The empirical literature includes varying terminologies aimed at operationalizing emotional well-being. For example, some working terminologies for emotional well-being have included terms such as flourishing, subjective well-being, and psychological well-being each of which may or may not include mental health as part of its definition (Diener et al., 1985; Ryff and Keyes, 1995; Seligman, 2012). However, there have been efforts to unite the field under the term emotional well-being, which specifically excludes mental health conditions such as depression and anxiety as indicators of emotional well-being (Feller et al., 2018). These efforts have resulted in a working definition of emotional well-being as “a multi-dimensional composite that encompasses how positive an individual feels generally and about life overall. [….] These features occur in the context of culture, life circumstances, resources, and life course.” (Park et al., 2023). Variables that encompass various dimensions of emotional well-being include happiness, life satisfaction, meaning in life, and self-esteem. For example, happiness is an emotionally positive state of being (Elfenbein and Ambady, 2002), life satisfaction and meaning in life are general appraisal of one’s existence (Pavot and Diener, 1993; Steger et al., 2006), and self-esteem is a positive evaluation of self-worth (Abdel-Khalek, 2016). These dimensions have all been previously included together in meta-analyses focusing on emotional well-being (e.g., Schmitt et al., 2014; Çikrıkci, 2016; Tang et al., 2020).

There are good conceptual reasons for why moral identity is related to each of these four dimensions of emotional well-being. One way how moral identity may relate to happiness, life satisfaction, meaning in life, and self-esteem is through engagement in prosocial behavior. Previous research has shown robust links between moral identity and prosocial behavior (Hertz and Krettenauer, 2016), which in turn has been linked with greater happiness (Aknin and Whillans, 2021), life satisfaction (Espinosa et al., 2022), meaning in life (Klein, 2017), and higher self-esteem (Fu et al., 2017). While these four dimensions of emotional well-being are intercorrelated (Lin et al., 2021), they differ in their cognitive, emotional, and reflective processes. Self-esteem and meaning in life both involve reflective self-evaluations with self-esteem involving internal beliefs toward oneself (Schmitt et al., 2014) and meaning in life involving cognitive reflections about the value one attributes to one’s existence (Dezutter et al., 2013). In comparison, happiness and life satisfaction are mainly based on lived experiences, with happiness being based on positive hedonic emotions and life satisfaction involving a cognitive assessment of the extent to which life experiences satisfy an individual’s desires (Schmitt et al., 2014).

These differences may affect the specific ways how moral identity relates to each dimension of emotional well-being. For example, moral identity may be related to higher self-esteem primarily through a more positive moral self-evaluation, which may be gained from a conscious sense of moral integrity that is present in individuals with high moral identity (Krettenauer et al., 2022). Conversely, moral identity may relate to greater purpose and meaning in life by providing a direction and ideals to strive for (Higgins-D’Alessandro and Power, 2005). In contrast, happiness and life satisfaction may require less self-evaluation and can be gained directly from positive experiences. While happiness and positive affect may be purely emotional responses, life satisfaction involves a conscious and reflective judgment of dispositional emotional well-being (Huebner and Dew, 1996). Having a conscious sense of moral integrity may therefore be more relevant for links between moral identity and life satisfaction, whereas moral identity may relate to happiness primarily through the emotional benefits of more frequently engaging in moral and prosocial behavior (Miles and Upenieks, 2022). Given these potential differences in the underlying mechanisms, it is possible that links between moral identity and these sub-dimensions differ in magnitude, which underlines the importance of examining heterogeneity in effects sizes of moral identity and emotional well-being across these dimensions. Thus, this study sought to explore moral identity and its relationship with overall emotional well-being as well as its individual relationship with happiness or positive affect, life satisfaction, meaning or purpose in life, and self-esteem.

