The role of death fears and attachment processes in social anxiety: a novel hypothesis explored

ABSTRACT

Objective: Research suggests an association between the fear of death and social anxiety, but the mechanisms through which these constructs are related remain unclear. From a socio-evolutionary perspective, abandonment and rejection are associated with premature death, and appraisals of the importance of social relationships for maintaining survival are influenced by one’s upbringing, reflected in individual differences in attachment patterns. We thus explored whether different attachment dimensions mediated the relationship between death anxiety and social anxiety to differing degrees, according to the value afforded by these dimensions to the importance of relationships for survival. Method: Self-report data on death anxiety, social anxiety, and two dimensions of attachment (anxiety and avoidance) was collected from 93 participants with varying levels of social anxiety. Results: Supporting the hypotheses, there was an indirect effect of death anxiety on social anxiety via attachment anxiety, even when fear of own death and fear of other’s death were analysed separately. No indirect effect via attachment avoidance was observed. Conclusions: These results clarify the mechanisms underpinning the association between social anxiety and the fear of death, and suggest that a socio-evolutionary framework provides utility in better understanding these clinical constructs.

KEY POINTS

What is already known about this topic:

(1) Research demonstrates a hitherto unclear association between the experience of social anxiety and underlying death fears.

(2) A socio-evolutionary approach highlights the intimate relationship between abandonment/rejection and premature death during our evolutionary history.

(3) Appraisals of the importance of social relationships for maintaining survival ought to be influenced by developmental experiences.

What this topic adds:

(1) We propose that variation in these appraisals are reflected in the different attachment dimensions (anxiety and avoidance).

(2) In support of this, we found that attachment anxiety mediated the relationship between death anxiety and social anxiety, whereas attachment avoidance did not.

(3) These results clarify the mechanisms underpinning the connection between social anxiety and underlying death fears and emphasise the utility of socio-evolutionary approaches to clinical syndromes.

KEYWORDS:

• Social anxiety
• death anxiety
• fear of death
• attachment
• attachment anxiety
• attachment avoidance

Introduction

A growing body of empirical evidence has recognized the important role of the transdiagnostic construct of death anxiety in psychological dysfunction (Iverach et al., 2014). In particular, the role of death fears appear to be most salient in existing understandings of anxiety-related disorders. For example, it is well-established that fears of death feature prominently in health anxiety (formerly, hypochondriasis) (Menzies et al., 2021; Noyes et al., 2002) and psychological treatment targeted towards addressing death fears has been shown to reduce hypochondriacal symptoms (Hiebert et al., 2005). Likewise, fears of dying due to cardiac arrest or suffocation are a central feature in Panic Disorder (Diagnostic and Statistical Manual of mental disorders [DSM-5]; American Psychiatric Association, 2013), with empirical investigation verifying that individuals with this diagnosis have elevated levels of death anxiety (Furer et al., 1997). The theme of death anxiety is similarly apparent in many specific phobias, in which the focus of anxiety is palpably related to survival threats (e.g., snakes, spiders, heights) (Marks, 1987; Strachan et al., 2007). The strong connection between death anxiety and Obsessive Compulsive Disorder (OCD) is also well established, with death reminders having been demonstrated to exacerbate compulsive symptoms (Menzies & Dar-Nimrod, 2017; Strachan et al., 2007). Moreover, across anxiety disorders generally, greater levels of death fears have been shown to be broadly associated with increased severity of dysfunction, distress/impairment, and intensity of intervention (Menzies et al., 2019).

