Inquiry on Threats of Evil within the Hostile-World Scenario: Emerging Content and Mental Health Concomitants Among Holocaust Survivors

ABSTRACT

Exposure to human evil, referring to malevolent deeds that deliberately inflict suffering or death, can be psychologically traumatic. This study examined self-perceived evil-related threats within the conception of hostile-world scenario (HWS) that signifies one’s mental representation of major threats in life. The study explored whether evil-related threats, along with HWS, differentiated Holocaust survivors from comparisons, and how these concepts related to mental health. Participants were Israelis aged 58–93, including 220 Holocaust survivors and 205 non-Holocaust survivors. Participants completed measures of mental health (neuroticism, positive and negative affect, life satisfaction, depressive symptoms) and HWS. Evil-related threats were assessed by a 7-item Evil Scale separated from the HWS Questionnaire. Evil-related threats contained two factors, fear of human violence and interpersonal distrust, and correlated with lower mental health. With age and gender controlled, Holocaust survivors were higher than comparisons on evil-related threats and negative engagement with HWS. Beyond associations of HWS (in its negative and positive engagement modes) with mental health, evil-related threats related to higher depressive symptoms among survivors. Results suggest that early traumatic experiences related to perceived threats in later life. Evil-related threats complemented the HWS in associating with mental health. The findings bear implications in approaching victims of evil and trauma.

KEYWORDS:

• Evil
• Holocaust survivors
• hostile-world scenario
• mental health
• trauma

The phenomenon of evil, referring to malevolent deeds that deliberately inflict suffering or death, has haunted humanity since the dawn of religions and philosophy (Nys & de Wijze, 2019). People hardly ignore dangers of falling victim to human evil, whether in the social surroundings of their mundane lives or in extreme realities of atrocities such as those occurring by war and terrorism. This study sought to explore how older individuals perceived evil-related threats that could possibly affect them, and how such threats were related to their mental health. Threats of evil were examined within the conceptual model on the pursuit of happiness in a hostile world (Shmotkin, 2005), which explicates how individuals may manage their well-being in the face of actual or potential adversities in life. In this study, the notion of evil within a hostile world was addressed in a pertinent group, namely Holocaust survivors, who experienced an atrocious evil earlier in their lives.

Threats of evil and the hostile-world scenario

The definition of evil is uneasy for social scientists because this term is laden with metaphysical and moralistic positions about good and bad, which differ among faiths, cultures and historical periods (Rorty, 2001). Additional confusion lies in the split between perspectives of victims and perpetrators. Notoriously, most people who commit evil acts do not regard their actions as evil (Baumeister, 1997). The horrible experience of World War II suggested that deviant personalities of individuals, predisposed to unjustified cruelty, could not adequately explain large-scale evil (Berkowitz, 1999). Thus, historians showed that even the harshest evil behaviors, in certain social and psychological circumstances, were doable by ordinary people who represented large sections of the population (Browning, 2017; Goldhagen, 1996). Famous experiments on obedience to authority (e.g., Milgram, 1974) and endorsement of cruelty through role-playing (e.g., Zimbardo, 2007) exposed conditions where regular students were ready to inflict suffering on their fellows. The contention about the “banality of evil” (Arendt, 1963), originally referring to top Nazi perpetrators of the Holocaust, explicated that sane and reasonable people committed evil out of a breakdown of personal judgment or mere thoughtlessness, thus bringing evil from the arena of historic catastrophes into everyday life. This was indeed the rationale for exploring the link between past trauma perpetrated by extreme evil and present threats of evil as perceived in regular life. This link appears viable as people who were traumatized by evil may live in a mind-set of “afterwardsness,” where the inconceivable trauma urges them to restore order and mastery by monitoring threats that resemble their trauma, thus trying to reprocess the event within their current life (Lazar, 2016).

The current study did not focus on general conceptions of evil, but rather on perceptions that trauma survivors had about threats that human evil might still impose on them. Thus, evil-related threats came up as a distinct theme in mapping the composition of the Hostile-World Scenario (HWS), defined as an image of actual or potential self-perceived threats to one’s life, or more broadly, to one’s physical or mental integrity. This concept of HWS is part of a model (Shmotkin, 2005; Shmotkin & Shrira, 2012, 2013) that presents the dialectical imperative of pursuing happiness in a hostile world, postulating that individuals address this imperative by reciprocal relations between the HWS and positive psychological systems, mainly subjective well-being and meaning in life. These systems prove to complement, or compensate for, each other in enhancing positive functioning in the face of challenge (Keyes et al., 2002; Shrira et al., 2011b). The HWS represents the multiplicity of the human condition, including threats to vital life domains (health, close relationships, work and economic status), harsh afflictions (war, terrorism, disasters, crime) and challenging life outcomes (aging, disability, loss and death). As a dynamic construct of self-beliefs, the HWS serves to scan for dangers, and was thus found to detect changes in physical and mental functioning among older adults as occurring in the past recent years (Lifshitz et al., 2020) as well as to predict changes in health outcomes occurring in subsequent years (Shmotkin et al., 2016). The HWS also showed sensitivity to vulnerabilities of particular populations, mainly in a context of experienced adversities (Shenkman & Shmotkin, 2013; Shrira, 2015).

