Resilience to Stress and Adversity: A Narrative Review of the Role of Positive Affect

Abstract

The modern conception of mental health encompasses not only mental illness but also mental wellbeing, including positive emotional states and other forms of positive experience. Accordingly, research on resilience — that is, recovery or adaptation following adversity – has recently expanded to consider the roles of positive affect in the resilience process. To review this research, we performed a keyword search of all peer-reviewed journals within the American Psychological Association’s PsycInfo database, retrieving all studies of positive affect in the context of resilience. These studies measured positive affect either as the outcome of the resilience process or as a resilience resource in its own right. With positive affect as the outcome, the literature suggests that various resilience resources can promote positive affect following a stressor, especially positive personality traits (eg, hope, optimism, self-compassion) and supportive interpersonal connections. With positive affect as a resilience resource, the literature suggests that higher levels of positive affect may protect individuals from the impact of stress on a number of outcomes, such as depression and trauma symptoms. In all, the reviewed research showcases a wide range of stressors, resources, and outcomes, and there are numerous openings for future discoveries in this promising area of inquiry.

Keywords:

• positive emotion
• wellbeing
• happiness
• stress
• trauma
• adversity

Introduction

Historical Context

Until recent decades, mental health was measured and studied primarily in terms of mental illness, negative emotional states, and other forms of psychological distress or dysfunction, along with the risk factors that predict or exacerbate such conditions. But more recently, mental health has been increasingly recognised as comprising two related but distinct domains, namely mental illness and mental wellbeing.^1–4 Thus, the modern conception of mental health acknowledges not only negative symptoms but also positive outcomes, and with this broadened view have come a range of studies on protective factors that predict or accentuate favourable mental health trajectories. Consequently, there is now a large literature on resilience, and the various attributes that support positive mental functioning in the aftermath of stress, trauma, and other forms of adversity. Initially, even studies of protective factors examined them mostly in relation to negative outcomes (eg, studies of personality traits that may predict lower levels of depression), but more recent research has considered predictors of positive attributes too, especially positive affect and related variables (eg, happiness, subjective wellbeing). The overlapping fields of resilience and positive affect have produced numerous studies that shed light on how resilience can be both enabled by positive affect and measured in terms of it, and the present paper offers a narrative review of these connections.

Defining Resilience

The concept of resilience has been defined and studied in various ways, and we must begin by clarifying it for the present purposes. Across all definitions of resilience, there is reference to the phenomenon of regaining (if not maintaining or even increasing) one’s mental health following a stressor (eg, a challenge, hardship, or aversive experience).⁵ Here we are using the broad scientific definition of “stressor” (ie, any experience that involves a challenge, threat, or hardship), as well as the broadest definition of “mental health”, encompassing all aspects – positive and negative – of an individual’s psychological health and functioning. Therefore, studying resilience requires investigations into why some people rebound from or adapt to stressors more easily than others, or experience stressors as less challenging or distressing from the outset, and why certain individuals demonstrate improvements in their mental health even after highly stressful or traumatic experiences (eg, posttraumatic growth⁶). In addressing these research questions, a recent systematic review of resilience interventions notes that resilience has been operationally defined in one of three ways: as an outcome, as a process, and as a trait.⁷ The outcome approach defines resilience as favourable mental outcomes at a specified point in time following the stressor in question (eg, lower levels of anxiety and higher levels of wellbeing three months after suffering a violent crime).⁷ The process approach defines resilience in a similar way, but with multiple measurement occasions (either before and after the stressor, or at multiple timepoints following the stressor), allowing one to identify different trajectories of mental functioning following the stressor (eg, temporarily versus indefinitely reduced wellbeing after a major loss).⁷ In contrast to the outcome and process definitions, the trait approach emphasises the personality traits that define one’s ability to recover from or adapt to stressors.⁷ The present review draws upon studies that employ all three definitions of resilience. However, we note that using the term “resilience” to label a trait (or an ensemble of traits) can make it difficult to differentiate the phenomenon of resilience (ie, recovery from or adaptation to stressors) from its predictors (such as the various protective traits that some researchers label as resilience). Therefore, we will use the term “resilience” to refer to the outcome or process of recovery or adaptation following a stressor, and we will use the phrase “trait resilience” where other authors have used the term to refer to personal attributes that enable the phenomenon of resilience to occur.

As stated above, all definitions of psychological resilience pertain to regaining or maintaining (or even improving) one’s mental health (including both positive and negative outcomes) in the wake of a stressor. Therefore, resilience cannot be directly studied outside the context of stressors (ie, a challenge, threat, or hardship). There are many studies of trait resilience that examine this construct as a predictor, outcome, mediator, or moderator within models that do not contain any measure of perceived or actual stress, challenge, or adversity (eg, models that correlate trait resilience measures with other positive factors, such as wellbeing or cognitive performance). Such studies may be valuable in elucidating the complex interplay of the psychological variables in question, but they do not test whether the characteristics labelled as “resilience” do in fact mitigate the impact of stressors. Therefore, the present review considers only those studies of resilience that examine it in relation to some kind of stressor (moreover, it would be beyond the scope of the review to employ a more expansive usage of the term “resilience”). In this framework, a protective factor (or resilience resource) is necessarily one that moderates the impact of a stressor on one’s mental health.

The Present Review

With these definitional matters resolved, we turn to the subject of our review: the relationship between resilience and positive affect. In this context, the dual role of positive affect is evident from the very definition of resilience. Given that resilience refers to mental health recovery or adaptation, and that mental health involves not only negative symptoms but also positive outcomes, then clearly positive affect can be treated as a measure of resilience. Therefore, all else being equal, those who experience smaller declines in positive emotion following a stressor, and those whose positive affect returns more quickly or more fully to pre-stressor levels, have exhibited greater resilience. Likewise, in the context of adaptation (rather than recovery), those who experience greater increases in positive affect following a challenge or stressor have exhibited greater resilience. Thus, the outcome or process of resilience can be measured in terms of positive affect, as well as other factors whose measures incorporate positive affect (eg, subjective wellbeing, life satisfaction). Alternatively, positive affect can be treated as a resilience resource – a personal or psychosocial attribute that may enable resilience. In this approach, those with greater pre-stressor positive affect are hypothesised to be more resistant to the stressor or better able to recover from (or adapt to) it; in other words, positive affect enables resilience by moderating the impact of the stressor (specifically, by mitigating this impact). The present review acknowledges both roles that positive affect can play in the context of resilience; therefore, it includes studies that treat positive affect either as an outcome measure of resilience or as a resilience resource.

Having clarified both the concept of resilience and dual role of positive affect in relation to it, we may now begin the review proper. The purpose of our review was to summarise the findings from all studies of psychological resilience (within a relevant database) that involved positive affect in one of these two roles. Therefore, the review is divided into two main sections, with the first section covering positive affect as an outcome measure of resilience and the second section covering positive affect as a resilience resource (see Figure 1). Each section is itself divided into subsections according to the stressor or adversity involved, allowing researchers and practitioners to see which forms of hardship have already been studied and which ones may warrant greater attention. Although we did not seek to follow the format of a systematic review, we still conducted our literature search in a rigorous manner. We performed a keyword search (the broadest mode of search) of all peer-reviewed journals in the American Psychological Association’s PsycInfo database, using the following search terms: (resilien* AND (“positive affect” OR “positive emotion*” OR “positive mood” OR “positive feeling*” OR positivity OR wellbeing OR “well-being” OR “well being” OR wellness OR happy OR happi* OR “life satisfaction”)). Thus, the search terms covered all words derived from resilience (eg, resilient, resiliency) and all words or phrases that either are synonymous with positive affect (or treated as such by many researchers) or denote variables (eg, wellbeing) whose definitions and/or measures often incorporate positive affect in some way. From the results of this search, we identified every empirical, quantitative study that measured resilience directly (ie, as the moderation of the impact of a stressor on a given mental health outcome), including studies whose tests of moderation yielded null findings. Our synthesis of these studies comprises the body of this review. From this synthesis, we have aimed to identify the resilience resources that promote the outcome of positive affect following a stressor, as well as the ways in which positive affect can serve as a resilience resource in its own right.

Figure 1 Graphical summary of review structure. The current review focuses on the overall topic of “resilience and positive affect” which can be considered in two ways, with “positive affect as an outcome measure” (left-hand model) or “positive affect as a resilience resource” (right-hand model). Each model is independent and depicts the phenomenon of resilience, whereby the impact of a stressor on a mental health outcome is moderated by a resilience resource. In the left-hand model, the outcome variable is positive affect. In the right-hand model, the moderator variable is positive affect. Thus, the models illustrate the two roles that positive affect may play within the resilience process.

Positive Affect as an Outcome Measure of Resilience

Of the studies retrieved from our search, most pertained to positive affect as an outcome (ie, as a measure of resilience) rather than a resilience resource. This is unsurprising, because while affective states can certainly motivate action, they are typically studied as responses to antecedent thoughts, beliefs, or experiences. This is especially so in the mental health context, where affective states are frequently cited in describing and diagnosing various psychological conditions. Positive affect specifically has been studied as a response to numerous kinds of stressor, which we have divided into six categories. The most traumatic stressors (usually involving abuse, violent crime, or other forms of severe mistreatment) are placed in two categories, one encompassing adverse childhood experiences and the other encompassing traumatic experiences in adulthood. More common stressors (eg, daily hassles, everyday challenges, personal disputes) are placed in another two categories, one covering occupational stressors (including stressors not only in the workplace but also in academic and childrearing contexts) and the other covering experiences that are typically grouped under general labels such as “stressful life events”, “recent life events”, or “everyday stressors”. It should be noted that the events labelled in such ways are not necessarily minor; they can include events that are not usually considered to be traumatic or egregious but can still be significantly upsetting or distressing (eg, having a serious argument, losing one’s job), and some studies of stressful life events combine measures of less severe stressors with measures of potentially traumatic events (PTEs). The final two categories include adversity related to medical illnesses or other health conditions, with one category dedicated to COVID-19-related stressors (given the number of studies on this topic) and the other category covering all other medical or health conditions. Across all six categories, each study includes at least one outcome measure that either measures positive affect exclusively or incorporates positive affect along with other aspects of mental health or wellbeing. See Table 1 for a summary of these studies and the key variables in each one.