Previous meta-analyses focusing on moral identity noted the importance of examining cultural differences, especially differences between individualism and collectivism (Hertz and Krettenauer, 2016; Lefebvre and Krettenauer, 2019). While individualistic cultures stress independence and autonomous self-actualization, collectivistic cultures are characterized by interdependence and individuals putting greater emphasis on needs of their community when constructing a self-identity (Markus and Kitayama, 1991). Although the general concept of moral identity is similar across cultures, the degree to which specific moral values are accentuated varies between individualistic and collectivistic cultures (Jia et al., 2019). The existing and most-widely used measures of moral identity may neglect moral attributes that are specific to eastern-collectivistic cultures, which possibly introduces cultural-bias into the concept of moral identity (Jia and Krettenauer, 2017). This potential bias has been referred to as an explanation for weaker links between moral identity and moral behavior in collectivistic cultures as compared to individualistic cultures (Hertz and Krettenauer, 2016). However, a previous meta-analysis examining moral identity in relation to moral emotions showed no differences between individualistic and collectivistic cultures (Lefebvre and Krettenauer, 2019).

No previous studies have examined cultural differences in links between moral identity and emotional well-being. There is some theoretical framework suggesting that links between moral identity and emotional well-being may be stronger in collectivistic cultures as compared to individualistic cultures. Due to greater cultural emphasis on independence and individual achievement (Markus and Kitayama, 1991), individuals from individualistic cultures experience stronger positive emotions in response to personal success, whereas individuals from collectivistic cultures have stronger positive emotions in response to outcomes that benefit others (Stipek, 1998). Since moral identity implies integrating personal motives with other-oriented moral principles (Frimer and Walker, 2009), moral identity may play a more crucial role for emotional well-being in collectivistic as compared to individualistic cultures. Thus, this study sought to examine whether links between moral identity and overall emotional well-being differ between individualistic and collectivistic cultures.

The literature searches for studies that examined links between moral identity and emotional well-being were conducted through ProQuest with its 65 databases including APA PsycInfo, APA PsycArticles, ERIC, ProQuest Dissertations & Theses Global, and Sociological Abstracts, among others. A separate literature search was conducted through PubMed. The literature search included two groups of key-terms. Group A included moral identity and Group B included various terms and phrases indicating emotional well-being (see Table 1 for details). Articles that included moral identity and at least one of the emotional well-being terms were captured. Moral identity was included as a single search term, following methods of previous meta-analyses that examined links between moral identity and moral behavior or moral emotions (Hertz and Krettenauer, 2016; Lefebvre and Krettenauer, 2019). Similarly, emotional well-being terms were based on the working definition by Park et al. (2023) and adapted from Koslouski et al. (2022).

All searches were limited to peer-reviewed journal articles, book chapters, dissertations/theses, and conference papers. No restrictions were placed on the location of where a search term needed to appear in the manuscript. Articles that were not available in either English or Spanish were excluded. Manuscripts that were unavailable online were requested from the authors’ university library.

To be included in the meta-analysis, eligible studies needed to include an independent empirical assessment of moral identity defined as the degree of how central moral values and beliefs are to one’s self-identity. Eligible studies also needed to include an independent empirical assessment of emotional well-being. Studies only measuring domain-specific forms of emotional well-being (e.g., career satisfaction) were excluded. Studies were excluded if moral identity and/or emotional well-being were measured following an experimental manipulation. Other inclusion criteria were that studies needed to include quantitative results on the relationship of mortal identity and at least one dimension of emotional well-being. If results were not available in the original paper, the corresponding authors were contacted to either provide the result or statistics that allow computing it. Finally, eligible studies had to use original data. There were no restrictions on year of publication or study populations.

