The Association Between Prosocial Behavior and Mental Health: A Three‐Level Meta‐Analysis

ABSTRACT

Introduction

Although numerous studies have reported a significant association between prosocial behavior and mental health, other studies have indicated a weak or non‐significant, and in some cases even negative link.

Method

To clarify these inconsistencies, the present study employed a three‐level meta‐analytic approach to examine the overall association between prosocial behavior and mental health, as well as the moderating effects of various demographic and methodological variables.

Results

A comprehensive literature search yielded 366 effect sizes (N = 222,866) that were included in the meta‐analysis. The main effect revealed a significant positive correlation (r = 0.26, p < 0.001) between prosocial behavior and mental health. Further moderator analyses indicated that this relationship was moderated by culture (F _{(1, 362)} = 16.16, p < 0.001), mental health indicators (i.e., positive vs. negative indicators of mental health; F _{(1, 364)} = 60.86, p < 0.001), mental health measurement tools (F _{(7, 358)} = 2.52, p = 0.015), and research design (i.e., cross‐sectional vs. longitudinal studies; F _{(1, 364)} = 21.06, p < 0.001).

Conclusions

Overall, the findings highlight the positive association between prosocial behavior and mental health and underscore the importance of considering prosocial behavior in efforts aimed at supporting psychological well‐being.

1. Introduction

Prosocial behavior, encompassing voluntary actions intended to benefit others and contribute to societal well‐being, has become a prominent focus in psychological research. Such behaviors, including helping, sharing, and providing emotional support, are generally believed to foster positive psychological functioning (Wittek and Bekkers). Consistent with this view, numerous studies have linked prosocial engagement to better mental health outcomes, such as higher life satisfaction, stronger self‐efficacy, and fewer negative emotions (Feng and Zhang 2022; Hui 2022; Kakulte and Shaikh 2023; Li et al. 2023). However, empirical findings are not entirely consistent. Some studies have reported weak, non‐significant, or even negative associations between prosocial behavior and mental health (Dunn et al. 2014; Lam et al. 2023; Rinner et al. 2022). These inconsistencies suggest that the relation between prosocial behavior and mental health may be more complex than initially assumed and may be influenced by moderating factors, including demographic variables (e.g., gender and age) and methodological differences such as study design (e.g., cross‐sectional vs. longitudinal), measurement tools (e.g., different prosocial behavior or mental health scales), and publication characteristics (e.g., journal articles vs. unpublished studies).

While prior meta‐analyses have significantly advanced our understanding of the relation between prosocial behavior and mental health, several limitations remain. First, much of the existing work has focused on specific populations or forms of prosocial behavior, such as older adults engaging in volunteering (Wheeler et al. 1998). This focus restricts the generalizability of findings across developmental stages. Moreover, these early reviews were published more than a decade ago, raising concerns about timeliness and the applicability of their conclusions to contemporary contexts. Second, prior research has mainly focused either on well‐being outcomes (e.g., Curry et al. 2018) or on psychopathology (e.g., internalizing and externalizing symptoms) in adolescence (e.g., Memmott‐Elison et al. 2020), without providing a comprehensive picture of mental health that includes both positive and negative indicators. Third, most previous meta‐analyses have relied primarily on traditional two‐level models, which treat all effect sizes as independent (e.g., Curry et al. 2018; Hui et al. 2020; Jenkinson et al. 2013). However, effect sizes in primary studies are often nested within the same sample or study, and ignoring this dependency can lead to biased estimates, increased risk of Type I error, and reduced accuracy of parameter estimation (Hox et al. 2017).

To address these limitations, the present meta‐analysis aims to provide a more comprehensive and methodologically rigorous examination of the association between prosocial behavior and mental health. First, it expands the population scope by including studies across different age groups, thereby enhancing the generalizability of the findings. Second, consistent with the dual‐factor model widely adopted in psychological research (Greenspoon and Saklofske 2001; Magalhães 2024; Suldo and Shaffer 2008), the present study conceptualizes mental health as a multidimensional construct encompassing both positive indicators (e.g., life satisfaction, happiness, subjective well‐being, positive affect, and meaning in life) and negative indicators (e.g., anxiety, loneliness, depression, and negative affect). These indicators were identified based on prior empirical and meta‐analytic studies on mental health (e.g., Chen et al. 2024; Padilla‐Walker et al. 2022; Q. Wang et al. 2023) and were further confirmed during the literature search process focusing on prosocial behavior. This approach provides a more comprehensive understanding of the link between prosocial behavior and mental health. Third, it applies a three‐level meta‐analytic model (Konstantopoulos 2011) that accounts for the nested structure of effect sizes within studies, reducing the risk of biased estimates and Type I error while improving the accuracy of parameter estimation.

2. Literature Review

2.1. The Association Between Prosocial Behavior and Mental Health

Social exchange theory (Homans 1958) provides a valuable framework for understanding the inconsistent association between prosocial behavior and mental health. According to this theory, individuals evaluate the perceived costs and benefits of their social interactions, and this appraisal shapes the psychological consequences of prosocial engagement. Although the theory was originally developed in the context of adult relationships, its central cost–benefit logic is not limited to a particular age group. The specific forms of perceived rewards (e.g., social approval, strengthened relationships, and enhanced self‐worth) and costs (e.g., time investment, emotional strain, and lack of reciprocity) may differ across social and developmental contexts, but the basic evaluative process remains consistent. When perceived rewards outweigh associated costs, prosocial behavior is likely to promote mental health. In contrast, when helping is experienced as burdensome or unreciprocated, it may contribute to psychological distress. This perspective helps explain why prior studies have reported mixed findings regarding the association between prosocial behavior and mental health.

Empirical evidence supports this dual‐effect perspective. On the one hand, a substantial body of research demonstrates that engaging in prosocial behavior leads to significant improvements in mental health, including enhanced well‐being, alleviation of depressive symptoms, and increased life satisfaction (Ju et al. 2025; Memmott‐Elison and Toseeb 2023; Son and Padilla‐Walker 2020). For instance, Ju et al. (2025) find that engaging in prosocial behavior toward family and friends, such as providing support or showing kindness, enhances adolescents' well‐being. On the other hand, some studies suggest a negative relation between prosocial behavior and mental health, particularly when engagement becomes excessive or compulsive (Alvis et al. 2023; Reimann et al. 2025). Persistent prioritization of others' needs over one's own can lead to emotional depletion, compassion fatigue, and even physical burnout (Adelman et al. 2014; Bjälkebring et al. 2021; Stern 2019). These risks were notably amplified during the COVID‐19 pandemic, in which individuals, especially young people, faced heightened pressures to support others despite their own resource constraints. In such contexts, altruistic efforts can transform into sources of stress, fostering feelings of inadequacy, helplessness, and emotional strain (Alvis et al. 2023). Furthermore, several studies have reported no significant correlation between prosocial behavior and mental health (Armstrong‐Carter et al. 2020; Zhang et al. 2023). A meta‐analytic synthesis is thus essential to provide a more systematic and integrated understanding of this association, identify key moderators, and offer clearer guidance for theory and practice.

2.2. Moderators of the Link Between Prosocial Behavior and Mental Health

2.2.1. Gender

Gender plays a pivotal moderating role in the relation between prosocial behavior and mental health, with evidence suggesting that this association is notably stronger among females (Armstrong‐Carter et al. 2020; Hui et al. 2020; Li et al. 2020). Although both males and females experience psychological benefits from engaging in prosocial behavior, the nature and intensity of these benefits appear to differ, largely due to gendered socialization processes and the distinct societal expectations placed on males and females. According to social role theory (Eagly 1987), individuals internalize and embody the behavioral patterns prescribed by societal gender roles. For females, traditional gender roles emphasize caregiving and nurturing, thereby promoting engagement in prosocial activities such as providing emotional support, caring for others, and expressing empathy. These behaviors, in turn, satisfy fundamental psychological needs for relatedness and competence (Deci and Ryan 1985), which are critical to maintaining psychological well‐being. In contrast, males are more likely to engage in instrumental or task‐oriented forms of helping, such as providing practical or material assistance (Armstrong‐Carter et al. 2020; Eagly 1987). Because widely used measures of prosocial behavior (e.g., PTM, PBS) tend to emphasize emotionally supportive and relational forms of helping, which align more closely with female helping patterns, these measures may underrepresent instrumental forms of helping that are more typical among males. Thus, both gendered helping patterns and measurement approaches may contribute to the gender differences reported in the literature.

Empirical evidence consistently supports this opinion. Numerous studies have reported that females who engage in prosocial activities exhibit significantly higher levels of positive affect compared to their male counterparts (Armstrong‐Carter et al. 2020; Li et al. 2020; Weinstein and Ryan 2010). Conversely, the provision of instrumental assistance to friends is associated with increased negative affect in young males (Armstrong‐Carter et al. 2020), suggesting that certain forms of prosocial engagement may entail emotional costs for males. Furthermore, a comprehensive meta‐analysis corroborates these observations, with findings revealing that females derive a greater enhancement in well‐being from prosocial behavior than males (Hui et al. 2020). Collectively, these results highlight the critical role of gender as a moderating variable when examining the psychological benefits of prosocial behavior.