Data of HWS showed that individuals could react in two modes: negative engagement with representations of their HWS (HWS-NE; i.e., losing competence in view of life adversities, such as expressing fear or preoccupation with particular threats) and positive engagement with such HWS representations (HWS-PE; i.e., gaining competence in view of life adversities, such as expressing resolve or proaction regarding particular threats). Although these HWS engagement modes appear conversely related, they may complement each other (Bergman et al., 2021; Shrira, 2015; Shrira et al., 2011b). Hence, this study targeted evil-related threats as germane to survivors of the Holocaust trauma, but examined this theme of threats within a larger perspective where the HWS represented other major threats and exposed engagement modes reactive to these threats.

Holocaust survivors and their coping with the hostile world

Holocaust survivors were victims of an extreme realization of a hostile world, where the Nazi regime in World War II acted to exterminate the Jewish people by a vastly organized scheme of dehumanization, persecution and mass murders. The Holocaust was framed as an utmost eruption, and indeed a paradigm, of human evil (Garrard, 1998). Hence, it was intriguing to explore how survivors considered evil and the HWS in their current lives, even decades after their trauma.

In studies during the first decades after World War II, Holocaust survivors showed a variety of pathological dysfunctions and psychological suffering, mainly in clinical settings (e.g., Eitinger, 1964; Krystal, 1968). Later studies, mainly in community samples from the 1980s onwards, yielded a more complex picture. Overall, evidence showed that Holocaust survivors, long after the trauma, still suffered from various forms of emotional distress and posttraumatic reactions, and yet they showed remarkable coping and adjustment in major functioning domains (for reviews, see Barel et al., 2010; Kahana et al., 2005; Kellermann, 2009; Zimmermann & Forstmeier, 2020).

In reviewing research from large community studies, Shmotkin (2003) and colleagues (Shmotkin et al., 2011) characterized the adaptational status of Holocaust survivors within a model of general resilience alongside specific vulnerabilities. In this view, most survivors manifested effective capabilities that enabled them to ultimately reach positive adaptation and lead essentially normal lives. Yet, these resilient outcomes did not preclude specific vulnerabilities that acted simultaneously, sometimes covertly, in distinct psychosocial domains. This model proved useful in interpreting long-term traumatic effects among aging Holocaust survivors who were compared to controls that did not experience the Holocaust (Carmel et al., 2016; Shmotkin & Lomranz, 1998; Shmotkin et al., 2003; Shrira et al., 2010, 2011a). Specific vulnerabilities of survivors typically emerged in highly stressful situations such as war and terrorism (Dekel & Hobfoll, 2007; Solomon & Prager, 1992) and health crises (Hantman & Solomon, 2007; Kimron & Cohen, 2012). Such an integrative view of the survivors’ adaptation indicated that posttraumatic distress and posttraumatic growth coexisted (Greenblatt-Kimron et al., 2019; Lev-Wiesel & Amir, 2003).

Research found variability in the way Holocaust survivors kept their trauma psychologically distanced in their past, or, rather, had the trauma merging into present life (Shmotkin & Barilan, 2002; Shrira & Shmotkin, 2008). This issue appeared acute while facing the challenges of old age (Palgi et al., 2015), as increasing vulnerabilities exacerbated susceptibility to being abused by evil deeds (Lowenstein, 2010). Thus, despite scarcity of longitudinal studies on Holocaust survivors, there was evidence suggesting the notion of reactivation, where older survivors reacted with expanded stress when confronting situations that reminded them of their trauma (Palgi et al., 2021; Shrira et al., 2020). Reactivation of traumatic remnants involves coping mechanisms, which had previously suppressed the traumatic memories, but later gave access to such memories, often joined by a sense of a current threat (Hiskey et al., 2008).

The aforementioned research warrants an expectation that Holocaust survivors, when relating to their current life, would express stronger evil-related threats as well as, in a broader view, a stronger negative engagement with their HWS. Such expectation is pertinent to studies that found Holocaust survivors, compared to controls, inclined to less positive assumptions about their world, even though they could be similar or even more positive in particular worldviews (Brom et al., 2002; Cassel & Suedfeld, 2006; Palgi et al., 2011; Prager & Solomon, 1995). These mixed results were mainly based on positively-phrased worldviews, such as Janoff-Bulman’s (1989) world assumptions (comprising beliefs about benevolence, meaningfulness and self-worth) and Antonovsky’s (1987) sense of coherence (comprising beliefs about comprehensibility, manageability and meaningfulness). The current study enlarges this body of research by relating to negatively-phrased themes whereby survivors describe their world in terms of threats stemming from human evil or otherwise subsumed within the HWS at large. The study examines whether these threats are associated with both positive and negative indicators of mental health.

Overview and study hypotheses

Among the numerous ramifications of evil, this study addresses a vital need to assess how evil-related threats are associated with individuals’ mental health, especially when evil had a major part in a trauma endured earlier in life. Conceptually, evil-related threats are presented as a pertinent theme within the HWS model, which refers to self-perceived threats to one’s life or integrity. The study examines whether evil-related threats exert any unique effects beyond the ones exerted by the more inclusive construct of HWS. The study’s design compared Holocaust survivors with counterpart older adults who did not experience the Holocaust. The first objective in this study was to delineate evil-related threats out of the HWS framework and explore whether such self-perceived threats formulated a meaningful Evil scale. Such a scale was expected to serve as a workable assessment tool among Holocaust survivors and comparisons.