Table 1 Summary of Reviewed Studies on Positive Affect as an Outcome of the Resilience Process

Adverse Childhood Experiences (ACEs)
Author	Stressor	Moderator	Outcome
Armitage et al, 2022^Citation8	Peer victimisation Bullying and Friendship Interview Schedule	Mental health genetics Polygenic scores	Mental wellbeing Warwick-Edinburgh Mental Well-Being Scale (WEMWBS)
Bagci et al, 2014^Citation9	Perceived ethnic discrimination Unnamed scale	Cross-ethnic friendships Ratio of cross-ethnic to total friendships	Mental wellbeing WEMWBS
Cipriano et al, 2011^Citation10	Family violence Maltreatment Classification System (MCS), Conflict Tactics Scale (CTS)	Vagal suppression Respiratory Sinus Arrhythmia (baseline minus task)	Positive affect Observed Child Temperament Scale (OCTS)
Cohen et al, 2021^Citation11	Emotional maltreatment Childhood Experience of Care and Abuse Questionnaire (CECA-Q) Community violence Screen for Adolescent Violence Exposure (KID-SAVE)	Trait emotional intelligence Trait Emotional Intelligence Questionnaire – Short Form (TEQ-SF)	Psychological wellbeing Psychological Well-Being Posttraumatic Change Questionnaire (PWB-PTCQ)*
DiClemente et al, 2018^Citation12	Past-year violence exposure Exposure to Violence – Revised scale (EV-R)	Family cohesion Family Environment Scale (FES)	Positive affect Custom scale
Kang et al, 2023^Citation13	ACEs Custom checklist	Trait resilience Resilience Research Centre Adult Resilience Measure (RRC-ARM)	Quality of life World Health Organization Quality of Life brief form (WHOQOL-BREF)
Nishimi et al, 2022^Citation14	Childhood adversity Custom checklist	Healthy lifestyle factors Custom measures, Body Mass Index (BMI)	Positive affect Mental Health Inventory (MHI)
Nurius et al, 2015^Citation15	ACEs Custom checklist	Sense of community Custom scale	Perceived wellbeing Custom scale
Piña-Watson et al, 2015^Citation16	Bicultural stress Bicultural Stressors Scale (BSS)	Caregiver connectedness Custom scale	Life satisfaction Brief Multidimensional Students’ Life Satisfaction Scale (BMSLSS)
Riley et al, 2020^Citation17	Juvenile victimisation Juvenile Victimization Questionnaire – 2nd Revision (JVQ-R2)	Problem-solving coping Self-report Coping Scale*	Life satisfaction Students Life Satisfaction Scale (SLSS)* Positive affect Positive and Negative Affect Schedule – Child (PANAS-C)**
Seon & Smith-Adcock, 2023^Citation18	Bullying victimisation Custom scale	Meaning in life Custom scale	Life satisfaction Single-item
Sleijpen et al, 2019^Citation19	Potentially traumatic events Custom checklist	Trait resilience Resilience Scale (RS)	Life satisfaction Satisfaction with Life Scale (SWLS)
Somers et al, 2017^Citation20	Childhood maltreatment Childhood Trauma Questionnaire (CTQ)	Heart rate reactivity Heart rate (baseline minus task)	Positive affect Positive and Negative Affect Schedule (PANAS)
Sorek, 2020^Citation21	Parental conflict Children’s Perception of Interparental Conflict scale (CPIC)	Closeness to grandparents Custom scale	Life satisfaction SLSS, single-item
Sorek et al, 2019^Citation22	Parental conflict CPIC, custom scale	Self-blame, active coping CPIC	Life satisfaction SLSS, single-item
Tung et al, 2022^Citation23	Family/household adversity Risky Family Questionnaire (RFQ)	Trait resilience Brief Resilience Scale (BRS)	Positive affect Positive and Negative Affect Schedule – Expanded (PANAS-X)
Turiano et al, 2017^Citation24	Childhood misfortune Custom checklist	Control beliefs Custom scale	Positive affect Custom scale
Adult trauma
Author	Stressor	Moderator	Outcome
Fetter et al, 2023^Citation25	Perceived historical loss Adolescent Historical Losses Scale (AHLS)	Ethnic identity Multigroup Ethnic Identity Measure – Revised (MEIM-R)	Mental wellbeing Mental Health Continuum – Short Form (MHC-SF)
Lee et al, 2021^Citation26	Vietnam War service	Psychosocial resources Multiple measures	Mental wellbeing MHC-SF
Liu et al, 2023^Citation27	Perceived discrimination Custom scale	Belief in a Just World (BJW) Belief in a Just World Scale (BJWS)	Life satisfaction SWLS
Yubero et al, 2023^Citation28	Chronic bullying victimisation Bullying Harassment and Aggression Receipt Measure (BullyHARM), custom scale	Trait resilience Connor Davidson Resilience Scale 10-item version (CD-RISC-10)	Mental wellbeing MHC-SF
Stressful life events
Author	Stressor	Moderator	Outcome
Alriksson-Schmidt et al, 2007^Citation29	Life stressors Adolescent Disability-Related Life Events Survey, Adolescent Life Events Survey – Revised	Social/family functioning Social Skills Rating System (SSRS), Offer Self-Image Questionnaire – Revised (OSIQ-R)	Quality of life Quality of Student Life Questionnaire (QSLQ)
Arampatzi et al, 2020^Citation30	Greek bailout referendum	Positive expectations Custom scale	Happiness Single-item
Awaworyi Churchill & Smyth, 2022^Citation31	Crime rate Crimes per capita per postcode Crime victimisation Single-item	Internal Locus of Control Mastery Scale	Mental health Mental Health Inventory (MHI-5)
Bucknor & Derringer, 2022^Citation32	Stressful life events Multiple measures	Psychosocial resources Multiple measures	Positive minus negative affect PANAS-X
Chong et al, 2023^Citation33	Negative experiences Custom measure	Trait self-compassion Self-Compassion Scale	Positive affect PANAS-X
Corral-Frías et al, 2016^Citation34	Social stress Trier Social Stress Test (TSST)	Trait reward sensitivity Behavioral Activation Scale (BAS)	Positive affect PANAS
Dooley et al, 2017^Citation35	Lifetime acute stress exposure Stress and Adversity Inventory (STRAIN)	Lifetime acute stress exposure STRAIN	Positive affect PANAS
Ertl et al, 2019^Citation36	Perceived stress Perceived Stress Scale (PSS)	Trait resilience BRS	Mental health Short Form Health Survey (SF-36)
Gerstberger et al, 2023^Citation37	Current stress Single-item	Physical activity Single-item Chronic stress Trier Inventory of Chronic Stress (TICS)**	Positive affect Custom scale
Goodman et al, 2017^Citation38	Negative life events Custom scale	Trait hope Adult Hope Scale (AHS)	Mental health Multiple measures
Greenfield & Marks, 2004^Citation39	Major role-identity absences Custom checklist	Volunteering Custom measure	Positive minus negative affect Custom scales
Hatzichristou et al, 2020^Citation40	Recession-related difficulties Economic Crisis Difficulties Questionnaire (ECDQ)	Positive peer relations California Healthy Kids Survey (CHKS)	Joy in life Berne Questionnaire of Subjective Well-Being/Youth form (BQSW/Y)
Hodzic et al, 2016^Citation41	Perceived stress PSS	Trait emotional repair Trait Meta-Mood Scale (TMMS)**	Life satisfaction SWLS
Huang et al, 2019^Citation42	Low socioeconomic status Monthly household income	Trait resilience Connor Davidson Resilience Scale 25-item version (CD-RISC-25)	Life satisfaction SWLS
Jagtiani et al, 2019^Citation43	Excessive SNS use Hours per day of SNS use	Evening meals with family Custom measure	Mental wellbeing WEMWBS**
Johnson et al, 2016^Citation44	Daily stressors Survey of Recent Life Experiences	Cognitive reappraisal Emotion Regulation Questionnaire (ERQ)	Positive mood Custom scale
Lazić et al, 2023^Citation45	Negative life events Serbian Life Events Check-List – Student Form (SLEC-SF)	Childhood adversity Revised Questionnaire for Attachment Assessment (QAA–R)	Life satisfaction SWLS Positive affect PANAS
Mutz et al, 2019^Citation46	Excessive screen time Hours per day of screen time	Outdoor group recreation	Life satisfaction Single-item Positive minus negative affect Custom scales
Noor & Alwi, 2013^Citation47	Negative life events Life Events Checklist, The Problem Questionnaire	Psychosocial resources Multiple measures	Life satisfaction BMSLSS
Philippe et al, 2018^Citation48	Negative mood induction Experimental task	Cognitive reappraisal Experimental intervention	Positive emotionality Custom scale
Seery et al, 2010^Citation49	Lifetime and recent adversity Custom checklists	Lifetime adversity Custom checklist	Life satisfaction SWLS
Occupational stressors
Author	Stressor	Moderator	Outcome
Annor & Amponsah-Tawiah, 2020^Citation50	Workplace bullying Negative Acts Questionnaire – Revised (NAQ-R)	Trait resilience CD-RISC-10	Subjective wellbeing World Health Organization Well-Being Index (WHO-5)
Bäccman et al, 2016^Citation51	Naval deployment	Prior combat-related stress Custom scale	Positive affect PANAS
Gabriel et al, 2011^Citation52	Job dissatisfaction Custom checklist	Trait resilience CD-RISC-25* Nurse-physician collegiality Revised Nursing Work Index	Positive affect Custom scale
Lai & Mak, 2009^Citation53	Daily hassles Inventory of College Students’ Recent Life Experiences (ICSRLE)*	Trait resilience Multiple measures	Mental health General Health Questionnaire (GHQ-30)
Leipold et al, 2019^Citation54	Academic work-related stress TICS	Coping styles Coping Orientation to Problems Experienced inventory (COPE)**	Subjective wellbeing Custom scale
Lutz et al, 2020^Citation55	Work-home boundary violations Experimental task	Trait resilience BRS	Positive affect PANAS
Sharda, 2022^Citation56	Parenting stress Parental Stress Scale	Social support Social Provisions Scale (SPS)	Mental wellbeing MHC-SF
Sharda et al, 2019^Citation57	Parenting stress Parental Stress Scale	Social support Multidimensional Scale of Perceived Social Support (MSPSS)	Quality of life Assessment of Quality of Life (AqoL-6d)
Shi et al, 2021^Citation58	Emotional dissonance Frankfurt Emotional Work Scales*	Trait resilience BRS	Positive affect PANAS**
Udayar et al, 2020^Citation59	Work stress General Work Stress Scale (GWSS)	Big Five personality traits NEO Five-Factor Inventory – Revised (NEO FFI-R)	Life satisfaction SWLS
van Erp et al, 2015^Citation60	Bystander conflict Experimental tasks	Trait resilience Custom scale	Positive affect Job Affective Well-being Scale (JAWS)**, custom scale
Wepf et al, 2022^Citation61	Caring responsibilities Custom measures	Trait benefit-finding General Benefit Finding Scale (GBFS)	Mental wellbeing WEMWBS
Health conditions (non-COVID)
Author	Stressor	Moderator	Outcome
Bos et al, 2016^Citation62	Psychological symptoms Depression Anxiety Stress Scales – 42-item version (DASS-42)	Self-defeating humour Humor Style Questionnaire (HSQ) Having a romantic partner Single-item	Life satisfaction Single-item Happiness Single-item
Delfabbro et al, 2011^Citation63	Body image dissatisfaction Single-item	Family/social functioning Family Assessment Device (FAD), custom measures	Life satisfaction Unnamed scale** Self-esteem Rosenberg Self-Esteem Scale (RSES)
Hajek & König, 2019^Citation64	Negative health comparisons Single-item	Self-efficacy Generalized Self-Efficacy scale** Optimism Unnamed scale Self-esteem RSES	Positive affect PANAS Life satisfaction SWLS
Kratz et al, 2007^Citation65	Pain severity Single-item	Pain acceptance Custom scale	Positive affect PANAS
Okun et al, 2011^Citation66	Chronic health conditions Custom measure	Past-year volunteering Single-item	Positive affect WHO-5
Plexico et al, 2019^Citation67	Stuttering Self-identified by participants	Trait resilience CD-RISC-25 Coping styles Brief COPE**	Life satisfaction SWLS
Windle et al, 2010^Citation68	Chronic physical illnesses Older American’s Resources and Services questionnaire (OARS)	Trait resilience Unnamed scale	Life satisfaction Life Satisfaction Index Z (LSI-Z)
COVID-19 stressors
Author	Stressor	Moderator	Outcome
Barni et al, 2020^Citation69	COVID-19 exposure and fear Single-item measures	Sense of coherence Sense of Coherence Scale	Psychological wellbeing Short-Form Health Survey (SF-12)**
Cunningham et al, 2021^Citation70	Social isolation Custom scale	Age Years	Positive affect PANAS
Faul & De Brigard, 2022^Citation71	Concern about pandemic Custom scale	Trait nostalgia Nostalgia Inventory	Positive mood Profile of Mood States (POMS)* Mood change Single-item measures
Haghayeghi et al, 2023^Citation72	Pandemic-related disruptions Single-item	Physical activity Single-item	Psychological wellbeing Psychological Well-Being Scale (PWB-18)
Hu et al, 2023^Citation73	Perceived stress Unnamed scale	Trait resilience Unnamed scale	Life satisfaction Unnamed scale
Paredes et al, 2021^Citation74	Perceived COVID threat Custom scale	Trait resilience CD-RISC-25**	Positive affect PANAS**
Shoshani, 2023^Citation75	Living through pandemic Pre-pandemic to post-fifth wave	Social support MSPSS Daily routine Adolescent Routines Questionnaire (ARQ)	Positive affect PANAS-C Life satisfaction BMSLSS
Wang et al, 2021^Citation76	Health-related stress Multicultural Events Schedule for Adolescents	Coping styles Response to Stress Questionnaire (RSQ)**	Positive affect PANAS-C

