The role of meaning in life in psychological distress during the COVID-19 pandemic

ABSTRACT

Background/Objective:

Meaning in life may function as a protective factor in the context of potentially traumatic experiences, such as the COVID-19 pandemic. We investigated the associations between meaning and psychological distress (i.e., depression, anxiety, COVID-19-related PTSD) prospectively and cross-sectionally. We hypothesized that meaning inversely predicts peri-pandemic distress and that meaning moderates the association between being negatively affected by the pandemic and distress. We additionally explored cross-sectional associations between meaning subcomponents and distress and a meaning violations perspective.

Methods:

Undergraduate students (N = 109) completed questionnaires before (October 2019 to March 2020; meaning, anxiety) and during the pandemic (April to June 2020; meaning, meaning subcomponents, depression, anxiety, PTSD).

Results:

Correcting for family-wise errors, meaning prospectively predicted less depression and anxiety, but not PTSD. Correcting for family-wise errors, peri-pandemic meaning was consistently related with peri-pandemic distress. Meaning did not moderate the link between being affected by the pandemic and distress. The meaning subcomponent comprehension was most strongly related with distress and a meaning violations perspective was partly supported.

Conclusion:

Meaning emerged as a significant correlate of peri-pandemic distress. Current findings should be replicated longitudinally and experimentally to establish their robustness and to examine the causal influence of meaning on distress.

KEYWORDS:

• Meaning in life
• psychological distress
• depression
• anxiety
• PTSD
• COVID-19

Introduction

For much of the world’s population, the coronavirus (COVID-19) pandemic abruptly disrupted normal life in 2020. The pandemic has been conceptualized as a potentially traumatic stressor (Horesh & Brown, 2020; Salari et al., 2020) which might elicit an increase in psychological distress, potentially involving high costs for affected individuals and society at large. A meta-analysis of international population-based studies published since the virus outbreak until May 2020 found 31.9% of participants affected by anxiety and 33.7% affected by depression (based on self-reported symptoms above a clinically relevant severity threshold; Salari et al., 2020). These estimates significantly exceed pre-pandemic point prevalence estimates (anxiety, globally: 3.6%, depression, globally: 4.4%; World Health Organization, 2017). COVID-related prevalence rates of posttraumatic and acute stress (i.e., above-threshold self-reported symptoms), ranged from 4.6% (in Wuhan in January/February; Sun et al., 2021) to 16.8% (in the UK in March; Shevlin et al., 2020), and even 37.0% (in Italy in March/April; Rossi et al., 2020). During the pandemic, younger adults reported particularly high psychological distress including anxiety, depression, and PTSD (Rossi et al., 2020; Shevlin et al., 2020; Varma et al., 2021). In addition to the psychological costs associated with the pandemic, it also affords the opportunity to identify protective factors which provide resilience. This information would be important as it may inform preventive and acute interventions. The aim of the current paper is to examine the role of meaning in life as a candidate protective factor. This research adds to the literature in several ways. For example, examining the relation between meaning and distress during the course of the pandemic allows insights into the directionality of effects. Moreover, the current study is among the first to investigate the three meaning subcomponents of comprehension, purpose, and mattering in a stressor context, furthering the evidence regarding beneficial meaning effects and offering relevant insights for intervention development.

Meaning in life

Meaning in life has been defined as a subjective sense of meaningfulness including three components: (1) a sense of comprehension (coherence and understanding of one’s life and experiences), (2) purpose in life (direction and motivation derived from aims and life goals), and (3) mattering (finding one’s life significant, worth living; e.g., George & Park, 2016; Martela & Steger, 2016). Several scholars have proposed that meaning is derived from meaning frameworks – mental structures which are comparable to schemas and include goals and beliefs (e.g., George & Park, 2016).

Meaning in life as a protective factor

Meaning in life may function as a protective factor against the effects of stressful and traumatic experiences on psychological distress (e.g., Frankl, 1985; Lightsey, 2006). Empirical support for this idea comes from longitudinal research showing that high baseline meaning predicts weaker aversive reactions to subsequent stressors. For example, greater baseline meaning inversely predicted less incident-specific rumination after a flood (Ostafin & Proulx, 2020, Study 2), and less distress in subsequent writing about one’s interpersonal traumatization (Ostafin & Proulx, 2020, Study 3). Moreover, greater baseline purpose has been shown to predict better emotional recovery (indexed by a lower startle response) after exposure to aversive pictures (Schaefer et al., 2013). Related, another study showed that participants in a meaning-reducing condition reported higher distress than control group participants in a subsequent social stress test (Park & Baumeister, 2017, Study 3).