Given that an Evil scale could be substantiated, the second objective was to test hypotheses associating evil-related threats and the prototypical, negative and positive engagement modes of HWS (HWS-NE and HWS-PE) with mental health among Holocaust survivors and comparisons. Mental health was assessed by essential dimensions of dispositional vulnerability (neuroticism), well-being (positive affect, life satisfaction), and distress (negative affect, depressive symptoms). It was hypothesized that: (1) Stronger evil-related threats would be conversely associated with positive mental health; (2) Holocaust survivors would reveal stronger evil-related threats than comparisons; (3) Holocaust survivors would differ from comparisons in a larger array of pertinent variables, having higher HWS-NE, lower HWS-PE, and less positive mental health; (4) In addition to main effects of HWS-NE and HWS-PE on mental health (in negative and positive directions, respectively), evil-related threats would reveal a particularly negative association with mental health among Holocaust survivors rather than comparisons.

Method

Sample

The sample included 425 older adults. Participants were recruited and interviewed in a wide array of sites such as community institutions, workplaces, commercial centers, sheltered housing and private homes in various parts of Israel. Interviews took place during 2002–2014. The age range was 58–93 years (M = 74.5, SD = 8.5). Of the sample, 64.5% were women and 54.4% were married. Participants were classified into two groups. (1) Study group (n = 220) included Holocaust survivors who reported to have lived under Nazi or pro-Nazi rule in Europe at some time between 1939–1945, and later immigrated to Israel (median year of immigration 1948). Queries about their Holocaust-related experiences indicated multiple categories, mostly concentration camps, forced-labor camps, ghettos, and living in hiding. (2) Comparison group (n = 205) included individuals that did not live under Nazi or pro-Nazi rule between 1939–1945, who had been born either in pre-state Israel (55%) or elsewhere (median year of immigration 1949). Excluded from the comparison group were individuals who reported their parents to have been Holocaust survivors.

Sociodemographic characteristics of the two groups are presented in Table 1. As shown, the survivor group was older and included relatively more women. While the sample was set to include participants born up to the end of World War II, more comparisons, than survivors, were born during the War, thus making the comparison group relatively younger. Both age and gender (that possibly indicated the higher longevity of women in the older group) served as covariates in subsequent analyses. Apparently reflecting the older age as well as characteristic life histories of the survivors, Table 1 shows that survivors, relative to comparisons, were almost entirely born in Europe (a few were born elsewhere but moved to Europe prior to the Holocaust), had a lower level of formal education, were more frequently unmarried (mostly widowed), and gave lower ratings of health status. No differences were found in economic status and religiousness.

Table 1. Background characteristics of the study groups.

Variable	Holocaust survivors (n = 220)	Comparison group (n = 205)	Difference test t/χ^2
Age M SD Range	79.3 6.4 61–93	69.3 7.5 58–93	t(423) = 14.8***
Gender (%) Women Men	69.1 30.9	59.8 40.2	χ^2(1,424) = 4.0*
Geographic origin (%) Pre-state Israel Europe Asia/Africa/Australia	2.3 95.5 2.2	55.0 24.8 20.2	χ^2(1, 422) = 146.6***
Education^a M SD	3.1 1.5	3.9 1.3	t(420.1) = 5.6***
Marital status (%) Not currently married Married	56 44	34.1 65.9	χ^2(1,423) = 20.3***
Economic status^b M SD	3.3 0.9	3.3 0.8	t(421) = 0.6
Religiousness (%) Not religious Religious	63.5 36.5	70.2 29.8	χ^2(1,424) = 2.2
Health status^c M SD	3.1 1.1	3.7 1.0	t(422) = 6.1***

^aRanging from elementary (1) to academic (5). ^bSelf-rated on a 5-point scale from bad (1) to very good (5). ^cSelf-rated on a 5-point scale from bad (1) to very good (5).

*p < .05. **p < .01. ***p < .001.

Measures

Hostile-World Scenario (HWS) Questionnaire

This measure, based on Shmotkin’s (2005) conceptualization, contains 72 items assessing one’s beliefs on hostile-world representations (e.g., mass disasters, accidents, illness, war, loss of loved ones, economic deprivation, aging, death). Items had been refined from in-depth, open-ended interviews about common beliefs relating to the HWS concept. The measure reflects two dimensions of HWS: negative engagement (HWS-NE) and positive engagement (HWS-PE). Thus, HWS-NE implies the weakening of the individual’s sense of competence as a result of encountering HWS representations (e.g., “I always expect the possibility that the worst things will happen;” “Every encounter with the death of human beings arouses great fear in me”). In parallel, HWS-PE implies the maintenance or strengthening of the individual’s sense of competence as a result of encountering HWS representations (e.g., “Successful coping with dangerous situations strengthens me;” “I am not afraid of growing old”). In both HWS-NE and HWS-PE, items may reflect affective, cognitive, or behavioral reactions to HWS themes. Respondents rated how well each item applied to them on a scale from 1 (I strongly disagree) to 5 (I strongly agree). The respondent’s scores were the mean ratings in reduced sets of 19 items for HWS-NE (16 in the current analyses as 3 items were included in the Evil Scale, next) and 21 items for HWS-PE. These sets had proved reliable and valid in assessing the two respective dimensions. A shorter, 40-item version of the HWS Questionnaire, containing only these two item sets, was occasionally used for the sake of brevity (in the current sample it was completed by 92 Holocaust survivors). In the current data, Cronbach’s alpha coefficients for HWS-NE and HWS-PE were 0.86 and 0.78, respectively. For further examples of using the HWS Questionnaire, see Shrira (2015), Shrira et al. (2011b), and Shenkman and Shmotkin (2013).