Notes: As in the body of the review, the studies are grouped under headings according to the type of stressor involved. For each study, the key variables (stressors, moderators, and outcomes) are listed, and the measures used to assess these variables (where applicable) are shown in italics. Where established measures have been employed, they are listed by name; otherwise, the type of measure (eg, custom checklist, unnamed scale, single-item) is stated. *Modified/adapted from original. **Shortened/reduced/abbreviated version.

Adverse Childhood Experiences (ACEs)

Our search identified seventeen studies that examined resilience in terms of the effects of ACEs (occurring during childhood or adolescence) on positive affect. In this context, the most commonly studied resilience resources are internal ones, especially traits. Three studies have investigated the role of trait resilience specifically, with mixed results. In a study of adolescents residing in the Netherlands, Sleijpen et al¹⁹ found that trait resilience moderated the relationship between PTEs (eg, loss of parent, abuse, lack of shelter) and satisfaction with life in a sample of 148 non-refugee secondary school students but not in a sample of 117 refugees residing in asylum seeker centres (who had experienced a significantly higher number and variety of PTEs than the non-refugees). In the non-refugees, number of PTEs was negatively associated with satisfaction with life, with this association being non-significant in those with above-average levels of trait resilience and increasingly negative with increasingly below-average levels of trait resilience. The absence of a similar moderation effect in the refugees suggests that individual differences in trait resilience may not always be enough to moderate the effects of the PTEs typically experienced by asylum seekers, who are at a markedly higher level of risk and experience different stressors from those typically encountered by non-refugees. The other two studies investigated trait resilience in relation to retrospectively reported ACEs. In a sample of 255 Korean unwed mothers, Kang et al¹³ found that a higher number of recalled ACEs was associated with greater perceived discrimination, which in turn was associated with lower wellbeing, especially among those with lower trait resilience. In addition, in a study of 98 Australia-based university students, Tung et al²³ divided the participants into three categories: control (few ACEs, low depression and anxiety), vulnerable (more ACEs, low trait resilience), and resilient (more ACEs, high trait resilience). The vulnerable group (who had lower trait resilience) reported lower positive affect than the control group and higher negative affect than the other two groups, while the resilient group (who had higher trait resilience) did not differ from the control group in positive affect.

Five studies have examined internal resilience resources other than trait resilience. Cohen et al¹¹ used clustering analyses to divide 584 US adolescents into distinct profiles, representing differing levels of psychological distress and mental wellbeing in response to previous experiences of PTEs (community violence and emotional maltreatment). Participants who had both high wellbeing and low distress (accounting for PTE exposure) reported the highest levels of trait emotional intelligence. Moreover, among those who had low wellbeing, emotional intelligence was higher in those who reported low (rather than high) distress. In a study of over 6000 US adults, Turiano et al²⁴ found that perceived control beliefs moderated the association between recalled childhood misfortune and current affective outcomes, with childhood misfortune predicting lower positive affect and greater negative affect more strongly among those with a lower level of perceived control over their lives. Similarly, Seon and Smith-Adcock¹⁸ found a buffering effect of meaning in life in a nationally representative sample of over 4800 US 15-year-olds, whereby the negative association between bullying victimisation (via verbal, physical, or relational bullying) and life satisfaction was weaker among those with a greater sense of life meaning. Likewise, in a sample of 454 Polish adolescents, Riley et al¹⁷ found that problem-solving coping strategies (but not social support-seeking or avoidant strategies) significantly moderated the negative relationship between victimisation and emotional wellbeing. Those who had experienced more types of victimisation (across categories such as bullying, crime, and child maltreatment) reported lower life satisfaction and lower positive affect, but less so if they reported greater use of problem-solving coping. Coping strategies were also examined by Sorek et al²² in a study of 122 Israeli children of divorced parents. It was found that child-rated parental conflict was associated with lower life satisfaction only among children with low levels of active coping. However, this moderation effect was observed for only one of the two measures of life satisfaction employed by the authors, and such effects were not observed when happiness or psychological wellbeing were the outcomes. When mother-rated parental conflict was considered, there was only one interaction effect, with reduced life satisfaction (on only one of the two measures) only in children with high self-blame.

A further five studies have examined the role of external resilience resources, specifically social or interpersonal connections. In an additional analysis from the aforementioned study of children of divorced parents, Sorek²¹ found that parental conflict was associated with lower life satisfaction only among children with less-close relationships with their grandparents. However, this interaction effect was found for only one of the two child-rated measures of parental conflict (feeling caught between one’s parents), and it was not found when happiness or psychological wellbeing was the outcome. DiClemente et al¹² examined the effects of violence exposure on the positive affect of 269 Black American adolescents, finding no evidence that family cohesion, neighbourhood cohesion, or school cohesion moderated the relationship between violence exposure (over the previous 12 months) and current positive affect in cross-sectional analyses of the participants in the seventh and eighth grades. In fact, violence exposure had no main effect on positive affect in either grade. In longitudinal analyses, violence exposure in seventh grade negatively predicted positive affect in eighth grade (controlling for seventh grade positive affect), and this relationship was moderated by family cohesion (but not neighbourhood or school cohesion), with the violence-affect relationship becoming less negative with higher levels of cohesion. However, this longitudinal result was found only in boys, and while the authors tested innumerable two-way and three-way interaction effects, they did not report any correction of the significance threshold to account for multiple testing. More straightforward results were obtained by Nurius et al¹⁵, who investigated retrospectively reported ACEs in a sample of over 13,000 US adults. The authors found a weaker negative correlation between number of recalled ACEs and perceived wellbeing among those with a greater sense of community, referring to adequate social support and satisfaction with one’s neighbourhood. External resilience resources have also been examined in the context of adolescent adversity related to ethnic or cultural differences. In a study of 247 British secondary school students of South Asian descent, Bagci et al⁹ found that perceived ethnic discrimination was negatively related to psychological wellbeing only among those with fewer cross-ethnic friendships. In addition, in a study of 524 US adolescents of Mexican descent,¹⁶ bicultural stress (ie, stress related to maintaining one’s heritage culture while adapting to the surrounding culture) had a stronger negative association with life satisfaction among those with stronger connections to their female caregiver (while male caregiver connectedness did not moderate this association). This may have been due to a floor effect, because those with lower female caregiver connectedness had lower life satisfaction even at low levels of bicultural stress.

Apart from research on internal and external resilience resources, four studies have considered physiological or physical health factors in the context of resilience to ACEs. In a study of over 3000 US adults, Nishimi et al¹⁴ divided the participants into four categories: resilient (adversity-exposed, good mental health), non-resilient (adversity-exposed, poor mental health), positive functioning (not adversity-exposed, good mental health), and unfavourable functioning (not adversity-exposed, poor mental health). These groups were then compared in terms of physical health behaviours (non-smoking, moderate alcohol consumption, regular exercise, healthy nutrition) and body weight. The resilient group did not differ from the positive functioning group except on non-smoking (there being significantly more non-smokers in the latter group). Furthermore, the resilient group was significantly healthier than the non-resilient group in terms of physical activity, nutrition, and body weight. Two studies have employed cardiac measures as indicators of resilience resources, given that emotional and behavioural regulation have been linked with variables such as respiratory sinus arrhythmia and heart rate. In a study of 92 US preschool children, Cipriano et al¹⁰ investigated whether vagal suppression moderated the relationship between violence exposure and three outcome variables (emotional problems, frustration, and positive affect), finding an interaction effect on emotional problems only. In contrast, in a sample of 150 US undergraduate students with a history of childhood maltreatment, Somers et al²⁰ found a moderating effect of heart rate reactivity, whereby maltreatment was negatively associated with positive affect in those with above-average or average heart rate reactivity, but not in those with below-average reactivity. One key difference between the Cipriano et al¹⁰ and Somers et al²⁰ studies (aside from the different cardiac measures employed) lies in their sample demographics, as the former recruited children with current or recent adversity whereas the latter employed an adult sample with recalled ACEs. Lastly, in a study of over 4800 UK adults, Armitage et al⁸ employed polygenic scores for wellbeing and depression (ie, genetic correlates of mental health) as potential moderators of the effect of peer victimisation at age 13 on mental wellbeing at age 23. The authors found no evidence of moderation, leading them to conclude that genetic profiling may not be useful for identifying those who are more or less resilient to bullying.