Perhaps more interesting is the role that underlying concerns about death may play in Social Anxiety Disorder (SAD), in which the focus of anxiety (negative social evaluation and interpersonal rejection) appears seemingly unrelated to the theme of mortality. Even so, recent research suggests that there does appear to be an evident relationship between the two ostensibly unrelated anxiety domains. Self-report data reveals an apparent association, with socially anxious individuals exhibiting elevated levels of death anxiety (Furer et al., 1997; Menzies et al., 2019). Further, Strachan et al. (2007) and Finch et al. (2016) demonstrated that individuals with high levels of social anxiety displayed significantly greater socially anxious behaviour when reminded about death. Subsequent research by Zuccala and Abbott (2021) sought to address the limitations of these prior studies by determining diagnostic status via clinical interview, in order to eliminate potential contamination of comorbid anxiety related to physical threat. The pattern of results failed to support Terror Management Theory, the currently prevailing model for understanding the role of death anxiety in psychopathology (Iverach et al., 2014), such that death reminders did not exacerbate social anxiety symptoms for individuals diagnosed with SAD. Instead, symptoms were exacerbated for these individuals on a task designed to elicit physical anxiety. Nevertheless, the overall findings of Zuccala and Abbott (2021) did suggest that death fears are an important process in social anxiety, albeit not by the mechanisms purported by Terror Management Theory. Thus, as the literature currently stands, it is evident that the mechanisms through which these constructs are related remains relatively unclear, necessitating new theoretical frameworks to explain the role of mortality concerns in SAD.

A socio-evolutionary perspective may provide a theoretically coherent and intuitively comprehensible framework for understanding the previously unclear connection between social fears and death anxiety, potentially offering greater utility towards understanding this relationship than currently employed paradigms. At its core, the theory of natural selection (Darwin, 1859) inherently addresses the systems that organisms have evolved to minimize the risk of premature death. The psychological construct of death anxiety can be considered the conscious representation of this elementary process – namely, how humans respond to (and emotionally manage) physical and psychological threats that may increase the likelihood of death (Landau et al., 2007).

Life History Theory (LHT; see Del Giudice et al., 2016) and attachment theory (Bowlby, 1969/1982) are both derivatives of natural selection theory, and have been equally employed to understand the significant role of evolutionary constructs in psychopathology (for example, see Del Giudice, 2018). These complementary theoretical frameworks both highlight the pertinence of the uniquely and exceptionally prolonged period of juvenility in humans, during which time children remain highly dependent upon caregivers for resources and safety (Kaplan et al., 2000). In accordance with this premise, Bowlby (1969/1982) concluded that separation or abandonment from caregivers would have dramatically increased a child’s probability of death throughout humanity’s ancestral history and, as a result of this, children are acutely motivated to maintain proximity to caregivers in order to ensure their survival. A broader socio-evolutionary perspective further highlights that, even in adulthood, humans would have remained highly reliant upon social connection for continued survival in our environment of evolutionary adaptiveness (Baumeister & Leary, 1995), and thus rejection or ostracism from one’s social network even into adulthood represents a significant survival threat. As a consequence, the attachment system (that is, the psychological system designed to maintain close relationships) remains highly relevant throughout the human lifespan, and indeed contemporary research has demonstrated how attachment patterns developed during childhood have a pervasive influence on the strategies used to develop and maintain interpersonal relationships even into adulthood (Holmes & Johnson, 2009).

It is clear how, from this perspective, the connection between fears of interpersonal rejection and fears of dying are intricately related. Throughout our evolutionary history, social rejection, ostracism, or abandonment across the lifespan significantly increases one’s chances of dying. Importantly, this perspective highlights that the attachment system plays an integral role in regulating how one utilizes social relationships in order to maintain survival (and consequently, assuage death anxiety). A novel hypothesis that can consequently be derived from this paradigm is that attachment processes ought to mediate the relationship between death fears and social anxiety.