Evil Scale

This measure was derived from the full HWS Questionnaire for assessing evil-related threats. It consists of seven items concerning perceived threats of being exposed to evil or evil-laden conditions (see items in Table 2). Three judges, acquainted with the HWS model, agreed upon the choice of items that corresponded to operational criteria of perceived evil. Criteria included the perpetration of indiscriminate aggression (war, terrorism), imposing unlivable conditions, and hurtful wrongdoings conducted deliberately (e.g., cruelty, betrayal, injustice). Notably, the items in the Evil Scale were not part of the item subsets that yielded, for this study, the scores of HWS-NE and HWS-PE. Items were rated on a scale from 1 (I strongly disagree) to 5 (I strongly agree). Alpha coefficient in the current data was 0.64. As this scale is a focal measure in this study, its factorial analysis and consequent scores are reported in Results, next.

Table 2. Exploratory factor analysis of the Evil Scale.

Item	Fear of human violence	Interpersonal mistrust
29. I always live in fear that there will be more and more wars	0.84	−.06
48. Sometimes I think it’s not worth living in this horrible world	0.55	0.22
58. I’m very scared of terrorism	0.80	−.03
67. It scares me to think how human beings’ cruelty can hurt me	0.61	0.32
17. It occurs to me that somebody I like might betray our relationship	−.08	0.78
36. There is a possibility that the police – or other government authorities – will sue me with false allegations	0.08	0.66
71. I know from personal experience that I can’t count on getting help from others, even close family members or friends	0.25	0.52
Explained variance (%)	32.27	18.44

Note. N = 299 from both groups, with listwise deletion of missing values. Item numbers are taken from the original HWS Questionnaire.

Big Five Inventory (BFI)

This measure allows an efficient assessment of the Big Five personality dimensions: neuroticism, extraversion, openness to experience, conscientiousness and agreeableness (John & Srivastava, 1999). It consists of 44 items, each with one or two prototypical adjectives corresponding to one of the traits measured. Respondents rated their agreement as to how each item depicted someone like themselves, on a scale ranging from 1 (strongly disagree) to 5 (strongly agree). Out of this Big Five assessment tool, this study used the 8-item Neuroticism scale, which taps negative emotionality (e.g., “worries a lot;” “gets nervous easily”), thus serving as a major negative indicator of mental health (Lahey, 2009). The score was the respondent’s mean rating for the Neuroticism items. Alpha coefficient in the current data was 0.82.

Satisfaction with Life Scale (SWLS)

This measure was designed by Diener et al. (1985) to assess the cognitive aspect of subjective well-being. It contains five items with general judgments of one’s life (e.g., “The conditions of my life are excellent”). Respondents rated their agreement with each item on a scale from 1 (strongly disagree) to 7 (strongly agree). The score was the respondent’s mean rating. Alpha coefficient in the current data was 0.84.

Affect Balance Scale (ABS)

This measure was designed by Bradburn (1969) to assess the affective constituents of subjective well-being, namely positive affect (PA) and negative affect (NA). Each is measured by five items that describe recent emotional experiences, such as “pleased” and “proud” (PA) or “upset” and “lonely” (NA). Using a scale from 1 (never) to 4 (often), respondents rated how often these feelings occurred to them in the past week. The respondent’s scores were the mean ratings of the respective PA and NA items. In the current data, alpha coefficients for PA and NA were 0.69 and 0.70, respectively.

Center for Epidemiological Studies – Depression Scale (CES-D)

Constructed by Radloff (1977), this is a widespread measure of depressive symptoms, involving factors of depressed affect, lack of positive affect, somatic symptoms and interpersonal distress. Respondents rated the frequency they experienced, in the past week, each of 20 symptoms (e.g., “I felt depressed”), using a scale from 1 (rarely or none of the time) to 4 (most or all of the time). After reversing four items’ ratings about positive feelings (thus indicating lack of positive affect), the respondent’s score was the mean rating for all items. Alpha coefficient in the current data was 0.86.

Procedure

As part of a larger research on well-being, university students conducted a convenience sampling in heterogeneous sites. The interviewers presented the aim of the research to the participants, explaining that participation was anonymous and voluntary, and obtained the participants’ informed consent. Debriefing was performed after completion of the questionnaires. The research was approved for ethical requirements by the institutional review board at the authors’ university. All measures that had originally appeared in English were administered in Hebrew versions following a systematic translation and validation in previously published studies. The current study included only participants who were administered the HWS Questionnaire.

Results

Factorial structure of the Evil Scale

Exploratory factor analysis (EFA) of the Evil Scale, employing principal component extraction with Varimax rotation (by SPSS 25), yielded a 2-factor solution (eigenvalue >1). Table 2 shows the items’ loadings on the two factors. Factor 1, labeled fear of human violence, included four items referring to participants’ dread of being hurt by violence or horror as inflicted by war, terrorism, human cruelty, and the horrible world in general. Factor 2, labeled interpersonal mistrust, included three items reflecting inability to trust people’s goodness, whether of close ones as family and friends or of authoritative agents in police and government. Factor 1 explained 32.27% of the variance and Factor 2 additionally explained 18.44% (their respective alpha coefficients were 0.70 and 0.40). When merely a 1-factor solution was imposed, factor loadings above .40 were found for items 29, 48, 58, 67 and 71, and the explained variance was 32.27%.

Next, confirmatory factor analysis (CFA) was conducted with maximum likelihood estimates (by AMOS 25). Following common guidelines (Hu & Bentler, 1999; Sun, 2005), goodness of fit was assessed by gathering absolute (chi square, GFI, AGFI), incremental (IFI, CFI), and noncentrality-based (RMSEA) indices (see the fully spelled terms in Table 3). Missing values were replaced by the respective item’s mean. As shown in Table 3, the least acceptable fit was found for model A, which included one underlying dimension with no associations allowed between measurement errors. A considerably better fit was obtained when two associations between measurement errors were included (between items 17 and 36 and between items 17 and 71). Model C and D examined the two factors obtained in the EFA. An acceptable fit was found for model C (without associations of measurement errors). The best fit was obtained for Model D that included an association between the measurement errors of items 48 and 58, yielding acceptable values close to 0.95 for the indices of AGFI, IFI and CFI, and close to .06 for RMSEA.