Together, studies of the effects of ACE on positive affect have shown that internal resilience resources such as trait resilience and external resources such as social support may serve to mitigate the negative impact of ACEs in both childhood and adulthood, although the evidence is mixed in some areas. There is also limited research involving physiological indices of resilience, with evidence that physical and cardiovascular health factors may be influential.

Adult Experiences of Mistreatment or Trauma

When shifting the lens to adult experiences of adversity and their effects on positive affect, our search yielded only four studies of resilience, all of which examined internal resilience resources. Two of the studies focused on minority populations at greater risk of suffering discrimination. In a sample of 242 Native American and Alaskan undergraduates, Fetter et al²⁵ examined a culturally specific stressor (perceived historical loss) as a predictor of mental wellbeing. The results showed a significant interaction effect whereby there was a weaker negative relationship between historical loss and wellbeing among those with a stronger sense of ethnic identity (although this resource did not moderate the effect of historical loss on psychological distress). Similarly, in a study of 872 Chinese impoverished college students, Liu et al²⁷ found a weaker negative association between perceived discrimination and subjective wellbeing among those with higher levels of belief in a just world (ie, the belief that the world is ultimately fair). This study also found that trait resilience moderated the impact of discrimination on self-esteem, but not as expected: Those with higher trait resilience had a larger self-esteem advantage at lower rather than higher levels of discrimination, possibly indicating a floor effect.

The other two studies examined stressors related to conflict or mistreatment. In a sample of over 1100 Spanish university students, Yubero et al²⁸ found a weaker negative relationship between chronic bullying victimisation (at school and university) and psychological wellbeing among those with higher levels of trait resilience, suggesting a buffering effect. Finally, Lee et al²⁶ employed a clustering analysis of posttraumatic stress symptoms and mental wellbeing in a sample of 367 South Korean male veterans of the Vietnam War, which yielded five distinct classes, including groups labelled “resilient”, “moderate distress”, and “severe distress”. The resilient group, who reported the highest levels of wellbeing, scored higher than the other groups on a range of resilience resources, including optimism, positive appraisals of military service, and social support.

Stressful Life Events or Everyday Stressors

Our search returned twenty-one studies that examined resilience in terms of the effects of stressful life events on positive affect, with the majority investigating how these effects may be moderated by internal resilience resources such as personality traits. Two studies have examined trait resilience specifically, with mixed findings. In a study of 95 Mexican informal caregivers, Ertl et al³⁶ observed a weaker correlation between perceived stress over the past month and mental quality of life (but not physical health-related quality of life) among those with higher trait resilience. In contrast, in a study of 486 Chinese rural-to-urban migrant school students, Huang et al⁴² found that family socioeconomic status (SES) was positively correlated with life satisfaction only in those whose participating parent was low in positive affect. The children’s own levels of trait resilience did not moderate the relationship between SES and life satisfaction (although trait resilience moderated the relationship between SES and emotion regulation).

Five studies have looked at personality traits other than trait resilience. In a longitudinal study of 797 community adults across 42 countries, Goodman et al³⁸ found that higher trait hope at each timepoint predicted a weaker negative association between stressful life events and subjective wellbeing over the following three months. However, such interaction effects were not observed when other traits, including grit, meaning in life, curiosity, gratitude, perceived control, and use of personal strengths, were analysed as moderators. In another longitudinal study, involving over 1100 Greek university students, Arampatzi et al³⁰ found that those with more positive expectations at baseline experienced smaller drops and larger rebounds in happiness over the subsequent timepoints, which coincided with major socioeconomic upheaval in the target population (the Greek bailout referendum). Similarly, in a nationwide study of over 15,000 Australian adults,³¹ the negative associations between positive affect and both neighbourhood crime levels and being a victim of crime were smaller among those with a more internal locus of control (ie, those who perceived themselves as being more in control of the events in their lives). Additionally, in a study of 235 US adults who identified as LGBQ (lesbian, gay, bisexual, and/or queer), Chong et al³³ found that daily negative experiences related to the participant’s sexual orientation were associated with lower positive affect, but not in those with high self-compassion (these experiences were also associated with higher negative affect, but this relationship was not moderated by self-compassion). Furthermore, in a laboratory study of 130 US university students, Corral-Frías et al³⁴ found that an experimentally contrived stressor resulted in reduced positive affect and increased negative affect in participants, but those with higher reward sensitivity reported smaller decreases in positive affect.

Three studies have examined cognitive reappraisal (an emotion regulation strategy) or related practices as resilience resources. Hodzic et al⁴¹ investigated emotional intelligence (labelled “trait meta-mood” by the authors) in a study of 835 university students across three countries (Spain, Portugal, and Brazil). Perceived stress over the past month was negatively correlated with life satisfaction, but less so among those with higher scores on a trait meta-mood subscale measuring “emotional repair” (ie, using positive thinking to overcome negative moods), a construct comparable with cognitive reappraisal. Furthermore, in a daily-diary study of 236 English undergraduate students, Johnson et al⁴⁴ found that greater daily stressors were associated with lower positive mood and greater negative mood, with the latter association being weaker among those who were higher in cognitive reappraisal and lower in brooding (a rumination style that contrasts with reflection). However, the relationship between daily stressors and positive mood was not moderated by any of the emotion regulation strategies or rumination styles under study. Conversely, in an experimental intervention involving 83 Canadian undergraduates, Philippe et al⁴⁸ observed smaller reductions in positive emotion when participants had been randomly allocated to a 10-minute resilience intervention (guided cognitive reappraisal) prior to a negative mood induction, rather than one of the control conditions (neutral music or guided relaxation).

Five studies have assessed external resilience resources, especially resources related to social or interpersonal factors. Two of these studies examined moderators of the relationship between excessive use of digital technology (a common source of everyday stress) and positive affect. Jagtiani et al⁴³ assessed the use of social networking sites (SNS) in a nationally representative sample of over 2200 UK young adults aged 16–21 who still lived with their parents. There was a significant interaction effect whereby heavy SNS use was associated with low mental wellbeing primarily among those who did not share any evening meals with their family. Mutz et al⁴⁶ evaluated the subjective wellbeing of 76 adolescents from German-speaking countries before and after they participated in 10 days of outdoor group recreational activities (eg, hiking, climbing, canoeing) in Southern France. For those who reported ≤3 hours per day of screen time (ie, leisure time spent in front of a television or computer), life satisfaction was high prior to the outdoor program and did not significantly increase over time, but for those who reported >3 hours of screen time per day, life satisfaction was initially low but then increased significantly. However, this moderation effect was not observed for hedonic balance (ie, positive affect minus negative affect), with hedonic balance increasing over time regardless of baseline screen time levels. It should also be noted that the resilience resource in this study (outdoor group recreation) comprised not only social interaction but also physical activity and exposure to natural environments, which may bring their own benefits with regard to resilience.

The other three studies of external resources also returned mixed results. In a nationally representative sample of 373 US older adults, Greenfield and Marks³⁹ measured the stressor of “major role-identity absences”, referring to a lack of important interpersonal roles (ie, among those without a job, spouse, or child). Greater role-identity absences were associated with a lower sense of purpose in life only among those who did not engage in monthly volunteering, but such interaction effects were not found when the outcome was either positive or negative affect. Furthermore, in a study of 726 Athenian high school students, Hatzichristou et al⁴⁰ found that everyday life difficulties (due to the contemporaneous Greek economic recession) were negatively correlated with joy in life and self-esteem, but these negative associations were lessened by positive peer relations only among those students who were academic high-achievers. Similarly, in a study of 159 US adolescents with a mobility disability,²⁹ Alriksson-Schmidt et al found that everyday life stress (including disability-related stress) was negatively correlated with quality of life, but this association was not moderated by social competence, peer social engagement, or family functioning.

Two studies have examined multiple resilience resources, covering both internal and external factors. In a study of 197 Malaysian students from disadvantaged households, Noor and Alwi⁴⁷ divided the participants into three groups, labelled “adapted” (low life stress and high life satisfaction), “resilient” (high stress but high satisfaction), and “maladapted” (high stress and low satisfaction). The adapted and resilient students reported better mother-child communication and higher levels of support from their teachers; they also had higher scores on the Big Five personality traits (with Neuroticism reverse-scored) than the maladapted students. Similarly, in a nationwide study of almost 10,000 US adults, Bucknor and Derringer³² found that life stress had a smaller negative association with positive affect among those who reported greater levels of social support; lower levels of loneliness; higher levels of subjective wellbeing; higher scores on the Big Five (with Neuroticism reverse-scored); higher scores on trait measures of optimism, mastery, purpose, and religiosity/spirituality; lower scores on pessimism and hopelessness; and higher levels of educational attainment. The same study also examined genetic data (European ancestries only), finding equivalent moderation effects with the polygenic scores for extraversion, neuroticism, subjective wellbeing, and educational attainment.

Six studies have examined whether the effects of life stress can be moderated by other stressful experiences. Three of these studies have suggested that previous adversity can lead to increased resilience against later stressors, provided the earlier adversity was not too severe. In a nationally representative study of over 2000 US adults, Seery et al⁴⁹ observed a curvilinear relationship between lifetime adversity and current life satisfaction, whereby satisfaction was highest among those who had experienced low levels of adversity, rather than no adversity or high levels of adversity. This study also found that lifetime adversity moderated the relationship between recent adversity (in the last 6 months) and life satisfaction, with recent adversity having the weakest negative association with life satisfaction among those who reported low levels of lifetime adversity. Similarly, in a study of 122 US women who had survived breast cancer, Dooley et al³⁵ found that the number of acute stressors in the one’s lifespan (pre-cancer diagnosis) predicted current levels of positive affect and cancer-related intrusive thoughts in a curvilinear fashion, whereby those who had experienced a moderate number of stressors (rather than a low or high number) reported greater positive affect and fewer intrusive thoughts. However, this study also found a straightforward linear relationship between acute stressors and negative affect, with greater stress predicting greater negative affect at a fixed rate. Mixed results were also obtained by Lazić et al⁴⁵ in a longitudinal study of 293 Serbian undergraduate students. The authors found that recent life stress was associated with lower life satisfaction in those who reported low or high but not moderate levels of perceived negative childhood experiences (PNCE); however, PNCE did not moderate the negative relationship between recent stress and positive affect (nor the positive relationship between recent stress and negative affect).

The other three studies have suggested that prior stressful experiences can compound rather than mitigate the effects of subsequent stressors. In a longitudinal experience-sampling study of 156 German adults, Gerstberger et al³⁷ found that current stress was negatively associated with current positive affect, and that this association was larger in those who reported greater ongoing chronic stress. This study also found that greater physical activity (since the last experience sample) predicted a weaker correlation between current stress and current negative affect (but not current positive affect), but this interaction was evident only for those who rated their past experiences of major life events as less severely stressful. In addition, the aforementioned studies by Nurius et al¹⁵ and Tung et al²³ also shed light on the interaction between past and present stressors. Nurius et al¹⁵ observed a stronger association between adult adversity and psychological distress among those who recalled greater childhood adversity, although this interaction effect was not found with perceived wellbeing as the outcome. Similarly, Tung et al²³ found that higher daily stress was associated with greater negative affect more strongly in the vulnerable group than in the resilient or control groups, but there was no such moderation of the association between daily stress and positive affect.