Crucially, attachment theory provides a robust and detailed framework from which to understand these complex processes, as is evidenced by its insight into individual differences in patterns (or styles) of attachment. Attachment theory proposes that when one’s upbringing is characterized by stable relationships with one’s attachment figures, then a secure attachment is likely to develop. In contrast, if the conditions of one’s upbringing are harsh and relationships with attachment figures are characterized by instability and/or abuse, attachment orientation is more likely to be insecure, resulting in the development of distinct strategies to address the challenges of the child’s unique developmental environment (Chisholm, 1996). Whilst various taxonomies for classifying these discrete insecure attachment styles have been proposed, the anxiety-avoidance continuum remains the most commonly employed due to its relevance from infancy through to adulthood (Del Giudice, 2009).

Individuals who develop an anxious attachment style (also known as anxious-ambivalent) are more likely to have been raised by inconsistent or neglectful caregivers (Ainsworth, Blehar, Waters & Wall, 1978), thus resulting in relationship patterns characterized by fervent attempts to elicit connection from attachment figures (Youngblade & Belsky, 1989). Such individuals are also likely to suffer from intense fears of separation and abandonment related to existing relationships. The anxious attachment style is often referred to as a “hyperactivation” of the attachment system (Fraley et al., 2006) and, in the context of the present study, could be considered an “over-emphasis” on the importance of relationships for survival.

On the other hand, individuals who exhibit an avoidant attachment style (also known as dismissive) are more likely to have been raised in physically or emotionally abusive environments (Ainsworth, Blehar, Waters & Wall, 1978). As such, they have developed relationship patterns characterized by self-reliance, independence, and avoidance of attachment interactions with caregivers (Youngblade & Belsky, 1989). The avoidant attachment style is often referred to as a “deactivation” of the attachment system (Fraley & Shaver, 1997), and in the context of the present study, could be considered an “under-emphasis” on the importance of relationships for survival.

In this way, we can see how one’s experiences during childhood fundamentally shape appraisals of the importance of social relationships for maintaining survival. It has also been well established that attachment patterns strongly influence psychological functioning throughout the lifespan, and we can observe the manifestation of this in the profound and multifaceted association that the attachment system has with psychopathology across many domains (Mikulincer & Shaver, 2012). For example, attachment processes have been proposed to be core components of the conceptualisation of many anxiety disorders, such as agoraphobia (De Ruiter & Van Ijzendoorn, 1992), Panic Disorder (Pini et al., 2014) and hypochondriasis (Noyes et al., 2003), even when these disorders seem ostensibly unrelated to interpersonal relations.

It has also been proposed that the attachment system comprises a central element of the formulation of social anxiety syndromes (Vertue, 2003), with extensive empirical evidence supporting this. Developmental models of SAD highlight that parenting practices and attachment processes in parent-child dyads play a critical role in the later development of social anxiety and avoidance (Ollendick & Benoit, 2012). Attachment insecurity has been consistently demonstrated to be related to social anxiety experienced in middle childhood and adolescence (Brumariu & Kerns, 2013). Even into adulthood, an anxious attachment style has been found to be associated with elevated rates of social interaction anxiety, fear of negative evaluation, and social avoidance (Eng et al., 2001; Manes et al., 2016). In a systematic review of 30 studies investigating adult populations, 28 studies evidenced a significant association between attachment insecurity and social anxiety (Manning et al., 2017), which was particularly strong for the attachment anxiety dimension. The authors of this review concluded that the literature compellingly supports an attachment-based theoretical conceptualisation of social anxiety, which ought to incorporate underlying evolutionary and cognitive factors.

Given the importance of the attachment system in generally regulating psychological functioning, it is similarly unsurprising that attachment processes have been found to significantly mediate the relationship between various clinically-relevant constructs. Anxious attachment beliefs have predictably been shown to mediate the relationship between maternal anxiety and child anxiety (Costa & Weems, 2005). Moreover, such processes have also been demonstrated to mediate the relationship between negative life experiences and the manifestation of psychological dysfunction later in life (Bifulco et al., 2006; Sandberg et al., 2010).

