Abstract
Research in Terror Management Theory finds that close interpersonal relationships (e.g., parents, romantic partners) mitigate threat reactions to reminders of mortality. Parasocial relationships (imagined relationships with media personalities) afford many of the same benefits as interpersonal relationships. Do these benefits extend to mortality concerns? We investigated whether those with strong parasocial attachments were differentially influenced by reminders of death. Results showed that those with strong parasocial relationships had more defensive reactions to a mortality prime, suggesting that such attachments may not afford the same existential benefits given by close human others and may instead indicate a heightened vulnerability.
Why do individuals form such strong emotional bonds with celebrities, fictional characters, and other media personae? These one-sided or parasocial relationships serve a number of important psychological functions, including mitigating feelings of loneliness (Derrick et al., 2009), offering a lens for constructing personal identity (Boon & Lomore, 2001), and (perhaps counterintuitively) offering a source of additional emotional support that supplements interpersonal relationships (Giles & Maltby, 2004). In other words, parasocial relationships fulfill many of the important functions of interpersonal relationships.
The present paper considers a relatively unacknowledged possibility suggested by this similarity between interpersonal and parasocial relationships. Past research (reviewed below) finds that interpersonal relationships play an important role in defending the individual from concerns about death. However, if close relationships with human others serve this important existential function, and parasocial relationships offer many of the same psychological benefits as these relationships, we are left with an unanswered question: Can these media figures offer the same existential solace that people find in supportive human relationships? In this study, we tested whether parasocial attachment moderates the established effect of mortality salience in increasing resistance to the idea of human creatureliness, or the view that humans are merely biological beings (Goldenberg et al., 2001).
Terror Management and Close Relationships
Terror management theory (TMT) asserts that both overt and subtle awareness of the inevitability of death exerts a considerable influence in daily life (for a review, see Pyszczynski et al., 2015). Inspired by Becker’s (1973) work, the theory posits that people can assuage death anxiety by reassurances of literal immortality (e.g., belief in an afterlife) and, perhaps more commonly, by aspiring to symbolic immortality. This latter notion refers to the fact that individuals feel that some part of them will go on if they meaningfully contribute to a shared set of cultural beliefs and practices (i.e., a cultural worldview) that persists after they are gone.
To date, decades of research have provided support for this broad hypothesis (for a comprehensive meta-analysis, see Burke et al., 2010). Typically, these studies employ a mortality salience (MS) induction designed to experimentally remind individuals of their own death prior to assessing reactions. MS studies commonly measure related variables after the manipulation such as death-thought accessibility (DTA), the amount of salient death-related thoughts a person experiences, and worldview defense, individuals’ desire to defend their worldviews after being threatened with mortality (Burke et al., 2010).
Research shows that MS increases efforts at immortality seeking in a variety of ways. For instance, MS increases the desire to have children (e.g., Fritsche et al., 2013) and to be famous (e.g., Greenberg et al., 2010) ostensibly out of that desire to perceive each as an opportunity for founding some enduring legacy that can mitigate death anxiety. Immortality need not always be so direct: Other research shows that MS increases people’s desire to defend cultural worldviews that afford symbolic immortality. For example, MS increases bias against individuals of other religions (e.g., Greenberg et al., 1990) and causes a marked increase in nationalism (Castano, 2004; Jonas et al., 2005).
The effects of MS can be attenuated in ways that further support the core thesis of TMT. For example, MS effects are diminished among individuals with a particularly high self-esteem (Harmon-Jones et al., 1997). According to the theory, this is because self-esteem is a product of accepting a cultural worldview, which might afford symbolic immortality, and seeing oneself as particularly successful in living up to the standards of that culture (Pyszczynski et al., 2004).
Self-esteem is not, however, the only thing that operates as an effective shield against the threat of death. Other research on the anxiety-buffer hypothesis tests whether specific features of the individual or their environment mitigate the psychological consequences of death anxiety (Florian et al., 2002). If so, then strengthening that psychological structure would be expected to make a person less likely to display death anxiety or anxiety-driven behaviors like expressing greater hostility toward members of other cultures.
