Introduction

All people experience interpersonal hurts or offenses at some point in life1. From a stress-and-coping perspective, interpersonal transgressions can precipitate a perceived sense of injustice to which they might respond with unforgiveness—a complex cognitive, emotional, and/or behavioral response that may involve (among other factors) bitterness, resentment, hostility, hatred, anger, fear, and revenge-seeking or avoidance motives2. It is not uncommon for individuals who have been wronged to struggle with unforgiveness. To illustrate, recent nationally representative data from Wave 1 of the Global Flourishing Study (GFS) indicated that approximately 25% of people across the 22 countries analyzed reported ‘rarely/never’ forgiving those who have hurt them, with national estimates ranging from 8% in Nigeria to 59% in Turkey3. When unforgiveness persists or remains unresolved, it can have detrimental effects on well-being4.

Individuals may attempt to cope with their unforgiveness through various strategies (e.g., seeking revenge, deferring to divine judgment, excusing the offender’s behavior). From a stress-and-coping perspective, forgiveness is considered an adaptive strategy for reducing the negative effects of unforgiveness on well-being5,6. Given how common interpersonal hurts or offenses are, the frequent struggles that people have with unforgiveness, and the potential benefits of forgiveness (especially when practiced consistently across time and situations), the implications of forgiveness for population well-being may be considerable7. In the present study, we use longitudinal data from 23 national representative samples to explore the population-level associations of dispositional forgivingness—the general tendency to forgive others—with a wide range of well-being outcomes assessed approximately one year later.

While scholars do not completely agree on the definition of forgiveness8, there is broad consensus that forgiveness involves a prosocial shift in a person’s thoughts, emotions, motivations, and/or behavioral intentions toward someone who has hurt or offended them4,9. Aggregation of specific instances of forgiveness can develop into a more enduring tendency to forgive others across situations and over time (i.e., dispositional forgivingness). Whereas a single incidence of forgiveness can temporarily boost well-being, the practice of forgiveness over time (i.e., forgivingness) can affect well-being more generally. However, forgivingness—despite being a disposition—is not perfectly stable. Recent events or highly salient transgressions can shift people away from a forgiving disposition. For example, offenses that re-trigger interpersonal traumas (e.g., a survivor of child sexual abuse learning that a child in their family has been abused), serious and ongoing abuse, or repetitions of earlier wrongs (e.g., a subsequent spousal infidelity after a couple reconciles) can influence the stability of forgivingness. In these types of cases, the cognitive and emotional demands associated with processing the transgression may temporarily suppress forgivingness so that a person’s score on a measure of forgivingness may be lower than usual10. However, people will typically return to their usual level of forgivingness after the immediate hurt or offense they are dealing with has been resolved11. Events like reconciliation of warring groups or individual experiences of reconciliation in broken important relationships also have the potential to move low levels of forgivingness higher.

Well-being has been conceptualized in different ways over the years. One increasingly popular framing positions well-being as a multidimensional concept that is embedded in a broader notion of human flourishing defined as “the relative attainment of a state in which all aspects of a person’s life are good, including the contexts in which that person lives” (p. 19)12. Compared to other frameworks that adopt a narrower view of well-being, human flourishing offers a holistic approach by considering “the quality of [an individual’s] personal subjective state” across the various dimensions of human existence (p. 4)13. Although there are many ways to carve up the broad conceptual terrain covered by this framework, in alignment with prior work14,15 the scope of the well-being outcomes that are central to the present study might be organized thematically into psychological (i.e., psychological well-being, psychological distress), social (i.e., social well-being, social distress, social participation), physical (i.e., physical health and health behavior), volitional (i.e., character and prosocial behavior), and material (i.e., socioeconomic) dimensions of well-being. If forgiveness—or more broadly forgivingness—is a ‘whole person experience’16, its influence may extend across these dimensions of well-being.

An abundance of observational (mostly cross-sectional) studies (and to a lesser extent experimental intervention studies) have reported evidence on the association between forgiving others and different aspects of well-being16. The psychological dimension has received the lion’s share of attention in this area of research, in part because theorizing (e.g., stress-and-coping theory) suggests that the emotion-regulation function of forgiveness has more direct and proximal benefits for reducing psychological distress and improving psychological well-being17. Several recent reviews4,6 and meta-analyses18,19 have shown that forgiving others is negatively related to indicators of psychological distress (e.g., anxiety symptoms, depression symptoms) and positively related to indicators of psychological well-being (e.g., life satisfaction, sense of purpose). While most existing studies are based on correlational data, findings for the psychological dimension of well-being tend to replicate in more rigorous longitudinal and experimental research10,20,21,22.