In sum, it is evident that attachment processes play an important role in mediating the manifestation of clinical symptoms. This role ought to be particularly salient in social anxiety, in which the attachment system comprises a central element of its formulation. In the context of the reviewed literature, in which we highlight the fundamental connection between underlying death fears and social anxiety, as well as the importance of the attachment system in regulating this relationship, it is thus proposed that attachment processes ought to mediate the relationship between death fears and social anxiety.

The purpose of the present study is to examine the mediating role of the two dimensions of adult attachment in the relationship between death anxiety and social anxiety. Specifically, it was hypothesized that attachment anxiety would significantly mediate this relationship, given that this dimension is characterized by an amplified emphasis on the importance of positive social evaluation (and social acceptance) for survival. On the other hand, it was hypothesized that attachment avoidance would fail to demonstrate a mediating effect, as the importance of relationships for survival is undervalued in this pattern of attachment. We also sought to explore whether the mediating effects of these attachment processes differed according to the subject focus of death anxiety – that is, whether such fears are related to fears of one’s own death, versus fears of other’s deaths – as a socio-evolutionary perspective would suggest that fears of interpersonal rejection are equally related to an underlying fear of one’s own death, as well as the fear of close others’ dying.

Method

Participants

Data for this study was collected from a larger experimental investigation conducted at The University of Sydney, Sydney, Australia (see Zuccala & Abbott, 2021). Participants consisted of 82 undergraduate psychology students from The University of Sydney (Mean age = 20.26; SD = 2.88) who were provided with course credit for their participation, as well as 20 participants recruited from local community and online noticeboards (Mean age = 23.45; SD = 4.39) who were provided with two movie vouchers for participating in the study.

All participants were assessed using the Anxiety and Related Disorders Interview Schedule for DSM-5 (ADIS-5; Brown & Barlow, 2014) to determine whether they met exclusion criteria for the study. To avoid contamination of physical-type anxieties, participants were excluded from analysis if they received any diagnosis related to physical threat (e.g., Panic Disorder; n = 8) or were experiencing a manic/psychotic episode at the time of testing (n = 1). The ADIS-5 was administered by a postgraduate clinical psychology student trained in the ADIS-5 procedure and under the supervision of a senior clinical psychologist. A second rater (Author 3) who was blind to diagnostic status randomly coded 10% of all interviews in order to establish interrater agreement, which was 87.5%, κ = .83 (agreement for principal diagnosis was 100%, κ = 1). Participants in the present study most commonly met criteria for a diagnosis of Social Anxiety Disorder (n = 43), followed by Generalized Anxiety Disorder (n = 17), Major Depressive Disorder (n = 12), and Specific Phobia (n = 11). Forty-two participants did not meet criteria for any diagnosis.

Overall, 63 women and 30 men (N = 93) with a mean age of 20.92 (SD = 3.57) were included in the final analysis. There were no significant differences between mean age of women (M = 20.44; SD = 3.17) and men (M = 21.93; SD = 4.17). 30.1% of the sample were employed and 69.9% were enrolled either full-time or part-time in a university degree.

Measures

Trait social anxiety

The Social Outcome Questionnaire (SOQ; Lovibond & Rapee, 1993) is a brief instrument which requires participants to rate how frequently they worry about 12 negative outcomes related to social threat on a 5-point scale. Research has evidenced that the SOQ comprises sound discriminant validity and reliability (Campbell et al., 2001). Internal consistency of the SOQ in this study was found to be excellent (α = 0.95).