The limited longitudinal and experimental evidence is complemented by extensive cross-sectional research in stressor-exposed samples demonstrating inverse associations between meaning and depression, anxiety, and posttraumatic stress disorder (PTSD). For example, greater meaning was associated with lower concurrent depression symptoms in earthquake survivors (Feder et al., 2013), combat-exposed veterans (Blackburn & Owens, 2015), cancer patients (Gravier et al., 2019; Jaarsma et al., 2007), and trauma-exposed elderly (Krause, 2007). Recent research suggests similar associations between meaning and depression during the COVID-19 pandemic (Chao et al., 2020; Schnell & Krampe, 2020). Moreover, greater meaning coincided with lower anxiety in cancer patients (Gravier et al., 2019; Jaarsma et al., 2007), participants in a smoking cessation program (Steger et al., 2009), and during the pandemic (Schnell & Krampe, 2020). Similarly, meaning was inversely related with PTSD symptoms in earthquake survivors (Feder et al., 2013), trauma-exposed students (Kashdan & Kane, 2011), and veterans (Blackburn & Owens, 2015).

These previous findings largely suggest that meaning inversely predicts distress, though most of the available evidence is cross-sectional and thus prohibits causal inferences. Several mechanisms underlying beneficial meaning effects have been proposed, such as transcendence of self and momentary suffering when committing to and focusing on personally relevant sources of meaning (e.g., raising children, work, religion/ spirituality; Frankl, 1985). Together with the motivation associated with purpose, this may counteract maladaptive behaviors associated with distress such as rumination and behavioral disengagement (García-Alberca et al., 2012; Nolen-Hoeksema et al., 2008).

Another potential reason for beneficial effects of meaning is the relation between meaning and uncertainty. Life during the pandemic has been posed to involve great uncertainty (e.g., Rutter et al., 2020), including for issues such as virus threat and treatment (e.g., Koffman et al., 2020) and economic matters (e.g., Baker et al., 2020). Uncertainty, especially when related to threat, may induce anxiety (Carleton, 2016; Grupe & Nitschke, 2013; Hirsh et al., 2012; Papenfuss et al., 2021). Additionally, uncertainty and uncontrollability about aversive outcomes may foster hopelessness and helplessness (Maier & Seligman, 1976), futile sense-making attempts (i.e., rumination), and thus depression (e.g., Nolen-Hoeksema et al., 2008). Moreover, uncertainty may be involved in PTSD, for example when the traumatic event is unexpected or uncontrollable (Mineka & Zinbarg, 2006), induces perceptual or behavioral uncertainty, its outcomes are uncertain (Goto et al., 2006) or when schemas are lastingly disrupted (Janoff-Bulman & Frieze, 1983; C. L. Park et al., 2012). Therefore, meaning may make pandemic stressors less aversive by reducing their associated uncertainty. Recent research supports the idea that response to uncertainty acts as a mechanism of the inverse relation between meaning and distress (Ostafin et al., 2021).

A sense of comprehension may be particularly important in the pandemic as it may counteract uncertainty. Comprehension is conceptualized as the opposite of uncertainty which stems from adopting clear meaning frameworks (George & Park, 2016) and from leading a life that is coherently aligned with personal values and goals. In this way, comprehension may reduce behavioral and perceptual uncertainty (Hirsh et al., 2012), guide and limit attention (for a review, see Dijksterhuis & Aarts, 2010), and may allow individuals to focus on goal-relevant rather than distressing environmental cues. In previous research, a strong sense of coherence (including comprehension but also manageability and meaningfulness; Antonovsky, 1979) has been linked to lower distress in cancer patients (for a meta-analysis, see Winger et al., 2016), lower anxiety and depression in care givers (for a systematic review and meta-analysis, see Del-Pino-Casado et al., 2019) and lower PTSD (for a meta-analysis see Schäfer et al., 2019).

While most earlier studies assessed life meaning as a unidimensional construct, investigating the subcomponents of meaning may offer more differentiated insights regarding its beneficial effect. For example, research with a novel multidimensional meaning measure found that while all subcomponents were inversely associated with distress in students, comprehension was the strongest predictor (George & Park, 2017; Ostafin et al., 2021). To our knowledge, the individual contributions of the subcomponents have not been studied in stressor-exposed populations.

In addition to associations between meaning and psychological distress, another way of investigating the protective effect of life meaning is examining the extent to which it moderates the relation between aversive experiences (e.g., trauma load, impact severity) and distress. Meaning has shown moderation effects on the relation between stressors and negative outcomes in various contexts, including trauma exposure and depression in trauma-exposed elderly (Krause, 2007), bullying victimization and suicidal ideation in male students (Henry et al., 2014), and COVID-19-specific stress and distress (i.e., anxiety and depression; Schnell & Krampe, 2020). However, research in female bullying victims (Henry et al., 2014) and combat-exposed veterans (Blackburn & Owens, 2015) has shown no such effect. Reasons for this mixed evidence are unclear.

In addition to functioning as a protective role in the context of stressors, a meaning violations perspective suggests that meaning may also be weakened as a consequence of stressors, for example when the core beliefs or important goals in a meaning framework are violated (Janoff-Bulman & Frieze, 1983; C. L. Park, 2010). Such meaning violations may then increase distress (Janoff-Bulman & Frieze, 1983; C. L. Park, 2010). In sum, meaning may counteract distress during the COVID-19 pandemic but stressors may also reduce meaning.