2.2.2. Age

Age appears to act as a crucial moderator in the relation between prosocial behavior and mental health, with evidence suggesting that the benefits of prosocial behavior tend to increase with age. As individuals mature, they develop enhanced cognitive and emotional capacities, including improved emotion regulation and a more optimistic outlook on life (Carstensen et al. 2000). These developmental changes facilitate the effective translation of prosocial actions into positive mental health outcomes. In addition, evolving social roles and relational needs across the life course foster a greater reliance on interpersonal support networks. Consequently, prosocial behavior becomes increasingly significant for sustaining psychological well‐being as individuals' happiness becomes more closely tied to the quality of their social connections (Caprara and Steca 2005). Empirical evidence supports this theoretical framework. Specifically, studies have consistently shown that older adults derive greater benefits from prosocial behavior and experience fewer negative emotions compared to younger individuals (Caprara and Steca 2005; Chi et al. 2021; Lam et al. 2023; Musick and Wilson 2003). Moreover, a broader assessment that includes multiple dimensions of mental health (e.g., life satisfaction, depressive symptoms, and stress) reveals that the overall association between prosocial behavior and mental health strengthens with age (Huang 2019; Kim and Pai 2010; Van Willigen 2000). This pattern suggests that the cumulative benefits of prosocial behavior on a wide array of mental health indicators tend to become more pronounced with increasing age.

2.2.3. Culture

Culture may moderate the link between prosocial behavior and mental health, with the association being stronger in collectivist cultures than in individualistic cultures. Two primary mechanisms may account for this differential association. First, collectivist cultures, often influenced by Confucian traditions, emphasize close‐knit family bonds, collective interests, and harmonious interpersonal relationships (Hofstede 1984; Markus and Kitayama 2014). In these contexts, prosocial behavior is not only socially expected but also deeply valued as a means of fostering interdependence and communal welfare. Engaging in prosocial actions aligns with normative cultural expectations and effectively fulfills individuals' psychological needs for belonging and social support (Bond 2010; Tsai and Kimel 2021). This cultural congruence amplifies the positive impact of prosocial behavior on mental health. A second explanatory mechanism involves culturally shaped patterns of self‐construal. In collectivist cultures, individuals typically adopt an interdependent self‐construal, where one's identity is closely linked to social relationships and group membership. Consequently, prosocial behavior can serve to reinforce a positive social identity and enhance feelings of connectedness, thereby contributing to improved mental health. Conversely, individualistic cultures emphasize personal autonomy and self‐actualization, rendering prosocial behavior relatively less central to one's identity and thus less strongly associated with mental health benefits. Supporting this dual‐mechanism perspective, a cross‐cultural comparative study indicates that individuals from collectivist cultures (e.g., Chinese participants) exhibit more pronounced positive emotional responses to prosocial behavior relative to those from individualistic cultures (e.g., American participants; Peng et al. 2024). These findings underscore the critical role of cultural context in shaping the magnitude of the prosocial behavior–mental health association.

2.2.4. The Type of Mental Health Indicators

The type of mental health indicators may serve as an important moderator of the association between prosocial behavior and mental health. Prior research consistently indicates that prosocial behavior tends to show stronger associations with positive mental health indicators than with negative indicators (Hui et al. 2020). This discrepancy may reflect fundamental differences in the conceptual nature and developmental mechanisms of different mental health constructs (Oberle et al. 2022). Positive mental health indicators typically reflect individuals' psychological resources and adaptive functioning, and may therefore be particularly sensitive to the social and emotional benefits derived from helping others. In contrast, negative mental health indicators such as depression and anxiety are often influenced by deeper and more persistent risk factors, including genetic vulnerability, early‐life adversity, and chronic stress (Abravanel and Sinha 2015; Feiler et al. 2023; Kuzminskaite et al. 2021). As a situational psychological resource, prosocial behavior may therefore have a relatively limited capacity to counteract these underlying vulnerabilities.

Moreover, negative emotional states can narrow individuals' cognitive and behavioral repertoires, thereby reducing their ability to benefit from positive social experiences (Tugade and Fredrickson 2004). When symptoms reach clinically significant levels, impairments in social motivation and executive functioning may further limit individuals' engagement in prosocial behavior or the psychological benefits derived from it. For example, research has shown that the daily mood‐enhancing effects of prosocial behavior are significantly weaker among adolescents with severe depressive symptoms than among their non‐depressed peers (Schacter and Margolin 2019). These theoretical and empirical considerations suggest that the type of mental health indicators may substantially shape the strength of the association between prosocial behavior and mental health.

2.3. The Current Research

Although several studies have explored the relation between prosocial behavior and mental health, inconsistencies remain regarding the strength and nature of this association. Variations in sample and study characteristics may have contributed to the heterogeneity of findings. Building upon prior research, this meta‐analysis aims to address two primary purposes. First, it seeks to provide an updated and comprehensive quantitative estimate of the overall association between prosocial behavior and mental health. Second, it aims to investigate potential moderating factors, such as demographic variables and methodological characteristics, that may account for variability in the observed relation.

3. Methods

The present meta‐analysis was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) guidelines (Moher et al. 2009; Page et al. 2021).

3.1. Literature Search Strategy

We employed four search strategies to retrieve relevant studies. First, we conducted a computer‐based literature search across multiple databases, including PsycINFO, PsycArticles, MEDLINE, Google Scholar, ProQuest Dissertations & Theses, Web of Science, Elsevier, Springer, CNKI, VIP database, and Wanfang database. Relevant publications were systematically searched with the following two groups of keywords restricted to titles and abstracts: (1) prosocial, altruis*, helping behavior, volunteer, cooperat*, compensate, share, consolation, and comfort*; (2) mental health, mental disorder, affect, distress*, anxi*, loneliness, depress*, life satisfaction, happiness, and well‐being. Second, we searched through review articles on similar topics. Third, we searched for articles by authors who had more than one study included in our initial search. Finally, we also examined reference lists of the retrieved articles manually to identify additional relevant publications. The literature search included studies published up to May 1, 2025.

3.2. Inclusion and Exclusion Criteria

The search yielded 613,193 records. After removing 237,789 duplicate records, 375,404 records remained for screening. The first author conducted the initial screening of titles and abstracts. During title screening, 373,119 clearly irrelevant records were excluded, and the remaining 2,285 records proceeded to abstract screening, where 1562 records were excluded. As a result, 723 articles were retained for full‐text assessment. These articles were independently assessed for full‐text eligibility by two researchers according to the predefined inclusion and exclusion criteria. Disagreements between the reviewers were resolved through discussion and consultation with the corresponding author until consensus was reached. The interrater agreement for the full‐text screening was 96%.

Studies were included or excluded based on the following criteria: (a) studies had to simultaneously report both prosocial behavior and mental health outcomes; (b) studies had to report Pearson correlation coefficients or provide sufficient data to compute effect sizes; (c) studies needed to be empirical and quantitative (i.e., review, theoretical, and qualitative studies were excluded); (d) studies focused on specific contexts (e.g., mental health in the workplace) were excluded; (e) studies were required to be published in English or Chinese. Following these criteria, a total of 121 studies contributing 366 effect sizes were included in the final meta‐analysis (see Figure 1).

FIGURE 1.

The flowchart for the included studies in the meta‐analysis.

3.3. Coding

Studies that satisfied the criteria for inclusion were coded for study descriptors and potential moderating variables. Study descriptors contained some basic information, such as first author, publication years, effect sizes, and sample size. Potential moderators were categorized into sample characteristics and study characteristics. A coding manual was developed to standardize the coding process. Two independent coders, including the first author and another trained researcher, coded all studies according to the manual (see Table 1). Intercoder agreement was achieved for 95% of the codes, and all discrepancies were resolved through discussion. It is important to note that in longitudinal studies, both the cross‐sectional effect sizes of prosocial behavior and mental health at the same time point, as well as the longitudinal effect sizes of prosocial behavior at earlier time points and mental health at later time points were included in the coding process.

TABLE 1.

Characteristics of the 121 studies included in the meta‐analysis.