In summary, there is evidence that a range of internal and external resilience resources may mitigate the impact of everyday life stress on positive affect and related outcomes. Interestingly, a few studies have suggested that stressful experiences themselves can foster resilience against subsequent stressors, provided the earlier stress was not excessive. Such stress-inoculation effects have been found chiefly in the context of everyday stressors, presumably because more severe stressors are typically too harmful to yield net benefits (at least in terms of positive affect). However, there is also evidence that stressful experiences can compound the negative impact of later adversity, so the question of stress-inoculation clearly warrants continued investigation.

Occupational Stressors

Our search identified thirteen studies that examined resilience in terms of the effects of occupational stressors (including academic and childrearing stressors) on positive affect, with the majority investigating how this impact may be moderated by internal resilience resources, especially traits. Six studies (across five papers) have examined trait resilience specifically, with mixed results. In an experience sampling study of 65 full-time, guest-facing employees of US hotels, Shi et al⁵⁸ found that emotional dissonance (needing to show guests a different emotion from one’s actual feelings) was associated with lower positive affect, lower job satisfaction, and greater intention to leave the job only in those with low trait resilience (although there was no such interaction in relation to negative affect). In a longitudinal study of 57 US nurses, Gabriel et al⁵² asked the participants to report the extent to which they had performed their patient-facing tasks (direct care) and other tasks (indirect care) to their satisfaction, after each of six consecutive shifts (thus, a lack of satisfaction with one’s caregiving constituted the occupational stressor). Direct care satisfaction predicted greater positive affect and lower negative affect (post-shift minus pre-shift), but neither of these associations was moderated by trait resilience or ratings of nurse-physician collegiality. However, while indirect care satisfaction did not have a main effect on positive affect, there were moderation effects whereby indirect care satisfaction predicted greater positive affect only at lower levels of trait resilience or collegiality (collegiality also moderated the association between indirect care satisfaction and negative affect). In a roleplay-based study of 66 Dutch paramedics, van Erp et al⁶⁰ had the participants respond to a simulated medical emergency involving interference from a bystander. Higher perceived interference was correlated with lower positive affect, both directly and indirectly via more-compromised cognition (ie, poorer concentration on the emergency), and the relationship between perceived interference and cognition was stronger in those with lower trait resilience. The authors published this roleplay-based study alongside a randomised experiment of bystander conflict in a university student setting. In this experiment, 47 Dutch students completed mathematical and grammatical exercises while a bystander was either silent or loud and rude. The latter condition resulted in lower positive affect, greater negative affect, and more unfavourable appraisals of the bystander, but only the effect on appraisals (not the effects on positive and negative affect) was especially strong among those with lower trait resilience. In another experimental study, of 337 German smartphone-owning employees, Lutz et al⁵⁵ had the participants react to imaginary scenarios wherein the work-home boundary was violated and they were pressured via text message to interrupt their current activity (by a friend/colleague requesting a call during/outside working hours). Higher pressure predicted greater negative affect but had no impact on positive affect, and trait resilience was not a significant moderator with either positive or negative affect as the outcome. In a study of 631 Ghanaian full-time employees, Annor and Amponsah-Tawiah⁵⁰ found that workplace bullying was negatively associated with mental wellbeing more strongly among those with higher trait resilience. The authors found this result contradictory to the claim that trait resilience is protective, although the data indicate a possible floor effect, whereby trait resilience correlates with wellbeing but not at high levels of bullying (in which case wellbeing bottoms out for everyone).

Four studies have examined occupational stressors in the context of personality traits other than trait resilience. In a longitudinal study of over 1200 Swiss employees, Udayar et al⁵⁹ found that work stress at baseline positively predicted work stress 12 months later across all participants, but baseline work stress negatively predicted 12-month life satisfaction only in those with an “oversensitive” personality profile, characterised by lower levels of extraversion and conscientiousness, and higher levels of neuroticism. In a study of over 2500 Swiss students in secondary school or vocational training, Wepf et al⁶¹ found that the experience of being a carer (for a relative or close friend with a health problem) was negatively correlated with mental wellbeing only in those with lower levels of “trait benefit-finding” (ie, the disposition to perceive positive changes following adversity). In a longitudinal study of 237 Hong Kong-based university students, Lai and Mak⁵³ found that daily hassles (in the university student context) over a one-month period predicted lower mental wellbeing only among those with low baseline levels of optimism, self-esteem, and perceived control. In a study of over 1600 German secondary school and university students, Leipold et al⁵⁴ found that perceived stress related to one’s academic work was negatively correlated with subjective wellbeing, but less so among those who were higher in either meaning-focused coping (involving acceptance and positive reinterpretation) or social support-seeking (ie, seeking both instrumental and emotional support). A third coping style (problem-focused coping) did not moderate the stress-wellbeing relationship.

Two studies, both in the childrearing context, have examined external resilience resources (specifically, forms of social support) as moderators of the effects of occupational stressors. In one study, of 139 US licensed foster parents, Sharda⁵⁶ found that self-reported parenting stress was negatively correlated with wellbeing, but less so among those who reported higher levels of social support (received and given). In the other study, of 152 US kinship caregivers (ie, non-parental caregivers such as foster parents), Sharda et al⁵⁷ found that parenting stress was negatively correlated with psychosocial quality of life, but this relationship was not moderated by received social support.

A single study has shed light on whether the effects of occupational stress can be moderated by prior stressful experiences. In this study, Bäccman et al⁵¹ assessed 129 Swedish marines before and after an anti-piracy naval deployment. Interestingly, those with prior experiences of combat-related stress had lower levels of positive affect and higher levels of negative affect than their colleagues before but not after the deployment. This finding was not reported as a test of moderation, and it might merely reflect regression to the mean, but it points to a possible stress-inoculation effect that may warrant further study.

In summary, a number of studies have found that trait resilience and other personality traits may mitigate the negative impact of occupational stressors on positive affect, but the evidence is limited and mixed with regard to the possible protective role of social support in this context. It would also be unsurprising if stress-inoculation played a role in this context, given that occupational stressors are typically similar in severity to everyday life stressors (ie, stressful but not traumatic).

Health Conditions (Other Than COVID-19)

Our search returned seven studies that examined resilience in terms of the effects of health conditions (other than COVID-19) on positive affect. Five of these studies focused on internal resilience resources. In a study of over 1800 British older adults, Windle et al⁶⁸ examined whether trait resilience significantly moderated the negative association between number of chronic physical illnesses and life satisfaction across four age groups (50–59, 60–69, 70–79, and 80–89 years). There was a significant moderation effect in each group except the youngest one, but the nature of the effect was not uniform. In the 60–69 and 70–79 year old groups, the association between illness and life satisfaction was more negative among those with lower trait resilience. However, in the 80–89 year old group, this association was steeper for those with higher trait resilience. This seemingly inconsistent finding may have been due to a floor effect, because those aged 80–89 with low trait resilience had low levels of life satisfaction even when they reported few chronic illnesses. In addition, in a US-based study of 94 adults with or without stuttering, Plexico et al⁶⁷ found that stuttering was associated with lower satisfaction with life scores at lower levels of trait resilience but not at higher levels of trait resilience; however, neither adaptive nor maladaptive coping style was found to moderate the negative association between stuttering status and satisfaction with life. In a longitudinal, nationally representative study of over 11,000 German adults aged 40 years and over, Hajek and König⁶⁴ assessed how favourably the participants rated their own health relative to others of the same age, and whether these health comparisons became more or less favourable over time. Among those whose comparisons worsened from “the same” to “much worse”, there was a corresponding reduction in life satisfaction and an increase in negative affect. However, the effect on life satisfaction was weaker among those with higher self-efficacy, while the effect on negative affect was weaker among those with higher optimism or higher self-efficacy. Furthermore, while there was no main effect of worsened health comparisons on positive affect, there was an interaction effect reflecting smaller reductions in positive affect among those with greater optimism or self-esteem. In a study of over 12,000 adults from the general Dutch population, Bos et al⁶² found that self-defeating humour moderated the negative association between symptoms of depression, anxiety, and stress and subjective wellbeing, with a weaker association among those higher in this form of humour. However, no such interaction effects were found when other factors were analysed as moderators, including trait empathy, religious belonging, occupation, and three other styles of humour (affiliative, self-enhancing, and aggressive). A null result was also obtained by Kratz et al⁶⁵ in a longitudinal study of 122 US women with osteoarthritis and/or fibromyalgia. The authors found that pain acceptance did not moderate the negative association between weekly worst pain severity and weekly positive affect (although the positive relationship between pain severity and negative affect was weaker among those with greater pain acceptance).

Three studies have investigated external resilience resources as moderators of the impact of illness on positive affect, with mixed findings. In a sample of over 4000 Arizonan adults (weighted to be representative of the statewide adult population), Okun et al⁶⁶ investigated whether volunteering moderated the association between number of chronic physical health conditions and positive affect. Those with more chronic conditions reported both lower positive affect and lower trait resilience on average, but these deficits were significantly smaller among those who had engaged in volunteering in the previous 12 months (compared with those who had not). Furthermore, in the aforementioned study by Bos et al⁶² it was found that for participants with a partner, psychological distress symptoms had a weaker negative association with subjective wellbeing. However, Delfabbro et al⁶³ investigated the moderating influence of interpersonal factors in a study of body image satisfaction versus dissatisfaction, with a sample of over 1200 Australian secondary schools students. A range of potential moderators, reflecting sociability and family functioning (eg, extraversion, neuroticism, relationship status, number of friends, family adjustment), were tested, but none significantly moderated the negative relationship between body image dissatisfaction and self-esteem. However, in a sub-analysis of only those who were dissatisfied with their physical appearance, participants above the 90th percentile on self-esteem were found to be more extraverted and have better-functioning families, relative to those below that threshold.

In summary, both internal and external resilience resources have been found to moderate the relationship between health conditions and positive affect, with most of the studies pertaining to personality traits and other internal attributes. However, this area of research has covered only a limited range of health conditions to date. There may be certain forms of physical or mental illness whose impact on positive affect is more (or less) amenable to mitigation by a given resilience resource. Clearly, further research is needed in this domain.

COVID-19

Our search identified nine studies on resilience in relation to the COVID-19 pandemic and its effects on positive affect. Almost all of these studies specified that data collection occurred under strict lockdown conditions (between the years 2020 and 2021). Given that individuals had little control over their external circumstances during the lockdowns, it is understandable that many of the studies focused on internal resilience resources (such as trait resilience or other traits). However, a few researchers examined how other factors may have served as resilience resources during the pandemic.

