story (Camus, 1958), taking approximately 5-8 min to complete. The no-distraction condition did not include these tasks.
Self/other prime. Following these manipulations, to prime thoughts about the self or other people, participants completed a modified version of the linguistic implications form (Wegner & Giuliano, 1983); this 10-item measure involved completing sentences by selecting either self-focused pronouns, or other-focused pronouns. example items include "Please don't do this to (me, my, I), it is just not fair" (self condition), and "After spreading fertilizer liberally over the flower bed, (he, she, they) watered the flowers" (other condition).

Inclusion of the hero in the self.
Participants then completed a modified version of the inclusion of other in self scale (IOS; Aron et al., 1992). The scale is designed to examine how much "the other is included in the self, " and has been used in prior identification studies (e.g. Tropp & Wright, 2001). Participants were instructed to think about a hero and their relationship with that hero, and then instructed to select one of seven Venn diagrams which overlapped to various degrees, from completely separate to near complete overlap. One circle contained the word "Self" and the other contained the word "Hero. " The more overlap, the more the hero is considered to be included in the self.

Inclusion of the hero in the self
A 2 (MS vs. uncertainty) X2 (no distraction vs. distraction) X2 (self vs. other prime) ANOVA (analysis of variance) on IOS revealed a main effect of MS, F(1, 152) = 13.5, p < .001, 2 p = .08, such that after reminders of mortality, inclusion of the hero in the self was greater. There was also a main effect of distraction, F(1, 152) = 6.65, p = .011, 2 p = .04, such that, after a distraction inclusion of the hero in the self was greater. As depicted in Figure 1 and Table 1, the predicted 3-way interaction between MS, distraction, and self-prime also emerged, F(1, 152) = 6.42, p = .012, 2 p = .04. 1 A series of planned comparisons revealed a pattern consistent with hypotheses. When participants were reminded and then distracted from mortality and primed with the self, they showed greater inclusion of the hero in the self than participants who were reminded of but not distracted from mortality and primed with self, t(39) = 3.55, p < .001, d = 1.14, participants who were reminded of mortality, distracted, but not primed with self, t(43) = 1.98, p = .049, d = .6, and participants who were distracted and primed with the self but reminded of uncertainty, t(36) = 4.31, p < .001, d = 1.44. In short, they evinced more hero inclusion than all other conditions.

Discussion
Study 1 tentatively suggests mortality reminders may motivate hero identification. When participants were reminded of, and then distracted from, thoughts of death, and the self was primed, participants evinced greater inclusion of the hero in the self.
In addition to the main finding that a reminder of mortality seemed to spur hero identification, Study 1 also makes three further tentative contributions concerning the role of mortality concerns in hero identification. First, greater inclusion of the hero in the self was only observed after MS when death thoughts were presumably outside of focal awareness, when previous research suggests cultural defenses are engaged (e.g. Greenberg et al., 2000). There were no differences in inclusion of the hero in the self when death thoughts were presumably in focal attention, when rationally-oriented defenses have been found to be operative. This attests to the cultural value of heroes and suggests identification with heroes may be ineffective in dealing with conscious death thoughts. Second, the results suggest observed inclusion effects are not due to existential threats broadly (as the control condition   was also an existential threat i.e. uncertainty). While the salience of uncertainty has important effects on social judgment and behavior (van den bos, 2001), in the present context, it did not produce the same effects as reminders of mortality. Third, the results suggest an important role of self-salience, in that inclusion effects did not emerge when participants were not primed to think of the self. An ambiguity of Study 1, however, stems from the omission of a non-hero comparison target. Although the focus of the study was on determining whether MS would engage heroic-self inclusion, and not with whether such inclusion was necessarily unique to heroes, this omission raises questions about the terror management function of heroic identification relative to other types of identification. One way to inform this ambiguity is to manipulate whether participants are given the opportunity to identify with a personal hero or with someone they do not consider a personal hero and assess the potentially differential effects on death-related cognition. Study 2 adopts this approach.

Study 2
Study 2 shifts focus from death concerns motivating identification with heroes, to examining how such identification may function to reduce mortality-related cognition. To test this, participants, after being distracted from reminders of mortality (vs. control), wrote about either a personal hero, someone else's hero, or an acquaintance. It was hypothesized that, after reminders of death (vs. control) and bringing to mind a personal hero (with whom they presumably identify), participants would be shielded from mortality concerns as indexed by lower DTA. The inclusion of the other person's hero condition allowed a test of whether bringing to mind heroes generally decreases mortality-related cognition, or if such decreases are the result of personal (identified) heroes uniquely. Similarly, the personal acquaintance condition was included to discern whether bringing to mind anyone (hero or not), tethered to the self, might serve to placate death-thoughts, perhaps as such facilitates feelings of not being alone (Wisman & Koole, 2003).
The self-prime was not included for this study as it was reasoned that writing about a personal hero (vs. other's hero vs. personal acquaintance) was sufficient to activate thoughts of a hero tethered to the self, i.e. an identified hero. Further, the distraction task was included across conditions as Study 1 found the existential value of heroes was only evinced when death thoughts were outside of focal awareness, when the cultural value buffer was presumably engaged. Finally, Study 2 changed the control condition from uncertainty to pain to inform whether hero identification results from eliciting threatening thoughts generally.

Participants
One hundred thirty-six participants (age: M = 36.97, SD = 11.95; 60 males; 76 female) completed the survey posted on MTurk in exchange for financial compensation ($.50). Participants were randomly assigned to conditions in a 2 (MS vs. pain) X3 (personal hero vs. other's hero vs. personal acquaintance) experimental design.

Materials and procedure
The MS manipulation and distraction materials were included as in Study 1, with the exception that the control condition was changed from uncertainty to pain ("briefly describe the emotions that the thought of pain arouses in you" and "Jot down, as specifically as you can, what you think happens to you physically when you experience pain").
Hero writing task. Following MS and the distraction task, participants were presented with an ostensible "writing task. " The accompanying instructions read We want you to think of someone you consider a hero (an actual person)/someone who others consider a hero but you do not/someone you consider an acquaintance. In the space below please write at least a paragraph about this person such as what they look like and some of the things they have done.
Participants then responded to a manipulation check question "When you think about this person, how heroic do you consider them?" on a 10-point Likert type scale (1 -Not at all heroic, 10 -Very heroic).
Death thought accessibility (DTA). DTA was assessed after the writing task with a word fragment completion task used in many previous TMT studies (Hayes et al., 2010). Participants completed 25 word fragments, 7 of which could be completed with either a death-related word or a non-death word (e.g. the fragment CO_ _ _ _ could be completed as COFFIN [a death-related word] or as COFFee [a non-death word]). Demographic information was then assessed.

Discussion
Study 2 is consistent with the hypothesis that bringing to mind a personal (identified) hero serves as a buffer against the psychological threat of mortality. When participants were reminded of death and then wrote about a personal hero, participants filled in fewer word completions pertaining to death. This finding has the potential to inform the role of identification with heroes as mitigation of death thoughts did not result from writing about other people's heroes, or a personal acquaintance, but only from writing about one's personal hero.
However, one limitation of this study was the different levels of perceived heroism elicited by the writing prompt. Recall that participants saw other people's heroes as less heroic than their own heroes, and even less heroic than a mere acquaintance. It may be that having participants think of a person others saw as heroic, but that they did not, led them to try and justify this difference of opinion by indicating, on the manipulation check, their view that the person was not actually that heroic. It may also be the other people's hero condition elicited thoughts of someone participants considered a "false hero" whose ruse they had pierced e.g. a politician of an opposing political party. Although the mean rating of heroism in this condition was at the midpoint of the scale (M = 4.76) and thus suggests a moderate level of perceived heroism rather than the piercing of a ruse, the possibility is still open that the results do not reflect the influence of thinking about a hero with which one identifies, but rather could reflect thinking about someone who was simply seen as more (truly) heroic. Further, participants' own heroes more frequently referred to family members, and this may contribute to the observed effect on DTA given that close relationships can serve terror management functions (Mikulincer, Florian, & Hirschberger, 2003). To clear up such ambiguities, and probe whether hero identification was a necessary ingredient for the reduction of death-cognition, a task was needed that presents people with the same hero, but allows for variation in identification with the hero. This was the focus of Study 3.