To test the aforementioned conceptualizations, we conducted a longitudinal study and examined three hypotheses. H1: higher pre-pandemic meaning in life prospectively predicts less severe symptoms of depression, anxiety, and PTSD during the first months of the COVID-19 pandemic. H2: higher peri-pandemic meaning concurrently predicts less severe peri-pandemic depression, anxiety, and PTSD symptoms during the first months of the COVID-19 pandemic. H3: life meaning moderates the association between being negatively affected by the pandemic and distress outcomes. We also conducted exploratory analyses to examine (i) the extent to which the three meaning subcomponents accounted for unique variance of distress and (ii) a meaning violations perspective of the extent to which the pandemic reduced meaning in life.

Method

Participants

From 21 October 2019 to 3 March 2020, 347 first year psychology students (262 females; age: range = 17–29; M = 19.79; SD = 1.98) participated in a screening study (T0). Between April and June 2020 (T1), n = 112 follow-up questionnaires were completed. On average, T0 and T1 participation were 164 days apart (SD = 57.38; Range = 47–252). Four duplicate entries were removed from T0 and two from the T1, retaining the initial responses. One participant was excluded at T1 due to a missing identification code. The final study sample consisted of 109 participants (82 females; age: range = 17–29; M = 19.93; SD = 2.01), offering adequate statistical power to detect moderate effect sizes (e.g., for r = .3; at α = .05, power = .89; at α = .01, power = .73), such as found or exceeded in prior, partly prospective, life meaning and distress studies (e.g., Ostafin & Proulx, 2020; Schnell & Krampe, 2020). Most participants were Dutch (n = 57) or German (n = 32). Other nationalities included European (n = 14) and non-European nationalities (n = 6).

Materials

Pre-pandemic (T0) measures

Meaning in life. We assessed meaning in life with the meaningfulness subscale of the Sources of Meaning Questionnaire (SoMe; Schnell, 2009; current sample: alpha = .86). Participants rated their agreement with five statements (e.g., “I have a task in life”) on a five-point scale from “Not at all” (0) to “Very much” (4).

Anxiety. We measured pre-pandemic anxiety severity with the sum score of the General Anxiety Disorder Questionnaire (GAD-7; Spitzer et al., 2006; current sample: alpha = .89). Participants indicated the frequency of anxious symptoms (e.g., “Feeling nervous, anxious, or on edge”) over the two previous weeks on a four-point scale from “Not at all” (0) to “Nearly every day” (3). Different cut-off values for clinical relevance exist (≥8 in a meta-analysis by Plummer, Manea, Trepel, and McMillan, 2016; ≥12 in Dutch adults; Donker, van Straten, Marks, & Cuijpers, 2011).

Peri-pandemic (T1) measures

Meaning in life. Peri-pandemic meaning was assessed with two measures adjusted to refer to the last month. We used the same scale as at T0, the SoMe (Schnell, 2009; current sample: alpha = .86; e.g., “In the last month, I thought that there is meaning in what I do,” “In the last month, I had a task in life”), and added the Multidimensional Existential Meaning Scale, assessing purpose, comprehension, and mattering separately (MEMS; George & Park, 2017). In the MEMS, participants rated their agreement with 15 statements (five per subscale; e.g., “Over the last month, my life made sense” for comprehension; “Over the last month, I had certain life goals that compelled me to keep going” for purpose; “Over the last month, I was certain that my life is of importance” for mattering) on a seven-point scale from “Very strongly disagree” (1) to “Very strongly agree” (7). The entire scale (alpha = .89), and the subscales, were highly reliable (alpha_(purpose) = .89; alpha_{(comprehension)} = .88; alpha_(mattering) = .84).

Anxiety. As in the pre-pandemic questionnaire, we estimated anxiety severity over the two previous weeks with the sum score of the GAD-7 (Spitzer et al., 2006; current sample: alpha = .85).

COVID-19-related PTSD severity was measured with the PTSD Checklist 5 (Weathers et al., 2013) in which participants report how much 20 symptoms related to a specified index trauma had bothered them over the last month from “Not at all” (0) to “Extremely” (4). To study the impact of the pandemic, index events were limited to pandemic-specific experiences (e.g., “Feeling very upset when something reminded you of a specific experience related to the corona pandemic?”, “Repeated, disturbing, and unwanted memories of a specific experience related to the corona pandemic?”, “Trouble remembering important parts of a specific experience related to the corona pandemic?”). Clinically relevant cut-off values were 33 in German adults (Krüger-Gottschalk et al., 2017) and up to 37 in college students (Blevins et al., 2015).

Depression. Depression severity was assessed with the sum score of the Major Depression Inventory (Bech et al., 2001; current sample: alpha = .88). Participants indicated how commonly they experienced 10 symptoms (e.g., “Have you felt low in spirits or sad?”) over the last two weeks from “At no time” (1) to “All of the time” (6). Scores of at least 26 indicated clinically relevant depression in Danish adults (Bech et al., 2001) and Dutch outpatients (Cuijpers et al., 2007).