Trait death anxiety

The Revised Collett-Lester Fear of Death Scale (CLFD-R; Lester & Abdel-Khalek, 2003) is a 28-item instrument that is used extensively within the death anxiety literature (see Neimeyer, 2015). The scale consists of four subscales, each assessing a different aspect of death anxiety: Fear of Own Death, Fear of Own Dying, Fear of Other’s Death, and Fear of Other’s Dying. The total combined score can also be used as a measure of overall death anxiety. The CLFD-R has been shown to be psychometrically valid and reliable, and the subscales have evidenced good discriminant validity (see Zuccala et al., 2019). The CLFD-R was chosen for the present study due to the unique nature of its subscales, which allow for distinction between anxiety related to one’s own death versus the death of others (e.g., loved ones), which was considered an important distinction given the attachment models utilized in the rationale for the present study. For analysis of these two distinct constructs, two categories were created by combining the subscales related to fear of own mortality (CLFD-Own) and the subscales related to fear of others’ mortality (CLFD-Others). In the current study, the two categories exhibited excellent internal consistency (CLFD-Own: α = 0.91; CLFD-Others: α = 0.92) as did the overall scale reliability (α = 0.94).

Trait attachment anxiety and avoidance

The Revised Experiences in Close Relationships scale (ECR-R; Fraley et al., 2000) is a 36-item instrument designed to measure the anxiety and avoidance dimensions of adult attachment. Each subscale consists of 18 items. Participants are required to rate how strongly each statement reflects their feelings in close relationships on a 7-point Likert scale. The ECR-R has demonstrated excellent psychometric properties in previous studies (Sibley et al., 2005; Sibley & Liu, 2004) and is used extensively as a measure of adult attachment within the relevant literature. In the present study, internal reliability was found to be excellent for both the attachment anxiety (α = 0.92) and attachment avoidance (α = 0.94) scales.

Procedure

All participants provided informed consent prior to participating in the study. Participants were first administered the ADIS-5 to ensure that diagnostic criteria were met for inclusion in the study. Participants then completed the SOQ, CLFD-R and ECR-R as part of a larger battery of self-report questionnaires. Following completion of the entire study (see Zuccala & Abbott, 2021), participants were fully debriefed on the nature and purpose of the research.

Statistical analyses

Simple Mediation models were tested using the PROCESS macro for SPSS (Hayes, 2018). Model 4 (mediation) in Process was used with 5000 bootstrapped samples to estimate the 95% percentile bootstrap confidence intervals of the indirect effect. Evidence of an indirect effect was determined if the lower and upper bootstrapped confidence intervals are entirely above or below zero.

In the first mediation model (N = 93) the independent variable was “Death Anxiety”, comprising the total score on the CLFD-R. The dependent variable was “Social Anxiety” as measured by the SOQ. The investigated mediator variable was “Attachment Anxiety” as measured by the attachment anxiety subscale of the ECR-R. In the second mediation model (N = 93) the independent and dependent variable remained the same, but the investigated mediator variable was instead “Attachment Avoidance”, as measured by the attachment avoidance subscale of the ECR-R. Further mediation analyses were undertaken only for models that were found to have evidence of an indirect effect. These additional analyses involved investigating whether the indirect effect remained when the independent variable was divided into “Death Anxiety-Own” and “Death Anxiety-Others”.¹

Results

Descriptive statistics

Means and standard deviations for all measures are reported in Table 1.

Table 1. Means and standard deviations of variables included in mediation analyses

	Independent Variables		Mediator Variables		Dependent Variable
Measure	M	SD	M	SD	M	SD
Death Anxiety (CLFD-R)	90.73	24.88
Death Anxiety-Own (CLFD-Own)	40.10	14.02
Death Anxiety-Others (CLFD-Others)	50.63	14.14
Attachment Anxiety (ECR-Anxiety)			62.46	19.31
Attachment Avoidance (ECR-Avoidance)			60.65	20.03
Social Anxiety (SOQ)					36.92	12.31

Tests of normality and visual analysis of QQ plots and box plots indicated that all variables met normality assumptions (Ghasemi & Zahediasl, 2012), with the exception of social anxiety (W(93) = .96, p = .003). The distribution of the SOQ was found to be platykurtic (kurtosis = −1.13) and visual analysis of the frequency distribution (histogram) indicated a roughly even distribution of scores (Chissom, 1970). This was not considered problematic given that normality is not a required assumption for the bootstrapping method employed (Hayes, 2018).