Table 3. Goodness-of-fit indices obtained in confirmatory factor analysis for four models of Evil Scale.

Model	χ ^2	df	p	GFI	AGFI	IFI	CFI	RMSEA
Model A (1 factor)	72.89	14	.000	.95	.89	.84	.84	.100
Model B (1 factor with correlated errors)^a	38.59	12	.000	.97	.94	.93	.93	.072
Model C (2 factors)	43.84	13	.000	.97	.94	.92	.92	.075
Model D (2 factors with correlated errors)^b	33.41	12	.001	.98	.95	.94	.94	.065

Note. N = 425. Goodness of fit indices are chi-square (χ ²), Goodness of Fit Index (GFI), Adjusted Goodness of Fit Index (AGFI), Bollen’s Increment Fit Index (IFI), Bentler’s Comparative Fit Index (CFI), and Root Mean Square Error of Approximation (RMSEA).

^aCorrelation of measurement errors was allowed between items 17 and 36, and between items 17 and 71 (see items in Table 2). ^bCorrelation of measurement errors was allowed between items 48 and 58 (see items in Table 2).

In sum, the conjunction of EFA and CFA supported a 2-factor structure of the Evil Scale. In the final model, the correlation between the two latent factors was 0.33. Nevertheless, the results did not preclude the use of the full Evil Scale by an overall score, although its internal consistency was only marginally satisfactory. Hence, we opted to use the overall score for the Evil Scale while considering also the separate scores for the two respective factors, mainly in order to capture a core sense as well as more nuanced themes of evil-related threats. In support of the first hypothesis, the Evil Scale scores largely maintained the expected relations with mental health indicators. That is, overall Evil positively correlated with neuroticism (rs were 0.48 and 0.43 in the respective groups of Holocaust survivors and comparisons, p < .001), negative affect (rs were 0.39 and 0.36, respectively, p < .001), and depressive symptoms (rs were 0.46, p < .001, and 0.28, p < .05, respectively). Only in the comparison group, significantly negative correlations of overall Evil were obtained with life satisfaction (r = −.27, p < .001) and positive affect (r = −.18, p < .05).

Comparing the study groups on evil-related threats, HWS, and mental health

A series of analysis of covariance (ANCOVA), controlling for age and gender, tested differences between the Holocaust survivors and their comparisons in the study variables (see Table 4). Supporting the second hypothesis, overall Evil was higher among Holocaust survivors than the comparison group. Yet, this result appeared due to a difference in Evil’s Factor 1 (fear of human violence) whereas no difference was found for Factor 2 (interpersonal mistrust). In partial support of the third hypothesis, Holocaust survivors, relative to comparisons, had a higher HWS-NE, yet they did not differ on the complementary HWS-PE. Still in line with that hypothesis, Holocaust survivors had worse scores on certain mental health indicators relative to their comparisons: They were higher on negative affect and depressive symptoms, and lower on life satisfaction. The survivors’ higher score on neuroticism reached only marginal significance (p < .10). Also, exceptionally, survivors scored higher on positive affect than comparisons.

Table 4. Univariate Analyses of Covariance (ANCOVAs) comparing the study groups on Evil Perceptions, Hostile-World Scenario (HWS), and Mental Health Indicators.

	Holocaust survivors		Comparison group
Variable	M	SD	M	SD	F Partial eta squared
Overall Evil^a	2.90	0.88	2.39	0.62	F(1,411) = 21.19*** ηp^2 = .049
Fear of human violence^b	3.14	0.93	2.54	0.83	F(1,411) = 25.14*** ηp^2 = .058
Interpersonal mistrust^c	2.12	0.75	2.18	0.70	F(1,319) = 1.76 ηp^2 = .005
HWS-NE^d	2.98	0.71	2.71	0.62	F(1,414) = 12.21*** ηp^2 = .029
HWS-PE	3.17	0.52	3.23	0.46	F(1,415) = 0.002 ηp^2 = .000
Neuroticism (BFI)	2.75	0.87	2.52	0.79	F(1.295) = 3.00 ηp^2 = .010
ABS-PA	2.65	0.84	2.56	0.63	F(1,352) = 4.21* ηp^2= .012
ABS-NA	1.88	0.73	1.58	0.60	F(1,353) = 8.25** ηp^2 = .023
SWLS	4.62	1.37	4.94	1.18	F(1,409) = 4.46* ηp^2 = .011
CES-D	1.83	0.55	1.47	0.21	F(1,121) = 6.03* ηp^2 = .047

Note. ANCOVAs were adjusted for age and gender. HWS-NE = Hostile-World Scenario, Negative Engagement; HWS-PE = Hostile-World Scenario, Positive Engagement; BFI = Big Five Inventory; ABS-PA = Affect Balance Scale, Positive Affect; ABS-NA = Affect Balance Scale, Negative Affect; SWLS = Satisfaction With Life Scale; CES-D = Center for Epidemiological Studies – Depression Scale.

^aIn the ANCOVA computed only for those who completed the full version of the HWS questionnaire, p < .01. ^bEvil scale, Factor 1. ^cEvil scale, Factor 2. ^dIn the ANCOVA computed only for those who completed the full version of the HWS questionnaire, p < .01.