The studies identified by the literature search were reviewed following the guidelines of the PRISMA 2020 flow diagram template for systematic reviews (Page et al., 2021a). First, the number of records found were reported for each database and duplicates were removed (n = 117). The literature review on the articles identified in the literatures search (n = 286) then consisted of three stages including title and abstract screening, full-text screening, and data extraction. The procedures were conducted by the first, second, fourth, and fifth author at each stage through Covidence (systematic review software; Veritas Health Innovation, 2023). Each record was reviewed independently by at least two reviewers at each stage of the review process to ensure blinding. At each stage, conflicts between reviewer assessments were resolved by the first and second author through discussion until consensus was reached. At the first stage, all records were screened by reviewing the title and abstract. Records were excluded when not reporting quantitative results or when not investigating any of the variables of interest meaning that records were only excluded in the first round if the title and/abstract neither mentioned that moral identity nor a dimension of emotional well-being were measured in the study. Inter-rater agreement between the reviewers at the title and abstract screening stage was fair (72% agreement, Cohen’s k = 0.38). As illustrated in Figure 1, the title and abstract screening resulted in 161 records proceeding to the full-text screening stage. At the second stage, full-text reviews were conducted and records were excluded if they did not include a quantitative measure of both moral identity and an indicator of emotional well-being in the same study. Inter-rater agreement between the reviewers at the full-text screening stage was almost perfect (94% agreement, Cohen’s k = 0.81). The full-text reviews resulted in 25 records proceeding to the third stage. At the third stage, the remaining records were assessed for eligibility and three records were excluded due to involving experimental manipulations prior to the assessments of moral identity and emotional well-being, which may have affected their association (Aquino et al., 2011; Furukawa et al., 2021; Semko, 2022). Moreover, two records were excluded for which results on links between moral identity and emotional well-being could not be obtained from the manuscript or received from the corresponding authors. Finally, the reference lists of the articles that were included after the full-text screening were examined. This citation searching resulted in one additional study being included in the review resulting in a total of 21 records being included in the review. Of these 21 records, three records included two eligible studies with separate samples (Giacalone et al., 2016; Wu et al., 2020; Wang and Hackett, 2022) and one record included four eligible studies with four separate samples (Jordan et al., 2015). Thus, the literature review yielded 27 eligible studies that were included in the meta-analysis.

From the eligible studies, the first and second author independently extracted key study characteristics. These include information on sample size and sample characteristics, information on the measures that were used for moral identity and emotional well-being, and effect sizes of relationships between moral identity and emotional well-being variables. The authors extracted correlation coefficients (r) from each individual study as the effect size. Three studies reported latent correlations between moral identity and emotional well-being, which were used for the analyses. When necessary, effect sizes were reverse coded so that positive correlation coefficients indicate that higher moral identity is associated with greater emotional well-being. Discrepancies between the two authors were resolved through discussion until consensus was reached.

The most commonly used measure of moral identity among the eligible studies was the Self-Importance of Moral Identity Scale (SIMI) (Aquino and Reed, 2002). This scale includes two subdimensions – internalization and symbolization. While the internalization dimension measures how central moral traits are to one’s self-concept, the symbolization dimension measures the degree to which individuals engage in public behavior to demonstrate their moral traits (Aquino and Reed, 2002). Since the focus of the present study is on moral identity defined as how central moral traits are to one’s self-identity, the internalization scale was prioritized. Therefore, if studies reported effects sizes for both the internalization and symbolization subscales separately, only the effect sizes for internalization were used in the analyses. However, if studies only reported effect sizes of the overall moral identity scale (i.e., combining internalization and symbolization), these effect sizes were used for the analyses. This approach is consistent with previous meta-analyses involving moral identity (Hertz and Krettenauer, 2016; Lefebvre and Krettenauer, 2019). Measures of emotional well-being were categorized into four groups (i.e., life satisfaction, sense of purpose/meaning, happiness/positive affect, and self-esteem). For studies that included multiple measures of emotional well-being, effect sizes of moral identity with the individual emotional well-being measures were combined to a single effect size if the measures of emotional well-being were in the same category. For example, effect sizes of the correlations of moral identity and happiness and of moral identity and positive affect in the same study were averaged.

Each eligible study was coded as either having involved sampled participants from individualistic or collectivistic cultures based on the country of data collection using data from a revised version of Hofstede’s quantitative measure of individualism (Minkov and Kaasa, 2022). This quantitative measure of individualism was initially developed from survey responses of employees of the same company from various countries to create country-level indicators of individualism (Hofstede, 1980). In countries that have a higher Individualism index score (IDV), people on average put greater emphasis on achieving personal goals, while in countries with lower IDV scores, people on average put greater emphasis on goals of larger groups and communities (Hofstede and McCrae, 2004). While the Hofstede measure has been criticized for being limited to the work environment, previous research has demonstrated convergent validity of the IDV and other measures of individualism (Schimmack et al., 2005).

