Assessing prosocial behavior in interventions: a scoping review of existing measures and their psychometric properties

1 Introduction

The past 20 years have witnessed a burgeoning increase in research examining the factors that promote and support prosocial behavior—voluntary actions intended to benefit others (Eisenberg et al., 2013)—and associated prosocial emotions such as empathy, compassion, and kindness (Binfet et al., 2025; Greenberg and Turksma, 2015; Khan et al., 2023; Malti and Speidel, 2024; Nowakowska et al., 2024; Peplak and Malti, 2022; Ricard, 2016; Roeser et al., 2018; Schonert-Reichl et al., 2012). Scientific advances in understanding the antecedents, developmental concomitants, and consequences of prosocial behavior (Kubzansky et al., 2023) have reinforced two key insights: humans possess an intuitive inclination toward prosociality (Rand et al., 2009; Warneken and Tomasello, 2006), and prosocial behavior is “malleable” and can be cultivated through interventions (Bayne and Shillington, 2025; Frazier et al., 2025; Hirani et al., 2022b; Jazaieri, 2018; Keltner et al., 2014; Malti et al., 2016; Schonert-Reichl et al., 2012, 2015). In this regard, the growing understanding of the factors that foster prosocial behavior has been accompanied by increased attention to and investment in the design and implementation of interventions aimed at promoting prosocial behavior. This work spans multiple disciplines, including developmental, social, and organizational psychology as well as medicine, and has been conducted in both experimental and applied settings, such as schools, workplaces, and community organizations.

Alongside this movement to promote prosocial behavior, researchers have increasingly focused on developing ways to define, operationalize, and measure prosocial behaviors across diverse contexts and developmental stages (Mascaro et al., 2020). Yet, there remains a relative paucity of research that systematically identifies, analyzes, and synthesizes the measures of prosocial behavior used in intervention research across disciplines. To advance both the science and practice of assessing prosocial behavior in intervention research, it is essential to distill existing evidence on how prosocial behavior has been defined and measured across a wide range of studies. Such synthesis is essential for advancing the science of prosocial behavior and informing the continued development of theory, measurement, and intervention research in the field. Accordingly, the purpose of this study was to conduct a scoping review to identify and characterize measures of prosocial behavior used in intervention research, evaluate their psychometric strengths and limitations, identify gaps in the existing evidence base, and provide guidance for future measurement development and application in intervention research.

1.1 Prosocial behavior

Prosocial behavior is foundational to the development and maintenance of social relationships and a fundamental component of healthy societal functioning (Batson, 2011). Prosocial behavior is an umbrella term referring to voluntary behaviors intended to benefit another, such as helping, donating, cooperating, sharing, and comforting (Eisenberg et al., 2013). Although these actions are often grouped together under the broad umbrella of prosocial behavior, they may differ markedly in their social context, developmental meaning, and underlying motivations. For example, sharing resources with a friend, comforting someone in distress, volunteering, and donating money to a charitable cause may all be forms of prosocial behavior, but they do not necessarily reflect the same skills, motives, or situational demands. This conceptual breadth has important implications for measurement, as different instruments may assess broad prosocial tendencies, specific behaviors, or prosocial behavior within specific contexts. Moreover, prosocial actions are widely valued across cultures because they promote cooperation, social cohesion, and harmonious interpersonal relationships, which are essential for group functioning and collective wellbeing (Batson, 2011). Given the centrality of helping, collaboration, generosity, and care for others to interpersonal and intergroup relations, improving the measurement of prosocial behavior is crucial to strengthening intervention research in social psychology and adjacent fields.

The definition of prosocial behavior does not assume a specific underlying motivation. Accordingly, such behaviors may be enacted for a variety of reasons, ranging from desires for social approval or tangible, self-interested rewards (e.g., a child sharing with another child to obtain a reward) to altruistic, other-oriented motives grounded in concern for others or internalized values and goals. The latter motives are often accompanied with other-oriented emotions and are not contingent on expectations of external or social rewards (Eisenberg et al., 2013).