*p < .05. **p < .01. ***p < .001.

The joint associations of HWS and evil-related threats with mental health

In a series of hierarchical multiple regression analysis, mental health indicators served, in turn, as dependent variables. The independent variables were introduced into the equation in consecutive steps. Step 1 included age and gender as covariates, and Group (Holocaust survivors versus comparisons) that represented the focal trauma. Step 2 additionally included HWS-NE and Step 3 additionally included HWS-PE. Step 4 additionally included, in turn, the Evil Scale scores: overall Evil, Factor 1, and Factor 2. Lastly, Step 5 additionally included, in turn, the respective interaction term of Group with any of the three Evil Scale scores.

Table 5 presents regression results where overall Evil was included in the equation. After adjusting for the covariates as well as for the main effects of Group, HWS-NE and HWS-PE, overall Evil had an additional main effect on negative affect and depressive symptoms. Beyond those effects, overall Evil revealed a significant interaction effect with Group on depressive symptoms (interactions affecting the other negative indicators of neuroticism and negative affect were marginally significant). Simple slope analysis, conducted by the PROCESS macro to probe moderation effects (Hayes, 2013), showed that the relationship between overall Evil and depressive symptoms was significant for Holocaust survivors (b = 0.24, 95% CI [0.09, 0.39], p < .01) but not for comparisons (b = 0.01, 95% CI [−0.13, 0.15], ns). Thus, net of the main effects of the two HWS engagements, a higher level of evil-related threats was uniquely associated with higher depressive symptoms among Holocaust survivors. These results provide partial support for the fourth hypothesis.

Table 5. Hierarchical multiple regression analyses predicting mental health indicators by negative and positive engagement with the Hostile-World Scenario (HWS) and by Overall Evil.

Step/predictor	Neuroticism	Positive affect	Negative affect	Life satisfaction	Depressive symptoms
Step 1 (R^2)	.06***	.02	.09***	.03*	.18***
Age	.04	−.11*	.05	.02	.03
Gender^a	.20***	.04	.19***	−.10*	.21*
Group^b	.15**	−.06	.22***	−.11*	.34***
Step 2 (ΔR^2)	.30***	.02**	.10***	.05***	.13***
HWS-NE	.56***	−.15**	.32***	−.24***	.37***
Step 3 (ΔR^2)	.07***	.09***	.03***	.10***	.03*
HWS-PE	−.27***	.32***	−.19***	.35***	−.17*
Step 4 (ΔR^2)	.003	.002	.03***	.003	.02*
Overall Evil	−.08	−.06	.25***	−.08	−.20*
Step 5 (ΔR^2)	.005+	.000	.01+	.003	.04**
Overall Evil X Group	.15+	.03	.11+	.13	.20**
R^2	.44	.13	.26	.19	.40
F	32.05***	7.53***	17.13***	13.21***	10.79***
df	(7,289)	(7,341)	(7,341)	(7,399)	(7,114)

Note. The table presents standardized beta coefficients for main effects, and unstandardized B coefficients for interaction effects. Results of each Step (2 to 5) are reported only for the additional variable in the particular Step. Continuous variables were standardized. HWS-NE = Hostile-World Scenario, Negative Engagement; HWS-PE = Hostile-World Scenario, Positive Engagement.

^aGender was coded 0 = man, 1 = woman. ^bGroup was coded 0 = non-Holocaust survivor, 1 = Holocaust survivor.

*p < .05. **p < .01. ***p < .001. +p < .10.

Parallel analyses that replicated the above regression model by alternately replacing overall Evil by the two factor scores of Evil, similarly found the interaction effect whereby higher level of Evil’s Factor 1 (fear of human violence) was uniquely associated with higher depressive symptoms among Holocaust survivors (B = 0.20, p < .05). Again, the simple slope analysis showed that this relationship was significant for Holocaust survivors (b = 0.19, 95% CI [0.07, 0.32], p < .01) but not significant for the comparisons (b = −0.01, 95% CI [−0.17, 0.14], ns). When the second factor was introduced instead into the equation, the regression showed an interaction effect whereby higher level of Evil’s Factor 2 (interpersonal mistrust) was uniquely associated with higher negative affect among Holocaust survivors (B = 0.17, p < .05). The simple slope analysis showed that this relationship was significant for Holocaust survivors (b = 0.19, 95% CI [0.08, 0.30], p < .001) but not significant for the comparisons (b = 0.06, 95% CI [−0.03, 0.14], ns).

Noteworthy, the main effects of the study variables in the regression model (see Table 5) deserve attention. That is, when HWS-NE was introduced in Step 2, after adjustment for covariates and the group separation (survivors versus comparisons), its beta coefficients revealed associations with all five indicators of mental health, all in the hypothesized directions. When HWS-PE was additionally introduced in Step 3, its beta coefficients also revealed unique associations with all five indicators of mental health, again in the hypothesized directions. When overall Evil was subsequently introduced in Step 4, its beta coefficients still revealed unique associations with negative affect and depressive symptoms. Also (not shown in Table 5), when Evil’s Factor 1 replaced overall Evil, it similarly revealed those unique associations with negative affect (β = 0.14, p < .05) and depressive symptoms (β = 0.21, p < .05). When Evil’s Factor 2 replaced overall Evil, it revealed a unique association with negative affect (β = 0.16, p < .01) as well as reverse associations with positive affect (β = −0.15, p < .01) and life satisfaction (β = −0.11, p < .05). These results indicated that both HWS-NE and HWS-PE maintained broad associations with mental health. Furthermore, even after adjusting for these HWS dimensions, all three Evil scores still maintained non-redundant associations with certain indicators of mental health.