Each study was assigned the country-specific scores as of 2023 that were retrieved from the website of Hofstede insights and are based on more recent replications of Hofstede’s initial study (Country Comparison Tool, 2023).² The majority of studies included in this meta-analysis were conducted in the United States (12 studies, IDV = 91) and China (6 studies, IDV = 20). Following Hofstede’s classification, the United States studies as well as three studies involving participants from both the United States and Canada (IDV = 81) were coded as individualistic. Studies conducted in China as well as a study conducted in Indonesia (IDV = 14), a study conducted in Iran (IDV = 41), and a study conducted in Georgia (IDV = 41) were coded as collectivistic. Two studies involved international business students and one study recruited participants via Amazon Mechanical Turk with no indication of geographical restrictions. These studies were not coded for cultural origin of the sample.

Data were extracted from the final included articles and uploaded into Comprehensive Meta-Analysis 4.0 (CMA) software (Borenstein et al., 2022) to conduct the meta-analysis. Before the calculation of the pooled effect size, we transformed correlation coefficients into Fisher’s z to stabilize variance in the analyses and then converted back into correlation coefficients (r) for ease of interpretation. Correlation coefficients can be interpreted using the following criteria: small (0.10), medium (0.24), and large (0.37) (Cohen, 1988). To address the first aim, a meta-analysis was conducted in two steps. First, an overall effect size was computed to analyze the association between moral identity and emotional well-being. Next, four separate meta-analyses were conducted across emotional well-being domains to examine the relation of moral identity with each of the four subdomains of emotional well-being. Positive effect sizes indicate that higher moral identity is related to better emotional well-being, whereas negative effect sizes indicate that higher moral identity is related to lower emotional well-being. Because it is implausible to assume that the true effect size is homogenous across the included studies, a random-effects model was used to conduct all analyses and interpret the effect sizes (Borenstein et al., 2021).

The literature search yielded a total of 27 eligible studies with a total of 24,733 participants that report at least one effect size of links between moral identity and emotional well-being (see Figure 1 for PRISMA flow diagram). Detailed information on the sample characteristics and measures of each eligible study are reported in Table 2. Most studies involved adolescents (n = 6), emerging adults (n = 11), and mid-aged adults (n = 17). Only a few studies included older adults (n = 3) and no studies specifically focused on older adult populations. As a result, one initially proposed and pre-registered hypothesis that links between moral identity and emotional well-being would be stronger in older adult populations could not be examined in the present study. Most studies included a similar number of males and females. Most eligible studies were conducted in either the United States (n = 12 studies) or China (n = 6). All but one study assessed moral identity using the SIMI (Aquino and Reed, 2002). Of these studies, most assessed internalization separately (n = 18) but few studies combined the internalization and symbolization subscales (n = 7). One study used a modified version of the SIMI (Yukhymenko-Lescroart and Sharma, 2020) and one study assessed moral identity with the Civic and Moral Identity Questionnaire (Han et al., 2019). All studies that assessed self-esteem utilized the Rosenberg’s Self-Esteem Scale (Rosenberg, 1965). All of the included studies that measured life satisfaction used either the Satisfaction with Life Scale (Diener et al., 1985) or its modified version the Temporal Satisfaction with Life Scale (Pavot et al., 1998). Studies that measured purpose and meaning in life utilized more diverse measures with the most frequently used measure being the Meaning in Life Questionnaire (Steger et al., 2006). Similarly, the included studies that measured positive affect and happiness used a variety of measures with the most frequently used being the Positive and Negative Affect Scale (Watson et al., 1988).

Each study reported correlation coefficients as the effect size to describe the relationship of moral identity and emotional well-being with three studies reporting latent correlations (Hardy et al., 2013, 2014; Yukhymenko-Lescroart and Sharma, 2020). In 23 instances (85%), the correlation coefficients were included in the published version of peer-reviewed journal articles. In four instances, authors of published peer-reviewed journal articles were contacted via email and provided the correlation coefficients or data to compute these (Cohen et al., 2011, 2014; Hardy et al., 2014; Hofmann et al., 2018). All correlation coefficients included in the present study were cross-sectional. Only a single study included a longitudinal analysis examining reciprocal links between moral identity and meaning in life over time to better understand temporal precedence of these variables (Han et al., 2019). This study found that moral identity predicts increased meaning in life over time, whereas meaning in life did not predict changes in moral identity suggesting that moral identity is a predictor of greater meaning in life but not vice versa.