Planned mediation analyses

The first mediation analyses investigated whether “Attachment Anxiety” mediated the relationship between “Death Anxiety” and “Social Anxiety”. The statistical diagram of this model is presented in Figure 1, with regression coefficients further outlined in Table 2. There was evidence of an indirect effect of death anxiety on social anxiety via attachment anxiety, b = .085, (Bootstrapped CI: .037, .139).

Table 2. Statistical information for mediation model 1

	Consequent
	Attachment Anxiety (M)			Social Anxiety (DV)
Antecedent	Coeff.	SE	p	Coeff.	SE	P
Death Anxiety (IV)	.260	.077	.001	.120	.043	.006
Attachment Anxiety (M)				.326	.055	<.001
Constant	38.890	7.211	<.001	5.661	4.353	.197
Variance	R^2 = .112 F(1,91) = 11.479, p = .001			R^2 = .403 F(2,90) = 30.440, p < .001

IV = Independent Variable, M = Mediator, DV = Dependent Variable.

Coeff = Regression Coefficient, SE = Standard Error.

Figure 1. Statistical diagram of mediation model 1

The second mediation analyses investigated whether “Attachment Avoidance” mediated the relationship between “Death Anxiety” and “Social Anxiety”. The statistical diagram of this model is presented in Figure 2, with regression coefficients further outlined in Table 3. There was no evidence of an indirect effect of death anxiety on social anxiety via attachment avoidance, b = .001, (Bootstrapped CI: −.023, .026).

Table 3. Statistical information for mediation model 2

	Consequent
	Attachment Avoidance (M)			Social Anxiety (DV)
Antecedent	Coeff.	SE	p	Coeff.	SE	p
Death Anxiety (IV)	.008	.084	.920	.204	.046	<.001
Attachment Avoidance (M)				.122	.058	.037
Constant	59.877	7.939	<.001	11.049	5.558	.050
Variance	R^2 = .000 F(1,91) = .010, p = .920			R^2 = .210 F(2,90) = 11.998, p < .001

IV = Independent Variable, M = Mediator, DV = Dependent Variable.

Coeff = Regression Coefficient, SE = Standard Error

Figure 2. Statistical diagram of mediation model 2

Additional mediation analyses

As the first mediation analyses found evidence of an indirect effect via attachment anxiety, a further two simple mediation models were tested with the independent variable (“Death Anxiety”) now split into “Death Anxiety-Own” and “Death-Anxiety Others”, respectively. The mediator (“Attachment Anxiety”) and dependent variable (“Social Anxiety”) remained unchanged.

The first additional mediation analyses investigated if “Attachment Anxiety” mediated the relationship between “Death Anxiety-Own” and “Social Anxiety”. The statistical diagram of this model is presented in Figure 3, with regression coefficients further outlined in Table 4. There was evidence of an indirect effect of death anxiety-own on social anxiety via attachment anxiety, b = .145, (Bootstrapped CI: .053, .252).

Table 4. Statistical information for mediation model 3a

	Consequent
	Attachment Anxiety (M)			Social Anxiety (DV)
Antecedent	Coeff.	SE	p	Coeff.	SE	p
Death Anxiety-Own (IV)	.415	.138	.003	.134	.077	.085
Attachment Anxiety (M)				.349	.056	<.001
Constant	45.822	5.844	<.001	9.776	4.030	.017
Variance	R^2 = .091 F(1,91) = 9.087, p = .003			R^2 = 0.372 F(2,90) = 26.798, p < .001

IV = Independent Variable, M = Mediator, DV = Dependent Variable.

Coeff = Regression Coefficient, SE = Standard Error

Figure 3. Statistical diagram for mediation model 3a

The second additional mediation analyses investigated whether “Attachment Anxiety” mediated the relationship between “Death Anxiety-Others” and “Social Anxiety”. The statistical diagram of this model is presented in Figure 4, with regression coefficients further outlined in Table 5. There was evidence of an indirect of death anxiety-others on social anxiety via attachment anxiety, b = .130, (Bootstrapped CI: .037, .227).