A similar pattern of findings has been observed for other dimensions of well-being, though fewer and less consistent evidence for non-psychological dimensions may reflect more indirect and downstream effects of forgiveness compared to the psychological dimension. For example, the psychological relief that forgiveness provides to the forgiver is thought to lay the foundation for the possibility of reconciling (when safe and appropriate) with the transgressor, experiencing deeper social connectedness more generally, and building trust in others23. Providing some support for this theorizing, prior (mostly cross-sectional) studies have positively linked forgiving others to indicators of social well-being (e.g., social connectedness, trust) and social participation (e.g., marriage, religious service attendance), and reported negative associations with social distress outcomes (e.g., perceived bullying victimization, loneliness)3,24,25,26,27.

Viewed as a moral virtue, forgiveness represents an intentional response to wrongdoing that is guided by an aspiration to act in accordance with one’s ethical commitments1. This suggests that forgiveness of others should be related to volitional types of well-being outcomes that reflect choices, actions, or experiences that are guided by a person’s agency, will, and commitment to upholding their moral convictions or personal values6. Evidence in support of this idea comes from (largely cross-sectional) research that has reported positive associations between forgiveness of others and character strengths or virtues, such as hope, gratitude, and temperance21,28,29, as well as prosocial behaviors like conciliatory behavior, making charitable donations, and volunteering6,20,30. While these findings suggest a pattern, further work is required to expand and strengthen the robustness of existing evidence and better understand how forgiveness of others is related to various indicators of volitional well-being (e.g., helping strangers, showing love to others).

A growing number of (mostly cross-sectional) studies have identified links between forgiveness of others and indicators of better physical health, physiological markers associated with health, and health behavior (for meta-analyses, see refs. 20,31, for reviews, see refs. 5,32). For example, prior work has shown that forgiveness of others is related to more favorable physiology (e.g., lower blood pressure, lower cortisol), subjective experiences or perceptions of physical health (e.g., lower bodily pain, higher self-rated health), and health-related behavioral outcomes (e.g., lower risk of substance abuse, higher quality dietary intake). Forgiveness of others generally tends to exhibit weaker associations with outcomes corresponding to the physical dimension of well-being compared to other dimensions (e.g., psychological, social), although disproportionately fewer studies have attended to the physical dimension of well-being16. While many hundreds of studies focusing on psychological outcomes have been published in the last two decades, a somewhat recent review of studies addressing physical health over an 18-year period prior to the database search date yielded just 55 studies33. With more than 80% of those studies based on cross-sectional data, there is a need for more rigorous research to draw firmer conclusions about the potential implications that forgiving others may have for the physical dimension of well-being.

Forgiveness of others may also influence material well-being through indirect pathways over time, such as by lowering chronic stress that can be disruptive to socioeconomic aspects of life (e.g., maintaining steady employment), supporting stability in the network of relationships that can contribute to socioeconomic standing (e.g., connections to potential employment opportunities), or freeing up cognitive-emotional resources to focus on important goals (e.g., educational or vocational pursuits) tied to material well-being34,35. Although research has examined forgiveness of others in relation to selected indicators of material well-being, evidence is generally quite limited and mixed. For example, some cross-sectional studies have reported that forgiveness of others is positively associated with educational attainment36, whereas in other studies this association has been more negligible37. In one of the largest studies to date, a recent analysis of Wave 1 GFS data from 22 countries showed that the socioeconomic subgroups with the lowest forgivingness in the greatest number of countries included those who were unemployed (in eight countries out of 22) and those with eight or fewer years of education (in 11 of 22), indicating that these subgroups were not consistently lowest in forgivingness across all countries3. While these findings represent a step forward in identifying vulnerable subgroups on selected socioeconomic indicators across a wide swath of countries, the evidence is insufficient for addressing questions of causality because the data were cross-sectional and the analysis did not include adjustment for potential confounders. Hence, additional work is needed to strengthen insights into the potential causal effects of forgiving others on outcomes representing the material dimension of well-being.

Prior research has contributed considerably to strengthening our understanding of the relationship between forgiveness of others and well-being, but some important gaps remain. First, most observational studies have relied on cross-sectional data that are typically not informative for drawing potential causal inferences. For example, when forgiveness of others and indicators of well-being are assessed contemporaneously, temporal sequence often cannot be discerned. Longitudinal studies can allow for clearer examination of temporal dynamics.