First author (year)	K	N	QS	Gender	Age (SD)	Age range	Culture	Country	IVD	Design	MH terms reported	PB scale	MH scale	Publication status
Affuso et al. (2024)	16	587	17	0.53	14.23 (0.58)		individualism	Italy	53	longitudinal	well‐being	Others	Others	Published
Almeida et al. (2021)	4	440	17	0.48	13.14 (1.26)	11–17	individualism	Portugal	59	cross‐sectional	loneliness	Others	Others	Published
Armstrong‐Carter (2020)a	4	411	17	0.34	18.62 (0.37)		individualism	USA	60	longitudinal	negative affect	Others	Others	Published
Armstrong‐Carter (2020)b	4	411	17	0.34	18.62 (0.37)		individualism	USA	60	longitudinal	positive affect	Others	Others	Published
Arslan (2024)a	1	408	16	0.31	15.80 (1.27)	13–18	collectivism	Turkey	46	cross‐sectional	well‐being	Others	Others	Published
Arslan (2024)b	1	408	16	0.31	15.80 (1.27)	13–18	collectivism	Turkey	46	cross‐sectional	meaning in life	Others	Others	Published
Aydinli‐Karakulak et al. (2019)a	1	195	16	0.60	16.50 (1.23)	14–19	collectivism	Bulgaria	50	cross‐sectional	life satisfaction	PTM	SWLS	Published
Aydinli‐Karakulak et al. (2019)b	1	195	16	0.60	16.50 (1.23)	14–19	collectivism	Bulgaria	50	cross‐sectional	negative affect	PTM	PANAS	Published
Aydinli‐Karakulak et al. (2019)c	1	195	16	0.60	16.50 (1.23)	14–19	collectivism	Bulgaria	50	cross‐sectional	loneliness	PTM	CLS	Published
Boundenghan et al. (2012)a	1	148	16	0.36	30.74 (16.20)		individualism	France	74	cross‐sectional	well‐being	Others	PANAS	Published
Boundenghan et al. (2012)b	1	148	16	0.36	30.74 (16.20)		individualism	France	74	cross‐sectional	positive affect	Others	PANAS	Published
Boundenghan et al. (2012)c	1	148	16	0.36	30.74 (16.20)		individualism	France	74	cross‐sectional	negative affect	Others	PANAS	Published
Boundenghan et al. (2012)d	1	148	16	0.36	30.74 (16.20)		individualism	France	74	cross‐sectional	affect	Others	PANAS	Published
Caprara (2005)	1	512	15	0.51	35.70	20–87	individualism	Italy	53	cross‐sectional	life satisfaction	Others	SWLS	Published
Chang et al. (2014)	1	901	16	0.50	12.50 (0.60)	10–14	collectivism	China	43	cross‐sectional	loneliness	nomination	CLS	Published
Chang et al. (2021)	1	1984	17	0.44			collectivism	China	43	cross‐sectional	depression	PBS	CES‐D‐10	Published
Chen et al. (2000)a	1	470	16	0.53	13.90 (0.92)		collectivism	China	43	longitudinal	loneliness	Others	CLS	Published
Chen et al. (2000)b	1	470	16	0.53	13.90 (0.92)		collectivism	China	43	longitudinal	depression	Others	CDI	Published
Chen (2018)a	1	941	13	0.48	16.41 (0.89)		collectivism	China	43	cross‐sectional	life satisfaction	IABS	SWLS	Unpublished
Chen (2018)b	1	941	13	0.48	16.41 (0.89)		collectivism	China	43	cross‐sectional	life satisfaction	ABCS	SWLS	Unpublished
Chen et al. (2020)	1	270	12	0.35	20.60 (4.56)		collectivism	China	43	cross‐sectional	happiness	IABS	SWLS	Published
Chen (2025)a	1	76,897	16		43.02 (16.37)	16–103				cross‐sectional	happiness	Others	Others	Published
Chen (2025)b	1	76,897	16		43.02 (16.37)	16–103				cross‐sectional	life satisfaction	Others	Others	Published
Cheng (2016)	1	313	16	0.49		10–13	collectivism	China	43	cross‐sectional	loneliness	PTM	CLS	Unpublished
Chu (2019)	1	777	13	0.58			collectivism	China	43	cross‐sectional	positive affect	IABS	PANAS	Unpublished
Corral‐Verdugo et al. (2011)	1	606	16	0.40	21.61 (7.22)	18–44	collectivism	Mexico	34	cross‐sectional	happiness	Others	Others	Published
Cui (2022)	1	657	16	0.44			collectivism	China	43	cross‐sectional	well‐being	SRAS‐DR	PANAS	Unpublished
Dai (2017)a	1	355	17	0.43			collectivism	China	43	cross‐sectional	mental health	PTM	Others	Unpublished
Dai (2017)b	1	355	17	0.43			collectivism	China	43	cross‐sectional	depression	PTM	Others	Unpublished
Dai (2017)c	1	355	17	0.43			collectivism	China	43	cross‐sectional	anxiety	PTM	Others	Unpublished
Deng et al. (2023)	1	779	16	0.26	18.46 (0.63)		collectivism	China	43	cross‐sectional	loneliness	Others	CLS	Published
Di Tata (2025)	1	458	17	0.45	11.45 (1.53)	9–14	individualism	Italy	53	cross‐sectional	loneliness	SDQ	Others	Published
Dong et al. (2021)	1	487	16	0.42	(1.67)	12–18	collectivism	China	43	cross‐sectional	loneliness	ABCS	Others	Published
Dou et al. (2019)a	1	1009	16	0.49	14.72 (1.27)	18–25	collectivism	China	43	cross‐sectional	life satisfaction	SDQ	SWLS	Published
Dou et al. (2019)b	1	2620	16	0.23	21.48 (5.27)	22–40	collectivism	China	43	cross‐sectional	life satisfaction	PTM	SWLS	Published
Dou et al. (2019)c	1	500	16	0.40	28.91		collectivism	China	43	cross‐sectional	life satisfaction	Others	SWLS	Published
Duan et al. (2022)	1	351	17	0.49		18–50	collectivism	China	43	cross‐sectional	well‐being	PTM	IWB	Published
Durrani (2023)	1	200	13	0.51	(1.67)	12–18	collectivism	Pakistan	5	cross‐sectional	life satisfaction	Others	SWLS	Published
Eli et al. (2021)	1	11160	17	0.49	14.34 (1.77)	10–17	collectivism	China	43	cross‐sectional	depression	SDQ	CES‐D‐10	Published
Feng and Guo (2016)a	1	525	17	0.24	20.13 (1.97)		collectivism	China	43	cross‐sectional	well‐being	SRAS‐DR	Others	Published
Feng and Guo (2016)b	2	525	17	0.24	20.13 (1.97)		collectivism	China	43	cross‐sectional	anxiety	SRAS‐DR	Others	Published
Feng and Guo (2016)c	1	525	17	0.24	20.13 (1.97)		collectivism	China	43	cross‐sectional	depression	SRAS‐DR	CES‐D‐10	Published
Feng (2017)	1	681	17	0.31	18.75 (0.79)	17–22	collectivism	China	43	cross‐sectional	well‐being	SRAS‐DR	PANAS	Unpublished
Feng (2020)	2	884	16	0.45	14.10 (1.55)	12–18	collectivism	China	43	cross‐sectional	well‐being	VIA‐IS	SWLS	Unpublished
Feng (2022)a	1	553	17	0.38	21 (1.56)	18–25	collectivism	China	43	cross‐sectional	mental health	Others	Others	Published
Feng (2022)b	1	553	17	0.38	21 (1.56)	18–25	collectivism	China	43	cross‐sectional	mental health	IABS	Others	Published
Feng et al. (2023)a	1	821	17	0.40	20.36 (1.52)	18–24	collectivism	China	43	cross‐sectional	mental health	Others	Others	Published
Feng et al. (2023)b	1	821	17	0.40	20.36 (1.52)	18–24	collectivism	China	43	cross‐sectional	mental health	IABS	Others	Published
Geukens et al. (2021)	1	1594	17	0.53	9.43 (0.77)	8–11	individualism	Netherlands	100	cross‐sectional	loneliness	nomination	Others	Published
Griese (2011)a	2	249	16	1			individualism	USA	60	longitudinal	loneliness	nomination	Others	Unpublished
Griese (2011)b	2	262	16	0			individualism	USA	60	longitudinal	loneliness	nomination	Others	Unpublished
Hu (2009)a	4	721	17	0.56			collectivism	China	43	longitudinal	depression	Others	CDI	Unpublished
Hu (2009)b	4	721	17	0.56			collectivism	China	43	longitudinal	loneliness	Others	CLS	Unpublished
Huang et al. (2016)a	1	130	17	0			collectivism	China	43	cross‐sectional	loneliness	PTM	Others	Published
Huang et al. (2016)b	1	175	17	1			collectivism	China	43	cross‐sectional	loneliness	PTM	Others	Published
Huang et al. (2018)	1	412	16	0.47	20.74 (1.37)	18–23	collectivism	China	43	cross‐sectional	well‐being	Others	Others	Published
Hui (2022)	1	3452	18	0.51	18.21 (2.01)		collectivism	China	43	cross‐sectional	life satisfaction	SRAS‐DR	SWLS	Published
Jing (2021)	1	360	16	0.43			collectivism	China	43	cross‐sectional	life satisfaction	ABCS	IWB	Published
Ke et al. (2022)	1	881	16	0.52			collectivism	China	43	cross‐sectional	meaning in life	PTM	Others	Published
Kumar and Dixit (2017)	1	153	15	0.55		60–75	collectivism	India	24	cross‐sectional	happiness	Others	Others	Published
Kwok et al. (2017)	4	368	17	0.20	35.8 (7.35)		collectivism	China	43	longitudinal	anxiety	Others	Others	Published
Laguna (2022)	2	181	15	0.39	21.61 (1.67)		collectivism	Poland	47	longitudinal	positive affect	Others	PANAS	Published
Lee et al. (2022)	1	100	17	0.36	67 (8.70)	50–85	individualism	Canada	72	cross‐sectional	loneliness	Others	Others	Published
Li (2015)	1	273	13	0.54	20.5 (0.90)	19–23	collectivism	China	43	cross‐sectional	well‐being	ABCS	IWB	Published
Li (2016)	1	134	16	0.51			collectivism	China	43	cross‐sectional	mental health	Others	Others	Published
Li et al. (2017)	1	481	16	0.44			collectivism	China	43	cross‐sectional	well‐being	ABCS	Others	Published
Li (2019)	9	372	19	0.69			collectivism	China	43	longitudinal	depression	SDQ	CES‐D‐10	Published
Li (2021)a	1	239	18	1	(1.74)	10.5–17.5	collectivism	China	43	cross‐sectional	well‐being	PBS	SWLS	Published
Li (2021)b	1	273	18	0	(1.74)	10.5–17.5	collectivism	China	43	cross‐sectional	well‐being	PBS	SWLS	Published
Li (2022)a	1	689	17	0.57			collectivism	China	43	cross‐sectional	positive affect	PTM	Others	Unpublished
Li (2022)b	1	689	17	0.57			collectivism	China	43	cross‐sectional	negative affect	PTM	Others	Unpublished
Li (2023)	1	221	17	0.40	20.16 (1.57)	17–21	collectivism	China	43	cross‐sectional	well‐being	Others	PANAS	Published
Li (2025)a	3	554	15	0.52	12.70 (0.28)	10–13	collectivism	China	43	longitudinal	well‐being	PBS	SWLS	Published
Li (2025)b	3	554	15	0.52	12.70 (0.28)	10–13	collectivism	China	43	longitudinal	negative affect	PBS	PANAS	Published
Li (2025)c	3	554	15	0.52	12.70 (0.28)	10–13	collectivism	China	43	longitudinal	positive affect	PBS	PANAS	Published
Liang (2022)a	2	1064	15	0.39			collectivism	China	43	cross‐sectional	life satisfaction	ABCS	SWLS	Unpublished
Liang (2022)b	4	1064	15	0.39			collectivism	China	43	cross‐sectional	positive affect	ABCS	SWLS	Unpublished
Liang (2022)c	2	1064	15	0.39			collectivism	China	43	cross‐sectional	well‐being	ABCS	SWLS	Unpublished
Liang (2022)d	1	1064	15	0.39			collectivism	China	43	cross‐sectional	life satisfaction	IABS	SWLS	Unpublished
Liang (2022)e	2	1064	15	0.39			collectivism	China	43	cross‐sectional	positive affect	IABS	SWLS	Unpublished
Liang (2022)f	1	1064	15	0.39			collectivism	China	43	cross‐sectional	well‐being	IABS	SWLS	Unpublished
Liu (2024)	9	553	17	0.51	16.22 (0.70)	14–18	collectivism	China	43	longitudinal	loneliness	Others	Others	Published
Liu (2024)	9	606	17	0.54	13.80 (0.52)		collectivism	China	43	longitudinal	emotional problems	SDQ	Others	Published
Lu (2020)	1	1064	15	0.52			collectivism	China	43	cross‐sectional	well‐being	PTM	IWB	Unpublished
Luo (2022)a	1	708	15	0.70			collectivism	China	43	cross‐sectional	life satisfaction	PTM	SWLS	Unpublished
Luo (2022)b	1	708	15	0.70			collectivism	China	43	cross‐sectional	positive affect	PTM	SWLS	Unpublished
Lv (2012)a	1	1093	17	1	12.7 (0.36)		collectivism	China	43	cross‐sectional	loneliness	Others	CLS	Unpublished
Lv (2012)b	1	1093	17	1	12.7 (0.36)		collectivism	China	43	cross‐sectional	depression	Others	CDI	Unpublished
Lv (2012)c	1	1004	17	0	12.7 (0.36)		collectivism	China	43	cross‐sectional	loneliness	Others	CLS	Unpublished
Lv (2012)d	1	1004	17	0	12.7 (0.36)		collectivism	China	43	cross‐sectional	depression	Others	CDI	Unpublished
Lv (2021)	1	665	13	0.38			collectivism	China	43	cross‐sectional	well‐being	PTM	SWLS	Published
Lv (2021)a	1	186	13	0.44	22.20 (2.32)	18‐29	collectivism	China	43	cross‐sectional	positive affect	Others	PANAS	Unpublished
Lv (2021)b	1	186	13	0.44	22.20 (2.32)	18––29	collectivism	China	43	cross‐sectional	negative affect	Others	PANAS	Unpublished
Lv (2021)c	1	186	13	0.44	22.20 (2.32)	18–29	collectivism	China	43	cross‐sectional	life satisfaction	Others	SWLS	Unpublished
Lv (2021)d	1	186	13	0.44	22.20 (2.32)	18–29	collectivism	China	43	cross‐sectional	well‐being	Others	SWLS	Unpublished
Ma (2019)a	1	627	17	0.39	16.07	12–18	collectivism	China	43	cross‐sectional	depression	PBS	CES‐D‐10	Published
Ma (2019)b	1	627	17	0.39	16.07	12‐18	collectivism	China	43	cross‐sectional	loneliness	PBS	CLS	Published
Ma (2019)c	1	627	16	0.39	16.07	12–18	collectivism	China	43	cross‐sectional	life satisfaction	PBS	SWLS	Published
Ma (2019)d	1	627	16	0.39	16.07	12–‐18	collectivism	China	43	cross‐sectional	positive affect	PBS	Others	Published