Prosocial behavior is associated with a range of positive outcomes. Research has demonstrated that prosocial behaviors predict academic success (Gerbino et al., 2018; Wentzel, 1993), self-esteem and self-efficacy (Piferi and Lawler, 2006; Zuffianò et al., 2014), individual and social wellbeing (Luberto et al., 2018), and mental health (Hirani et al., 2022b; Miles et al., 2022). Additionally, in a recent meta-analysis of 72 studies, (Zheng and Zhang 2026) found that prosocial behavior was positively associated with peer acceptance and negatively associated with peer rejection in children and adolescents. Moral emotions also play a central role in prosocial behavior. Decades of research demonstrate a significant and positive relationship between prosocial behavior and empathy across childhood, adolescence, and adulthood, although the strength of this relation appears to vary as a function of the measure of prosocial behavior utilized (Eisenberg et al., 2015, 2016; Karnaze et al., 2022; Peng et al., 2024; Yin and Wang, 2023). Taken together, the current evidence base suggests that prosocial behavior is associated with cognitive, emotional, social, and health benefits (Byrne et al., 2023; Hui et al., 2020).

1.2 Global growth of interventions to promote prosocial behaviors

For much of the 20th century, research and intervention efforts in psychology and education were largely directed toward reducing deficits and treating aggression, conduct disorders, and mental illness, with the primary goal of preventing, remediating, or mitigating mental health and behavioral problems (Keyes, 2007). Over the past two decades, however, the field has experienced a paradigm shift, moving away from a preoccupation with decreasing weaknesses toward the intentional cultivation of positive human qualities, such as happiness (Diener and Seligman, 2002; Lyubomirsky, 2009, 2014). This emerging field of positive psychology emphasizes the scientific study of positive features of human development, including empathy, compassion, and prosocial behavior, alongside other human strengths (Seligman and Csikszentmihalyi, 2000). Collectively, this growing body of research reflects an increasing focus not only on reducing negative behaviors but also on intentionally promoting positive social functioning and human flourishing across multiple settings (Nelson et al., 2016).

As research in the field of positive psychology has grown, research on interventions designed to foster prosocial behavior has proliferated across age groups and cultural contexts. For instance, researchers have increasingly focused on promoting voluntary behaviors that benefit others—not only as a means to curb antisocial tendencies, but to cultivate more empathic and inclusive societies (Freitas et al., 2021). In schools worldwide, for example, social and emotional learning (SEL) programs—universal school-based interventions designed to promote students' social and emotional competencies, academic success, wellbeing, and prosocial behavior—have become increasingly prominent in the last two decades (Greenberg, 2023; Schonert-Reichl, 2023; Weissberg, 2019). Meta-analyses of hundreds of school-based SEL interventions demonstrate significant improvements in students' social and emotional skills, academic performance, positive social behavior (e.g., helping, sharing, cooperation), and concurrent reductions in conduct problems (Cipriano et al., 2023; Durlak et al., 2011). Beyond classrooms, community-based youth development initiatives and digital platforms have emerged as new avenues for cultivating compassion and cooperation. For instance, prosocial digital games and online training programs have started to target youth altruism and empathy—a clear shifting trend from earlier concerns about violent media toward harnessing technology for positive social outcomes (Saleme et al., 2020). Given this surge in interventions, understanding how prosocial outcomes are assessed is crucial for evaluating program effectiveness and comparing across interventions.

1.3 Measuring prosocial behavior

Given the surge in intervention studies examining the promotion of prosocial behavior in the past few decades (Bayne and Shillington, 2025; Dunn et al., 2014; Frazier et al., 2025; Kilmer et al., 2025; Laguna et al., 2020; Lee et al., 2025; Maggioni et al., 2018; Mesurado et al., 2019), a critical next step is to examine how prosocial behavior is measured across these varied interventions. The expansive and varied conceptualizations of prosocial behavior have made it a challenging construct to operationalize (El Mallah, 2020). In a systematic review of prosocial digital games, for example, Saleme et al. (2020) noted a lack of consistent outcome frameworks and widely divergent measures of prosocial behavior across studies, reflecting a lack of consensus in the field. This inconsistency reflects a broader challenge in psychological research: scholars often rely on different instruments to assess similar constructs.

Although several recent systematic reviews and meta-analyses have analyzed the effectiveness of interventions designed to promote prosocial behaviors (e.g., Freitas et al., 2021; Luberto et al., 2018; Mesurado et al., 2019; Shin and Lee, 2021), to our knowledge, only one other systematic review has examined measures of prosocial behaviors. Martí-Vilar et al. (2019) conducted a systematic review of measures of prosocial behavior in articles published from 1900 to 2017. Although their study was an important contribution to understanding how prosocial behaviors have been measured, there were several limitations. For example, only two databases were examined, and they utilized a relatively narrow definition of prosocial behavior. In addition, there are no previous scoping reviews that examine and evaluate the ways in which prosocial behaviors are measured and assessed in experimental studies of interventions specifically.