Discussion

This study outlined perceptions relating to the theme of evil, whereby participants expressed the extent of their concern about potential threats of being victimized by human violence or other social malice. These evil-related threats were derived from the conceptual model on the pursuit of happiness in a hostile world (Shmotkin, 2005), in which the key concept of HWS encapsulated individuals’ perceptions of critical threats to their life or integrity. While HWS was considered of high relevance to mental health, and particularly to people who survived a major trauma, perceived evil was supposed to be a distinct theme in this construct. In this study, evil-related threats were specifically examined in conjunction with the broader HWS, expecting to reveal unique associations of evil-related threats, additional to the other HWS-based threats, with mental health and the experience of trauma. The study addressed the Holocaust experience as an extremely massive trauma that had been inflicted on its survivors in their earlier lives, but plausibly lingered with its impact up to their old age.

In its first part, this study delineated a particular set of self-perceptions about evil-related threats in order to examine whether they provided an adequate tool for assessment. The items were pulled out of the HWS Questionnaire, and once clustered within the Evil Scale they were removed from any HWS-based measure, so that separate and complementary effects of evil-related threats and the HWS could be adequately examined. The factorial analyses suggested that the Evil Scale could be feasibly handled through either a unidimensional (overall) score or two distinct (factorial) scores, and hence it seemed useful to employ the scale in both these modalities. The factors pointed to two sub-themes by which evil posed its threats: human violence and interpersonal mistrust. These sub-themes conveyed two core facets of threats: physical victimization, entailing death and a sense of a horrific world, and psychosocial victimization, entailing the disruption of social solidarity and mutual trust (Staub, 2003). Thus, besides evil of disastrous and ruthless cruelty, “everyday evil” also inflicts personal harm on victims through notoriously widespread social behaviors (Welner et al., 2022).

The results substantially supported the first two study hypotheses, thus providing essential construct validity to the current Evil Scale. In this regard, stronger evil-related threats were moderately correlated with higher scores on negative indicators of mental health (neuroticism, negative affect, depressive symptoms) for both survivor and comparison groups. Stronger evil-related threats were correlated with lower scores on positive indicators of mental health only in the comparison group. Besides their affinity with negative aspects of mental health, evil-related threats were linked to trauma as overall Evil was higher among Holocaust survivors than comparisons. This finding was replicated for fear of human violence (Factor 1) but not for interpersonal mistrust (Factor 2). Hence, the results conform with key theoretical positions that psychological trauma demands a confrontation with the capacity for evil in human nature, and that a reconstructive approach in tackling the harm of evil necessitates a restoration of the victims’ mental health (Herman, 1992).

As evil-related threats were derived from the larger HWS construct, their effects should be inspected within the context of HWS. This pertains to the third hypothesis, which was largely confirmed albeit not in full. Notably, Holocaust survivors were higher than comparisons in negative engagement with the HWS; yet, the groups did not differ on positive engagement despite the greater vulnerability of the survivors, relative to comparisons, in mental health indicators (negative affect, depressive symptoms, life satisfaction). Interestingly, beyond group differences, the main effects of both HWS engagement modes revealed consistent and non-redundant associations with the mental health indicators. There is evidence, then, that both HWS-NE and HWS-PE constitute a solid link with mental health, but it is HWS-NE that appears more involved in the repercussions of trauma. As these two engagement modes may be regarded as basic coping patterns (Békés et al., 2021), it appears that trauma survivors continue to recruit HWS-NE for its function to scan for critical threats even when such threats are less feasible in their current lives; yet, survivors adopt HWS-PE not less than their counterparts for efficient management of life adversities (Palgi et al., 2015).

Regarding the fourth hypothesis, the affinity of evil-related threats with trauma appeared further pronounced through an interaction whereby higher overall Evil was related to higher depressive symptoms among Holocaust survivors, rather than comparisons, beyond the simultaneous effects of the two HWS engagements. This regression model also revealed the viability of the two Evil’s factors. Thus, fear of human violence maintained the same interaction found for overall Evil whereas interpersonal mistrust maintained a similar interaction with negative affect. Lending partial support to the fourth hypothesis, these interactions suggest that evil-related threats may still sustain a particular link to painful experiences among victims of evil-inflicted trauma (D’Ippolito et al., 2020). Such interactions may also further elucidate whether and how victims judge prototypes of evil differently from observers who apply more cognitive terms for evaluating evil without experiencing the particular evil themselves (Govrin, 2018).

Several more questions still await future research to follow this study. Thus, the current assessment of evil-related threats yielded two factors. Factor 1 represented harsher threats of human violence, and differentiated between survivors and comparisons. Factor 2 appeared weaker, but still generated viable results as a theme of interpersonal mistrust, thus relating to a pertinent vulnerability in the social integration of Holocaust survivors (Kahana et al., 2015). Further studies should extend these themes of evil into a more inclusive perceptive frame of evil.

Another aim for future research is to examine differences between the currently guiding model of HWS – as well as its derivative of evil-related threats – and other theories that also dwell on formulations of life dangers. Notable in these theories is the terror management theory (Pyszczynski et al., 2015), which considers the terrifying concern of death as a main generator of human behavior. A key postulation in this theory is that humans adopt mechanisms, mainly cultural worldviews and self-esteem, that can transcend the individual’s death and thus thwart death anxiety. In contrast to this theory, the current model assumes a broader sense of threats, not necessarily related to death, that activates the HWS. Moreover, the activation of the HWS (or particular themes such as evil-related threats) is conducted under the motivation to pursue happiness, thus raising interactive dialogs with positive systems of well-being and meaning (for these parts of the current model, see Shmotkin & Shrira, 2012, 2013).