Interpersonal relationships are thought to be one such existential anxiety buffer. Most of the research on this approach has taken the perspective of attachment theory (Bowlby, 1973; Mikulincer & Shaver, 2007), which proposes that individuals have an innate need for the security and support provided by close others. According to the theory, some individuals have a history of seeking and receiving support and that experience provides a degree of psychological resistance to stress and threat and some researchers have extended that model to include anxieties about mortality. In other words, individuals who have had supportive relationships in the past develop the expectation that others will support them, what the theory calls a secure attachment style, and this perception provides reassurance when the individual experiences distress.
In contrast, other individuals who experience neglect or inconsistent support develop persistent insecurities. According to attachment theory, individuals who experience chronic neglect in their relationships develop attachment avoidance, a deep mistrust of others and an unwillingness to pursue intimacy. People who instead receive inconsistent or unreliable support form a different kind of insecure attachment characterized by persistent uncertainty about relationship partners. This attachment style is referred to as attachment anxiety and it is marked by lingering fears of abandonment and jealousy in response to the anxieties that an individual has about their relationship partners.
Initial research employing attachment theory demonstrates that these personality differences in relationship style influence susceptibility to reminders of mortality. In one initial study, researchers found that individuals with more secure attachment (i.e., individuals who trust and feel comfortable depending on their partners) did not respond to MS with greater punishment of a violation of the norms of one’s cultural worldview (Mikulincer & Florian, 2000). In contrast, those who generally expect relationship partners to neglect them (i.e., an avoidant style) and those who have persistent doubts about their relationships (i.e., an anxious style) showed characteristic patterns of worldview defense following and MS prime (Mikulincer & Florian, 2000).
This finding that close relationships serve as an anxiety buffer against MS has been shown across different types of relationships. One series of studies found that MS elicited standard worldview defense unless individuals were reminded of their parents (Cox et al., 2008). Following the earlier work on attachment style, these effects were limited specifically to those individuals who had generally positive and supportive expectations about their parents. In addition to the security provided by these reminders of parents, other research has looked to the moderating effects of positive reminders of romantic partners (Cox & Arndt, 2012). When exposed to MS, people are more likely to overstate how positively their romantic partners view them and reminders of positive interactions with partners attenuated any effects of death anxiety. Importantly, not all positive relationships reduced thoughts of death; positive thoughts of parents and romantic partners have been shown to reduce the salience of death after MS, while positive thoughts of acquaintances do not, likely because people do not identify as closely with them and do not see them as a dependable long-term relationship (Cox et al., 2008; Cox & Arndt, 2012).
These studies, among many others, support the conclusion that relationships play a central role in the management of death concerns. In one initial model (Hart et al., 2005) proposed that secure interpersonal relationships represent one of three pillars in existential defense (the others being self-esteem and worldview defense). In short, individuals who achieve a sense of security through one of those sources seem ostensibly less motivated to pursue security through the others in response to existential threat. In the most comprehensive review of research on this topic to date, there is overwhelming support for the psychologically protective benefits of secure close relationships across 36 studies testing whether close relationships buffer death anxiety (Plusnin et al., 2018). Close relationships can serve as a source of attachment security by providing a sense that one is loved and supported, can provide self-esteem through the experience of personal value in the relational context, and can even provide individuals with worldview defense by reaffirming central beliefs that play a role in mitigating death concerns.
In sum, research on TMT shows that fear of death motivates a number of defensive processes and that these effects can be buffered by supportive positive relationships. Could one receive the same psychological benefit from reminders of psychologically important, but physically distant or even fictional, others?