Measurement lies at the very core of scientific progress in psychology, underpinning theory development, empirical testing, and the rigorous evaluation of intervention effectiveness (Flake and Fried, 2020). Measurement in psychology is often hindered by what Elson et al. (2023) term the “toothbrush problem,” in which researchers prefer to create their own measures rather than adopt existing ones. To investigate this phenomenon, Elson and colleagues reviewed over 250,000 studies indexed in APA PsycINFO from 1993 to 2022. They found that in the majority of studies, measures were created and used only once (43% of the time) or twice (13% of the time), and were rarely updated, refined, or critically evaluated over time. According to the authors, this “proliferation and variability threaten the credibility of research” and constitute “a serious barrier to cumulative science” (Elson et al., 2023, p. 1).

These findings underscore the need for systematic evaluation of psychological measures, particularly in domains such as prosocial behavior, where heterogeneity across measures can impede the comparability, reliability, and interpretability of intervention research. That is, in the case of intervention studies, without systematic examination of the measures used, it is difficult to (1) determine which measures are most appropriate for assessing specific prosocial behaviors, (2) understand their strengths and limitations across different age groups and contexts, or (3) compare results across studies. Moreover, Flake and Fried (2020) argue that even rigorously designed studies may fail to yield meaningful conclusions when the constructs of interest are not clearly defined and validly measured. In other words, when constructs are not clearly defined and explicitly operationalized in intervention research, the evaluation of potential threats to internal, external, statistical conclusion, and construct validity is obfuscated.

Additional concerns arise from replication challenges in psychological research (Open Science Collaboration, 2015). When interventions yield inconsistent effects across studies, it is often unclear whether discrepancies reflect genuine differences in intervention efficacy or are artifacts of measurement error. The use of different prosocial behavior measures across studies complicates synthesis and limits the generalizability of findings.

1.4 Current study

A scoping review of prosocial behavior measures used in intervention research offers a systematic approach to addressing these challenges. By providing a comprehensive overview of existing tools—including their conceptual focus, utility, and psychometric properties—such a review can clarify which measures are reliable, valid, and appropriate across different contexts and provide valuable information to researchers for deciding which measures to use. This, in turn, supports the selection of robust measures that are developmentally appropriate and psychometrically strong, enhances the comparability of intervention findings, and strengthens the interpretability of research outcomes. Importantly, given ongoing concerns about replicability in psychological science (Open Science Collaboration, 2015), a systematic synthesis of prosocial behavior measures used in intervention research is a critical step toward improving methodological rigor and ensuring that conclusions about intervention efficacy are valid and meaningful (Flake and Fried, 2020).

This scoping review expands on the current knowledge by (1) searching the literature from 2008 to 2023, (2) utilizing multiple interdisciplinary databases, (3) using a more expansive definition of prosocial behaviors, and (4) identifying and analyzing the psychometric properties (e.g., validity, reliability) and whether studies reported statistically significant intervention-related change on the prosocial behavior measure(s). Thus, the purpose of this review was not to evaluate intervention effectiveness or determine the strength of evidence for specific interventions, but to instead map how prosocial behavior has been conceptualized, measured, and psychometrically reported across intervention studies. Specifically, the review addresses four core research questions:

• What is the extent, volume, and distribution of intervention studies published between 2008 and 2023 that included measures of prosocial behavior?
• What are the key characteristics of these intervention studies, including publication trends over time, geographic location, and features of the intervention research?
• What measures of prosocial behavior have been used in intervention research, and how frequently have these measures been used?
• What psychometric properties (e.g., reliability and validity) have been reported for prosocial behavior measures used in intervention research? Which types of measures are associated with significant changes in prosocial behavior?

2 Methods

We elected to conduct a scoping review, as this approach is particularly well-suited to mapping broad and heterogeneous bodies of literature, especially in areas characterized by substantial variation in conceptualizations, populations, and methodological approaches (Munn et al., 2018; Vieten et al., 2024). Given the diversity in intervention contexts, target populations, and assessment strategies used to measure prosocial behavior, a scoping review enables a comprehensive synthesis of the literature that would not be appropriate for a more narrowly focused systematic review.