Ma (2019)e	1	627	16	0.39	16.07	12–18	collectivism	China	43	cross‐sectional	negative affect	PBS	Others	Published
Ma (2022)	1	2256	16	0.43			collectivism	China	43	cross‐sectional	well‐being	PTM	IWB	Published
Martela (2016)a	1	76	16	0.36	20.4		individualism	Australia	73	cross‐sectional	positive affect	Others	Others	Published
Martela (2016)b	1	76	16	0.36	20.4		individualism	Australia	73	cross‐sectional	negative affect	Others	Others	Published
Mouratidis (2009)	1	247	16	0.49			individualism	Belgium	81	cross‐sectional	loneliness	Others	CLS	Published
Nakamura (2025)a	1	9971	19				individualism	USA	60	longitudinal	depression	Others	Others	Published
Nakamura (2025)b	1	9971	19				individualism	USA	60	longitudinal	stress	Others	Others	Published
Nakamura (2025)c	1	9971	19				individualism	USA	60	longitudinal	happiness	Others	Others	Published
Nakamura (2025)d	1	9971	19				individualism	USA	60	longitudinal	loneliness	Others	Others	Published
Nastina (2025)	1	247	17	0.50	53.50 (12.8)		individualism	Canada	72	longitudinal	positive affect	Others	Others	Published
Padilla‐Walker (2015)a	12	465	17	0.48	15.29 (1.06)	14–16	individualism	USA	60	longitudinal	depression	VIA‐IS	CES‐D‐10	Published
Padilla‐Walker (2015)b	12	465	17	0.48	15.29 (1.06)	14–16	individualism	USA	60	longitudinal	anxiety	VIA‐IS	SCAS	Published
Padilla‐Walker (2025)a	6	472	17	0.48	18.37 (1.04)		individualism	USA	60	longitudinal	anxiety	Others	Others	Published
Padilla‐Walker (2025)b	6	472	17	0.48	18.37 (1.04)		individualism	USA	60	longitudinal	depression	Others	Others	Published
Rinner (2022)	3	222	18	0.51	32 (12.24)	18–75	individualism	Switzerland	79	cross‐sectional	well‐being	Others	Others	Published
Ripoll‐Núñe (2019)a	1	930	17	0.44	13.83 (1.32)		collectivism	Colombia	29	cross‐sectional	well‐being	Others	Others	Published
Ripoll‐Núñe (2019)b	1	930	17	0.44	13.83 (1.32)		collectivism	Colombia	29	cross‐sectional	life satisfaction	Others	Others	Published
Rosli (2021)	1	156	16	0.31	22.31 (1.33)	20–27	collectivism	Malaysia	27	cross‐sectional	well‐being	Others	Others	Published
Schacter (2018)a	1	99	17	0.55	18.01 (1.12)	14.92–21.30	individualism	USA	60	longitudinal	positive affect	Others	PANAS	Published
Schacter (2018)b	1	99	17	0.55	18.01 (1.12)	14.92–21.30	individualism	USA	60	longitudinal	negative affect	Others	PANAS	Published
Schacter (2018)c	1	99	17	0.55	18.01 (1.12)	14.92–21.30	individualism	USA	60	longitudinal	depression	Others	Others	Published
Schwartz (2002)	2	122	16	0.54	11	10‐12	collectivism	South Korea	18	cross‐sectional	loneliness	Others	Others	Published
Sharma (2015)a	1	100	12			18–21	collectivism	India	24	cross‐sectional	positive affect	PBS	PANAS	Published
Sharma (2015)b	1	100	12			18–21	collectivism	India	24	cross‐sectional	negative affect	PBS	PANAS	Published
Son (2019)a	1	230	17	1	18.4 (1.04)	16–21	individualism	USA	60	cross‐sectional	anxiety	Others	SCAS	Published
Son (2019)b	1	230	17	1	18.4 (1.04)	16–21	individualism	USA	60	cross‐sectional	depression	Others	CES‐D‐10	Published
Son (2019)c	1	230	17	1	18.4 (1.04)	16–21	individualism	USA	60	cross‐sectional	life satisfaction	Others	SWLS	Published
Son (2019)d	1	240	17	0	18.4 (1.04)	16–21	individualism	USA	60	cross‐sectional	anxiety	Others	SCAS	Published
Son (2019)e	1	240	17	0	18.4 (1.04)	16–21	individualism	USA	60	cross‐sectional	depression	Others	CES‐D‐10	Published
Son (2019)f	1	240	17	0	18.4 (1.04)	16–21	individualism	USA	60	cross‐sectional	life satisfaction	Others	SWLS	Published
Spithoven (2017)	1	1424	17	0.42	15.67 (1.25)		individualism	Belgium	81	cross‐sectional	loneliness	Others	Others	Published
Sun (2022)	1	1555	13				collectivism	China	43	cross‐sectional	well‐being	PTM	Others	Published
Tamura (2024)a	1	34,187	18				collectivism	Japan	46	longitudinal	depression	Others	Others	Published
Tamura (2024)b	1	34,187	18				collectivism	Japan	46	longitudinal	loneliness	Others	Others	Published
Tamura (2024)c	1	34,187	18				collectivism	Japan	46	longitudinal	happiness	Others	Others	Published
Tang (2019)	1	610	16	0.36			collectivism	China	43	cross‐sectional	well‐being	PTM	IWB	Unpublished
Valentiner (2017)	1	65	17	0.49	12.39 (1.64)		individualism	USA	60	cross‐sectional	loneliness	Others	Others	Published
Wang (2010)a	1	589	16	0.57		16–19	collectivism	China	43	cross‐sectional	positive affect	Others	Others	Published
Wang (2010)b	1	589	16	0.57		16‐19	collectivism	China	43	cross‐sectional	negative affect	Others	Others	Published
Wang (2010)c	1	589	16	0.57		16–19	collectivism	China	43	cross‐sectional	mental health	Others	Others	Published
Wang (2012)a	1	390	16	0.44			collectivism	China	43	cross‐sectional	life satisfaction	ABCS	Others	Published
Wang (2012)b	1	390	16	0.44			collectivism	China	43	cross‐sectional	positive affect	ABCS	Others	Published
Wang (2012)c	1	390	16	0.44			collectivism	China	43	cross‐sectional	negative affect	ABCS	Others	Published
Wang (2016)a	1	2949	16	0.44	15.51 (1.57)	12–18	collectivism	China	43	cross‐sectional	life satisfaction	PBS	Others	Published
Wang (2016)b	1	2949	16	0.44	15.51 (1.57)	12–18	collectivism	China	43	cross‐sectional	positive affect	PBS	PANAS	Published
Wang (2016)c	1	2949	16	0.44	15.51 (1.57)	12–18	collectivism	China	43	cross‐sectional	negative affect	PBS	PANAS	Published
Wang (2017)	1	248	14	0.49			collectivism	China	43	cross‐sectional	well‐being	PTM	Others	Unpublished
Wang (2018)	1	1008	17	0.51	13.1 (0.99)	11–17	collectivism	China	43	cross‐sectional	loneliness	Others	CLS	Published
Wang (2022)a	1	601	12	0.51			collectivism	China	43	cross‐sectional	life satisfaction	PTM	SWLS	Unpublished
Wang (2022)b	1	601	12	0.51			collectivism	China	43	cross‐sectional	positive affect	PTM	SWLS	Unpublished
Wang (2022)c	1	601	12	0.51			collectivism	China	43	cross‐sectional	negative affect	PTM	SWLS	Unpublished
Wang (2022)d	1	601	12	0.51			collectivism	China	43	cross‐sectional	well‐being	PTM	SWLS	Unpublished
Wang (2023)	1	239	16	0.40	17 (2.00)	15–19	collectivism	China	43	cross‐sectional	life satisfaction	IABS	SWLS	Published
Wang (2023)	1	208	16	0.66			collectivism	China	43	cross‐sectional	well‐being	Others	Others	Published
Wang (2024)a	1	283	17	0.38	65.26 (5.73)	60–85	collectivism	China	43	cross‐sectional	well‐being	Others	Others	Published
Wang (2024)b	1	283	17	0.38	65.26 (5.73)	60–85	collectivism	China	43	cross‐sectional	meaning in life	Others	Others	Published
Wei (2015)	2	2097	16	0.52	12.27 (0.36)		collectivism	China	43	cross‐sectional	loneliness	Others	CLS	Published
Wen (2023)	1	497	12	0.62			collectivism	China	43	cross‐sectional	negative affect	PTM	PANAS	Published
Wentzel (2007)	1	263	16	0.48		11–14	individualism	USA	60	cross‐sectional	depression	nomination	Others	Published
Woodhouse (2011)	1	2091	17	0.39	16.5	16–17	individualism	USA	60	cross‐sectional	loneliness	Others	CLS	Published
Wu (2017)	1	964	16	0.41			collectivism	China	43	cross‐sectional	well‐being	SRAS‐DR	Others	Unpublished
Xiong (2022)	1	275	16	0.31			collectivism	China	43	cross‐sectional	loneliness	PTM	Others	Unpublished
Xiong (2023)	9	1248	17	0.53	13.44 (0.65)	12–15	collectivism	China	43	longitudinal	well‐being	nomination	IWB	Published
Xu (2019)	1	1042	16	0.51			collectivism	China	43	longitudinal	negative affect	Others	Others	Published
Yang (2017)	1	293	17	0.49	16.5 (0.85)		collectivism	China	43	cross‐sectional	well‐being	PBS	SWLS	Published
Yang (2017)	1	898	17	0.69		15–18	collectivism	China	43	cross‐sectional	positive affect	PBS	PANAS	Published
Yang (2017)	1	898	17	0.69			collectivism	China	43	cross‐sectional	negative affect	PBS	PANAS	Published
Yang (2017)a	1	2082	12	0.43	15.32 (1.96)	13–18	collectivism	China	43	cross‐sectional	meaning in life	Others	Others	Published
Yang (2017)b	1	2082	12	0.43	15.32 (1.96)	13–18	collectivism	China	43	cross‐sectional	life satisfaction	Others	Others	Published
Yang (2017)c	1	2082	12	0.43	15.32 (1.96)	13–18	collectivism	China	43	cross‐sectional	positive affect	Others	Others	Published
Yang (2017)d	1	2082	12	0.43	15.32 (1.96)	13–18	collectivism	China	43	cross‐sectional	negative affect	Others	Others	Published
Yelpaze (2020)	1	559	15	0.35	21	18–36	collectivism	Turkey	46	cross‐sectional	life satisfaction	Others	SWLS	Published
Yu (2016)a	1	645	16	0.54	11.42 (1.17)	10‐15	collectivism	China	43	cross‐sectional	loneliness	Others	CLS	Unpublished
Yu (2016)b	1	645	16	0.54	11.42 (1.17)	10–15	collectivism	China	43	cross‐sectional	well‐being	Others	Others	Unpublished
Yu (2016)c	1	645	16	0.54	11.42 (1.17)	10–15	collectivism	China	43	cross‐sectional	depression	Others	CDI	Unpublished
Zeng (2022)	1	1488	15	0.34	19.84 (1.61)	17–25	collectivism	China	43	cross‐sectional	well‐being	IABS	IWB	Published
Zhang (2015)	1	532	15	0.69			collectivism	China	43	cross‐sectional	well‐being	PTM	Others	Published
Zhang (2018)	1	749	16	0.43	15.79 (1.54)	13–20	collectivism	China	43	cross‐sectional	life satisfaction	Others	SWLS	Published
Zhang (2021)	1	4959	17	0.49	14.12		collectivism	China	43	cross‐sectional	depression	SDQ	Others	Published
Zhang (2022)	1	174	17	0.44	21.48 (2.63)		collectivism	China	43	cross‐sectional	well‐being	SRAS‐DR	Others	Unpublished
Zhang (2022)	2	120	17				collectivism	China	43	longitudinal	well‐being	Others	Others	Published
Zhang (2022)	4	913	13	0.26	19.63 (1.04)		collectivism	China	43	longitudinal	meaning in life	PTM	Others	Published
Zhang (2022)a	1	862	16	0.48			collectivism	China	43	cross‐sectional	life satisfaction	PBS	Others	Published
Zhang (2022)b	1	862	16	0.48			collectivism	China	43	cross‐sectional	positive affect	PBS	Others	Published
Zhang (2022)c	1	862	16	0.48			collectivism	China	43	cross‐sectional	negative affect	PBS	Others	Published
Zhang (2023)a	1	3169	18	0.51	13.09 (1.31)	11–16	collectivism	China	43	cross‐sectional	anxiety	SDQ	Others	Published
Zhang (2023)b	1	3169	18	0.51	13.09 (1.31)	11–16	collectivism	China	43	cross‐sectional	depression	SDQ	CDI	Published
Zhao (2016)	1	1059	16				collectivism	China	43	longitudinal	anxiety	PTM	Others	Unpublished
Zhao (2019)a	1	2102	17	0.53	13.48 (1.10)		collectivism	China	43	cross‐sectional	life satisfaction	PBS	Others	Published
Zhao (2019)b	1	2102	17	0.53	13.48 (1.10)		collectivism	China	43	cross‐sectional	depression	PBS	CDI	Published
Zhao (2019)c	1	2102	17	0.53	13.48 (1.10)		collectivism	China	43	cross‐sectional	loneliness	PBS	CLS	Published
Zhao (2019)d	1	2102	17	0.53	13.48 (1.10)		collectivism	China	43	cross‐sectional	happiness	PBS	Others	Published
Zhao (2021)	18	514	15	0.5			collectivism	China	43	longitudinal	well‐being	VIA‐IS	SWLS	Unpublished
Zheng (2012)	1	496	16	0.45	20.69 (1.86)		collectivism	China	43	cross‐sectional	anxiety	IABS	Others	Published
Zheng (2017)	1	887	16	0.42	16.82 (3.01)		collectivism	China	43	cross‐sectional	well‐being	IABS	IWB	Published
Zheng (2018)	1	467	16	0.46	13.86 (1.58)	12–16	collectivism	China	43	cross‐sectional	well‐being	IABS	IWB	Published
Zheng (2018)	1	356	17	0.45	20.64 (1.31)	17–24	collectivism	China	43	cross‐sectional	well‐being	IABS	Others	Published
Zhou (2018)a	9	1710	16	0.5	(0.54)		collectivism	China	43	longitudinal	positive affect	PBS	IWB	Published
Zhou (2018)b	9	1710	16	0.5	(0.54)		collectivism	China	43	longitudinal	life satisfaction	PBS	IWB	Published
Zhou (2018)c	9	1710	16	0.5	(0.54)		collectivism	China	43	longitudinal	well‐being	PBS	IWB	Published
Zhou (2021)	1	309	16				collectivism	China	43	cross‐sectional	well‐being	SRAS‐DR	Others	Unpublished
Zhou (2022)	1	467	17				collectivism	China	43	cross‐sectional	loneliness	PTM	CLS	Unpublished
Zhu (2020)	1	601	16	0.53		13–18	collectivism	China	43	cross‐sectional	loneliness	SDQ	Others	Unpublished
Zhu (2022)	1	612	13	0.47			collectivism	China	43	cross‐sectional	loneliness	PTM	Others	Unpublished

Abbreviations: ABCS, The Altruistic Behavior of College Students; CDI, Child Depression Inventory; CES‐D‐10, 10‐item Center for Epidemiologic Studies Depression Scale; CLS, Children's Loneliness Scale; Gender, percentage of males; IABS, Internet Altruistic Behavior Scale; IVD, Individualism Index; IWB, Index of Well‐being; K, number of effect sizes; MH, mental health; N, number of participants; PANAS, Positive and Negative Affect Scale; PB, prosocial behavior; PBS, Prosocial Behavior Scale; PTM, Prosocial Tendencies Measure; QS, Quality Assessment; SCAS, Spence Children's Anxiety Scale; SD, Standard Deviation; SDQ, Strength and Difficulty Questionnaire; SRAS‐DR, Self‐Report Altruism Scale Distinguished by the Recipient; SWLS, Satisfaction with Life Scale; VIA‐IS, The Values in Action Inventory of Strengths.

3.4. Sample Characteristics

Three sample characteristics were coded. First, gender composition was recorded based on the percentage of male participants. Second, age was examined both as a continuous variable and as a categorical moderator across developmental groups. For the categorical moderator analyses, studies were classified into four developmental groups based on the mean sample age: childhood (< 12 years), adolescence (12–17 years), young adulthood (18–24 years), and adulthood (≥ 25 years). Third, cultural context was coded categorically as collectivistic versus individualistic based on the country's individualism index following Hofstede's Cultural Dimension Theory (https://www.hofstede-insights.com). Countries with an individualism score of 50 or lower were classified as collectivistic, and those above 50 were classified as individualistic.