Parasocial Relationships
Parasocial relationships are quasi-social bonds between a viewer and an actor (Horton & Wohl, 1956), and other relationships that form between viewers and media personae (Klimmt et al., 2006). Viewers sometimes establish a feeling of intimacy toward the parasocial target, which motivates them to invest effort into spending time with target media figures (Rubin et al., 1985). These relationships may seem superficial, but the loss of a television show elicits similar emotions to those experienced during a romantic breakup (Eyal & Cohen, 2006), indicating that people perceive these relationships as something deeper.
Parasocial relationships are a diffuse, although often unacknowledged feature of media consumption. For example, studies suggest that parasocial liking can occur in even short-term interactions with a novel media figure (Dibble et al., 2016). Additionally, studies show that online fan cultures promote parasocial relationships through diverse factors, including the similarity and attractiveness of the media figure and how entertaining they are (Lou & Kim, 2019).
In addition to these factors, the formation of parasocial bonds is sometimes motivated by the perception that one’s interpersonal relationships are insufficiently supportive (Keefer et al., 2014; Rosaen & Dibble, 2016). For example, anxiously attached individuals, who experience persistent fears of interpersonal abandonment, have been found to be the most likely to develop a parasocial bond with a media figure (Cole & Leets, 1999). Additionally, anxiously attached individuals experienced more negative reactions after thinking about their favorite character being taken off the air (Cohen, 2004). Ostensibly, the role of attachment anxiety here is driven by the fact that parasocial relationships can and do provide individuals with a sense of social connection that effectively reduces feelings of loneliness (Derrick et al., 2009).
Do these relationships afford other benefits for the individual? One series of studies found that individuals with low self-esteem who were reminded of a beloved (vs. neutral) celebrity subsequently reported lower discrepancies between their current and ideal selves (Derrick et al., 2008). This boost was so strong that self-discrepancy scores were equivalent for both low- and high self-esteem participants writing about a favorite celebrity (but differed significantly for the control group). More targeted research shows that images of ideal bodies (e.g., of celebrities, superheroes) decreased personal body esteem, but only for those individuals without a parasocial attachment to the target (Young et al., 2012, 2013). For individuals who had a bond with the idealized figure (e.g., a muscular superhero), that image ironically increased body esteem and even performance in a physical strength task (Young et al., 2013).
These self-esteem findings raise important questions about the role of parasocial relationships in terror management processes. As noted above, supportive interpersonal relationships buffer the effects of mortality salience. Increased levels of self-esteem have also been shown to reduce effects of mortality salience as higher levels of self-esteem imply the individual is living up to cultural ideals that can provide symbolic immortality. This led Plusnin et al. (2018) to propose that self-esteem is one specific route through which relationships offer individuals protection against concerns about death. Parasocial relationships have been shown to both provide feelings of social connection as well as enhance perceived self-worth, two established buffers of existential anxiety. It follows that parasocial relationships may therefore operate as an anxiety buffer against death awareness as well.
The Current Study
To date, a single study has been conducted to show that parasocial relationships are affected by terror management processes. Semmler (2016) found that death-thought accessibility predicted higher levels of parasocial attachment (after controlling for mood, interpersonal attachment, and sex). These results echo prior research showing that reminders of death increase individuals’ desire for stronger interpersonal relationships (Florian et al., 2002) and suggest a psychological parallel.
The current study sought to expand on these previous interpersonal findings by testing the hypothesis that stronger parasocial attachments might buffer the effects of mortality salience. As noted above, these one-sided relationships still provide feelings of social connection and bolster self-esteem, two established processes by which media personae could mitigate fears of death. We will refer to this as the parallel hypothesis, the expectation that parasocial attachment will yield similar buffering as that observed with interpersonal relationships.
To test this broad hypothesis, we first assessed dispositional attachment toward a favorite parasocial figure (e.g., character, superhero, celebrity). Participants were then randomly assigned to either a mortality salience induction or a matched negative affect control used in prior research. Following this priming, participants completed a delay task (shown to enhance MS effects; Burke et al., 2010) before completing our dependent measures assessing worldview defense and death-thought accessibility.