Six studies (across five papers) investigated internal resilience resources as moderators, two of which focused on trait resilience. In a study of 711 Spanish university students, Paredes et al⁷⁴ found that the perceived threat of COVID-19 was negatively associated with positive affect via heightened pandemic-related anxiety about the future, with this indirect negative effect being stronger among those with lower trait resilience. In addition, in a study of over 1000 Chinese university students, Hu et al⁷³ found that perceived stress had a stronger negative association with both social adaptation (to the pandemic) and life satisfaction among those with lower levels of trait resilience (called “emotional resilience” by the authors). However, Hu et al neglected to specify the names of the scales they employed to measure these variables, so we are limited in the extent to which we can interpret their findings (eg, it is unclear whether their measure of perceived stress pertained to stress in general or pandemic-related stress specifically).

The other four studies of internal resources yielded mixed results. In a sample of over 2700 adults residing in Italy during the 2020 lockdown, Barni et al⁶⁹ found that participants who knew someone that had contracted the coronavirus reported lower psychological wellbeing, but less so if they had a greater sense of coherence (SOC, ie, the view that one’s life and world are comprehensible, manageable, and meaningful). Participants with a greater fear of contracting the virus also reported lower wellbeing, but this association was stronger among those with greater SOC. However, this latter finding may indicate a floor effect, because those with greater SOC had higher wellbeing across all levels of fear. Faul and De Brigard⁷¹ examined whether trait nostalgia served as a resilience resource for US adults in the first wave (study 1, April 2020) and second wave (study 2, August 2020) of the pandemic. In study 1 (involving 134 participants), those who reported greater concern about the pandemic were more likely to report worsened mood over the past month, but not if they were high in nostalgia. However, this moderation effect was not found for ratings of mood change over the past week. Furthermore, in study 2 (involving 159 participants), nostalgia did not moderate the association between pandemic concern and ratings of mood change over the past month nor the past week. In addition, in study 1, it was found that pandemic concern had a stronger negative association with positive mood among those lower in trait nostalgia, but again, this moderation effect was not replicated in study 2. During the April 2021 period, in a nationally representative daily-diary study of 444 US adolescents, Wang et al⁷⁶ found no evidence that either primary engagement coping (similar to problem-solving coping) or secondary engagement coping (similar to emotion regulation) moderated the negative association between daily health-related stress (including pandemic-related stress) and either same-day or next-day positive affect. Primary engagement coping also had no moderation effects on negative affect, although secondary engagement coping mitigated the negative association between health-related stress and both same-day and next-day negative affect. Wang et al also measured whether an external resource – parental support – moderated the relationship between daily stress and positive or negative affect. Parental support did not moderate the relationship with positive affect, and although it significantly moderated the relationship with negative affect, the data showed that parental support was not a significant moderator at higher levels of stress.

Three other studies investigated moderators other than internal resources. In a longitudinal study of the mental health outcomes of over 5000 Israeli children and adolescents in the period from pre-pandemic (September 2019) to post-fifth wave (May 2022), Shoshani⁷⁵ observed smaller reductions in positive emotion, life satisfaction, and gratitude among participants who reported higher levels of social support and/or higher levels of daily routine (eg, more frequent engagement in social activities, family rituals, distance learning tasks). In a study of over 900 US adults during the first wave of the pandemic (March-April, 2020), Cunningham et al⁷⁰ found that those who reported greater social isolation had lower levels of positive affect. This negative association was stronger among older adults, possibly indicating a floor effect, because older age was associated with greater positive affect at all levels of social isolation. Indeed, the positive associations between social isolation and negative outcomes (including negative affect and depressive symptoms) were weaker among older adults. In a study of over 800 Iranian adults, Haghayeghi et al⁷² found that participants who rated the pandemic as more disruptive to their lives reported lower levels of psychological wellbeing, but this negative association was weaker among those who reported smaller pandemic-related reductions in physical activity.

In summary, a range of resilience resources have been investigated as moderators of the effects of COVID-19 on positive affect, including internal resources such as trait resilience and other factors such as physical activity and age. Only two studies examined external resilience resources (parental support, social support, daily routine) in this context, with mixed findings.

Positive Affect as a Moderating Resilience Resource

As mentioned earlier, most of the studies retrieved from our literature search were concerned with positive affect as an outcome of the resilience process, rather than as a resilience resource. However, of the few studies that examined positive affect as a moderator, there were at least two that fell under each of the six subheadings employed in the previous section of this review (except COVID-19, for which there was only one study), so we have retained these subheadings in the present section. See Table 2 for a summary of the key variables in these studies.

Table 2 Summary of Reviewed Studies on Positive Affect as a Resilience Resource

Adverse childhood experiences (ACEs)
Author	Stressor	Moderator	Outcome
Arslan, 2023^Citation77	Psychological maltreatment Psychological Maltreatment Questionnaire (PMQ-B)	Positive affect Scale of Positive and Negative Experience (SPANE)	Social wellbeing Mental Health Continuum – Short Form (MHC-SF)
Logan-Greene et al, 2014^Citation78	ACEs Custom checklist	Life satisfaction Single-item	Physical and mental health Single-item measures
Adult trauma
Author	Stressor	Moderator	Outcome
Gee et al, 2023^Citation79	Potentially traumatic events Australian Aboriginal Version of the Harvard Trauma Questionnaire (AAVHTQ)	Multidimensional strengths Aboriginal Resilience and Recovery Questionnaire (ARRQ)	Trauma symptom severity AAVHTQ
Kumar et al, 2022^Citation80	Posttraumatic stress PTSD Checklist for DSM-5 (PCL-5)	Optimism Life Orientation Test – Revised (LOT-R) Gratitude Gratitude Questionnaire (GQ-6)	Suicidal ideation Single-item
Stressful life events
Author	Stressor	Moderator	Outcome
Brailovskaia et al, 2020^Citation81	Stressful life events Brief Daily Stressor Screening (BDSS)	Mental wellbeing Positive Mental Health scale (PMH)	Suicidal ideation Single-item
de Vries et al, 2021^Citation82	Negative life events Shock-Processing Inventory List (SchIL)	Life satisfaction Satisfaction with Life Scale (SWLS)	Anxiety and depression symptoms Achenbach System of Empirically Based Assessment (ASEBA)
Kathryn McHugh et al, 2013^Citation83	Perceived stress Perceived Stress Scale (PSS)**	Positive affect Profile of Mood States (POMS)	Negative affect POMS Alcohol cravings Obsessive-Compulsive Drinking Scale (OCDS)
Riskind et al, 2013^Citation84	Negative life events Psychiatric Epidemiology Research Instrument (PERI)*	Positive life events PERI*	Depressive symptoms Beck Depression Inventory (BDI)
Occupational stressors
Author	Stressor	Moderator	Outcome
Maffoni et al, 2020^Citation85	Lack of managerial support Hospital Ethical Climate Survey (HECS)	Positive affect Positive and Negative Affect Schedule (PANAS)	Moral distress Moral Distress Scale – Revised (MDS-R)
Viola et al, 2016^Citation86	Low hardiness (trait resilience) Dispositional Resilience Scale (DRS)	Psychological wellbeing Psychological Well-Being Scale (PWB-18)	Lack of career readiness Career Decision-Making Difficulties Questionnaire (CDDQ)
Health conditions (non-COVID)
Author	Stressor	Moderator	Outcome
Siegmann et al, 2018^Citation87	Depressive symptoms Depression Anxiety Stress Scales – 21-itemversion (DASS-21)	Psychological wellbeing PMH Life satisfaction SWLS	Suicidal ideation Suicidal Behaviors Questionnaire – Revised (SBQ-R)
Strand et al, 2006^Citation88	Pain intensity Single-item	Positive affect PANAS	Negative affect PANAS
Yu et al, 2021^Citation89	Depressive symptoms Modified Depression Scale (MDS), Patient-Reported Outcomes Measurement Information System (PROMIS)	Life satisfaction Single item Optimism LOT-R	Suicidality Composite International Diagnostic Interview (CIDI)*
Zautra et al, 2005^Citation90	Average weekly pain Single-item Interpersonal stress Custom scale	Positive affect Positive and Negative Affect Schedule – Expanded (PANAS-X)	Negative affect PANAS-X
COVID-19
Author	Stressor	Moderator	Outcome
Venkatesh et al, 2023^Citation91	Pandemic-related stress Single-item	Positive affect Subcomponents of Affect Scale	Depressive symptoms Center for Epidemiologic Studies Depression Scale – Revised (CESD-R)

Notes: As in the body of the review, the studies are grouped under headings according to the type of stressor involved. For each study, the key variables (stressors, moderators, and outcomes) are listed, and the measures used to assess these variables (where applicable) are shown in italics. Where established measures have been employed, they are listed by name; otherwise, the type of measure (eg, custom checklist, unnamed scale, single-item) is stated. *Modified/adapted from original. **Shortened/reduced/abbreviated version.

Adverse Childhood Experiences (ACEs)

In contrast to the relatively large number of studies examining the effects of ACEs on positive affect as an outcome, when we considered positive affect as a resilience resource, only two studies fit our search criteria. Both of these studies involved adult populations with negative childhood experiences that were retrospectively reported. In one study, involving a representative sample of over 19,000 US adults residing in Washington State, Logan-Greene et al⁷⁸ found that recalled ACEs were associated with poor physical and mental health among those with low life satisfaction, but not among those with high satisfaction. In the other study, of 381 Turkish university students, Arslan⁷⁷ measured the relationship between recalled ACEs (in the form of psychological maltreatment by parents) and social wellbeing, with trait resilience as a mediator and positive affect as a moderator. The author found a stronger negative association between recalled maltreatment and social wellbeing among those with lower positive affect. Interestingly, Arslan also found a stronger negative association between maltreatment and trait resilience among those with higher positive affect, but this might indicate a floor effect, because those with higher positive affect reported greater trait resilience at all levels of maltreatment.

Adult Experiences of Mistreatment or Trauma

Our search identified only two studies of positive affect as a resilience resource in the context of adult experiences of mistreatment or trauma. Consistent with the ACE studies that suggested a buffering role of positive affect, both studies on adult trauma found significant effects of positive affect-related moderators. In a sample of 81 Aboriginal Australian community adults attending a family counselling service, Gee et al⁷⁹ found a moderating effect of resilience resources encompassing both personal and relational/cultural strengths (and including a measure of positive affect). Those who had experienced more potentially traumatic events in their lifetimes reported higher levels of posttraumatic stress symptoms, but this association was not significant among those with high levels of strengths. In a study of 425 US female undergraduates who had previously experienced sexual assault, Kumar et al⁸⁰ examined the potential buffering roles of optimism and gratitude (both involving positive affect) in the relationship between posttraumatic stress and suicide risk. Both optimism and gratitude acted as protective factors, where higher levels of either resource predicted a weaker association between posttraumatic stress symptoms and suicidal ideation. This is consistent with the idea that both optimism and gratitude function to pivot one’s focus toward favourable aspects of life, thereby promoting mental health.