Study 3
Although becker (1973), Sullivan and Venter (2005), and Study 1 suggest heroes can be bound up with the self, questions remain concerning whether identification is necessary for heroes to placate death thoughts. Study 2 presumed, based on prior research (Sullivan & Venter, 2005), that a personal hero was also an identified hero. Study 3 provides a more direct test of the role of hero identification and its capacity to mitigate death thoughts -that is, does the hero have to be "my hero" to reduce DTA? If so, how can a connection between the self and a hero be fostered? One way people feel connected to others is through shared personal information. Finch and Cialdini (1989) found when people ostensibly share a date of birth with a target person, they tend to evaluate that target person more positively; Study 3 borrowed this reasoning. Participants were reminded of death (or not), and exposed to a hero who shared a connection with the participant (in the form of shared date of birth), or not. DTA was then measured.
To expose all participants to the same hero, thus avoiding preexisting biases tied to people's extant heroes -a hero had to be constructed with which participants would (or would not) be led to identify. This necessitated a consideration of the core traits of heroes, with an eye toward identifying the facets of heroes that may protect against existential threat. As such, an additional focus of Study 3 was to probe the existentially valuable traits of heroes and the role they play in the potential placating of mortality-related cognition. While numerous characteristics and traits are associated with heroism (see e.g. Goethals & Allison, 2012;Kinsella, Ritchie, & Igou, 2015), becker highlighted the heroic characteristics of sacrifice and legacy, as often motivated by mortality concerns, and their importance in heroes' quelling such concerns. Accordingly, Study 3 focuses on these two aspects of heroes in relation to death cognition.
Sacrifice, which others have considered a trait prototypical of heroes (becker & eagly, 2004; Kinsella et al., 2015), may be existentially important because it involves navigating situations that require overcoming adversity, such as those that present potential harm, a defining feature of heroism (Stenstrom & Curtis, 2012), and because it is associated with confidence in the face of death (becker, 1973). Importantly, from a terror management perspective, one reason sacrifice may be existentially valuable is that it helps to advance the individual's legacy. Heroes often sacrifice to perpetuate a way of life, or collection of values, which proponents of that way of life or values will often then reward (e.g. soldiers who sacrifice their lives defending their country often gain a legacy through being memorialized, Durkheim, 1951;Lerner, 1991; recall also the Rosenblatt et al. [1989] study, in which, the "hero" was rewarded more money for helping apprehend the criminal).
Previous TMT work informs the link between death concerns, sacrifice, and legacy. Routledge and Arndt (2008) reminded british participants of death (vs. dental pain), before asking them to imagine working either for an organization that would perpetuate after their death, or, cease to exist once they had passed away. Participants then indicated willingness to physically self-sacrifice for england. Participants reminded of death and asked to imagine the scenario that did not provide symbolic immortality indicated greater willingness to sacrifice, presumably because it offered a death-transcending symbolic identity (or legacy; i.e. one's nation) that was not otherwise available. Recent work also informs the link between sacrifices in the defense of cultural values and perceptions of heroism. Schindler and Reinhard (2015) hypothesized edward Snowden, a whistleblower who revealed that the US Government was spying on its citizens, would be seen more heroically after a death reminder because he made a sacrifice (potential jail time/exile from his homeland) in order to defend a cultural value -truth. Findings revealed participants reminded of death did indeed perceive Snowden as more heroic than those not reminded of death. Collectively, these studies are consistent with becker's (1973) theorizing highlighting the integral role of sacrifice and legacy for heroism, and its potential terror management function.
Taken together, theory and research converge to suggest sacrifice and legacy achievement are bound up together and may be central to the existential function of heroes. However, questions arise concerning the necessity of the interconnectivity of sacrifice and legacy in terms of the identified hero's efficacy in reducing DTA. That is, do sacrifice and legacy individually, or in combination, have the capacity to reduce death-related thoughts when participants are led to identify with heroes who hold varying sets of such traits? Further, legacies can be more or less enduring. Some last for a short period, others for centuries. Therefore, Study 3 manipulates the traits of the hero with which participants will identify (or not) to test whether sacrifice and holding a legacy (of varying duration) is important in reducing DTA with the assumption being that longer legacies provide greater defense against mortality concerns than more fleeting ones.
In sum, for Study 3, we hypothesize that, after reminders of mortality, DTA will be lower when participants are led to identify with a hero through a shared connection to the self (i.e. date of birth). Thus, this study will address the concern in Study 2 regarding whether participants rated their own heroes as more heroic because they identified with them or because they were simply (in their eyes) more heroic than someone else's hero. However, given the ambiguity about whether the heroic traits of sacrifice and having a legacy confer existential security separately or in tandem, we advance no firm predictions on this question. Lower DTA may (or may not) occur when the hero is portrayed as separately sacrificing or having a legacy, or when displaying both of these attributes together.