2.1 Protocol and pre-registration

This scoping review was pre-registered with the Open Science Framework.¹ Our protocol adheres to the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) guidelines (PRISMA-P. Group et al., 2015) and the PRISMA Extension for Scoping Reviews (PRISMA-ScR) (Tricco et al., 2018). Consistent with scoping review methodology and PRISMA-ScR guidance, we did not conduct a formal risk-of-bias or quality appraisal of individual studies, as our primary aim was to map measures of prosocial behaviors used in intervention research rather than evaluate the strength of different interventions or estimate pooled intervention effects.

2.2 Eligibility criteria

This scoping review included peer-reviewed journal articles published between 2008 and 2023 that examined interventions targeting children, adolescents, or adults, and that incorporated at least one measure of prosocial behavior. Eligible studies employed randomized controlled trials (RCTs) or quasi-experimental designs and included a comparison group (i.e., waitlist control, placebo, business-as-usual, or another treatment group). We excluded studies that focused on infants or children under the age of four, lacked a control or comparison group, or used within-subjects single-group designs. Additionally, qualitative, correlational, observational, literature reviews, theoretical papers, protocols, and non-peer-reviewed sources (e.g., theses, dissertations) were excluded. Only articles available in English were included. See Table 1 for a detailed overview of our inclusion and exclusion criteria.

Table 1. Summary of inclusion and exclusion criteria.

^a“Related term” refers to the following: caring, sharing, helping, cooperating, altruism, acts of kindness, and comforting.

2.3 Search process

To develop a comprehensive search strategy, the research team identified ten exemplary articles to create an initial list of relevant search terms. The search strings were based on relevant eligibility criteria and included the following key concepts: (1) prosocial behavior, (2) intervention, and (3) evaluation. A university librarian was involved in developing the search strings and peer-reviewed the searches before they were conducted. We conducted searches in three databases: ERIC, PubMed, and PsycInfo, which were selected as they support the interdisciplinary aims of the scoping review. The general search strategy is presented in Table 2, and the specific searches for each database are provided in Supplementary Table 1.

Table 2. General search strategy.

All searches were conducted on December 6, 2023, and the following limits were applied: published between January 2008 to December 2023, English language, peer-reviewed journal articles. The search strategy was adapted for each database (see Supplementary Table 1 for the full search string for each database).

We conducted supplementary searches to identify additional eligible journal articles not captured in our database searches by utilizing a two-pronged approach adapted from recommendations by Godin et al. (2015). First, a custom search was conducted in Google Scholar, and the first 100 results were reviewed for relevant articles. Additionally, a snowballing procedure (i.e., hand searching) was used to screen the reference lists of relevant publications, including the ten exemplary articles and seven other relevant reviews that were identified (i.e., Freitas et al., 2021; Hirani et al., 2022a; Laguna et al., 2020; Luberto et al., 2018; Martí-Vilar et al., 2019; Mesurado et al., 2019; Shin and Lee, 2021). This included (1) backward snowballing, in which one team member scanned the reference lists of relevant articles, and (2) forward snowballing, where a team member used the “cited by” feature in Google Scholar to identify more recent relevant articles.

2.4 Selection of studies

All identified records were imported directly into the Covidence software for deduplication and screening (Kellermeyer et al., 2018). Each study was independently screened by two team members. In the first phase, team members screened the titles and abstracts to eliminate obviously irrelevant articles using a screening manual. In the second phase, the full texts were screened with an expanded screening tool co-developed by the research team. Before both phases, two pilot exercises with 25 randomly selected articles were conducted with all team members of the screening team, following the recommendations from Polanin et al. (2019), which served as a calibration exercise to refine the screening tool and ensure consistency among screeners. Any disagreements regarding the inclusion or exclusion of full-text records were resolved through full team discussion.

2.5 Data charting

One team member extracted data for each article using predetermined criteria via a data extraction form, following recommendations from Peters et al. (2020). To ensure consistency across reviewers, we conducted a pilot exercise with three of the included studies that were selected randomly. Discrepancies regarding data extraction were resolved through discussion with other screening team members. All extracted data were then independently checked for accuracy and completeness by a second reviewer in Covidence. Data were extracted using the standardized form that included: (1) general study details (e.g., study location, journal name), (2) method (e.g., study design, participant information), (3) intervention characteristics (e.g., intervention name, type of intervention), and (4) prosocial behavior measure characteristics (e.g., number of items, psychometric properties). See Supplementary Table 2 for the full extraction form.