Negative impact of the pandemic was assessed with the item “Over the last month, how much has your life in general been negatively affected by events related to the coronavirus?”. Answer options ranged from “Not at all” (1) to “Very much” (7). On the same scale, participants rated (1) to what extent the coronavirus threatened their livelihood, how much they thought the coronavirus situation was (2) harmful and (3) relevant to them personally, and (4) how strictly they adhered to social distancing/quarantine. Moreover, since it has been posed that aversive experiences can lower meaning (e.g., by violating beliefs and goals) and create distress (e.g., C. L. Park, 2010), a meaning violations perspective was explored additionally. To this end, participants indicated (5) to what extent the pandemic interfered with the pursuit and accomplishment of three self-chosen goals from “Not at all” (1) to “Very much so” (6).

Procedure

The Ethical Committee Psychology of the University of Groningen approved this study. At each virtual assessment (through Qualtrics), participants gave informed consent. Participants received compensation (T0: partial course credit; T1: partial course credit (n = 74) or a financial compensation (n = 35)). The assessments lasted 5 and 45 min, respectively. Both assessment points included additional instruments not presented here. The study was preregistered at aspredicted.org under no. 52568.

Statistical analyses

Statistical analyses were conducted using IBM SPSS Statistics for Windows, Version 26. Power calculations were conducted in G*Power. H1 (that pre-pandemic meaning is inversely associated with peri-pandemic distress) was first tested with Pearson correlations. We applied the Bonferroni–Holm correction to reduce the family-wise type-1 error rate (i.e., the adjusted p-values were .0167; .025; .05). We subsequently tested H1 with semi-partial correlations, controlling for T0 anxiety. Controlling for T0 depression was not possible, as depression was not assessed at this time point. However, given that depression and anxiety are strongly related (e.g., Lamers et al., 2011; in the current study, r = .73 at T1), we controlled for T0 anxiety. Similarly, for the cross-sectional H2 that meaning is inversely associated with distress, Pearson correlations (with the adjusted p-values of .0083; .01; .0125; .0167; .025; .05) and semi-partial correlations controlling for T0 anxiety were investigated. The H3 that meaning buffers the association between being negatively affected and distress was tested with regression analyses with T0 anxiety as a control variable. Here, predictor variables were standardized. For the exploratory analyses regarding the meaning subcomponents and a meaning violations perspective, we used an alpha of .05, two-sided, without multiple testing corrections due to their exploratory nature (Streiner & Norman, 2011). To investigate the differential associations of the meaning subcomponents with distress cross-sectionally, we conducted multiple linear regression analyses. We controlled for T0 anxiety and T1 SoMe in subsequent regression analyses. To examine the meaning violations perspective, i.e., that stressors can challenge and reduce meaning, which fosters distress, we first investigated meaning changes from T0 to T1 and correlations between pandemic stressors (i.e., goal obstruction, being affected) and (i) T1 meaning and (ii) meaning change. Next, we examined whether T1 meaning and meaning change mediate the link between pandemic stressors and distress outcomes with model 4 of the PROCESS Macro for SPSS with bootstrapping (5000; Hayes, 2013). All analyses were executed separately per outcome. Model assumptions were assessed with plots.

Missing data, dropouts, and outliers

One missing T0 anxiety item value was replaced with the participant’s scale mean. Three participants had particularly high PTSD symptoms (more than 1.5* the interquartile range above the third quartile). As they scored within the common range of the “meaning” and “being affected by the pandemic” variables, they were not excluded from analyses. Examinations of scatterplots and PP-plots indicated no substantial deviations from model assumptions, allowing the planned analyses (Ernst & Albers, 2017). Since the participation intervals varied greatly, we explored whether amount of time between assessments moderated the link between T0 meaning and T1 distress. The amount of time did not moderate any of the correlations between T0 meaning and T1 outcome variables (p’s ranged from .459 to .769). Neither age (b = 0.064; p = .281), nor gender (b = −0.021; p = .938) or T0 meaning (b = 0.007; p = .770) predicted continued participation (109 out of initially 347 participants) in a logistic regression.

Results

Descriptive statistics

Before the pandemic, participants reported moderate to high meaning in life. Between 18.3% and 45.9% (depending on the cut-off value) experienced clinically relevant anxiety symptoms. During the pandemic, meaning (SoMe) was significantly lower than before it (Table 1; d = .88; t = −9.159; p < .001). Peri-pandemic anxiety did not significantly differ from T0 levels (d = .04; t = 0.384; p = .702) with 17.4% to 42.2% reporting clinically relevant symptoms. Additionally, depending on the chosen threshold, 5.5%–9.4% reported clinically significant PTSD symptoms and approximately 60.6% experienced clinically relevant levels of depression. Participants reported being moderately to strongly negatively affected by the pandemic, which was significantly associated with all distress outcomes (see Table 1). Moreover, participants rated their livelihoods as somewhat threatened (M = 2.87; SD = 1.67), practiced quite strict social distancing (M = 5.14; SD = 1.45), found the pandemic moderately personally relevant (M = 4.32; SD = 1.51) and controllable (M = 3.94; SD = 1.17), quite harmful (M = 5.24; SD = 1.15), and reported moderate goal obstruction (M = 11.50; SD = 3.25). Except controllability (r = –.12; p = .228) and social distancing (r = .15; p = .109), these factors were significantly linked to being more negatively affected (r’s from .28 to .35; p’s from <.001 to .003).

Table 1. Pearson correlations among key study variables.

Variables	M (SD)	1	2	3	4	5	6	7	8	9	10
1. Meaning T0	11.96 (4.49)
2. Anxiety T0	7.67 (4.63)	−.19*
3. Meaning T1 (SoMe)	7.91 (4.43)	.46**	−.06
4. Meaning T1 (MEMS)	62.06 (13.71)	.50**	−.26**	.68**
5. Comprehension T1	20.14 (5.33)	.35**	−.27**	.60**	.86**
6. Purpose T1	23.92 (5.28)	.44**	−.17	.54**	.85**	.62**
7. Mattering T1	18.01 (5.52)	.48**	−.23*	.58**	.84**	.57**	.57**
8. Depression T1	28.39 (9.01)	−.28**	.35**	−.49**	−.67**	−.65**	−.56**	−.50**
9. Anxiety T1	7.48 (4.53)	−.22*	.38**	−.36**	−.45**	−.54**	−.32**	−.29**	.73**
10. PTSD T1	17.06 (10.54)	−.16	.33**	−.42**	−.52**	−.59**	−.40**	−.33**	.77**	.66**
11. Affected T	4.36 (1.39)	−.06	.01	−.23*	−.17	−.23*	−.13	−.08	.28**	.30**	.32**

Note: N = 109. *p < .05. **p < .01. T0 refers to the assessment before the pandemic, T1 to the assessment during the initial wave in Europe.

Hypothesis testing

H1 – the prospective link between pre-pandemic meaning and peri-pandemic distress

Bivariate associations (Table 1) support our first hypothesis that higher T0 meaning is associated with lower T1 depression (p = .003) and anxiety (p = .023). The correlation between T0 meaning and T1 PTSD was not statistically significant (p = .100). When controlling for baseline anxiety, T0 meaning significantly predicted T1 depression, but not anxiety (Table 2, step 1 of the T0 SoMe). To explore the directionality of the link between meaning and distress further, we tested the inverse model (i.e., T0 distress predicting T1 meaning). In a regression model, T0 anxiety did not significantly predict T1 meaning (sr = .03; p = .772) but T0 meaning did (sr = .46; p < .001), further supporting the effect of meaning on distress rather than the reverse.

H2 – the relationship between peri-pandemic meaning and concurrent distress

Supporting our second hypothesis, higher T1 meaning (both the SoMe and MEMS measures) was significantly related with less T1 distress (Table 1; all p’s < .001.) and remained significant when controlling for T0 anxiety (Table 2, step 1 of both the T1 SoMe and T1 MEMS measures).

H3 – meaning as a moderator on the link between being affected and distress

Neither T0 (SoMe) nor T1 meaning (SoMe and MEMS) significantly moderated the relationships between being directly negatively affected by the pandemic and any of the distress outcomes, with p-values ranging from .083 to .976 (Table 2, step 2 of the T0 SoMe, T1 SoMe, and T1 MEMS measure). The interaction effects only uniquely explained up to 2.3% of the outcome variables’ variances (i.e., sr², Table 2). To investigate whether statistical power issues may account for insignificant findings, post-hoc power analyses were executed and indeed, statistical power to detect the interaction effects was low (≤.44). Moreover, the confidence intervals around the semi-partial correlations were fairly wide (see supplementary file), indicating a large degree of uncertainty regarding their effect sizes.

Exploratory analyses on the differential associations between meaning subcomponents and distress

All T1 meaning components (MEMS) were significantly correlated with T1 outcomes. As can be seen in Table 1, the bivariate correlations with comprehension were the strongest. With all three meaning components entered in one regression model, only comprehension consistently predicted unique variance of all distress variables (see Table 3, step 1) and purpose predicted unique variance of depression. Mattering did not predict unique variance of any of the distress variables. We conducted another regression analysis that additionally included baseline (T0) anxiety and the T1 SoMe meaning measure to examine if the subcomponent effects would explain distress variance above and beyond baseline distress and if they could account for overall meaning effect. As can be seen in Table 3, step 2, the comprehension and purpose effects remained significant while T1 SoMe did not predict outcomes, indicating that subcomponent effects accounted for overall meaning effects.

Table 2. Select output from regression analyses with the different meaning in life measures.

Meaning measure	Step	Predictor	Outcome
			Depression		Anxiety		PTSD
			sr; p	R²	sr; p	R²	sr; p	R²
T0 SoMe	1			0.16		0.15		0.1
		Meaning	−.22; .015		−.15; .099		−.10; .290
		T0 Anxiety	.30; .001		.34; <.001		.31; .001
	2			0.24		0.22		0.19
		Meaning	−.20; .019		−.13; .122		−.08; .373
		T0 Anxiety	.27; .002		.34; <.001		.29; .001
		Affected	.28; .001		.28; .002		.31; <.001
		Moderation	.15; .083		−.04; .649		.08; .349
T1 SoMe	1			0.33		0.25		0.25
		Meaning	−.47; <.001		−.34; <.001		−.40; <.001
		T0 Anxiety	.32; <.001		.36; <.001		.31; <.001
	2			0.35		0.28		0.3
		Meaning	−.41; <.001		−.28; .001		−.34; <.001
		T0 Anxiety	.32; <.001		.35; <.001		.30; <.001
		Affected	.17; .027		.22; .010		.22; .008
		Moderation	.00; .976		−.04; .668		−.09; .267
T1 MEMS	1			0.47		0.26		0.3
		Meaning	−.60; <.001		−.36; <.001		−.45; <.001
		T0 Anxiety	.18; .011		.27; .002		.20; .014
	2			0.49		0.3		0.35
		Meaning	−.55; <.001		−.32; <.001		−.40; <.001
		T0 Anxiety	.19; .007		.28; .001		.21; .008
		Affected	.17; .016		.22; .008		.21; .008
		Moderation	−.01; .937		−.03; .719		−.08; .316

Note: N = 109. R² stands for adjusted R². T0: pre-pandemic assessment. T1: peri-pandemic assessment. Moderation: interaction between meaning and being affected.

Table 3. Select output from regression analyses with the meaning subcomponents.

Step	Predictor	Outcome
		Depression		Anxiety		PTSD
		sr; p	R²	sr; p	R²	sr; p	R²
1			0.46		0.28		0.33
	Comprehension	−.34; <.001		−.42; <.001		−.41; <.001
	Purpose	−.16; .031		.02; .839		−.06; .454
	Mattering	−.09; .193		.02; .847		.02; .776
2			0.49		0.34		0.36
	Comprehension	−.26; <.001		−.32; <.001		−.32; <.001
	Purpose	−.14; .040		.03; .707		−.04; .571
	Mattering	−.05; .497		.07; .355		.07; .334
	T1 SoMe	−.08; .279		−.11; .168		−.11; .148
	T0 Anxiety	.18; .009		.26; .001		.20; .010

Note: N = 109. R² stands for adjusted R². T0: pre-pandemic assessment. T1: peri-pandemic assessment.

Exploratory analyses on a meaning violation perspective

As mentioned above, meaning decreased from pre-pandemic to pandemic assessment points (d = .88), indicating that the pandemic may have violated meaning. Similarly to being affected (see Table 1), pandemic-related goal obstruction was linked to lower T1 meaning (SoMe: r = −.46; p < .001; MEMS: r = −.35; p < .001), with larger effect sizes than being affected (Table 1). The relation between goal obstruction was especially strong with the comprehension subscale (r = −.43; p < .001; purpose: r = −.17; p = .084; mattering: r = −.30; p = .002). Goal obstruction was also related with greater distress (depression: r = .49; p < .001; anxiety: r = .46; p < .001; PTSD: r = .40; p < .001). Meaning change was not significantly linked with goal obstruction (r = .16; p = .105) or being affected (r = .17; p = .086) or anxiety (r = .14; p = .152) but was linked with depression (r = .19; p = .043) and PTSD (r = .24; p = .011). Subsequent mediation analyses investigating if meaning variables (change magnitude, T1 meaning) could account for link between pandemic stressors and distress, indicated that while T1 meaning was a significant mediator between being affected and all three distress outcomes as well as between goal obstruction and the three distress outcome variables, meaning change (T0-T1) was not (see supplementary file). Since comprehension was the meaning component most closely related to stressor and distress variables, comprehension was also investigated as a mediator. Also here, all indirect effects were significant. To investigate the directionality of these effects, we also tested models with meaning change, T1 SoMe, and T1 comprehension as predictors and stressors (i.e., being affected, goal obstruction) as mediators. Stressors partly mediated the links between T1 SoMe and comprehension and distress outcomes (see supplementary file), indicating that the relationship between the variables might be multidirectional.

Discussion

The current study investigated whether meaning in life had beneficial prospective (H1), concurrent (H2), and stressor-buffering (H3) effects on psychological distress during the first wave of the COVID-19 pandemic. Our findings partly supported the first hypothesis, fully supported the second, but did not support the third.

The link between pre-pandemic life meaning and peri-pandemic distress

Higher pre-pandemic meaning prospectively predicted lower peri-pandemic depression and anxiety. These results align with findings demonstrating a beneficial link between pre-incident meaning and aversive responses to other naturalistic stressors such as city flooding (Ostafin & Proulx, 2020; Study 2). Conversely, there was no prospective relation with PTSD. Secondarily, pre-pandemic anxiety was controlled for in the analyses because (a) pre-incident distress may predispose for post-incident distress (Brewin et al., 2000) and (b) anxiety is commonly inversely related with meaning cross-sectionally (e.g., Blackburn & Owens, 2015; Gravier et al., 2019). That is, the prospective meaning effects could potentially be inflated due to the relation between T0 meaning and T0 distress. Controlling for T0 anxiety, the results showed that pre-pandemic meaning continued to significantly predict peri-pandemic depression but not anxiety. This finding suggests that the prospective relation between T0 meaning and T1 anxiety might have been overestimated. Next, we tested the inverse model where T0 anxiety predicted T1 meaning to gain additional insight into the directionality of effects. Anxiety did not significantly prospectively predict meaning and the effect was smaller than of the prospective effect of meaning on anxiety, suggesting that a model in which meaning influences subsequent distress is more likely than the inverse model. Future research should include prospective analysis for all distress outcomes as well as implement experimental studies to complement these findings.

Meaning in life did not prospectively predict T1 PTSD and anxiety when controlling for T0 anxiety. One possible explanation is that compared to pre-incident meaning, meaning during and after a crisis might be more closely associated with these outcomes. A reason for this is that aversive experiences can at times violate meaning frameworks (by contradicting beliefs or obstructing goals), thereby reducing meaning and eliciting distress (Janoff-Bulman & Frieze, 1983; C. L. Park, 2010). In such cases, post-event distress may be more strongly related with post-event meaning compared to pre-event meaning because both post-event measures are elicited by the same event. This perspective is supported by findings that the associations between meaning and distress in peri-pandemic studies (Schnell & Krampe, 2020; current study: anxiety r = −.36) are often stronger than associations in pre-pandemic studies (Schnell, 2009; Sørensen et al., 2019; current study: anxiety r = −.19).

Given that depression was the only distress variable predicted prospectively when controlling for baseline anxiety, it may be that meaning is particularly related to depression-related distress. Previous meaning research using the SoMe (e.g., Schnell, 2009; Sørensen et al., 2019) and other meaning measures (e.g., Gravier et al., 2019; Jaarsma et al., 2007) has similarly found stronger links with depression than anxiety, indicating that meaning might be more closely related with some forms of distress than others. One possible explanation why meaning may be particularly strongly related with depression might be that a sense of comprehension counteracts maladaptive rumination and thereby depression (Nolen-Hoeksema et al., 2008). Additionally, individuals experiencing meaning may have a “why to live” (Frankl, 1985), a purpose that provides direction, motivation and rewarding experiences (McKnight & Kashdan, 2009) and thereby counteracts depression (e.g., Lejuez et al., 2001). Indeed, in the current study, the meaning components of comprehension and purpose explained unique variance in depression in cross-sectional models including the other meaning components while for other distress outcomes, only comprehension did (Table 3).

The concurrent relationship between life meaning and distress

Meaning in life during the pandemic was consistently inversely related with distress for both meaning measures, even when correcting for multiple testing and controlling for pre-pandemic anxiety. Since cross-sectional results do not allow causal inferences, it may, however, alternatively be that meaning and distress are associated in other ways. For example, a depressed individual’s mental state may contradict prior held beliefs about the self (e.g., being a happy, fun-loving person) or negative affect and behavioral inactivation may hinder purposeful goal pursuit, hence decreasing meaning. The current findings align with previous studies linking meaning to distress in stressor-exposed populations (e.g., Blackburn & Owens, 2015; Feder et al., 2013; Jaarsma et al., 2007).

Meaning as moderating the effects of aversive events on distress

The current results did not support a model in which life meaning moderates the effects of aversive events on psychological distress. Investigating stressor-moderating effects may be relevant because the severity of aversive events has been linked to symptom severity (for a meta-analysis on PTSD, see Brewin et al., 2000). In the current sample, being negatively affected by the pandemic was only moderately related to participants’ distress. Since the pandemic stressors are multifaceted (e.g., health threat, economic implications, personal restrictions, and losses) and hence difficult to comprehensively assess, the current open phrasing (“Over the last month, how much has your life in general been negatively affected by events related to the coronavirus?”) was chosen. The open phrasing may, however, have introduced measurement error due to interpretation differences, especially since few participants were native English speakers and may have led to the comparatively weak associations. Related, being affected ratings were significantly related with some potential pandemic-related stressors (livelihood threat, goal obstruction, negative evaluations), but not with social distancing, another potentially important pandemic stressor (e.g., Rossi et al., 2020; Varma et al., 2021). Hence, the current impact assessment may not fully reflect pandemic stressors. The potential of using a suboptimal stressor assessment may qualify the interaction effect results. More objective, reliable assessment of the negative effects of stressors might have increased the statistical power (Aguinis, 1995), altered results and is recommended for future research. In contrast with earlier findings (Chao et al., 2020; Krause, 2007; Schnell & Krampe, 2020) but aligning with others (e.g., Blackburn & Owens, 2015), the current results showed no interaction between being negatively affected and meaning on the distress variables. However, stressor exposure assessment differed greatly between studies, impeding comparisons of the findings. In sum, evidence for this effect is mixed and future research is needed to clarify under which conditions meaning may moderate the association between stressors and distress.