Method
Participants
We collected data from 128 undergraduate students at the primary investigator’s institution (M age = 19.94, SD age = 3.01). The sample consisted of primarily women (73% Women, 25% Men, 2% Other Gender) and was majority Black (51% Black or African-American, 41% White or Caucasian, 4% Asian or Pacific Islander, 2% Hispanic or Latinx, 2% Multiracial). Students participated in a study described as research on media and personality.
Although established effect sizes for the death anxiety-buffering effects of close relationships are extremely large (d = .91; Cox & Arndt, 2012), we considered it unlikely that parasocial relationships would have equally powerful effects. In the absence of a more appropriate a priori effect size estimate, we aimed to collect a sample sufficient to detect a moderate (partial R2 of .07) interaction between parasocial attachment and MS priming condition in a regression model. This yielded a target sample of 107 in a G*Power analysis (Faul et al., 2009).
Data, materials, and analysis syntax are openly available at: https://osf.io/kdb7e/
Materials and Procedure
After providing informed consent and general demographic information (see Supplemental Materials on OSF for specific items), participants were asked to complete the following materials in a fixed order, with DTA always measured last. Recent research, including a pre-registered replication, finds that completion of the DTA measure increases explicit death-thought accessibility and thereby eliminates any effects of MS priming (Hayes & Schimel, 2018). Ostensibly this occurs because MS effects are strongest after a short delay, when death concerns are not directly salient (Burke et al., 2010) and completion of the DTA measure alone brings death directly to mind again. Participants also completed the Rosenberg Self-Esteem scale (Rosenberg, 1965) and these data are available on OSF. We exclude them from our discussion here to focus on our primary hypothesis but note that controlling for self-esteem has no impact on our reported results (see OSF for full reporting of primary analysis after controlling for self-esteem, as well as syntax for reproducing this analysis).
Parasocial Relationships
To assess individual differences in the strength of individuals’ parasocial relationships, we first asked participants to complete a subset of items from the PSR scale (Tukachinsky, 2010) for their favorite media persona. This scale assesses the extent to which individuals perceive a parasocial target as either a friend or a romantic partner with separate subscales for each.
We included only the 13-items of the measure that assess para-friendship to avoid the possibility that some participants may be unable to report romantic bonds with a parasocial target. TMT studies on close relationships have shown buffering effects for both romantic and parental relationships. This suggests that romantic feelings toward a target may be sufficient, but are not necessary, for an anxiety buffer. Since the PSR romantic items showed the lowest endorsement in previous research (Tukachinsky, 2010, Study 3), we decided that the PSR friendship items better reflect perceptions of intimacy and support more akin to standard measures of interpersonal attachment and thus present an appropriate assessment of our moderator.
Responses on this and all other self-report measures were made using a 7-point Likert-type scale (1 = Strongly Disagree; 7 = Strongly Agree) and ratings of the 13 PSR items were averaged to compute a composite score (α = .92, M = 5.45, SD = 1.22).
Mortality Salience Condition
Following previous research on MS (Greenberg et al., 1994), we then randomly assigned participants to one of two priming conditions. In the MS condition, participants were asked to write two open-ended essays about their reactions to their own death (see OSF Supplement Material for specific wording). In the control condition, participants received a similar prompt about the experience of dental pain (a common negative affect control condition in TMT research; Burke et al., 2010). After the written task, both groups completed a 20-item Positive and Negative Affect Schedule (PANAS; Watson & Clark, 1994) as a distractor before moving on to the dependent measures. This specific distractor has been used in past research on the anxiety-buffer hypothesis (Cox & Arndt, 2012).
DV 1: Worldview Defense
Following previous research (Goldenberg et al., 2001) we then presented participants with two essays in a randomized order. In the threatening essay, the writer expresses the view that humans are merely biological organisms without special significance. In the control essay, the author argues that humans are in fact uniquely valuable. Past studies have found that under MS, participants defensively reject the view that humans are merely animals, ostensibly because perceptions of human significance afford the opportunity for literal (e.g., a soul) or symbolic (cultural) immortality through one’s cultural worldview.