3.5. Study Characteristics

First, research design was coded as cross‐sectional or longitudinal. Second, in terms of publication types, we examined both publication years and status. Publication years were coded as a continuous variable. The publication status was coded categorically as published (peer‐reviewed journal articles) or unpublished (dissertations and conference papers). Third, mental health outcomes were classified as positive (e.g., subjective well‐being, positive affect, and happiness) or negative (e.g., depressive symptoms, anxiety, and loneliness). Finally, the instruments used to assess prosocial behavior and mental health were coded to examine whether measurement tools moderated the association. Due to small frequencies (fewer than 10 instances) for some instruments, they were grouped into an “Others” category. Prosocial behavior measures were categorized into nine groups, including: (1) Prosocial Behavior Scale (PBS), (2) Values in Action Inventory of Strengths (VIA‐IS), (3) Prosocial Tendencies Measure (PTM), (4) Altruistic Behavior of College Students Scale (ABCS), (5) nomination, (6) Internet Altruistic Behavior Scale (IABS), (7) Strengths and Difficulties Questionnaire (SDQ), (8) Self‐Report Altruism Scale Distinguished by Recipient (SRAS‐DR), and (9) Others. Mental health measures were grouped into eight categories: (1) Satisfaction with Life Scale (SWLS), (2) Index of Well‐Being (IWB), (3) 10‐item Center for Epidemiologic Studies Depression Scale (CES‐D‐10), (4) Positive and Negative Affect Schedule (PANAS), (5) Children's Loneliness Scale (CLS), (6) Spence Children's Anxiety Scale (SCAS), (7) Child Depression Inventory (CDI), and (8) Others.

3.6. Quality Assessment

We used the 10‐item Joanna Briggs Institute (JBI) critical appraisal checklist to assess the quality of eligible studies (Munn et al. 2020). Each item was scored as “information present = 2,” “present but with limitations = 1,” or “information absent = 0.” An example item is “Was statistical analysis appropriate?” The maximum possible score was 20. Following previous research (Fan et al. 2021), studies scoring above 10 were considered to be of acceptable methodological quality. The quality of the included studies was independently assessed by two trained researchers, and any disagreements were resolved through discussion.

3.7. Data Analyses

Since both prosocial behavior and mental health were treated as continuous variables, Pearson's correlation coefficients (r) were used as the effect size metric. To ensure directional consistency, correlations were recoded when necessary so that higher scores uniformly indicated better mental health (e.g., correlations involving depressive symptoms, anxiety, and loneliness were reverse‐coded). Because Pearson's r is not normally distributed, all effect sizes were first transformed to Fisher's z‐scores prior to analysis (Card 2015; Hedges and Olkin 2014). Subsequently, the Fisher's z‐scores were retransformed into Pearson's correlation coefficients for interpretation.

The traditional meta‐analysis approach assumes that effect sizes are independent of each other. However, some studies included in our meta‐analysis reported multiple effect sizes from the same sample, thereby violating this assumption. Ignoring such dependency could bias the results (Hox et al. 2017; Lipsey and Wilson 2001). Thus, we employed a three‐level meta‐analysis model to estimate the relation between prosocial behavior and mental health.

The three‐level meta‐analysis model contained three sources of variance: sampling variance at level 1, variance across effect sizes within the same study at level 2, and variance across effect sizes between studies at level 3 (Cheung 2014; Ran et al. 2021; Yan et al. 2020). Compared to traditional meta‐analytic approaches, the three‐level model explicitly accounts for the dependency among effect sizes originating from the same study (Lei et al. 2020). By modeling these dependencies, this approach maximizes the use of available information and improves statistical power (Assink and Wibbelink 2016; Hox et al. 2017; Van den Noortgate et al. 2013).

We conducted the three‐level meta‐analysis in RStudio (R Core Team 2021) with the Metafor package (Viechtbauer 2010). All model parameters were estimated using the restricted maximum likelihood method (Viechtbauer 2005), and a two‐tailed p‐value smaller than 0.05 was considered statistically significant. The present three‐level meta‐analysis was performed in the following steps. First, we calculated the overall mean effect size to assess the direction and magnitude of the relation between prosocial behavior and mental health. Cohen (1992) suggested criteria were used to interpret the mean effect sizes, which demonstrates that r < 0.01 means trivial, r = 0.10 means small, r = 0.30 means medium, and r = 0.50 means large. Second, we applied the log‐likelihood ratio test to determine heterogeneity at levels 2 (within‐study) and 3 (between‐study heterogeneity; Assink and Wibbelink 2016). Third, we employed four methods to assess potential publication bias: (a) visual inspection of funnel plots of effect size standard errors; (b) calculation of fail‐safe N (Hunter and Schmidt 2004; Rosenthal 1991); (c) Egger's regression test; and (d) the p‐curve analysis (Simonsohn et al. 2014). Fourth, sensitivity analyses were conducted by sequentially excluding each study and recalculating the overall effect size to examine the robustness of the results. Finally, to further address potential dependency among effect sizes drawn from the same study, we conducted robust variance estimation (RVE) with cluster‐robust standard errors as a robustness check (Assink and Wibbelink 2024). The RVE analyses were based on the fitted three‐level meta‐analytic model, with standard errors clustered at the study level. A small‐sample correction was applied to improve the accuracy of statistical inferences.

4. Results

4.1. Sample Description

In total, 366 effect sizes from 121 studies were included in the current meta‐analysis, encompassing a combined sample of 222,866 participants (see Table 1). Sample sizes ranged from 65 to 76,897, with a mean participant age of 19.20 years. Of the included studies, 98 employed a cross‐sectional design, while 23 utilized a longitudinal design. The number of effect sizes contributed by each study ranged from 1 to 27. Regarding cultural background, 100 effect sizes were derived from samples in individualistic cultures and 264 from collectivist cultures, while two effect sizes were based on cross‐national samples. In terms of publication status, 283 effect sizes came from peer‐reviewed journal articles, and 83 from non‐journal sources such as dissertations and conference papers. With respect to mental health indicators, 195 effect sizes assessed positive indicators and 171 assessed negative indicators.

4.2. Quality Assessment

In the quality assessment, all 121 studies scored above 10 points (see Table 1), with a mean quality score of 16.56, indicating that the overall methodological quality of the included studies was acceptable.

4.3. The Overall Relation Between Prosocial Behavior and Mental Health

A random‐effects model was employed to estimate the overall relation between prosocial behavior and mental health. Results showed a significant correlation between these two variables (r = 0.26, p < 0.001, 95% CI = [0.24, 0.29]). RVE yielded results highly consistent with those obtained from the three‐level meta‐analysis (r = 0.26, 95% CI = [0.24, 0.29]), indicating that the overall effect size estimate is robust to within‐study dependence. Moreover, the log‐likelihood ratio test revealed significant heterogeneity (p < 0.001) at both the within‐study level (level 2) and the between‐study level (level 3). Follow‐up variance component analyses showed that variances at the sampling, within‐study, and between‐study levels were 2.10%, 28.93%, and 68.97%, respectively. The substantial heterogeneity observed across effect sizes suggested that examining potential moderators would be meaningful (J. B. Li et al. 2019). Thus, we conducted moderator analyses to examine whether the association between prosocial behavior and mental health varied by sample or study characteristics.

4.4. Moderating Analyses

4.4.1. Sample Characteristics

Gender did not significantly moderate the association between prosocial behavior and mental health (F _{(1, 349)} = 2.92, p = 0.088).

Neither the continuous nor categorical analysis of age revealed a significant moderating effect. The continuous analysis yielded F _{(1, 214)} = 0.50, p = 0.481, and the categorical analysis across childhood, adolescence, young adulthood, and adulthood yielded F _{(3, 212)} = 0.23, p = 0.875.

Culture significantly moderated the association (F _{(1, 362)} = 16.16, p < 0.001), with a stronger effect observed in collectivist cultures (r = 0.30, 95% CI = [0.27, 0.32]) than in individualistic cultures (r = 0.15, 95% CI = [0.08, 0.22]).

4.5. Study Characteristics

We examined whether the type of mental health indicators moderated the association between prosocial behavior and mental health. As shown in Table 2, a significant moderating effect of mental health indicators was found (F _{(1, 364)} = 60.86, p < 0.001). Specifically, the association was stronger when mental health was assessed using positive indicators (r = 0.33, 95% CI = [0.29, 0.36]) than when using negative indicators (r = 0.19, 95% CI = [0.15, 0.22]). Further analyses examining 11 specific mental health (i.e., well‐being, loneliness, life satisfaction, positive affect, depression, negative affect, anxiety, happiness, meaning in life, mental health, and others) constructs revealed significant differences among them (F _{(10, 355)} = 8.63, p < 0.001). Among these constructs, meaning in life showed the strongest association with prosocial behavior (r = 0.33, 95% CI = [0.22, 0.44]). In contrast, negative affect (r = 0.09, 95% CI = [0.04, 0.15]) and anxiety (r = 0.18, 95% CI = [0.11, 0.24]) exhibited comparatively weaker associations.

TABLE 2.

Results of categorical and continuous moderators for the association between prosocial behavior and mental health.