Stressful Life Events or Everyday Stressors

Five studies (across four papers) have examined positive affect as a resilience resource in the context of stressful life events. In a longitudinal study of 126 German university students, Brailovskaia et al⁸¹ found that stressful life events at baseline predicted greater suicidal ideation both at baseline and 24 months later, but these associations were far weaker among those with higher baseline levels of mental wellbeing (including positive affect and other aspects of wellbeing). Similarly, in a cross-sectional study of over 1300 US alcohol-dependent patients, Kathryn McHugh et al⁸³ observed a smaller positive association between perceived stress and negative affect among those with higher positive affect (however, although positive affect was negatively correlated with alcohol cravings, it did not moderate the positive association between perceived stress and cravings). Together, these results suggest that positive affect may serve as a resilience resource even in the context of more severe outcomes such as suicidality and more severe stressors such as substance addiction.

However, other research has produced less consistent findings. Riskind et al⁸⁴ conducted two studies of resilience resources in relation to positive and negative affect, positive and negative life events, and depression and anxiety symptoms. In the first study, of 99 US undergraduate students, baseline trait negative affect positively predicted depressive symptoms six weeks later in those with low baseline trait positive affect, but not in those with high baseline trait positive affect. However, trait positive affect did not have such an interaction effect on anxiety symptoms. In the second study, of 107 US undergraduates, there was an interaction between number of positive events (in the past year) and number of negative events (in the past year) in predicting depression symptoms (but again, not anxiety) six weeks later. For those who had experienced a low number of positive events, there was a positive association between number of negative events and depression. However, for those who had experienced a high number of positive events, there was a negative association between negative events and depression. Thus, experiencing fewer negative events was seemingly a liability for those who reported a high number of positive events. This counter-intuitive finding warrants further exploration.

Finally, a single study has considered genetic factors in the context of positive affect as a potential moderator of the impact of stressful life events. In a sample of over 13,000 members of the Netherlands Twin Register (including pairs of twins and their biological siblings), de Vries et al⁸² investigated the longitudinal association (over periods of up to 10 years) between wellbeing and resilience to negative life events. Resilience was operationalised as the residuals from regressing a measure of anxiety and depression symptoms on a measure of the number of negative events reported by the participants in their lifetimes. Thus, higher resilience scores indicated lower levels of anxiety and depression than would have been expected given the number of reported negative events. After accounting for genetic overlap, the authors found evidence of a causal effect of life satisfaction on resilience, with those higher in life satisfaction reporting lower levels of anxiety and depression for a given number of negative events. However, there was also evidence of a causal effect in the opposite direction, whereby resilience positively predicted wellbeing at follow-up.

Occupational Stressors

Two cross-sectional studies, both conducted in Italy, have examined positive affect as a moderator of the impact of occupational stressors. In one study, of 222 patient-facing healthcare professionals, Maffoni et al⁸⁵ found that managerial support (in dealing with ethical issues at work) was negatively correlated with moral distress (and thereby emotional exhaustion) only in those with higher levels of positive affect. This finding suggests that positive affect accentuates the benefits of managerial support, but conversely, it can be interpreted as suggesting that positive affect worsens the distress caused by a lack of managerial support. Clearly, this domain warrants further research. In the other study, of 131 never-employed young adults, Viola et al⁸⁶ found that psychological wellbeing was negatively associated with lack of career readiness (ie, career indecision and lack of motivation to find employment) more strongly in those with lower levels of hardiness (ie, lower trait resilience). This interaction effect suggests that wellbeing might be a more impactful resource for those with lower trait resilience, at least in some contexts. However, it could also be interpreted in the reverse direction, with lack of career readiness as the stressor, wellbeing as the outcome, and hardiness as the resilience resource. Again, further research is required.

Health Conditions (Other Than COVID-19)

Four studies have examined positive affect as a possible protective factor for those with various health conditions. Two of these studies found that positive affect (or related wellbeing measures) significantly moderated the association between depression and suicidality. In a longitudinal study of over 1900 US adolescents, Yu et al⁸⁹ classified the participants into three profiles: low levels, mild levels, and moderate-to-high levels of depressive symptoms. Regardless of life satisfaction, those in the low profile had a low risk of suicide, and those in the moderate-to-high profile had a high risk of suicide. But in the mild profile, life satisfaction was negatively associated with suicide risk. This suggests that while life satisfaction may serve to mitigate the impact of mild depression, it may not be so effective in the context of more severe depression. However, Yu et al also found that there was a negative association between optimism and suicide risk only for the moderate-to-high profile, which suggests that positive affect-related traits (such as optimism) may serve as resilience resources even for those with severe depression. Furthermore, Siegmann et al⁸⁷ examined whether moderators of the association between depression and suicidal ideation were the same cross-culturally in large samples of German (601 participants) and Chinese (over 2600 participants) university students. In both populations, the negative association between depression severity and suicidal ideation was weaker for participants who reported higher levels of mental wellbeing (including positive emotions). Life satisfaction played a similar buffering role, but only for the German participants.

In addition, two studies have found evidence of the possible protective role of positive affect for adults experiencing chronic physical pain. In one study, of 43 rheumatoid arthritis (RA) patients, Strand et al⁸⁸ made weekly measurements of the severity of the participants’ most intense RA-related pain. Those with more severe pain reported higher levels of negative affect, but this association was weaker among those with higher positive affect. In the other study, of 124 women with fibromyalgia and/or osteoarthritis, Zautra et al⁹⁰ measured the participants’ average weekly levels of both physical pain (due to their health condition) and perceived interpersonal stress. Both pain and interpersonal stress were positively associated with negative affect, but less so among those with higher positive affect. Together, these studies suggest that positive affect can be an effective resilience resource even in the context of severe physical pain.

COVID-19

A single study investigated positive affect as a resilience resource in the context of the COVID-19 pandemic. In a longitudinal study of 292 US university students, Venkatesh et al⁹¹ administered five surveys in the period May-December 2020. Higher levels of pandemic-related stress at a given timepoint predicted higher levels of depression symptoms at the next timepoint, but this positive association was weaker among those with greater positive affect at the earlier timepoint. However, this moderation effect was not found when anxiety, physical symptoms, or overall health were the outcomes. Furthermore, although the stress-depression link was stronger among those with lower positive affect, those with higher positive affect tended to report greater depression across almost all levels of stress, suggesting that positive affect was mostly a liability rather than a buffer. Despite this, the authors interpreted the interaction effect as if it revealed a protective effect. At most, this conclusion was warranted only at the highest levels of pandemic-related stress.

Discussion

Summary of Review

In the context of resilience, mental health research has examined positive affect either as an outcome of the resilience process or as a resource that enables or promotes this process. In both roles, positive affect has been studied in relation to six broad categories of stressor: ACEs, PTEs in adulthood, general life stress, occupational stress, health-related stressors (except COVID-19), and COVID-19-related stress.

Studies of positive affect as an outcome have suggested that the negative impact of these various stressors can be mitigated by a diverse range of resilience resources. These include internal resources such as personality traits and other personal attributes, as well as external resources such as interpersonal connections and social factors. Numerous studies (eg,^{19,28,36,58,68,74}) have found that those with higher trait resilience experience smaller reductions (or larger rebounds) in positive affect following a stressor. The possible protective roles of several other traits have also been revealed. These traits typically involve a positive attitude toward the future (eg, hope,³⁸ optimism,^53,64 positive expectancy,³⁰ reward sensitivity³⁴), a positive attitude toward oneself (eg, self-efficacy,⁶⁴ self-esteem,^53,64 self-compassion,³³ internal locus of control,³¹ perceived control,^24,53 problem-solving coping style,¹⁷ sense of ethnic identity²⁵), or the tendency to see positives or meaning in one’s life (eg, trait benefit-finding,⁶¹ sense of coherence,⁶⁹ meaning-focused coping style,⁵⁴ sense of life meaning,¹⁸ belief in a just world²⁷). The Big Five personality traits have also been examined as resilience resources, with evidence that Neuroticism exacerbates the negative impact of stressors on positive affect, while the other four traits (especially Conscientiousness and Extraversion) mitigate this impact.^32,47,59 A few studies^41,44,48 have also examined the possible protective role of emotion regulation, finding that those who employ cognitive reappraisal are less negatively affected by adversity.

When we consider the psychological processes by which each of the aforementioned traits may enable resilience, the most plausible mechanisms are usually discernible from the name of the trait itself (or the category of traits to which it belongs). For example, we may readily hypothesise that the future-oriented traits^38,53,64 promote resilience via future-directed thoughts and behaviours. One may be more motivated to overcome adversity if one can envision a positive future awaiting oneself (making the present hardship worth one’s while), and one’s responses to adversity may be more effective if one is better able to conceive of viable paths linking present actions with future outcomes (see Snyder’s Hope Theory⁹²). However, when we consider trait resilience specifically, the relevant psychological mechanisms are not immediately evident. This is because the term “trait resilience” is used in disparate ways across the literature. For example, one of the most popular trait measures of resilience – the Connor-Davidson Resilience Scale (CD-RISC)⁹³ – comprises five dimensions, covering a wide range of characteristics such as personal competence, acceptance of change, relationship security, perceived control, and spiritual attitudes. In contrast, another popular trait measure – the Brief Resilience Scale (BRS)⁹⁴ – does not refer to any personal resources at all. Instead, the BRS simply asks the respondent whether or not they tend to recover quickly and easily from hardship (via six essentially synonymous questions). Thus, scales such as the CD-RISC actually measure one or more traits (eg, perceived control) that may serve as resilience resources, whereas scales such as the BRS pose autobiographical questions about the extent to which the respondent typically recovers from adversity, without inquiring about the personal attributes that may promote or impede such recovery. Therefore, we advise the reader that any findings pertaining to trait resilience (in the present review and elsewhere) must be interpreted with reference to the specific measures used. Where multidimensional measures such as the CD-RISC have been used, the reader must speculate as to which of the incorporated traits might have been responsible for any observed resilience (unless sub-analyses of the individual scale dimensions were performed). Where autobiographical measures such as the BRS have been used, the reader is left to wonder whether any differences in resilience were due to genuine differences in (unmeasured) personality traits or instead due to non-personality differences that covary with self-reported recovery from adversity.