For each study, the individual and pooled effect sizes of the association between moral identity and emotional well-being and its sub-dimensions are reported in Table 2. A Forest plot depicting the effect sizes of all included studies together with 95% confidence intervals is shown in Figure 2. The meta-analysis showed a significant overall correlation indicating that higher moral identity is associated with greater emotional well-being (r = 0.27, p < 0.001; CI_95% [0.22–0.33]). This indicates a medium effect size according to conventional effect size interpretation guidelines (Rice and Harris, 2005). There was a large proportion of heterogeneity (I² = 93.42) and the variance of true effect sizes (τ²) was 0.02. Relatedly, the Q-test for heterogeneity indicated that effect sizes were heterogenous across studies supporting the use of a random effects model (Q₂₆ = 395.15, p < 0.001). The I² statistic had a score of 93, which indicates that 93% of the variance in observed effects reflected variance in true effects rather than sampling error. According to conventions, this I² value is considered high (Higgins et al., 2003), meaning that the identified association is substantially due to true effects.

Results from the sensitivity analysis excluding three studies that reported latent correlations replicated the main results. The meta-analysis on the remaining 24 studies showed a significant overall correlation of medium size linking higher moral identity with greater emotional well-being (r = 0.25, p < 0.001; CI_95% [0.20–0.31]). As in the main analysis, effect sizes were heterogenous across studies (Q₂₃ = 208.52, p < 0.001). However, heterogeneity decreased upon removal of latent effect sizes (I² = 88.97, τ² = 0.01). Results from a post-hoc sensitivity analyses showed that moral identity is related to greater emotional well-being even if the symbolization subscale of the SIMI is used with the effect size being similar (r = 0.22, p < 0.001) to when only the internalization subscale is used (r = 0.24, p < 0.001).

Results from the meta-analyses examining links between moral identity and each dimension of emotional well-being (i.e., happiness/positive affect, sense of purpose/meaning, life satisfaction, and self-esteem) are reported in Table 3. The results showed that moral identity is associated with more happiness and positive affect, a greater sense of purpose and meaning in life, more life satisfaction, and higher self-esteem. The effect sizes were medium for happiness/positive affect (r = 0.28, p < 0.001), sense of purpose/meaning in life (r = 0.29, p < 0.001), and self-esteem (r = 0.25, p < 0.001), and small for life satisfaction (r = 0.15, p = 0.011). Effect sizes were heterogenous between studies for each domain and between 84% and 95% of variance in observed effects reflected variance in true effects rather than sampling error. Since the sense of purpose/meaning and self-esteem domains both included two studies that reported latent correlations as the effect size, a sensitivity analysis was conducted excluding the latent effect sizes. The results from this sensitivity analysis replicated the original results by showing a medium effect size for moral identity and meaning/purpose (r = 0.26, p < 0.001) as well as a small to medium effect size for moral identity and self-esteem (r = 0.22, p < 0.001).

Results from a subgroup analysis showed a slightly larger effect size of moral identity and emotional well-being in studies conducted in more collectivistic countries (r = 0.30, p < 0.001) compared to studies conducted in more individualistic countries (r = 0.27, p < 0.001). However, the difference between these two effect sizes was not statistically significant (Q₁ = 0.38, p = 0.536). It is important to note that all studies that report latent correlations were in the individualistic group, which may have attenuated differences between collectivistic and individualistic cultures. Therefore, a sensitivity analysis was conducted excluding these latent effect sizes. The results showed a larger difference in effect sizes with moral identity being more strongly related to emotional well-being in collectivistic (r = 0.30, p < 0.001) as compared to individualistic cultures (r = 0.23, p < 0.001). However, this difference was also not statistically significant (Q₁ = 2.15, p = 0.142). A second sensitivity analysis was conducted that excluded two studies that were conducted in Iran and Georgia to enable a comparison between studies conducted in North-America (United States and Canada) and studies conducted in East and Southeast Asia (China and Indonesia). The results from this sensitivity analysis replicated the original results by showing slightly larger effect sizes in collectivistic Asian cultures (r = 0.30, p < 0.001) compared to individualistic Western cultures (r = 0.27, p < 0.001) but this difference was not statistically significant (Q₁ = 0.32, p = 0.572). Similarly, when excluding latent correlations as well as the studies from Iran and Georgia, effect sizes are larger in the collectivistic group (r = 0.30, p < 0.001) versus in the individualistic group (r = 0.23, p < 0.001). However, this difference was not statistically significant (Q₁ = 2.94, p = 0.087).