# Moderator variables	# ES	Intercept/mean z (95% CI)	β1 (95% CI)	Mean r	F (df1, df2) a	p b	Levels 2 variance	Levels 3 variance
(1) Sample characteristics
a. Gender	351	0.32 (0.26, 0.37)***	−0.087 (−0.186, 0.013)	—	F (1, 349) = 2.92	0.088	0.009***	0.021***
b. Age	216	0.21 (0.14, 0.29)***	0.001 (−0.002, 0.005)	—	F (1, 214) = 0.50	0.481	0.008***	0.018***
c. Age groups					F (3, 212) = 0.23	0.875	0.008***	0.019***
Childhood	20	0.21 (0.09, 0.32)***		0.20
Adolescence	121	0.25 (0.20, 0.31)***	0.046 (−0.084, 0.177)	0.25
Young adulthood	60	0.23 (0.16, 0.29)***	0.021 (−0.122, 0.155)	0.22
Adulthood	15	0.23 (0.12, 0.34)***	0.024 (−0.135, 0.184)	0.23
d. Culture					F (1, 362) = 16.16	< 0.001	0.009***	0.018***
Collectivist culture (RC)	264	0.30 (0.27, 0.32)***		0.29
Individualistic culture	100	0.15 (0.08, 0.22)***	−0.149 (−0.222, ‐0.076)***	0.15
(2) Study characteristics
a. Mental health indicators
Overall					F (1, 364) = 60.86	< 0.001	0.007***	0.019***
Positive (RC)	195	0.33 (0.29, 0.36)***		0.32
Negative	171	0.19 (0.15, 0.22)***	−0.141 (−0.176, ‐0.105)***	0.18
Sub−indicator details					F (10, 355) = 8.63	< 0.001	0.007***	0.022***
Well−being (RC)	100	0.31 (0.27, 0.35)***		0.30
Life satisfaction	37	0.31 (0.26, 0.36)***	−0.005 (−0.053, 0.043)	0.30
Positive affect	32	0.30 (0.25, 0.35)***	−0.011 (−0.065, 0.043)	0.29
Happiness	7	0.27 (0.18, 0.36)***	−0.044 (−0.140, 0.052)	0.26
Meaning in life	8	0.33 (0.22, 0.44)***	0.017 (−0.095, 0.130)	0.32
Mental health	7	0.31 (0.18, 0.43)***	−0.006 (−0.133, 0.121)	0.30
Loneliness	52	0.25 (0.20, 0.30)***	−0.059 (−0.120, 0.003)	0.25
Depression	49	0.25 (0.20, 0.30)***	−0.062 (−0.124, ‐0.000)*	0.24
Anxiety	30	0.18 (0.11, 0.24)***	−0.135 (−0.207, ‐0.063)***	0.17
Negative affect	24	0.09 (0.04, 0.15)***	−0.219 (−0.279, ‐0.159)***	0.09
Others	20	0.33 (0.26, 0.40)***	0.020 (−0.045, 0.086)	0.32
b. Publication years	366	−4.43 (−17.45, 8.60)***	0.002 (−0.004, 0.009)	—	F (1, 364) = 0.50	0.479	0.009***	0.022***
c. Publication types					F (1, 364) = 0.35	0.554	0.009***	0.022***
Journal (RC)	283	0.26 (0.23, 0.30)***		0.26
Unpublished studies	83	0.29 (0.22, 0.35)***	0.022 (−0.050, 0.093)	0.28
d. Research design					F (1, 364) = 21.06	< 0.001	0.008***	0.021***
Cross−sectional study (RC)	233	0.28 (0.25, 0.31)***		0.28
Longitudinal study	133	0.21 (0.17, 0.25)***	−0.073 (−0.105, ‐0.042)**	0.21
e. Measurement for PB					F (8, 357) = 1.26	0.265	0.009***	0.020***
PBS (RC)	63	0.36 (0.27, 0.45)***		0.35
VIA−IS	44	0.26 (0.10, 0.43)**	−0.096 (−0.287, 0.096)	0.26
PTM	36	0.30 (0.23, 0.37)***	−0.058 (−0.170, 0.055)	0.29
ABCS	16	0.26 (0.16, 0.36)***	−0.096 (−0.231, 0.038)	0.26
nomination	16	0.12 (−0.03, 0.26)	−0.243 (−0.413, ‐0.073)**	0.12
IABS	15	0.26 (0.17, 0.34)***	−0.104 (−0.227, 0.020)	0.25
SDQ	25	0.27 (0.16, 0.38)***	−0.092 (−0.232, 0.048)	0.26
SRAS−DR	10	0.27 (0.14, 0.40)***	−0.094 (−0.250, 0.063)	0.26
Others	141	0.25 (0.20, 0.29)***	−0.111 (−0.211, ‐0.011)*	0.24
f. Measurement for MH					F (7, 358) = 2.52	0.015	0.009***	0.021***
SWLS (RC)	61	0.31 (0.25, 0.37)***		0.30
IWB	44	0.29 (0.19, 0.39)***	−0.004 (−0.128, 0.119)	0.29
CES−D−10	29	0.21 (0.11, 0.30)***	−0.104 (−0.225, 0.016)	0.20
PANAS	28	0.23 (0.16, 0.31)***	−0.078 (−0.158, 0.003)	0.23
CLS	20	0.29 (0.21, 0.36)***	−0.014 (−0.109, 0.080)	0.29
SCAS	14	0.11 (−0.00, 0.22)	−0.230 (−0.369, ‐0.092)**	0.08
CDI	10	0.28 (0.20, 0.37)***	−0.013 (−0.123, 0.098)	0.29
Others	160	0.26 (0.22, 0.30)***	−0.045 (−0.116, 0.026)	0.26

Note: #ES, number of effect sizes; 95% CI, 95% confidence interval; β1, estimated regression coefficient; ABCS, The Altruistic Behavior of College Students; CDI, Child Depression Inventory; CES‐D‐10, 10‐item Center for Epidemiologic Studies Depression Scale; CLS, Children's Loneliness Scale; df, degrees of freedom; For MH scale: SWLS, Satisfaction with Life Scale; For PB scale: PTM, Prosocial Tendencies Measure; IABS, Internet Altruistic Behavior Scale; IWB, Index of Well‐being; Levels 2 variance, variance between effect sizes extracted from the same study; Levels 3 variance, variance between studies; mean z, mean effect size (Fisher's z); MH, mental health; PANAS, Positive and Negative Affect Scale; PB, prosocial behavior; PBS, Prosocial Behavior Scale; r, mean effect size expressed as a Pearson's correlation; SCAS, Spence Children's Anxiety Scale; SDQ, Strength and Difficulty Questionnaire; SRAS‐DR, Self‐Report Altruism Scale Distinguished by the Recipient; VIA‐IS, The Values in Action Inventory of Strengths. ***p < 0.001. ^aOmnibus test of all regression coefficients in the model. ^b p‐value of the omnibus test. Bold values indicate statistically significant moderators in the meta‐analysis.

We further examined other study‐related moderators, including mental health measurements, prosocial behavior measurements, publication years, publication types, and research design. As shown in Table 2, mental health measurements significantly moderated the association between prosocial behavior and mental health (F _{(7, 358)} = 2.52, p = 0.015). More specifically, the association was strongest in studies using the SWLS (r = 0.31, 95% CI = [0.25, 0.37]). Research design also emerged as a significant moderator (F _{(1, 364)} = 21.06, p < 0.001), with cross‐sectional studies reporting a stronger association (r = 0.28, 95% CI = [0.25, 0.31]). However, no significant moderating effects were observed for publication years (F _{(1, 364)} = 0.50, p = 0.479), publication types (F _{(1, 364)} = 0.35, p = 0.554), or prosocial behavior measurements (F _{(8, 357)} = 1.26, p = 0.265).

4.6. Publication Bias Analyses

Multiple methods were employed to assess potential publication bias, none of which indicated substantial risk. First, visual inspection of the funnel plot revealed a symmetrical distribution (see Figure 2), indicating that sampling errors were random and that statistically significant results were not selectively favored. Second, the fail‐safe N (k ₀  > 5k + 10) yielded a sufficiently large value, suggesting that 2,244,771 additional studies with null effects would be required to reduce the overall effect size to non‐significance. Therefore, concerns about the true effect being null were unwarranted. Third, the Egger's test results also indicated that publication bias should be ignored since the p‐value of this test exceeded 0.05 (b = 0.24, p = 0.741). Fourth, the p‐curve showed significant right skew and no flatter than 33% (Figure 3), suggesting that there were more extremely significant p‐values (e.g., p < 0.001) than marginally significant ones (e.g., p = 0.049). Thus, the likelihood of extensive p‐hacking was low. Taken together, these findings suggested that the present meta‐analytic results were unlikely to be substantially affected by publication bias.

FIGURE 2.

Funnel plot of the association between prosocial behavior and mental health.

FIGURE 3.

P‐curve for meta‐analysis of prosocial behavior and mental health.

4.7. Sensitivity Analysis

A sensitivity analysis was conducted to assess the stability of the main effect size between prosocial behavior and mental health by systematically excluding individual studies from the dataset. Results showed that the average effect size ranged from 0.259 to 0.266, which was consistent with the overall effect size (r = 0.26). It indicated that the exclusion of any individual study did not significantly alter the overall effect size estimate. Thus, the findings of this meta‐analysis were stable and reliable.

5. Discussion

In this meta‐analysis, we reviewed 121 related studies encompassing a total of 222,866 participants to examine the relation between prosocial behavior and mental health, as well as the moderating effects of gender, age, culture, and certain methodological variables. Results indicated a significant positive association, suggesting that individuals who engage more frequently in prosocial behaviors tend to report better mental health. Furthermore, the strength of this association varied depending on cultural context, research design, as well as specific indicators and measurement tools used to assess mental health. These findings highlight the important role of both sample characteristics and methodological factors in shaping the observed relationship between prosocial behavior and mental health.