Second, prior work in this area has typically examined one or a few well-being outcomes, often within a single domain6. Longitudinal cohort studies that include multidimensional assessments of well-being are needed to address temporal dynamics between forgiveness of others and outcomes of interest, develop more precise estimates of associations, and build a more integrative understanding of how forgiveness of others might be implicated in human flourishing. A few recent studies have attempted to address some of these gaps10,20,25,28. For example, Chen et al.20 examined associations of forgivingness with 25 well-being outcomes assessed between 3 and 6 years later in a large United States sample of young adult children of nurses. Forgivingness was associated with better well-being on 11 of the outcomes, with stronger and more consistent associations observed for psychosocial outcomes (e.g., depression symptoms, life satisfaction, self-esteem) compared to physical health and health behavior outcomes. Findings have been similar in other studies that have examined a wide range of well-being outcomes10,25.

Third, existing research on forgiveness of others and well-being has tended to focus on population segments (e.g., university students, young adults, nurses). Reliance on selective samples can make it difficult to determine if findings apply to the general population due to potential sampling bias (e.g., convenience sampling), the influence of shared characteristics or context-specific features within a sample on forgiveness-related processes and outcomes, or the inability to account for broader socioecological variation that may be relevant. For example, one large longitudinal study of young adults found little evidence of association between forgivingness and physical health20, but this may mask population-level effects because young adults tend to have comparatively fewer physical health problems. To more accurately estimate the potential impact of forgiving others on well-being at the population level, studies with representative samples of the general population are needed.

Fourth, most previous research reporting associations between forgiveness of others and indicators of well-being has been concentrated in more Western, educated, industrialized, rich, and democratic (WEIRD) societies. For example, one recent review on forgiveness of others and well-being reported that about two-thirds of the samples were from North America or Europe6. Comparatively fewer studies have been conducted in less WEIRD contexts. Although the findings of studies involving samples from less WEIRD contexts generally align with the broader empirical research linking forgiveness of others with well-being, methodological weaknesses of this literature (e.g., cross-sectional designs) constrain its utility. Principles of socioecological systems theory38 would suggest that contextualized historical events, cultural dynamics, and systemic factors have the potential to impinge upon experiences of forgiving others and its relationship to well-being outcomes39,40, which implies that we should not assume forgiveness of others will have the same effects on different aspects of well-being in different countries or regions around the world. While some prior research points to the possibility that the benefits of forgiving others for well-being may vary across different socioecological contexts22, no prior study has used nationally representative data from multiple countries to explore this possibility across a diverse range of cultures and geographic contexts.

To help address some of these gaps in knowledge, this preregistered study uses two waves of nationally representative data from the GFS to examine longitudinal associations of self-reported forgivingness assessed in Wave 1 with two subsequent composite well-being indicators and an array of specific indicators of well-being across eight domains: psychological well-being, psychological distress, social well-being, social distress, social participation, character & prosocial behavior, physical health & health behavior, and socioeconomic outcomes. Based on previous research, we expected that forgivingness would generally be associated with improved subsequent well-being across the 23 countries included in the GFS, although associations were anticipated to be somewhat stronger and more consistent for outcomes in some domains (e.g., psychological well-being) compared to others (e.g., physical health & health behavior). Due to the diversity of socioecological contexts represented in the GFS, we expected that the pattern of associations between forgivingness and the outcomes would vary to some extent across the different contexts.

Methods

The study design, sampling, and survey development for the GFS are described elsewhere15,41,42,43. The methodology for the analyses outlined below is described in detail in Padgett et al.44 and VanderWeele et al.45. The GFS was ruled exempt by the Baylor University Institutional Review Board (#1841317-2) because it met the criteria for exemption according to Baylor’s Institutional Review Board guidelines (e.g., minimal risk to participants, adherence to specific federal regulations). Ethical approval for all data collection activities was also obtained from the Institutional Review Board at Gallup Inc. Data collection activities were performed in accordance with relevant ethical regulations, and informed consent was obtained from all participants. All direct common identifiers were removed from the data used in the present study by Gallup Inc.

Study sample

Wave 1 of the GFS included nationally representative samples from 23 countries: Argentina, Australia, Brazil, China, Egypt, Germany, Hong Kong (Special Administrative Region of China), India, Indonesia, Israel, Japan, Kenya, Mexico, Nigeria, the Philippines, Poland, South Africa, Spain, Sweden, Tanzania, Turkey, the United Kingdom, and the United States (N = 207,919). The countries were selected to (1) maximize coverage of the world’s population, (2) ensure geographic, cultural, and religious diversity, and (3) prioritize feasibility based on Gallup Inc.’s existing data collection infrastructure. Data for Wave 1 were collected from March 2022 to January 2024, except in China where data were collected in March and April 202446. Data for Wave 2 were collected from January 2024 to December 2024, with data in China collected at least six months after Wave 1. The rate of nonresponse to the Wave 2 survey was 38% in the total sample. Descriptive statistics for the total sample by retention status are reported in Tables S3–S4. Attrition was 38% or below in 11 samples (lowest in China at 9%), whereas attrition in the other 12 samples ranged from 41% to 80% (highest in Hong Kong). Descriptive statistics for each country by retention status are reported in Tables S9c, d to S31c, d.