The concurrent relationship between meaning subcomponents and distress

In line with earlier findings (George & Park, 2017; Ostafin et al., 2021) comprehension consistently emerged as a significant predictor of distress when examining the unique variance of mental distress accounted for by the three meaning components in exploratory analyses. This finding aligns with meta-analyses demonstrating lower distress in stressor-exposed individuals with a strong “sense of coherence” (for meta-analyses, see Del-Pino-Casado et al., 2019; Schäfer et al., 2019; Winger et al., 2016). In a regression model with comprehension, the other scales of purpose, mattering and overall meaning (i.e., SoMe) no longer predicted anxiety and PTSD. This suggests that cross-sectional links between meaning and distress may have been largely driven by comprehension.

Additionally, purpose predicted lower depression, even in models including comprehension, overall meaning (SoMe), and pre-pandemic anxiety, supporting the idea that purpose and the associated goal pursuit and motivation may counteract depression. In line with this, pursuing meaningful goals is seen as a core working mechanism against depression in behavioral activation therapy (Lejuez et al., 2001) and behavioral activation and reward sensitivity coincided with higher purpose in students (George & Park, 2017). Future research on long-term trajectories of meaning, meaning subcomponents, and distress might elucidate the causality of these associations, advancing the insights into the current findings.

A meaning violations perspective

In addition to the current study’s main model of meaning reducing distress, we also explored a meaning violations perspective in which stressors may reduce meaning, which fosters distress (Janoff-Bulman & Frieze, 1983; C. L. Park, 2010). A meaning violations perspective is supported in this sample by meaning declining from pre-pandemic to pandemic assessment points (though there was not a concomitant increase in anxiety) and the amount of decline in meaning correlating with depression and PTSD severity. Moreover, pandemic stressors (i.e., being generally negatively affected, goal obstruction) were related with greater meaning declines, though these correlations were not statistically significant. Links between stressors, peri-pandemic meaning, and distress further support this perspective. Moreover, that peri-pandemic meaning and comprehension partially mediated the link between being negatively affected by the pandemic and psychological distress further supports this perspective. Conversely, meaning changes did not mediate the link between being affected and the distress outcomes, drawing a meaning violations perspective into question. Potentially, the stressor-related meaning violation results (i.e., low T1 meaning) might play a larger role in distress than the violation magnitude (i.e., meaning decrease). These findings add to the meaning violations literature and point to the importance of examining meaning during and after a stressor. If, as suggested by these findings, meaning violations may partly account for pandemic-related distress, incorporating meaning making into supportive interventions may counteract distress. Moreover, (post-pandemic) meaning restoration, their link with distress, and meaning violations (e.g., results vs magnitude) during other stressors may be promising topics for future studies.

Strengths and limitations

This study was, to the best of our knowledge, among the first to prospectively and cross-sectionally test the links between meaning and psychological distress during a pervasive, persistent stressor. Moreover, it investigated the separate meaning subcomponents, offering novel insights regarding their relative importance. Furthermore, controlling for pre-pandemic anxiety and implementing false-positive error correction reduced bias as well as the chance of committing Type I errors.

In addition to the study strengths, the research has several limitations. Reliance on a student sample prevents generalizing the results to other populations. However, as students appear vulnerable to peri-pandemic distress (e.g., Elmer et al., 2020; Fried et al., 2021) and younger adults reported greater pandemic distress than older ones (Varma et al., 2021), this might represent a particularly vulnerable group and thus one worth studying. Additionally, the breadth of the pre-pandemic assessment was limited. If future research replicated current results with all relevant predictors assessed at both timepoints, stronger inferences could be made (e.g., when controlling for baseline depression when prospectively modeling depression). Moreover, causal claims cannot be made due to the correlational design of the study. Lastly, the sample size may have been insufficient to detect small effects (Cohen, 1992). For example, the interaction effects were small – uniquely explaining only up to 2.3% of distress outcomes’ variances. Therefore, especially the interaction findings must be interpreted with caution and should be examined in larger samples.

Recommendations for future research and clinical practice

The current study offers several promising insights that would benefit from replication and extension. For example, since the pandemic’s societal impact and scope were rather unique, it is unclear if current findings will replicate in the context of other naturalistic stressors. Especially the prospective relationship between meaning and distress and the distinct associations between comprehension and purpose and distress contribute to and may inspire future longitudinal and experimental research that could ultimately inform clinical practice. Future research may benefit from examining the role of meaning in the development of and the recovery from stressor-related mental health issues.

Conclusion

Experiencing life as meaningful was inversely related with psychological distress during the COVID-19 pandemic. Prospectively, greater pre-pandemic meaning in life was most strongly associated with less depression during the pandemic. The prospective association between meaning and anxiety during the pandemic became insignificant when accounting for pre-pandemic anxiety. Cross-sectionally, meaning and particularly its subcomponent comprehension was inversely associated with depression, anxiety, and COVID-19-specific PTSD. On the other hand, meaning levels decreased between assessments and peri-pandemic meaning partly mediated the link between being affected by the pandemic and distress, indicating that the pandemic might have entailed meaning violations. The current research shows that meaning may function as a protective factor in times of crisis yet may also be decreased by aversive experiences.

Disclosure statement

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