Future research may also examine the affinity of the HWS in general, and evil-related threats in particular, with other theoretical conceptions that assume superordinate beliefs by which individuals characterize their world. Such conceptions were related to trauma as well, as in the notions of world assumptions (Janoff-Bulman, 1989), worldviews (Park et al., 2012) and primal world beliefs (Clifton, 2020). Following applications of this vein of investigation to Holocaust survivors (see introduction), this study employed the concepts of HWS and evil-related threats in a way that converted general world assumptions into tangible threats that people were aware of and reacted to. This model exposed vulnerability among Holocaust survivors through the negative engagement mode of HWS and its derived theme of evil, but it also attested to the resilience of the survivors, whose positive engagement mode of HWS and positive affect were at levels similar to those of comparisons (Shmotkin, 2003). Generally, the current model anchors one’s view of the world within an adaptational perspective that helps to link potential adversities, present mental health, and past traumatic experiences. Further research may examine this network of constructs.

This study should be considered in view of various limitations. First, the construction of the Evil Scale was based on the 72-item HWS Questionnaire whereas other relevant item pools may represent themes of evil differently. This problem notwithstanding, the HWS Questionnaire had initially reflected a broad spectrum of self-perceived threats as emerging from systematic in-depth interviews. Thus, this Questionnaire was also a basis for delineating other HWS-related themes such as interpersonal vulnerability, also in a context of trauma (Ifrah & Tsipper, 2022; Shenkman et al., 2018). Second, data were drawn from convenience samples in both survivor and comparison groups. Thus, the study groups were not representative of their respective populations, although an effort was made to obtain heterogeneous samples. Moreover, as Holocaust survivors endured highly complex experiences, research in this area has typically revealed difficulties in the formation of comparison groups (Shmotkin et al., 2003). Third, while the Holocaust survivor group represented a massive trauma, this study did not examine effects of different experiences during the Holocaust, nor did it consider critical experiences (e.g., migration) or other forms of trauma (e.g., bereavement) that could feasibly impact the study’s constructs. Such refined analyses of different kinds of traumas are particularly needed among aging participants (Palgi et al., 2015; Shrira et al., 2010). Lastly, the cross-sectional nature of this study does not allow for causal inferences. Thus, while evil-related threats (or HWS engagements at large) may affect mental health conditions, variability in mental health may also modify perceptions of major threats. Longitudinal designs may facilitate examinations of such paths. Such designs may also indicate whether perceptions of evil-related (or more broadly, HWS-based) threats are steady dispositions or, rather, situation-bound reactions.

In conclusion, this study demarcated self-perceived threats related to the theme of evil. These threats were usually part of multiple themes that composed the construct of HWS. The current results indicated that evil-related threats appeared as a distinctive theme within the HWS and could separately manifest unique links with poorer mental health among survivors of the Holocaust trauma. In juxtaposing results pertaining to evil-related threats with the larger engagements with HWS-based threats, the study found that Holocaust survivors were higher than comparisons in negative, but not in positive, engagement with the HWS. Also, Holocaust survivors were higher on the Evil’s factor representing fear of human violence, but not on the Evil’s factor of interpersonal mistrust. Distinctive effects of evil-related threats emerged when main effects of HWS engagements were adjusted for, and also through an interaction whereby Holocaust survivors showed stronger associations between evil-related threats and negative indicators of mental health, notably depressive symptoms. Beyond the focus on the trauma survivor group, the constructs of evil-related threats and HWS engagements proved generally relevant to the comparison participants as well, with substantial links to both negative and positive facets of mental health.

This study bears an array of implications for therapists, counselors and caregivers addressing survivors of trauma at large and human evil in particular. The findings suggest a differential approach to the psychological capabilities of the survivors: They appear more resilient in restoring their positive engagement (i.e., proactive coping) with the HWS rather than ameliorating their negative engagement (i.e., anxiety-laden reactions). Regarding their perceptions of evil, the issue of interpersonal mistrust appears more repairable than fear of human violence. Hence, practitioners may take advantage of the differential levels of perceived threats, addressing either a more inclusive outlook on the threatening world (the HWS) or the specific threats of evil alone. As both levels proved in this study to be related to trauma and mental health, practitioners may have a certain space in sequencing a treatment process, where resilience can be salient even in the midst of vulnerability (Ayalon, 2005). Practically, the present approach, besides previous HWS measures (full or short), also offers a potentially usable, brief tool to assess the presence of evil-related threats, whether for a preliminary inquiry or for tracking treatment effects. As the generation of Holocaust survivors is disappearing, further inquiries should specify how the present approach can be expanded to the survivors’ offspring who may have absorbed gloomy notions of evil and a hostile world (Shmotkin et al., 2011; Shrira, 2015) or, more broadly, to any trauma survivors who experienced the perpetration of human evil.

Ethical approval

This study was approved for ethical requirements by the IRB at the authors’ university.

Acknowledgments

We are grateful to Amit Shrira, Sharon Avidor, Kfir Ifrah, Geva Shenkman, Roy Haziza, and Ido Averbuch for their dedicated help in preparing datasets and conceptual materials for this study.

Disclosure statement

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

Additional information

Funding

This work was partially supported by a research grant (# 3-12067) of the Israeli Ministry of Science and Technology.