2.6 Data synthesis

A narrative synthesis was utilized to qualitatively synthesize the data (Peters et al., 2020; Popay et al., 2006). Specifically, the findings were synthesized through rich textual descriptions, tables, and figures that summarized relevant data. Descriptive statistics were also calculated to represent frequencies and other relevant descriptive data (Peters et al., 2020).

2.7 Researcher positionalities

The authors' identities and professional experiences inform how prosocial behavior is conceptualized and interpreted throughout this work. Four authors identify as U.S. White Americans, two as people of Indian heritage, and one as a Black American. In terms of gender identity, six authors identify as cisgender women and one as a cisgender man. Four authors have experience as classroom teachers across developmental levels from elementary through secondary school, and all team members have been involved in research projects that involve the promotion and assessment of prosocial behavior.

3 Results

3.1 Search results

In total, 4,446 articles were retrieved from the electronic database searches and supplemental searches. After removing duplicates and screening articles based on pre-determined criteria, a total of 531 peer-reviewed journal articles were included in our final review (see Figure 1 for the PRISMA diagram detailing each step of the process). Of the 531 publications, we extracted a total of 622 unique studies, as some publications reported multiple studies that differed in their interventions, samples, or measures. All included studies had to meet the established criteria and had to have at least one quantitative measure of prosocial behavior.

3.2 Characteristics of included studies

The 531 articles included in our review were published in over 300 academic journals. Of the included articles, 29 (5%) were from Frontiers in Psychology, and 28 (5%) were published in PLoS ONE. The other academic journals with the most articles were Computers in Human Behavior, Journal of Economic Behavior & Organization, and Mindfulness, each featuring 10 included articles. The other 444 (84%) articles came from 301 academic journals—all with fewer than 10 included articles for this review (see Supplementary Table 3 for a full list of journals).

Of the 622 studies that met our inclusion criteria and reported their location, we identified 63 unique countries where the research was conducted, with nine (1%) studies taking place in more than one country. The five most frequent countries where studies were conducted were the United States (35%), United Kingdom (9%), China (7%), Germany (5%), and the Netherlands (5%), together accounting for approximately 61% of the studies that reported their location (see Supplementary Table 4 for a full list and Supplementary Figure 1 for a visualization of included countries). In terms of global region, included studies were conducted in North America (39%), Western Europe (33%), East Asia (10%), Middle East and North Africa (5%), Latin America and Caribbean (3%), Australia and New Zealand (3%), Central and Eastern Europe (3%), Sub-Saharan Africa (2%), Southeast Asia (2%), South Asia (1%), and unknown countries or world region (< 1%).

Finally, regarding the publication period of the 531 articles included in this scoping review, we observed a steady upward trend, with the number of publications rising from 11 in 2008 to 54 in 2023—a nearly five-fold increase over the 15-year period (see Figure 2).

3.3 Characteristics of included interventions

Considerable heterogeneity was observed in the intervention sites of the 622 included studies, with many interventions implemented across multiple locations. Schools (33%) were the most common intervention site, followed by university labs (22%), community-based settings (9%), online survey platforms (e.g., MTurk) (8%), online (e.g., website) (6%), and medical/clinical settings (6%). The remaining 14% of intervention sites varied widely (e.g., university classrooms, workplaces, prisons). Two studies involved policy interventions that were not location-specific. Moreover, 6% of studies did not report the intervention site. See Supplementary Table 5 for a full list of intervention sites.