The GFS survey assesses various aspects of well-being, including happiness, health, meaning, character, relationships, and financial security, along with a range of other demographic, social, economic, political, religious/spiritual, personality, childhood, community, health, and well-being variables. Gallup Inc. translated the GFS survey into multiple languages following the TRAPD (translation, review, adjudication, pretesting, and documentation) model for cross-cultural survey research47. Additional details about the translation, cognitive interviewing, and pilot testing phases of the GFS can be found elsewhere42,48,49.

Sampling design

The precise sampling design varied by country to ensure samples were approximately nationally representative43,47. In most countries, local field partners implemented a probability-based face-to-face or telephone methodology to recruit panel members. Recruitment involved an intake survey gathering basic sociodemographic information and details for recontacting participants. Following recruitment, participants received invitations to participate in the annual survey via phone or online. Follow-up for Wave 2 data collection relied on the respondent-provided contact information. A minimum of three contact attempts were made on different days of the week and times of the day to maximize the possibility of retention. Post-stratification and nonresponse adjustments to the Wave 1 sampling weights were performed separately within each country, using either census data or a reliable secondary source. Additional information about the sampling design and weighting scheme for Wave 2 is available elsewhere44,46.

Outcome-wide analytic design

An outcome-wide analytic approach for longitudinal designs50 was employed to examine the associations of a single exposure (i.e., dispositional forgivingness) with a range of subsequent outcomes. Compared to traditional analytic strategies focused on a single outcome, this approach provides a more holistic assessment of an exposure’s possibly differential relations with multiple life outcomes. The outcome-wide analytic design has several strengths, including (1) reducing researcher subjectivity or degrees of freedom in the analysis by ensuring a consistent analytic strategy and the same set of covariates across models for all outcomes; (2) mitigating publication bias by reporting results for all examined outcomes simultaneously; and (3) providing insights into beneficial, detrimental, and null associations with the exposure51,52. Further details about the outcome-wide approach can be found elsewhere50,51.

Measures

The exposure was taken from Wave 1. The measure of dispositional forgivingness used in the GFS survey comes from the Forgiveness-Short Form measure that is part of the Brief Multidimensional Measure of Religiousness/Spirituality53. The Forgiveness-Short Form contains three items, one for each of three types of forgiveness: forgiveness of others, self-forgiveness, and divine forgiveness. The original forgiveness of others item (i.e., “I have forgiven those who hurt me”; response options: ‘Never,’ ‘Seldom,’ ‘Often,’ ‘Always or almost always’) has been used in previous research54, and other studies have employed variations of this item as well25. To strengthen the cross-cultural equivalence of the original item for use in the GFS survey, slight modifications were made to the phrasing and response categories based on the results of cognitive interviews conducted during the GFS survey development process42,43,48,49. The version of the item used in the GFS survey is: “How often have you forgiven those who have hurt you?” (response options: ‘Never,’ ‘Rarely,’ ‘Often,’ ‘Always’). Following prior studies using Wave 1 data from the GFS3,55,56, we dichotomized responses to this variable into (0) never/rarely and (1) often/always. In addition to maintaining consistency and comparability with earlier work in the GFS that supports the cumulative development of evidence in this area, our approach to grouping response options reflects a meaningful threshold for habitual forgiveness and facilitates clearer interpretation of results for non-technical audiences.

Country-specific analyses adjusted for 17 covariates (9 sociodemographic and 8 childhood variables) taken from Wave 1, unless data were not available (as described below and in relevant tables). Additional details for all variables can be found in the GFS codebook (https://osf.io/cg76b). Weighted descriptive statistics for the sociodemographic and childhood variables are reported in Table S1 for the total sample and Tables S9a–S31a for each country by wave (see also Table S32).

Year of birth (age) was classified into 1998–2005 (18–24 years), 1993–1998 (25–29 years), 1983–1993 (30–39 years), 1973–1983 (40–49 years), 1963–1973 (50–59 years), 1953–1963 (60–69 years), 1943–1953 (70–79 years), and 1943 or earlier (80 years or older). Gender was assessed as male, female, or other. Marital status was assessed as single/never married, married, separated, divorced, widowed, and domestic partner. Employment was assessed as employed, self-employed, retired, student, homemaker, unempl