5.1. The Overall Relation

The current meta‐analysis provides compelling evidence for a significant positive association between prosocial behavior and mental health, highlighting that engagement in prosocial activities is associated with enhanced psychological well‐being. This finding aligns with the principles of self‐determination theory (Deci and Ryan 1985), which posits that the fulfillment of basic psychological needs (i.e., autonomy, competence, and relatedness) is fundamental to mental health. Prosocial behavior may represent one potential pathway through which these basic psychological needs are satisfied, as it fosters autonomy by allowing individuals to voluntarily engage in helping actions, promotes competence through effective contributions to others' well‐being, and strengthens relatedness by facilitating the development of meaningful social bonds (Dunn et al. 2014; Li et al. 2023). Through these processes, prosocial behavior may be associated with higher levels of psychological well‐being. For instance, Aknin et al. (2011) conduct an experiment where participants are randomly assigned to spend money either on others or on themselves, with those spending on others reporting significantly greater happiness. Similarly, Thompson et al. (2022) find that individuals serving as peer supporters for patients with chronic diseases experience significant improvements in self‐esteem, confidence, and role functioning.

However, the present meta‐analysis is based on correlational evidence and therefore does not allow conclusions regarding the causal direction of this association. It is possible that prosocial behavior is associated with better mental health, while individuals with better mental health may also be more likely to engage in prosocial behavior. Individuals with higher levels of psychological well‐being typically possess greater cognitive, emotional, and social resources (Fredrickson 2001; Hobfoll 2002), which may facilitate prosocial actions such as helping, sharing, and volunteering (Grant and Gino 2010). In addition, the observed association may reflect shared underlying factors, such as personality traits, social support, or broader contextual influences. These findings highlight the robust association between prosocial behavior and mental health and underscore the need for future longitudinal and experimental studies to clarify the causal mechanisms underlying this relationship.

5.2. The Role of Culture

The present study revealed that culture significantly moderated the association between prosocial behavior and mental health, with a notably stronger effect observed in collectivist than in individualistic cultures. This finding underscores the critical role of cultural context in shaping the psychological benefits derived from prosocial actions. In collectivistic societies, where social harmony, interdependence, and communal goals are emphasized (Markus and Kitayama 2014), prosocial behavior is highly valued for maintaining strong social bonds and fulfilling social obligations. Consequently, individuals in these settings tend to derive greater psychological benefits from engaging in prosocial acts (Buchtel et al. 2018; Tsai and Kimel 2021). Conversely, individualistic cultures prioritize personal autonomy, self‐achievement, and independence, rendering prosocial behavior more voluntary and secondary to personal goals. Although these behaviors still contribute positively to mental health, their impact is less pronounced compared to collectivistic contexts where personal identity is more closely tied to collective well‐being. This interpretation is consistent with cross‐cultural research, which demonstrates that individuals from strongly collectivist backgrounds exhibit more robust positive emotional responses to prosocial behavior (Peng et al. 2024). Future research should further examine the mechanisms underlying these cultural differences, which can provide deeper insights into how cultural values and norms amplify or attenuate the mental health benefits of prosocial behavior.

5.3. The Roles of Gender and Age

Contrary to the initial hypotheses, the results demonstrated that neither gender nor age moderated the association between prosocial behavior and mental health. This finding suggests that the positive association between prosocial behavior and mental health appeared to be broadly similar across male and female participants and across different age groups. One possible explanation is that prosocial behavior satisfies fundamental psychological needs that are shared across demographic groups. Engaging in helping behaviors may strengthen social connectedness, generate positive emotions, and enhance a sense of purpose. These psychological processes may operate in similar ways across gender and age groups and may therefore contribute to comparable mental health benefits. Supporting this notion, Durrani (2023) find comparable correlation magnitudes between prosocial behavior and life satisfaction across genders.

However, while gender and age did not moderate the overall relation, they may still exert influence under specific conditions. The demographic composition of the included studies may have constrained our ability to detect moderation effects. The age composition of the included samples was concentrated among children, adolescents, and young adults, with relatively few studies involving middle‐aged or older adults. Moreover, many of the included studies also showed relatively balanced gender compositions, resulting in limited variability across demographic groups. This uneven age distribution and restricted variability in demographic characteristics may have limited the statistical power to detect potential moderation effects of age and gender. Indeed, prior research has reported mixed findings regarding the moderating roles of demographic factors in the association between prosocial behavior and mental health (e.g., Memmott‐Elison et al. 2020), suggesting that these effects may vary across contexts, developmental stages, or outcome domains. For example, adolescents may derive mental health benefits largely through enhanced peer acceptance and identity formation (Oberle et al. 2022), whereas adults may experience gains more via meaning‐making and social connectedness (Fritz et al. 2023). It is also important to note that the present analyses did not examine potential nonlinear age effects (e.g., quadratic effects). Moreover, adolescence and early adulthood involve distinctive developmental transitions that may introduce additional confounding influences. Future research should examine not only the magnitude but also the mechanisms of these associations across developmental periods, including potential nonlinear patterns that may emerge during transitional life stages.

5.4. The Role of Mental Health Indicators

The association was stronger when mental health was assessed using positive indicators than when using negative indicators. A possible explanation is that positive mental health indicators are more immediately influenced by the emotional and social rewards elicited by prosocial actions (Hui 2022). Engaging in helping behaviors can enhance self‐worth, foster social connectedness, and generate positive emotions, creating a reinforcing cycle that boosts well‐being (Coulombe and Yates 2021; Ju et al. 2025; Kakulte and Shaikh 2023). Consistent with this interpretation, further analyses showed that the strongest association was observed for meaning in life, suggesting that prosocial actions may be closely linked to individuals' sense of purpose and existential fulfillment (Dakin et al. 2022). In contrast, negative indicators typically reflect deeper and more chronic psychological vulnerabilities, which are often rooted in long‐standing cognitive patterns, emotional dysregulation, and adverse life experiences (Abravanel and Sinha 2015; Feiler et al. 2023; Kuzminskaite et al. 2021). As such, they may require more intensive or sustained interventions to observe meaningful change. Accordingly, relatively weaker associations were observed for indicators such as negative affect and anxiety, which may reflect more complex or context‐dependent psychological processes.

5.5. Other Moderating Variables

In addition to the theoretically relevant moderators discussed above, two methodological characteristics also emerged as significant moderators of the association between prosocial behavior and mental health. First, the strength of the association varied across measurement instruments, with the strongest effects observed in studies using the SWLS. This may reflect the alignment between life satisfaction and the interpersonal and socially valued nature of prosocial behavior (Diener et al. 1991; Wittek and Bekkers 2015). In contrast, the weakest association was observed in studies using the SCAS, which primarily assesses anxiety symptoms. This may be due to the more complex and potentially non‐linear relationship between prosocial behavior and anxiety, as some studies suggest that higher anxiety may co‐occur with increased prosocial tendencies (Alvis et al. 2023; Nantel‐Vivier et al. 2014). These findings highlight the importance of considering how mental health is operationalized when interpreting the association between prosocial behavior and mental health.

Second, research design also moderated the association, with cross‐sectional studies reporting larger effect sizes than longitudinal studies. This pattern may partly reflect shared method variance and concurrent measurement effects in cross‐sectional designs (Podsakoff et al. 2003). Longitudinal studies, although providing stronger evidence regarding temporal ordering, may yield smaller associations due to time lags and intervening factors (Breitling et al. 2017; Shrout et al. 2011). Future research may benefit from multi‐wave or experience sampling designs to better capture both short‐term and longer‐term psychological correlates of prosocial behavior.

6. Limitations and Future Directions

There are several limitations that should be recognized in the current study. First, the causal direction between prosocial behavior and mental health remains ambiguous. Although numerous studies suggest that prosocial behavior enhances mental health, alternative research indicates that better mental health may also promote greater prosocial behavior (Isen and Levin 1972; Lyubomirsky et al. 2005). Moreover, a bidirectional or mutually reinforcing relationship has been proposed (Aknin et al. 2011; Thoits and Hewitt 2001). Future studies should focus on experimental designs to clarify these directional effects and compare effect sizes across competing causal models. Second, our analysis primarily assumed a linear association between prosocial behavior and mental health. However, emerging evidence suggests that this relation may be non‐linear, with excessive prosocial engagement potentially undermining well‐being (Bjälkebring et al. 2021; Luoh and Herzog 2002). Due to the limited number of studies reporting non‐linear effects, a synthesis was not feasible in the current analysis. Advancing statistical methodologies to meta‐analyze non‐linear relationships is an important avenue for future research. Third, cultural context was classified using country‐level individualism scores, which may overlook subcultural variations within nations, particularly in multicultural societies. Cultural values and helping norms may differ across ethnic groups within the same country, potentially influencing the prosocial behavior–mental health association. Future research could incorporate ethnicity or race to better capture within‐country cultural heterogeneity. Fourth, the demographic composition of the included studies may have constrained our ability to detect moderating effects of age and gender. Most samples were gender‐balanced and concentrated in younger age groups, with limited representation of middle‐aged or older adults. Future research should include more diverse age ranges and gender compositions.

7. Conclusions

The present meta‐analysis, using a three‐level meta‐analysis model, demonstrates a positive association between prosocial behavior and mental health. Moderating analyses indicate that culture, mental health indicators, mental health measurement tools, and research design moderate this relation. However, the association remains consistent across age, gender, publication years, publication types, and prosocial behavior measures. These findings contribute to a better understanding of the association between prosocial behavior and mental health and provide directions for future research and practice.

Author Contributions

Sen Li: conceptualization, writing – original draft, funding acquisition, supervision, writing – review and editing, resources, methodology, project administration, visualization. Qingliang Ding: software, writing – review and editing, data curation, writing – original draft, investigation, formal analysis. Yijin Lin: software, formal analysis, data curation. Denghao Zhang: supervision, validation.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.