After reading each essay, participants rated (1 = Not at all, 7 = Very Much) their response to five questions about each author (e.g., “How intelligent did you think this person was?”). As in past research, these face-valid reaction items formed reliable composites for both the Threat (α = .89, M = 4.95, SD =1.36) and Control (α = .89, M = 5.48, SD = 1.16) essays. Additionally, we followed past research (Goldenberg et al., 2001) by calculating a difference score between the two essay ratings (Control Essay—Threat Essay) with higher scores indicating a greater preference for the control essay (i.e., worldview defense; M = .52, SD = 1.83).
DV 2: Death-thought Accessibility
Death-thought accessibility (DTA), the extent to which death is implicitly cued for participants, was measured by having participants complete 25 word fragments with the first word that comes to mind (for a review of this measure see Hayes et al., 2010). Six of the 25 fragments could be completed to form either neutral terms or death-related terms (e.g., coffee/coffin, grave/grand). The possible death-related words are buried, dead, grave, killed, skull, and coffin. Following standard measurement, each participants’ DTA score was determined by adding the total number of death (vs. neutral) words generated (M = 2.14, SD = 0.98).
Results
Given that there was considerable variability in our primary outcome measures (unadjusted Threat Essay range: 1–7; Control essay range: 1.6–7; Difference score range: −5.4 to 6), we first proceeded by Winsorizing all continuous variables at a 95% interval. This procedure replaces the 5% most extreme low/high values in each variable by the closest observed score within the 95% interval around the mean. This effectively reduces the effect of outliers on the analysis without sacrificing statistical power by dropping or excluding cases (which severely biases statistical analysis; Bakker & Wicherts, 2014). Statisticians argue that Winsorizing offers a valuable technique for reducing bias (Erceg-Hurn & Mirosevich, 2008), particularly when testing correlational analyses (Wilcox, 2016). Readers interested in reviewing the analysis without outlier management can find full reporting on OSF in an additional Supplemental Material, although we note that the pattern of results is identical albeit slightly weaker due to additional variance.
Regression analysis was employed throughout to test the interactive effects of experimental condition and initial levels of parasocial attachment. There was no evidence that any outcomes were substantially skewed (|all skew values| < .34) and kurtosis was acceptably low (|all kurtosis values| < 1.1). For the two significant models below, there was no significant evidence that data violated the normality assumption (both ps > .102 in a Shapiro–Wilk test).
Worldview Defense
To analyze worldview defense, we conducted a series of three regression models predicting essay evaluations on the basis of condition (dummy-coded: Control = 0/MS = 1), parasocial attachment, and their predicted interaction, additionally controlling for the order in which the essays were presented. Specifically, we first computed a difference score (Control Essay—Threat Essay) to assess participants’ relative preferences for the control (vs. threat essay); additionally, we analyzed each outcome separately to determine which essay specifically showed effects as a function of condition and parasocial attachment.
Regression models predicting worldview defense, essay ratings, and DTA as a function of predictors.
Note. MS Condition dummy-coded with 1 = MS/0 = Control prime.
Probing the observed interaction on worldview defense (Figure 1) indicated several important findings. At low (−1 SD) levels of parasocial attachment, there was no observed effect of the mortality salience manipulation (b = −.13, SE = 0.39, t = 0.34, p = .74). Although conditions indicated directionally more worldview defense under MS (vs. control) at mean levels of parasocial attachment, this difference was also not significant (b = 0.43, SE = 0.28, t = 1.54, p = .13). However, relative to the control, at high (+1 SD) levels of parasocial attachment, there was a large effect of MS in increasing worldview defense (b = 0.99, SE = 0.39, t = 2.51, p = .013). Within levels of condition, simple slopes analysis indicated no relationship between parasocial attachment and worldview defense for the MS group (b = 0.19, SE = 0.20, t = 0.96, p = .34) but a negative slope of attachment in the control group (b = −0.37, SE = 0.19, t = 1.92, p = .057). In other words, individuals with stronger parasocial attachments reported less defensive scores in the control group, but this attachment had no bearing on defensiveness under MS. Worldview defense as a function of condition and parasocial attachment. Note. Score represents difference between control and threat essay ratings with higher scores representing a greater pro-control/anti-threat position. Scores ranged from −1.80 to 3.95.
Next we probed the second interaction on threat essay ratings specifically (Figure 2) with patterns closely matching those for the difference score. Again there was no effect of the MS manipulation at low (−1 SD) levels of parasocial attachment (b = 0.11, SE = 0.31, t = 0.37, p = .72) nor at mean levels (b = −0.34, SE = 0.22, t = 1.55, p = .123), although the pattern indicated a trend toward more critical evaluations of the threat essay following MS. As with worldview defense generally, we saw a clear effect of condition among those highly attached to a parasocial figure (b = −0.79, SE = 0.31, t = 2.56, p = .011) indicating that MS (vs. control) resulted in more negative views of the threat essay. Within conditions, simple slopes analysis again showed no relationship between parasocial attachment and threat essay ratings in the MS condition (b = 0.04, SE = 0.16, t = 0.24, p = .81) but did indicate an association in the control group (b = 0.49, SE = 0.15, t = 3.27, p = .001). Specifically, individuals with stronger parasocial attachments reported more favorable evaluations of the threat essay in the control group; we return to this point below. Essay ratings as a function of condition and parasocial attachment. Note. Scale ranged from 1 to 7 with higher scores indicating a more favorable essay rating. Scores ranged from 2.25 to 7 (threat essay) and 3.60 to 7 (control essay).
Death-Thought Accessibility
Unexpectedly, we observed no main effects or interactions on death-thought accessibility. In the current sample, DTA was uncorrelated with parasocial attachment (r = −.02, p = .78) and control essay ratings (r = −.11, p = .21). Nevertheless, DTA correlated negatively with worldview defense (i.e., difference scores; r = −.24, p = .008) and positively with Threat Essay ratings (r = .22, p = .01). These relationships indicated that more defensive individuals reported lower levels of DTA. Although correlational, these findings suggest that the opportunity to affirm a worldview by criticizing the threat essay specifically may have subsequently reduced DTA levels, obscuring the effects of our manipulation.
Discussion
The current study tested a parallel hypothesis about the potential for parasocial attachments to buffer mortality concerns. Past research suggests that these relationships provide many key benefits of actual relationships, including providing feelings of social connection and self-worth.
Although we expected to find support for that prediction, the data unequivocally supported an alternative interpretation. Specifically, individuals with a strong parasocial attachment were instead particularly threatened by a mortality salience induction and these results were evident in their reactions to the threatening creatureliness essay. In contrast, our low parasocial attachment individuals showed little to no defensiveness difference between the pain and MS conditions.
Instead of the predicted parallel hypothesis, the data instead suggest that parasocial attachment is a marker of increased vulnerability. Research consistently demonstrates that individuals who form deep parasocial attachments are also particularly psychologically vulnerable. For example, people who form strong parasocial attachments tend to be lonelier (Greenwood & Long, 2009) and to have more insecure interpersonal attachments characterized by fears of abandonment and mistrust (Cohen, 2004). Additionally, there is evidence that individuals tend to use parasocial relationships with beloved characters as a means of compensating for insecure interpersonal relationships (Rosaen & Dibble, 2016) and a means of fulfilling needs unmet by one’s social milieu (Madison et al., 2016). This raises the possibility that those with strong parasocial attachments may be more sensitive and reactive to salient morality concerns. However, given the fact that parasocial relationships can be formed for many other reasons, this interpretation relies on strong assumptions about the motives for parasocial attachment in the sample that are difficult to answer without more information about participants’ interpersonal attachments.
In the absence of more information about these highly parasocially attached individuals, another parsimonious explanation for this observation is that individuals who were high in parasocial attachment may simply be more influenced by the experience of imagining their own death. Parasocial relationships rely in part on creative and imaginative abilities that allow the individual to feel connected to a distant media figure (Brown, 2015). Therefore, individuals high in parasocial attachment may be particularly imaginative and/or capable of more vivid fantasies that could explain the greater impact of imagining one’s death among those high in parasocial attachment.
There was however limited evidence for a benefit associated with parasocial attachment. We observed significantly lower defensiveness about the creatureliness essay among highly attached individuals in the control group and, again, this reaction was specific to the existentially threatening essay that diminished human significance. Thus, it may be that strong parasocial attachments do provide existential benefits under low threat conditions, but that these do not stand up to the threat of direct morality salience as our results show.
The null results we observed were only partially expected. Specifically, we had no expectation that parasocial attachment or MS prime would influence reactions to the control essay espousing human uniqueness. Results confirmed that pattern and provide compelling evidence for the specificity of the effects on the threatening creatureliness essay. In short, effects were specific to the threatening (v. control) essay, indicating that MS elicited specific resistance to the idea that humans are merely animals, but not a general reaction to philosophical reflections on human nature, etc. Analysis of DTA scores, unfortunately, was somewhat inconclusive beyond demonstrating that more defensive individuals demonstrated reduced DTA on the latter task. This may be an artifact of our design and perhaps a larger sample could counterbalance the order of the DVs to determine whether the effects we observed on threat essay ratings would also be observed for DTA if it were measured directly.
Limitations and Future Directions
One notable limitation of the current study is that it depends exclusively on participants’ pre-existing parasocial attachments. As a result, it leaves many questions about whether and when individuals could more strategically employ parasocial relationships to manage existential concerns. These relationships can provide a sense of self-esteem (e.g., Derrick et al., 2008), so there is still potentially good reason to consider them a potential refuge from at least some existential concerns. More complex experimental designs may be necessary to more clearly see the benefits of these relationships.
Additionally, it is noteworthy that the sample was a convenience sample of university students. The age of participants might be of particular importance. According to socioemotional selectivity theory, as individuals get older, they typically opt to reduce their social network to a few close relationships, instead of many relationships that are not as close (Lansford et al., 1998). These findings have been found to extend to online social networks. As individuals get older, they typically have less friends on social media, but these friends are more likely to be real-life friends (Chang et al., 2015). Accordingly, it may be that the younger sample was both more likely to form parasocial bonds and may be less likely to have salient anxieties about mortality. Older samples who may have more death anxiety and smaller social networks may respond quite differently. For instance, it is possible that parasocial attachments are more beneficial for older adults who have reduced social networks but more research is necessary to explore this point.
This sample also leaves open questions about the role of gender. Although we observed no effects of participant or target gender in the supplemental analysis (available on OSF), our sample was overwhelmingly women and they primarily chose women as parasocial targets. More diversity in gender representation is important as previous studies have found that gender moderates the effects of parasocial interaction (Wang et al., 2008).
This convenience sample also limits the effective range of observed parasocial attachment scores and additionally limits our ability to draw inferences about more extreme cases. Research shows that mortality concerns can exacerbate social anxiety (Finch et al., 2016) so it may be that individuals with extreme social anxiety may find comparatively safer parasocial relationships to be a vehicle for some degree of self-worth. Future research with more diverse populations may tell us considerably more about the existential uses of parasocial relationships.
Another crucial limitation of the current study is that other individual differences that could explain our findings were not measured. First, we did not collect data on creativity, imaginative ability, or anything else that might allow us to see whether these differences among those high in parasocial attachment can account for the findings. Centrally, we also did not measure interpersonal attachment style. It is possible that attachment style in a relational context may play an important role in shaping how parasocial attachment operates in this context; for example, securely attached individuals may receive enough support from close relationships that parasocial attachment is no longer a particular risk factor in the context of mortality salience. Alternatively, as noted above it may be that parasocial attachment was only a risk factor because it was a proxy for insecure attachment, loneliness, or some other metric of insufficient social relatedness. In short, more information about the personalities of our participants would be necessary to understand the specific processes that explain the current findings.
On the same note, the current study only measured individuals’ attachment toward one parasocial figure. Future work might want to ask participants to rate their attachment to multiple parasocial figures as attachment to more parasocial figures might act as a stronger anxiety buffer than attachment to a single parasocial figure. Furthermore, we are unable to definitively state that individuals’ parasocial relationships are distinct from their real-life interpersonal social networks as we did not measure individuals’ interpersonal social networks and compare that to their parasocial networks. Future research may seek to further examine potential differences between real-life social networks and parasocial networks.
Finally, we allowed participants to freely select parasocial targets based on personal significance. Although this has the advantage of ensuring that individuals are consistently thinking of personally relevant targets, it introduces considerable noise to the design. Future research with tighter experimental controls may be able to provide clearer data about the role of parasocial relationships in this context.
Conclusion
In this paper, we tested the possibility that parasocial attachments might offer a parallel buffer for mortality concerns akin to interpersonal relationships. In contrast, the data supported the opposite pattern, that in fact individuals with strong parasocial attachments were particularly defensive following a mortality salience induction. These results speak to the potential psychological vulnerability of individuals who rely on social proxies to fulfill psychological needs for belonging. Mortality concerns are thought to underlie a wide range of psychopathologies (e.g., Iverach et al., 2014) and the current data suggest that media offers only an ineffective replacement for meaningful social connections, although the data clearly point to a need for further investigation to understand the specific processes that account for the current findings.
Supplemental Material
sj-pdf-1-ome-10.1177_00302228221085173 – Supplemental Material for A Distant Ally?: Mortality Salience and Parasocial Attachment
Supplemental Material, sj-pdf-1-ome-10.1177_00302228221085173 for A Distant Ally?: Mortality Salience and Parasocial Attachment by Lucas A. Keefer, Faith L. Brown, Zachary K. Rothschild, and Kaitlyn Allen in OMEGA-Journal of Death and Dying
Supplemental Material
sj-pdf-2-ome-10.1177_00302228221085173 – Supplemental Material for A Distant Ally?: Mortality Salience and Parasocial Attachment
Supplemental Material, sj-pdf-2-ome-10.1177_00302228221085173 for A Distant Ally?: Mortality Salience and Parasocial Attachment by Lucas A. Keefer, Faith L. Brown, Zachary K. Rothschild, and Kaitlyn Allen in OMEGA-Journal of Death and Dying
Supplemental Material
sj-pdf-3-ome-10.1177_00302228221085173 – Supplemental Material for A Distant Ally?: Mortality Salience and Parasocial Attachment
Supplemental Material, sj-pdf-3-ome-10.1177_00302228221085173 for A Distant Ally?: Mortality Salience and Parasocial Attachment by Lucas A. Keefer, Faith L. Brown, Zachary K. Rothschild, and Kaitlyn Allen in OMEGA-Journal of Death and Dying
Supplemental Material
sj-csv-1-ome-10.1177_00302228221085173 – Supplemental Material for A Distant Ally?: Mortality Salience and Parasocial Attachment
Supplemental Material, sj-csv-1-ome-10.1177_00302228221085173 for A Distant Ally?: Mortality Salience and Parasocial Attachment by Lucas A. Keefer, Faith L. Brown, Zachary K. Rothschild, and Kaitlyn Allen in OMEGA-Journal of Death and Dying
Supplemental Material
sj-r-1-ome-10.1177_00302228221085173 – Supplemental Material for A Distant Ally?: Mortality Salience and Parasocial Attachment
Supplemental Material, sj-r-1-ome-10.1177_00302228221085173 for A Distant Ally?: Mortality Salience and Parasocial Attachment by Lucas A. Keefer, Faith L. Brown, Zachary K. Rothschild, and Kaitlyn Allen in OMEGA-Journal of Death and Dying
Footnotes
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) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