Table 5. Statistical information for mediation model 3b

	Consequent
	Attachment Anxiety (M)			Social Anxiety (DV)
Antecedent	Coeff.	SE	p	Coeff.	SE	p
Death Anxiety-Others (IV)	.396	.137	.005	.229	.073	.002
Attachment Anxiety (M)				.329	.054	<.001
Constant	42.406	7.20	<.001	4.750	4.337	.276
Variance	R^2 = .084 F(1,91) = 8.358, p = .005			R^2 = 0.415 F(2,90) = 31.897, p < .001

IV = Independent Variable, M = Mediator, DV = Dependent Variable.

Coeff = Regression Coefficient, SE = Standard Error

Figure 4. Statistical diagram for mediation model 3b

Multicollinearity analysis

High multicollinearity was considered unlikely to be a concern given that correlations across measures ranged between 0.02 and 0.60 (Tabachnick & Fidell, 2001). Nevertheless, multicollinearity was directly tested by calculating the Variance Inflation Factor (VIF) and tolerance levels. High multicollinearity is indicated by VIF levels greater than 5 and tolerance levels lower than 0.2 (Perini et al., 2006). Analyses indicated acceptable levels of multicollinearity in the present study as all regression models fell within these ranges.

Discussion

The present study sought to examine the mediating role of attachment processes in the relationship between death anxiety and social fears, within the broader context of exploring socio-evolutionary theoretical frameworks for understanding social anxiety. The findings of this study supported all of the novel hypotheses. Mediation analysis revealed that there was an indirect effect of death anxiety on social anxiety via attachment anxiety, but not for attachment avoidance. Additional analyses revealed that the indirect effect via attachment anxiety remained unchanged when fear of own death and fear of others’ death were analysed separately.

Employing a socio-evolutionary perspective, we underlined that the attachment anxiety dimension is characterized by an inflated emphasis on the importance of social connection for survival, and thus proposed that attachment anxiety ought to mediate the relationship between death anxiety and social anxiety. The present results supported this hypothesis. We also proposed that, on the other hand, attachment avoidance would not mediate the relationship between death anxiety and social anxiety, as the importance of social relationships for survival is characteristically undervalued in this dimension. Results again supported this hypothesis.

The present findings represent a uniquely interesting development in our understanding of the mechanisms underlying social anxiety, as they provide support for a novel approach towards conceptualising social fears. Previous literature has highlighted the connection between attachment processes in childhood and the development of social anxiety (e.g., Ollendick & Benoit, 2012), but relatively less research has focused on the role of specific attachment processes in adulthood and how these link to an overarching socio-evolutionary perspective. The current findings highlight how interpersonal acceptance is psychologically interconnected with survival value, and how this may then manifest overtly as social anxiety. These results also add value towards understanding how differences in developmental upbringing ubiquitously influence relationships amongst clinically-important constructs, and overall provide the foundations upon which to build an integrative theoretical approach towards understanding the role of socio-evolutionary processes in the development, maintenance, and treatment of social anxiety.

It is of particular interest to highlight that the results remained unchanged even when the “death anxiety” construct was divided into anxiety about “death of self” versus “death of others”. From a purely cognitive standpoint, the connection between the clinical constructs in the present study and the fear of others’ dying is relatively clear: attachment anxiety and social anxiety are both characterized by the fear of losing positive relationships with others. What is perhaps more thought-provoking is that attachment anxiety remained a mediator even for anxiety related to death of self exclusively. Whilst fear of own death would ostensibly appear to have little association with attachment or social anxiety, a socio-evolutionary approach would suggest that this phenomenon reflects the inherent survival value of social relationships. Within this context, it is unsurprising that a fear of being rejected, ostracized or abandoned by others, would also be related to a fear of premature death of self.