Interventions also varied in terms of type and format, with 13% being categorized into multiple types—for example, a school-based SEL program containing mindfulness practices or a naturalistic field experiment evaluating a community-based intervention. The most common type of intervention was school-based programs or practices (24%), which excluded experimental tasks conducted in schools. School programs/practices were further categorized as either SEL programs/practices or non-SEL programs/practices, given that many SEL programs aim to promote positive and prosocial behavior and to measure these outcomes (Durlak et al., 2011; Greenberg, 2023). The majority of the programs or practices that were interventions in schools were SEL programs or practices (21%). These SEL programs included more well-known, established programs such as MindUP (Crooks et al., 2020; Matsuba et al., 2021; Schonert-Reichl et al., 2015), PATHS (Calhoun et al., 2020; Hughes and Cline, 2015), Second Step (Espelage et al., 2016; Merrin and Low, 2023), Promoting Prosocial and Emotional Skills to Counteract Externalizing Problems in Adolescence (Italian acronym CEPIDEA) (Caprara et al., 2015; Cirimele et al., 2022), Roots of Empathy (Santos and Langill, 2020; Schonert-Reichl et al., 2012), the Hero Program (Mesurado et al., 2020, 2021), School-Wide Positive Behavioral Interventions and Supports (SWPBIS) (Bradshaw et al., 2015; Caldarella et al., 2009), and the Papilio-3 to 6 Program (Crayen et al., 2011; Zarra-Nezhad et al., 2024). The other school-based programs/practices that were non-SEL (4%) included interventions such as changes to teaching practices like a team charter (Cook et al., 2019), cooperative learning contexts (García-González et al., 2023) or programs such as Mission I'm Possible (Chong et al., 2014) and Journey of Hope (Powell and Thompson, 2016). These programs were not categorized as SEL programs as they did not clearly fall into the social and emotional competencies as defined by the Collaborative for Academic, Social, and Emotional Learning (CASEL) (n.d.).

The second most common interventions were lab-based experimental tasks (22%), such as memory recall to induce specific emotional reactions, cognitively demanding tasks such as the Stroop task, or variations in the framing of scenarios. Most of these lab-based experimental tasks were conducted in university laboratories with university students as participants. The other types of interventions we observed in studies included online-based experimental tasks (8%), therapeutic/psychosocial interventions (8%), mindfulness programs (8%), parenting or family interventions (6%), naturalistic field experiment or tasks (5%), and public goods game manipulations (5%). The other 30% of intervention types varied greatly (e.g., video games, medical treatment, teacher training). See Supplementary Table 6 for a full list of intervention types.

Regarding the target population, studies could include more than one type. Among the 622 studies, the majority involved students as participants, with 33% in pre-kindergarten to 12th grade and 30% at the university level. The next largest target populations were children and adolescents outside of school (13%), parents and caregivers (7%), community members (6%), and online participants, such as MTurk participants (6%). The remaining 13% of reported target populations varied widely across studies but were all adult-focused. 23 of the 622 studies (12%) had target populations that were categorized into one of the aforementioned categories but also included populations that faced systemic barriers, such as clinical populations, individuals with disabilities, those affected by war or trauma, youth in foster care, and people involved with the justice system. Five studies (< 1%) did not report the characteristics of their participant populations. See Supplementary Table 7 for a full breakdown of target populations, including those facing systemic barriers.

3.4 Measures of prosocial behavior: conceptualizations, types, psychometric properties, and intervention-associated change

We identified a total of 776 measures of prosocial behavior across the 531 publications and 622 studies included in our review, as some studies reported using multiple measures. Measures captured various aspects of prosocial behavior, ranging from a general, broad operationalization of the construct to more specific sub-types, such as helping behaviors or collaboration. Measures also spanned multiple types: self-report questionnaires, informant-reports (i.e., teachers and caregivers, peers, and other reports), economic allocation games (e.g., dictator game), structured and naturalistic observations, experimental tasks, and donation tasks. The frequency for each measure type, description, and examples is reported in Table 3. For self-report, informant-reports, and economic allocation games, we report the top measures, as these types of measures tended to be more standardized. In line with similar scoping reviews with a focus on measurement (e.g., Too et al., 2022), we report the most frequently used measures for each type, although we note that a measure's frequency of use should not be interpreted as an endorsement of its suitability for all interventions. For the purposes of this review, we also summarized the reporting of psychometric properties across the different types of measures and considered reliability and validity to be “reported” if authors provided evidence for their sample or referenced prior studies. Additionally, as an exploratory descriptive indicator of intervention-associated change, we coded whether each study reported a statistically significant improvement in prosocial behavior on the target measure. We emphasize that statistical significance depends on many factors, including study design, statistical power, publication bias, and intervention characteristics, and should not be interpreted as a direct psychometric estimate of responsiveness.

Table 3. Types of prosocial behavior measures.

CR, Caregiver Report; PR, Peer-report; SR, Self-report; TR, