Beyond the watery grave: death and water reminders as (un)expected ways to increase environmental identity / Más allá de la tumba acuática: recordatorios de muerte y agua como formas (in)esperadas para fomentar la identidad ambiental

Dual-process defence model

Mortality reminders can be explicit (e.g., asking someone to think of their own death) and evoke immediate, proximal responses, or reminders can be subtle (e.g., walking past a cemetery, the word ‘death’ flashed quickly as a subliminal message) and evoke subconscious, distal responses. Figure 1 (from Smith et al., 2022) displays proximal responses (denial, distraction and rationalization) and distal (self-esteem bolstering and worldview reinforcement; Kosloff et al., 2019; Pyszczynski et al., 1999; Wolfe & Tubi, 2019) responses.

OPEN IN VIEWER

Figure 1.

Dual-process model of mortality reminder defence mechanisms (Smith et al., 2022).

Terror Management Theory researchers posit that bolstering self-esteem defends against existential anxiety evoked by MRs — by strengthening cultural, societal identity, existential anxiety is offset via symbolic immortality. Our focus lies in distal defences involving identity and worldview defence.

Distal mortality-related defences are particularly significant since they can influence identity reinforcement, PEBs and one’s support for or against climate policy and solutions. Distal defences may result in strengthened beliefs and identities (Barth et al., 2018; Castano et al., 2002; Harmon-Jones et al., 1996; Stollberg & Jonas, 2021). Concerning EID, discussed in more detail below, those who already self-identify as highly environmental may even more strongly support these values in response to existential threats to reinforce identity and protect self-esteem. For individuals who do not identify as environmental, mortality threats may elicit behaviours aligned with dominant cultural identities. In a Western context, these include individualism, consumerism and capitalism, all of which are problematic for climate solutions that require collective action, reduced resource consumption and significant social changes (Metz, 2015). Climate-related water threats will only increase with predicted environmental changes, requiring urgent attention on human response to life-threatening water crisis communication, especially about EID, explained below.

EID and sociocultural connections

Since climate problems are rarely an individual’s sole responsibility and solutions require collective responses (Clayton, 2003), individual and socio-cultural concepts have been linked to EID and PEB (Dono et al., 2010; van der Werff et al., 2014; Vesely et al., 2021; Whitmarsh & O’Neill, 2010). Collective, cultural connection is particularly relevant since TMT researchers have emphasized the significance of self-esteem derived from society, culture and relationships as a mortality anxiety defence (Castano et al., 2002; Greenberg et al., 1990; Kosloff et al., 2019; Plusnin et al., 2018). These defences can exacerbate intergroup biases, potentially barring successful collective action efforts that require cooperation amongst diverse actors who do not necessarily hold similar worldviews. With increased life-threatening climate communications, this may be particularly true since MRs result in stronger adherence to one’s worldviews and identity (Greenberg et al., 1990; Pyszczynski et al., 2015; Schimel et al., 2019), perhaps making it even more difficult to convince low-EID individuals that action is needed. Moreover, identity’s role within social identity and collective action is important to understand from a TMT perspective: human responses to existential threats involve clinging more strongly to one’s ingroup and distancing from outgroups (Castano et al., 2002; Harmon-Jones et al., 1996). In these ways, the collective action required to address climate changes will necessitate an understanding of one’s social identity and working with those in — and adjacent to — one’s ingroups to achieve a mutual goal (Barth et al., 2018; Fritsche & Masson, 2021).

An individual’s EID is connected to their identified social group, including as characterized by political ideologies around environmental interests that may be linked to sense-of-self (Clayton, 2003, 2012). For example, someone who identifies as right wing or more politically conservative is less likely to support climate policies compared to left-wing or more liberal-identifying people (Currie & Choma, 2018; Huber, 2020; Unsworth & Fielding, 2014). Just as a person may reinforce their existing connections to other social groups following an MR (Barth et al., 2018; Castano et al., 2002; Harmon-Jones et al., 1996), they may also more strongly reinforce their EID — or lack thereof if they do not identify as pro-environmental (Stollberg & Jonas, 2021; Uenal et al., 2021; Uhl et al., 2018). While some TMT researchers have found greater death anxiety can increase negative nature evaluations due to creatureliness reminders — connections to our animal natures and, thus, mortality — these researchers did not differentiate participants’ responses according to pre-existing EIDs or values (Fritsche & Hoppe, 2019; Koole & Van Den Berg, 2005). It is possible that EID could have an unidentified moderating role in these analyses (Fritsche & Häfner, 2012).

Overall, attitudes, norms, values and identity are integrally involved in increased environmentalism (Li et al., 2019; Newell et al., 2021). The persistence of these features illustrates where TMT connects and how existential threats can influence EID, in both environmentally positive and negative ways.

Participants

Six hundred American and Canadian adult participants were recruited in 2020 via Amazon’s Mechanical Turk (MTurk) — a crowd-sourcing online platform frequently used in social psychology studies (Anderson et al., 2019; Crump et al., 2013). Details on MTurk and screening processes can be found in Appendix 1. For completing the ~30-minute study, participants received $2.50 USD — comparable remuneration to studies of similar length and format.

After incomplete or erroneous response removals, total N was 455 (attrition details in Appendix 2). The 23% attrition rate was within expected and acceptable range for an MTurk study (15–30%; Aguinis et al., 2021).

Procedure

Participants were recruited in the fall of 2020 (potential pandemic implications presented in the Discussion section) and informed they would be completing a study on ‘personality and sustainability’. We disguised full study purpose to avoid priming participants with mortality or confounding concepts. As part of the deception, participants were first asked to complete a 22-item personality inventory (Eysenck & Eysenck, 1964), a typical cover in TMT research (Cox et al., 2018). The subsequent study procedure was consistent with conventional TMT study design (Figure 2).

OPEN IN VIEWER

Figure 2.

Visual procedural flow diagram of study design.

Participants were randomly assigned to one of five experimental conditions (i.e., control, mortality salience, drowning, dehydration or contaminated water), where they were asked to respond to two open-ended questions about their own death (Mortality Attitudes Personality Survey; see Appendix 3). The water-specific conditions replaced the mortality salience version’s ‘death’ with either drowning, extreme thirst or contaminated water consumption (Appendix 3). The control condition was asked about a painful dental visit as an anxiety-inducing but non-life-threatening prompt — a well-established control used in approximately two-thirds of previous TMT research (Burke et al., 2010; Cox et al., 2018).

Next, two delay tasks were presented — the Positive and Negative Affect Schedule (PANAS; Watson et al., 1988) and the Rosenberg Self-Esteem Measure (Rosenberg, 1965), allowing MRs to reach subconscious levels, activating distal defences (Greenberg et al., 2000).

After delays, the Environmental Identity Scale — short version (EID; Clayton, 2003; Clayton et al., 2021) was presented. This 11-item, seven-point Likert scale (full scale in Appendix 4) includes items such as ‘I think of myself as a part of nature, not separate from it’ and ‘Behaving responsibly towards the earth — living a sustainable lifestyle — is part of my moral code’. The EID was chosen because we were not able to measure environmental behaviours remotely and the EID has been frequently used in PEB research (Balundė et al., 2019) and widely validated (Clayton et al., 2021; Olivos & Aragonés, 2011). The short version was used to avoid participant fatigue and shows reliable, consistent results compared with the full version (Chew, 2019; Clayton, 2012).

Finally, demographic information was collected, followed by a deception check and debriefing materials.

Descriptive statistics

Environmental Identity means and distribution were analysed via SPSS v.28 software to compare the experimental conditions’ responses spread to that of the control (data visualizations in Appendix 5). Descriptive statistics are displayed in Table 1, below.

OPEN IN VIEWER

Table 1.

Descriptive statistics for each group’s environmental identity responses.

Condition	n	M	SD
Control	116	4.91	1.16
Mortality salience	122	4.95	1.21
Drowning	71	4.83	1.29
Contaminated water	77	4.92	1.24
Dehydration	69	4.89	1.26
TOTAL	455	4.89	1.22

Cronbach’s alpha indicated high internal consistency across EID responses (11 items; α = .90). Levene’s tests of homogeneity of variance indicated overall distributions were not statistically significant when comparing experimental conditions (e.g., typical mortality salience or water mortality salience) to control condition. Results are shown in Table 2. To statistically determine whether responses were more drastically spread in experimental conditions, indicating stronger responses compared to control, the Moses Extreme Reactions Test was used (Table 2). Where Levene’s tests determine homogeneity of variance between conditions, Moses Extreme Reactions tests explicitly for response extremity. Typical means comparison statistical tests were not relevant to our main hypothesis; group means could be similar while maintaining different response spread and distribution, blurring effects. The Moses test determines difference in response extremity — in either direction — between two independent groups, designed intentionally to determine experimental variable influences when extreme reactions in opposing directions might be expected (e.g., an experimental condition compared to a control; Arnold & Briley, 1973; Colman, 2015). In this test, both groups’ scores are combined and become ranks (Moses, 1952). The test statistic is one group’s range plus one; the group is chosen at random. Span probability is calculated and then recalculated after removing a certain number of extreme scores from each end, before one-tailed probability is determined (Moses, 1952).

OPEN IN VIEWER

Table 2.

Levene’s and Moses Extreme Reactions Test outcomes: comparing experimental condition response spread to control.

Group	Levene’s test statistic	p	Moses test statistic	p
Control-Mortality salience	.37	.542	206	.042*
Control-Drowning	2.66	.105	160	.007**
Control-Contaminated water	2.13	.146	164	.007**
Control-Dehydration	2.03	.156	160	.019*

Note: p < .05 indicates significantly more extreme responses (in either direction) than control; *p < .05; **p < .01

Response distributions and results

Response distributions were statistically compared via the Levene’s and Moses Extreme Reactions tests, displayed in Table 2.

Results indicate all conditions had responses significantly more extreme than the control (p < .05).

COVID-19 potential implications

Before discussing the implications of our findings, we should acknowledge the context in which this research was completed. The COVID-19 pandemic brought frequent worldwide MRs and initially impacted US and Canadian populations around March 2020 (Courtney et al., 2020; Pyszczynski et al., 2021). The non-stop media and government COVID-19 information persisted throughout the spring and summer and our recruitment occurred in fall 2020, prior to widespread COVID-19 vaccine availability. Recurrent pandemic-related MRs in this time could have subdued observed effects’ strength since our control may have had above-average mortality awareness. To our knowledge, an average mortality anxiety score for the pandemic period has not been empirically established, but this information would be useful for future research.

Research results in empirical context

Overall, we found MS and all water crises conditions significantly increased polarization of the participants’ EID compared to the control. This was particularly true for the drowning and contaminated water conditions, both at p = .007 significances. Our findings support prior work by Vess and Arndt (2008) and Fritsche and Häfner (2012), who tested environmental concern and/or identity in the context of MS and TMT. Vess and Arndt (2008) found that for people who derived self-esteem from pro-environmental behaviours, MS increased their concern for the environment; for people who did not derive self-esteem from pro-environmental behaviour, the reverse relationship presented. While Vess and Arndt (2008) utilized the Environmental Contingencies of Self-Worth scale as a measure of self-esteem derived from environmental behaviour, rather than EID, the trend was replicated within our findings. Fritsche and Häfner (2012) noted that MS decreased motivations for participants to protect the environment for intrinsic reasons and investigated EID as a potential moderating variable. They found that when people did not define themselves as part of nature or an environmentalist group — thus having low EID — they had less motivation and concern for the environment as something to value for intrinsic reasons (Fritsche & Häfner, 2012).

While we did not have a pre-existing EID measure for our participants, groups can be assumed to have a normative mix of pro-environmental individuals. Although American and Canadian studies indicate growing awareness and concern regarding climate change, this awareness is not the population majority, nor is awareness equivalent to EID or, critically, nor does it translate to pro-environmental action (Leiserowitz et al., 2021; Mildenberger et al., 2016). Americans are increasingly identifying as ‘Alarmed’ about climate change — up to 24% of the population according to Yale’s Global Warming’s Six Americas (Leiserowitz et al., 2021). While encouraging for potential climate solutions, even within this segmentation, 20% are ‘Inactive’ — potentially less likely to engage in PEBs. Compared to the US, a greater proportion of Canadians believe climate change is happening, but belief that humans are responsible is mixed. Researchers have shown that only ~20% of people in the Prairies and Western provinces’ oil-economy areas believe humans are responsible for climate changes while ~70% of Canadians take responsibility on the East and West Coasts (Mildenberger et al., 2016).

Given these baseline findings, it is likely that some of our participants identified as pro-environmental, but for others who did not, EID would not be a self-esteem source. For those participants with low EID, their identity and self-esteem were likely more strongly linked to traditional consumerist values, which remain dominant in Western society (Metz, 2015).

While it could be worthwhile for future studies to include an initial EID measure before a mortality reminder occurs, it would be essential to do so carefully since this could prime participants with environmentalism, potentially blurring and undermining results. A measure obtained in a pre-screening study could provide this specific insight, rather than assuming a normative mix of environmental identities. This would allow for a fuller understanding of effects and implications of existential threats. Smith et al. (2022) provide additional environmental operationalizations worth considering in future environmental-TMT research.

Implications for environmental change and water management

Our findings inform future water crises communications for those who aim to increase pro-environmental identities and behaviours. On an individual level, it would be valuable to understand individuals’ values, worldviews and identity related to water and environmentalism before implementing a communication strategy; such context is important to consider when seeking to understand PEBs. For a more widespread campaign, it may be beneficial to know the audience’s majority EID, as has been aptly exemplified via consumer segmentation efforts by industry and academia. For example, Yale’s Global Warming’s Six Americas assesses awareness and concern among the public, and Słupik et al.’s (2021) European energy-user analysis based on behavioural and socioeconomic factors, although many other examples exist. A campaign designed around the majority’s worldview would admittedly result in some individuals not receiving messages that would most effectively increase their EID, pro-environmental values and potentially behaviours, but it may be the best option given budgetary constraints that are common in governments’ and non-governmental organizations’ environmental efforts.

But while individual behaviour change can offset some forecasted water insecurities (IPCC, 2018), various structural barriers such as class, gender and race will often prevent individual actions from having more powerful or sustained change impacts (Kenis & Mathijs, 2012; Newell et al., 2021; Uzzell & Räthzel, 2009). Given this reality, significant water management power remains with government and policymakers; also considering how existential threats may influence these expert decision-makers is crucial.

Whether working in a group or alone, we consider decision-makers as individuals and recognize that they are all influenced by their own implicit and explicit values, worldviews and identities (Staerklé, 2015; van der Werff et al., 2014), as explicated above through TMT perspectives (Pyszczynski et al., 2015). Despite what historical economics espouse, humans are not merely rational machines (Johnson, 2021; Staerklé, 2015); it is necessary to consider these individualized factors that influence decisions — particularly in water-related contexts (Wolfe, 2017). These individual factors are not the only influences, but they are important to recognize to understand how they are changed by existentially threatening scenarios or communications and what that means for desired outcomes. In environmental efforts these desired outcomes are often increased PEB. While some economists have argued that groups can exercise more rational decision-making than individuals (Kugler et al., 2012), we must remember that these groups consist of individuals with their own backgrounds, beliefs, worldviews and values.

Further, recent environmental scholars have found that identity-diverse groups (e.g., diverse genders, races and/or classes) make more ethical, environmentally protective and equitable decisions (Craft, 2013; Glover et al., 2002; Vollan & Henry, 2019). Time and again, researchers have argued that environmentally responsible decisions are best made when processes involve diverse members affected by the issue, are open, transparent and built on trust, and are participatory, collaborative and flexible (Bolderdijk & Jans, 2021; Decaro et al., 2017; Henshaw, 2018; Powell et al., 2017; Vollan & Henry, 2019). However, when discussing life-threatening crises, MRs that arise may evoke worldview and group identity reinforcement as defences against death anxieties (Castano et al., 2002; Harmon-Jones et al., 1996; Uhl et al., 2018), as we observed with EID.

Future research opportunities

Our findings provide important and novel insights into (1) how individuals may experience EID changes following life-threatening communication and (2) represent a necessary first step to understanding how these messages could also influence powerful decision-makers. Future research on life-threatening communication among water experts specifically would be essential. Through our findings, we further show that inquiry with specific sub-populations (e.g., among water decision-makers) would be worthwhile to help determine how water crisis decisions are impacted by life-threatening communications. Future research with a collectivist lens would provide a more extensive understanding of the influence of both EID and threatening messaging on pro-environmental outcomes. Emerging social science research has highlighted the need for collectivist perspectives, and social identity theory notes usefulness in embedding EID within these views (Fritsche & Masson, 2021; Kenis & Mathijs, 2012; Klein et al., 2022). It would also be valuable to replicate our research in different social contexts and with participants who have the power to directly influence environmental and climate outcomes.