Furthermore, even when a unidimensional trait measure has been used (such as the 10-item version of the CD-RISC⁹⁵), the reader is not able to discern any plausible psychological mechanisms without inspecting the individual items comprising the scale. Upon doing so, the reader may discover that many of the items are autobiographical rather than personality-related (at least at face value). For example, in the 10-item CD-RISC, some of the items arguably refer to traits such as self-efficacy (eg, “I see myself as a strong person”) or a sense of humour (eg, “I try to see the humorous side of problems”), but many of them refer only to the phenomenon of resilience itself (eg, “I tend to bounce back after illness or hardship”, “I can achieve goals despite obstacles”, “I am not easily discouraged by failure”). Such items may serve as useful measures of resilience insofar as they correlate strongly with the phenomenon, but they do not shed any light on how the process of resilience unfolds nor the psychological mechanisms that enable it. Indeed, to answer the question of why a given individual tends to bounce back from adversity, it would be both pointless and circular to mention that the individual strongly agrees with the statement “I tend to bounce back from adversity”. In short, we caution the reader to be mindful of circular reasoning (ie, “He’s resilient because he’s high in trait resilience”) and to examine the specific scales and items subsumed under the label “trait resilience” in order to interpret the relevant studies accurately.

Shifting from personality characteristics to interpersonal or social factors, a range of resilience resources have been found to be potentially protective of positive affect. These include social support (given and received),⁵⁶ having a romantic partner,⁶² teacher support (for students),⁴⁷ positive peer relations,⁴⁰ parent-child communication,⁴⁷ sense of community,¹⁵ family meals,⁴³ outdoor excursions,⁴⁶ volunteering,⁶⁶ and cross-ethnic friendships.⁹ Other relevant moderators include personal characteristics that relate to social factors, such as loneliness (exacerbating stress)³² and social support-focused coping style (mitigating stress).⁵⁴

As the preceding summary demonstrates, studies of positive affect as a resilience outcome have focused primarily on resilience resources of a personal or interpersonal nature, especially personality traits and various forms of social support or social connection. Aside from these resources, we note that only a single study has examined each of the following moderators of the link between adversity and positive affect: physical exercise,¹⁴ educational attainment,³² daily routines,⁷⁵ vagal suppression,¹⁰ and low heart rate reactivity.²⁰ In addition, a small number of studies have examined whether a given form of adversity can moderate its own impact (or the impact of other forms of adversity) on positive affect, with mixed findings. There is evidence that the negative effects of one stressor can be compounded by those of another stressor,^15,23,37 but there is also evidence of curvilinear relationships whereby a low or moderate level of adversity (either concurrently or in one’s past) can be protective of positive affect, relative to no adversity or high levels of adversity.^35,45,49 The latter findings concur with the literature on stress inoculation and post-traumatic growth,^96,97 reminding us of the formative role that hardship can play in the development of mental health and fortitude.

Although dozens of studies have examined positive affect as an outcome of the resilience process, far fewer studies have treated positive affect as a resilience resource that can mitigate the impact of stressors on other outcomes. Moreover, the latter studies pertain to a relatively narrow range of mental health outcomes, predominantly comprising symptoms related to depression^82,84 or suicide,^80,81,87,89 or other negative outcomes such as anxiety, trauma, or negative affect.^82,83,88,90 Other outcomes, including moral distress,⁸⁵ career readiness,⁸⁶ social wellbeing,⁷⁷ and physical health,⁷⁸ have each been addressed by a single study only.

Despite the relative lack of studies on positive affect as a resilience resource, a number of key findings are evident from the literature. In the domain of ACEs, positive affect may mitigate the impact of early life stress on both physical and mental health outcomes.⁷⁸ Regarding PTEs in adulthood, positive affect may reduce the risks of posttraumatic stress and suicidality.^79,80 Similarly, in the context of stressful life events, positive affect may be protective against negative affect, depression, anxiety, and suicidal ideation.^81–84 Indeed, when depression itself is the stressor, positive affect may be protective against suicidality.^87,89 Although these key findings are promising, far more research is needed in order to understand the full scope of positive affect as a resilience resource.

Theoretical Implications

A key theoretical implication of our review is that it reinforces the conception of mental and emotional health as comprising two overlapping but distinct areas of functioning, pertaining to positive emotions and mental wellbeing on the one hand versus negative emotions and mental illness on the other. Where appropriate, we have noted throughout the review those studies which yielded divergent results with regard to positive versus negative constructs (eg, a resilience resource that mitigated a stressor’s impact on positive but not negative affect). Such findings corroborate our modern understanding of mental health, often termed the dual-continua model.⁹⁸

Another implication pertains to the nature of moderation effects and how they are interpreted. Without a graph of an interaction effect (or inspection of the corresponding conditional means), the effect cannot be interpreted merely from the associated regression coefficient. For example, a positive coefficient might reveal that the negative association between a stressor and positive affect becomes less negative at higher levels of the resilience resource in question. However, while this seems to indicate that the resource is protective, such a coefficient could instead originate from a floor or ceiling effect. For instance, those with high levels of the supposedly protective resource might have low levels of positive affect across all values of the stressor, in which case the resource would actually confer no advantage at high levels of stress and a disadvantage at low levels of stress. For most of the studies in our review, the authors provided sufficient graphical or numerical summaries for us to determine the exact nature of the interaction effects tested. Where appropriate, we have offered alternative interpretations of results whose authors may have omitted the necessary data or neglected to consider all explanations of their findings. Overall, we recommend that future scientists not only perform moderation analyses of candidate resilience resources (rather than testing only direct effects or mediations) but also publish all information needed to interpret the analyses accurately (ideally in graphical form).

We raise the issue of moderation analyses not only as a methodological critique but also to highlight some of the interesting findings of our review. In almost every study that yielded a floor or ceiling effect, the purported resilience resource conferred little or no advantage at high levels of adversity and a large advantage at low levels of adversity. In these cases, while there is no evidence that the resource is protective against the stressor in question, the resource can still be considered a net asset, being associated with positive outcomes at least at low levels of the stressor. Such results invite researchers to speculate on the possible causes of the ceiling or floor effect. It may be that the stressor is especially impactful, reducing everyone’s wellbeing to a minimal level (at high levels of the stressor) regardless of individual differences in the purported resilience resource. Or it may be that the resource actively compounds the benefits that exist at low levels of adversity (eg, it might accentuate the mental health benefits of relaxation or contentment) while being inert at high levels of adversity. These theoretical questions warrant exploration in future research.

Gaps and Future Directions

Our literature review has encompassed various studies covering a wide range of stressors, resilience resources, and outcomes. These studies reveal an evidence base with many gaps, but each gap represents not only missing knowledge but also opportunities for future research.

One major gap is evident from the body of this review: Relatively few studies have treated positive affect as a resilience resource. As discussed earlier, it is understandable that many researchers would treat affective states as outcomes of antecedent experiences, but it is also important to remember the motivational role of these states. From an evolutionary perspective, it is obvious that affective states serve to motivate productive behaviours in countless settings, so it is incumbent on resilience researchers not to neglect the possible protective role of positive emotional responses in the context of stress or adversity.

Another major gap pertains to the ways in which positive affect has been measured in the resilience context. As shown in Table 1 and 2, many resilience studies have employed outcome measures that incorporate positive affect but do not measure it as a standalone variable. If we had restricted our analysis only to those studies with a “pure” measure of positive affect, our review would have covered far fewer studies and areas of interest. Future scientists may wish to employ specific measures of positive affect in replications of those studies which did not include such measures originally, to evaluate whether the effect still holds for positive affect when taking into consideration related constructs.

Our review has covered numerous stressors, resources, and outcomes, as well as a wide range of countries, populations, and settings, but this diversity is both a strength and a limitation. Obviously, it is desirable to study positive affect and resilience across varying contexts, but the variability of the literature is also a measure of the lack of replications therein. Direct replications are missing from the literature, and few studies have examined the same populations, stressors, moderators, or outcomes in the same ways. Even where, for example, a resilience resource (such as trait resilience) has been investigated by multiple studies, the studies employ differing measures of the resource or differing measures of the associated outcomes. Again, this is not merely a limitation but also a substantial opening for future inquiry. In addition, there have been relatively few studies that employed experimental or clinical trial designs in the area of positive affect and resilience. This is understandable, given the ethical constraints involved in research on stress and trauma, but there is still scope for innovative experiments and interventions to test resilience resources in more rigorous and realistic ways. Another gap relates to the kinds of stressors examined in the literature to date. Understandably, much of the research has focused on more severe or salient stressors (such as child maltreatment), but we must remember that resilience pertains to not only recovery from adverse events but also adaptation to stressors more broadly, including life challenges that may be demanding but fulfilling (eg, competitive sport). Future resilience researchers may wish to explore the roles of positive affect in the context of such stressors.

A key strength of our review is that it employed a rigorous definition of resilience, whereby resilience resources moderate the impact of a given stressor on a given mental health outcome. We also included all null results of the moderation analyses reported in the included studies, to give the most complete portrayal of the literature possible. However, we must stress that while a given variable may not have an interactive effect within the resilience process, it may still have an additive effect on the outcome in question. Thus, a given attribute might not serve as a resilience resource (in that it does not mitigate the impact of any stressor) but it might still be worth promoting such attributes because of other benefits they may bring beyond the context of recovery or adaptation. Furthermore, even when a significant interaction effect is present, the reader must exercise caution in interpreting the nature of the effect. We recommend consulting graphical depictions of such interactions, in order to assess whether any ceiling or floor effects may be influencing the interpretation (see our earlier discussion in Theoretical Implications).

Finally, we should note that some authors may be studying the resilience process without always using the terminology of resilience. To take a deliberately extreme example, if an author had studied recovery from adversity, but referred to this process exclusively with the term “amelioration”, then their study would not have been included in our review. This is not a weakness of the review because it would have been beyond the scope of this paper to include all synonyms of resilience. We simply wish to alert the reader to the possibility of alternative conceptualisations, and to recommend that future researchers use the terminology of resilience when studying the process of overcoming adversity or adapting to stressors, in order for their findings not to be overlooked by the mainstream of the field.

Concluding Remarks

The field of research on resilience and positive affect is relatively young, but the literature has already provided a wide array of studies covering numerous forms of stress and adversity. Most of the studies measured positive affect as an outcome of the resilience process, highlighting various internal and external resources that may be protective of positive affect and mental wellbeing in the face of stressors. Although relatively few studies have treated positive affect as a potential resilience resource, their findings show that positive affect may be protective of mental health following adversity in childhood, adulthood, and everyday life. In all, there are countless opportunities to build upon the knowledge base covered in this review. Numerous stressors, moderators, and outcomes have yet to be examined in the context of positive affect and resilience. Indeed, an inspection of the years of publication of the reviewed papers shows how many were published only in the last few years. Clearly, this field is expanding quickly, which aligns with the broader acceleration of modern psychological research on the positive dimensions of mental health. The potential for new discoveries, and thereby new approaches to helping people recover from or adapt to stress and trauma, seems immense.

Disclosure

LAE, HRPP, and JMG were supported by a National Health and Medical Research Council (NHMRC) Project Grant (1122816) awarded to JMG. JL was supported by a University of New South Wales Research Training Program PhD scholarship and a Neuroscience Research Australia PhD Pearl Program award. All authors declare that they have no other conflicts of interest in this work.