Participants
One hundred sixty undergraduate students (age: M = 19.34, SD = 2.06; male = 70, female = 90) were randomly assigned to conditions in a 2 (salience: mortality vs. failure) X4 (heroic traits: legacy linked to self vs. sacrifice linked to self vs. legacy and sacrifice linked to self vs. Legacy and sacrifice together not linked to self ).We opted for only one (legacy and sacrifice) control condition in which the heroic portrayal was not linked to the self (i.e. through birthdate) because we reasoned that if these traits are existentially valuable (in the sense that they reduce DTA), then this would be most likely when they were presented in tandem. As such, a condition in which both legacy and sacrifice are highlighted but not linked to the participants represents the strongest control condition. Stated differently, the identification variable was not fully crossed with the other factors and only manipulated in the legacy and sacrifice condition. Table 4 clarifies the different features of the conditions.

Materials and procedure
Participants completed the MS/control manipulation similar to those in Study 1 and 2 with two changes. First, the distraction task utilized the expanded, 60-item version of the PANAS  (Watson & Clark, 1999). Second, the control topic was again changed, this time from pain to failure, to assess the potential unique effects of the psychological threat of mortality vs. threat generally ("briefly describe the emotions that the thought of failure arouses in you" and "Jot down, as specifically as you can, what you think happens to you physically when you experience failure"). Participants were then exposed to the hero manipulation before completing the DTA measure as in the previous study.

Hero exposure.
To expose participants to a hero with the traits of sacrifice, legacy, legacy, and sacrifice (connected to the self vs. not) participants read an ostensible excerpt from a newspaper (generated via http://www.fodey.com/generators/newspaper). The headline read either "Hero talks of sacrifice, " "Hero talks of legacy, " or "Hero talks of sacrifice and legacy. " The main body of text then described the acts of a hero (gender matched to participant) who rescued a baby from a burning car. Additional aspects that were varied included whether the hero was injured by flames during the rescue (sacrifice) or not and whether the rescue conferred a longer legacy, taking place three years ago (legacy) vs. fleeting legacy, taking place a week ago (see Table 4). For those in the legacy and sacrifice condition the hero was both injured and executed the rescue three years ago. For the conditions in which a connection with the hero was fostered, following Finch and Cialdini (1989), we had the hero share the participant's date of birth (obtained from a mass pretest completed at the beginning of the semester). For those in the no connection condition the hero's date of birth was not reported. After the article, participants completed the DTA and demographics measures.

Death thought accessibility
A 2 (MS vs. failure) X4 (heroic traits: legacy linked to self vs. sacrifice linked to self vs. legacy and sacrifice linked to self vs. legacy and sacrifice not linked to self ) ANOVA was conducted. This revealed a main effect for MS, F(1, 152) = 4.31, p = .039, 2 p = .03, such that after reminders of mortality, DTA was higher. There was also a main effect for heroic traits, F(3, 152) = 3.82, p = .011, 2 p = .07, such that participants exposed to the hero who both sacrificed and had a legacy but who had a different birthday had greater DTA. As depicted in Figure 3 and Table  5, the predicted 2-way interaction between salience and heroic traits also emerged, F(3, 152) = 3.69, p = .013, 2 p = .07. 6 After being reminded of mortality, when a link to the hero who both sacrificed and had a legacy was not fostered (i.e. participants did not share a birthdate), the accessibility of death-related thought was greater relative to the failure salience condition, t(38) = 3.27, p < .001, d = 1.06; this demonstrates that absent fostered identification with the hero, MS did indeed increase DTA. However, when identification with the hero was fostered via the shared birthdate, participants reminded of mortality no longer showed increased DTA. Not only did they not differ from their control condition counterparts (all pairwise ts < 1, ps > .318), but those with a link to the hero who sacrificed, t(38) = 3.01, p < .001, d = .97, had a longer legacy, t(38) = 3.72, p < .001, d = 1.21, and both sacrifice and longer legacy, t(38) = 3.72, p < .001, d = 1.21, had lower DTA than MS participants for whom a link to the hero was not fostered. Moreover, there were no differences between MS participants who were linked to a sacrificing hero, a hero with longer legacy, or sacrifice and legacy (all pairwise ts < 1, ps > .318), suggesting that each feature separately or together is sufficient to mitigate thoughts of mortality.

Discussion
Study 3 builds on the previous studies to suggest that after mortality reminders fostered hero identification functions to lower DTA. Further, the results suggest the heroic traits of sacrifice and legacy (separately, or in combination) may reduce mortality-related cognition, but importantly, this may only be possible to the extent that an individual is able to draw a connection between the self and the hero e.g. via shared date of birth. Finally, given the control topic was again shifted from pain to another threat i.e. failure, additional evidence is provided for the unique effects of death cognition.

General discussion
The present research utilized a terror management perspective to investigate the relationship between mortality concerns and hero identification. Three studies explored two questions about motivations and functions related to hero identification. First, does awareness of mortality motivate hero identification? Second, does such identification help to manage the cognitive activation of death-related thought? Study 1 found that when reminded of mortality, given a distraction task, and then primed with the self, participants indicated greater hero identification as demonstrated by more overlap between the self and a hero. These findings provide a tentative extension of existing research, noting heroes are integrated with the self (e.g. Sullivan & Venter, 2005), and also highlights a potential underlying motivation that can spur such inclusion. Further, the study found that hero identification was motivated specifically by non-conscious thoughts of mortality. Greater levels of inclusion of the hero in the self were not observed in response to the salience of uncertainty, or consistent with the dual process model of TMT, when participants still presumably had death-related thoughts in focal attention (Pyszczynski et al., 1999). As noted previously, the lack of a non-hero target leaves unanswered questions about the specificity of the effect to heroes.
Study 2 switched focus to the existential function of hero identification, specifically examining whether such identification facilitates the management of the psychological threat of mortality. After reminders of mortality and writing about a personal hero, participants reported lower DTA. Study 2 distinguished that the link between a hero and the self was important in placating mortality-concerns as the reduction in DTA only occurred when writing about one's own (identified) hero and not other's heroes, or an acquaintance. Thus, Study 2 helps to address some of the ambiguity from Study 1, albeit indirectly.
Study 3 continued the focus on the existential function of hero identification, fostering hero identification (or not), to probe the necessity of identification in reducing the accessibility of death-related thoughts. Further, the study extended the investigation to probe characteristics of the hero that may contribute to this functionality. Specifically, after reminders of mortality and reading about an ostensible hero who sacrificed and/or had a legacy, participants reported lower DTA, but only when they were led to identify with the hero via a shared personal detail (date of birth). It seems that for heroes to help thwart death-cognition, the hero has to be "my hero. " The study also informs becker's (1973) and others ' (e.g. see S. becker & eagly, 2004) theorizing that legacy and sacrifice are existentially valuable traits of heroes, as participants displayed lower DTA when they could draw a link between themselves and a hero who had these traits, either separately or in tandem.
Taken together these studies offer advances on four fronts for terror management and heroism research. First, Study 1 is the first to illustrate the motivational properties of death-cognition for hero identification. Second, Studies 2 and 3 highlight a hitherto unrecognized function of hero identification i.e. defense against mortality concerns. Third, Study 3 suggests characteristics of identified heroes that may be influential in their capacity to offer existential solace i.e. legacy and sacrifice. Fourth, these studies contribute to extant terror management literature demonstrating various routes to existential security via social identification. For example, research indicates MS increases women's perceived similarity to other women (Arndt, Greenberg, Schimel, Pyszczynski, & Solomon, 2002), greater optimism for Dutch participants' national soccer team to succeed (Dechesne, Greenberg, Arndt, & Schimel, 2000), and that Americans' worldview defense was mediated by identification with America (Hohman & Hog, 2015). The present research builds on these studies, uncovering a previously unrecognized identification pathway to quell death concerns -identification with our heroes.
Despite the conceptual contributions and insights offered by the present studies, they also raise issues that may merit future research. One concern with the present studies may relate to statistical power and sample size. Although, as noted, these studies followed conventions at the time they were conducted, in recent years there has been considerable discussion about the merits of large samples and issues of power. To justify the current sample sizes one could point to meta-analyses of MS effects which implicate a medium to large effect size, especially when longer delays are included in the experimental procedure as was done here (burke et al., 2010;Martens et al., 2011;Steinman & Updegraff, 2015), or one could appeal to effect sizes from previous research that has investigated, of present relevance, monetary rewards to heroes after MS (Rosenblatt et al., 1989) or different types of cultural identifications (e.g. Arndt et al., 2002). Power analyses (G*Power; Faul, erdfelder, buchner, & Lang, 2009) with such estimates (assuming r = .42 (d = .93), suggested approximately 20 participants per cell to detect effects of similar magnitude. However, we do not find these justifications compelling, as such effect sizes (and thus inferences about power and sample size) are as much if not more about the sensitivity of the particular laboratory measures and manipulations as they are about the strength of the underlying psychological process. More importantly, such effect sizes are ultimately of limited utility insofar as the present studies explored effects related to a novel psychological process (i.e. heroic identification) with novel manipulations and or measures and previous effect sizes are of questionable relevance. The field clearly has differing and passionate views on this broader issue, and thus from some contemporary perspectives, the present studies may be seen as underpowered and the present effects critiqued as unreliable. At this point, it may be most prudent to regard the present findings as existing within a tentative context of discovery and awaiting further research to establish robustness and reliability (baumeister, in press; Sakaluk, in press).
Secondly, although hero identification was measured in Study 1, there was no comparable measure to determine if MS leads to identification (inclusion of other in the self ) with anyone (vs. heroes specifically). Though we did not make the prediction that MS leads to hero identification uniquely (which would ignore previous research TMT on identification noted above), future research might examine whether MS results in greater or similar levels of identification with heroes vs. worldview-supporting non-heroes. As becker noted (1973), heroes are central to, and often embody worldview values. As such, heroes may be identified with to a greater degree after MS as they may wield greater "cultural power" compared to non-heroic cultural worldview supporters. Indeed, irrespective of whether MS can motivate a general tendency to identify with others or not, when considered alongside extant literature such as Sullivan and Venter (2005), finding quicker response rates on an IOS self-description test for hero (vs. non-hero others) traits, McCabe et al. (2015) finding lower reported pain in the hero (vs. positive personality) condition, and Kut et al. (2007) finding that hero identification via role-playing a hero (vs. control) led to greater pain tolerance, it is plausible to suggest that hero identification (vs. identification broadly) has some distinct psychological properties.
In relation to the above point, intriguing questions also remain concerning the nomological network of mortality cognition, hero identification, cultural worldview defense, and self-esteem. becker (1973) suggested that cultural significance provides feelings of selfworth, which protects an individual from overwhelming fear of death, and that heroism was the idealization of this process. Though the present research had people bring to mind their own hero, and fostered identification with a hero, consideration about how hero identification relates to cultural worldview defenses and implications for self-esteem remain to be studied. Locating the present research within the broader landscape of TMT literature may suggest one hypothesis. Consider previous TMT studies finding mortality reminders exacerbate religiosity for those partisan to different religions (Vail et al., 2010), enhance risky driving for those who gain self-esteem from being a fast driver (Taubman ben-Ari, Florian, & Mikulincer, 1999), and result in greater political affiliation (Greenberg & Kosloff, 2008). Future research might examine whether figures central to these value systems e.g. Jesus, Michael Schumacher, and political party leaders, are seen as more heroic, and are more closely integrated with the self, perhaps enhancing self-esteem, after mortality reminders.
Finally, it is unclear how terror management processes are implicated, if at all, in the bridge between hero identification and heroic enactment. While the present studies suggest mortality reminders can result in greater inclusion of the hero in the self, and prior research (McCabe et al., 2015) found death reminders can spur heroic enactment, little is known about whether existential concerns play a role in the potential shift from hero identification to heroic enactment. However, previous research, from beyond the purview of TMT, indicates there may be a link. yoon and Vargas (2014) found participants who first played as, and presumably identified with, a hero (vs. villain) in a videogame, in a later ostensibly unrelated study, allocated less unpleasant hot chili sauce and more chocolate to an apparent future participant. Though one might reasonably contest that offering chocolate to an unknown person does not constitute the most heroic gesture, this research does suggest hero identification may translate to enactments consistent with the identified hero. Juxtaposing this research with the current studies opens the intriguing possibility that mortality cognition may serve as the lubricant to accelerate the transition from identification to enactment. Indeed, as alluded to in Don Quixote's transition from merely consuming heroic tales to actually becoming the infamous hero knight-errant, it may be that hero identification precedes heroic enactment, and that terror management efforts can be an important part of this process. Using becker's (1973) theorizing and the terror management framework (Greenberg et al., 1986), three studies informed the bi-directional relationship between mortality concerns and hero identification. Two questions guided the exploration: Does awareness of mortality motivate hero identification? And, does such identification placate mortality concerns? Taken together, theoretical advances were made in how researchers might understand hero identification from an existential perspective, identifying a potential motivational catalyst underlying hero identification (awareness of death), highlighting important existential attributes of heroism (sacrifice and legacy) and how these attributes may aid in reducing thoughts of death. These studies thus help inform the motivational underpinning and function of hero identification.

Notes
1. because of unequal sample sizes, weighted means analyses were pursued, this did not significantly alter the patterns of results reported. 2. All α refer to Cronbach's alpha. 3. The effect of MS on fear reflects that after MS (vs. uncertainty) fear was greater. Notably, the IOS effect held controlling for fear. extant research suggests that while MS manipulations such as that used here do not generally increase negative affect (burke et al., 2010), it can produce effects on affect (both positive and negative) that are typically relatively small (Arndt, Allen, & Greenberg, 2001;burke et al., 2010). Further, recent research suggests that the affective consequences of MS may indeed lie more with fear than generalized negative affect, though these small increases in fear do not mediate effects on outcomes related to worldview defense (Lambert et al., 2014), and have not been found in a number of other studies. because this MS effect on fear did not replicate in Studies 2 and 3, we do not discuss it further. 4. It might expected that after reminders of death (vs. control), participants would perceive heroes as more heroic as a way to bolster their cultural worldview. That this was not exhibited may reflect a couple of factors. First, it may be that writing about one's hero was enough to mitigate existential threat, negating the need for further worldview bolstering. Second, because responses in the personal hero condition were near the top end of the scale, this might suggest a ceiling effect where there was simply not enough room to detect further increases in perceived heroism as a function of MS. 5. because the Levene's test indicated unequal error variance (p = .001), likely a function of the unequal cell sizes, we also ran the pairwise comparisons using Welch's t-tests. This did not alter the patterns of significance reported. Further, because of unequal cell sizes, weighted means analyses were also pursued. This also did not significantly alter the patterns reported. 6. A Levene's test indicated unequal error variance (p = .01). However, as this study had cells with an equal number of participants, the ANOVA is robust to violations of this assumption.