Importantly, these findings support a general theoretical framework that highlights how the construct of “death anxiety” is deeply intertwined with social processes. Expert consensus on the environment of evolutionary adaptedness generally emphasises the important survival value associated with the formation and maintenance of robust interpersonal relationships, particularly for sapiens (Flinn et al., 2005; Herrmann et al., 2007). Such relationships were important for the regulation of survival in childhood (as outlined in attachment theory; Bowlby, 1969/1982), but also throughout adulthood, due to humanity’s reliance on social cohesion for collective resource acquisition, protection from predators, and alloparenting, amongst other functions (Baumeister & Leary, 1995; Hrdy, 2007). Thus, it is unsurprising that the fear of death is apparently assuaged by social processes, (in this study specifically, attachment processes) as interpersonal relationships functionally served to prevent premature death throughout our evolutionary history.

These results also implicate the role of early life experiences in the manifestation of death anxiety and social anxiety later in life. As emphasised in both life history and attachment theory, the environment of developmental upbringing profoundly shapes an individuals’ psychological organisation throughout their lifespan (Bowlby, 1969/1982; Del Giudice et al., 2016). Current understandings of the attachment system propose that children who are reliant upon inconsistent caregivers are susceptible to becoming highly anxious about maintaining close relationships later in life (represented in the attachment anxiety dimension). The present results provide direct evidence that this dynamic is related to the associated consequences on survival (as is represented in the death anxiety construct), and that such an association even remains into adulthood. This approach would perhaps also suggest that the concept of “death anxiety” could be better framed as a general drive to maintain survival and avoid premature death, rather than a distinct type of clinical anxiety per se. Nevertheless, that is not to say that death anxiety does not have utility as a clinical construct within certain populations/frameworks (such as palliative care), especially with the consideration that the present results were sourced from a relatively general and young cohort.

It should be acknowledged that this dataset was derived from a sample of individuals who were screened and included according to whether they received a diagnosis of Social Anxiety Disorder. Further, individuals diagnosed with any diagnosis related to physical anxiety (e.g., Panic Disorder or Illness Anxiety Disorder) were excluded from participating. There are several advantages associated with this recruitment strategy, including the fact that strong conclusions can be made regarding the role of death anxiety in social anxiety specifically, as there is no “contamination” of physical anxiety (where conscious death fears are cognitively more proximal to the subject focus of anxiety). However, on the other hand, these inclusion criteria constrain the external validity of these results, as the sample is limited in its representation of physical-type anxieties in the general population. This limitation should be addressed in future research by examining these important research questions in broader, and larger, samples of participants. Nonetheless, it also deserves noting that a strength of the present study was the breadth of social anxiety scores, suggesting that these results represent the full spectrum of social anxiety severity.

Previous investigations have emphasised the need for more robust theoretical frameworks for understanding the construct of death anxiety (Zuccala & Abbott, 2021). The present study marks an important step towards the development of a novel theoretical approach towards understanding the role of the fear of death in social anxiety (as well as in other clinical syndromes), grounded in the principles of existing, well-established theoretical paradigms – specifically, socio-evolutionary theory, life history theory, and attachment theory. Whilst the methodology used in the present study may be somewhat conventional and relatively modest, the confirmation of the novel hypotheses derived using these frameworks suggests that this approach is a promising direction for future investigation.

Ethics approval and participant consent

This study was approved by The University of Sydney Human Ethics Research Committee (2017/073).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data is available from authors upon reasonable request.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Notes

^1. Given the gender imbalance in this study, we conducted One-Way ANOVAs to determine whether outcome variables differed according to gender. Death Anxiety-Others was the only variable that significantly differed by gender (p = .044). Mediation analyses involving Death Anxiety-Others was repeated with gender as a covariate, and no differences in the patterns of significance were observed.