Based on a funnel plot (see Figure 3), the distribution of effect sizes of links between moral identity and overall emotional well-being was not completely symmetrical around the average overall effect size. Egger’s regression intercept further indicated funnel plot asymmetry (t = 2.77, p = 0.010). However, the rank correlation test was not significant suggesting no asymmetry (τ = 0.09, p = 0.246). In the presence of publication bias, smaller studies would be expected to have larger effect sizes due to requiring larger effect sizes to reach statistical significance and smaller studies being more likely to get published if they produce significant results (Sterne and Harbord, 2004). However, the funnel plot does not provide any indication of a tendency that smaller studies would have larger effect sizes as the effect sizes of studies with larger standard errors (at the bottom of the plot) are scattered both below and above the average overall effect size. It is important to note that asymmetry within the funnel plot can be due to other factors beyond publication bias such as between-study heterogeneity and methodological differences between studies (Page et al., 2021b). Additionally, upon removal of the latent effect sizes, the funnel plot appeared symmetrical (see Figure 4) and Egger’s regression test was not significant indicating that there was no asymmetry (t = 0.28, p = 0.390). Likewise, the rank correlation test was not significant suggesting no asymmetry (τ = 0.11, p = 0.228). The results of the fail-safe N calculation showed that an additional 9,478 studies reporting no correlation of moral identity and emotional well-being (r < 0.01) would be needed for the estimated effect size to become non-significant (p < 0.05). Finally, publication bias may also have been reduced by including studies that did not examine links between moral identity and emotional well-being as a primary objective. In fact, more than half of the included studies did not formulate a hypothesis on links between moral identity and emotional well-being (n = 14). Thus, the included effect sizes from these studies should not have affected the study’s chance of publication, which has likely been more affected by whether the primary study aims were supported with significant results.

Several limitations need to be considered when interpreting the results of the present study. Most importantly, the meta-analysis was based on correlational effect sizes and therefore did not examine temporal precedence. Thus, any directional inferences cannot be made based on the present results. A second limitation is that only a few studies involved older adult samples and the results may therefore not be generalizable to older adult populations. Even though this review included studies from diverse cultural backgrounds, the vast majority of studies were conducted in the United States and China, which may suggest that results are less generalizable to individuals from other parts of the world. Relatedly, the analysis examining the moderating role of culture is limited by focusing on individualism and collectivism, which by no means captures every aspect of cultural difference between the studies in the individualistic and collectivistic groups (Brewer and Chen, 2007). Similarly, the present study could not consider cultural differences between the countries within the individualistic and collectivistic groups or at the participant level. Another limitation may be that the sample size did not allow for separate analyses of culture as a moderator in each individual relationship of moral identity and the four dimensions of emotional well-being.

Since this study involved study-level data with many studies showing little variability in gender composition, gender differences in links of moral identity and emotional well-being could not be examined. A previous meta-analysis showed that females have higher moral identity compared to males (Kennedy et al., 2017) but none of the empirical studies included in this review conducted a formal statistical analysis on gender differences in links between moral identity and emotional well-being. However, one study examined links between moral identity and life satisfaction separately by gender reporting the same effect size (r = 0.28) for males and females (Wu et al., 2020). Nevertheless, a more thorough investigation on whether links between moral identity and emotional well-being differ by gender is an interesting direction for future research.

Although this meta-analysis was based on correlational results and existing longitudinal evidence is limited it is possible that promoting moral identity may be a pathway to achieve greater emotional well-being. Thus, the present findings highlight the need of longitudinal studies to better understand the directionality of this relationship and examine if existing programs designed to promote moral identity also benefit individuals’ emotional well-being. Examples of successful interventions include promoting moral identity through stories about behaviors of moral exemplars (Čehajić-Clancy and Olsson, 2023). Additionally, moral identity may be fostered through community engagement and service learning (Narvaez and Lapsley, 2009) which have also been found to increase emotional well-being (Lawton et al., 2021). If moral identity indeed predicts better emotional well-being, promoting moral identity through these programs may become a relevant target for interventions designed to improve various dimensions of emotional well-being (Kubzansky et al., 2023).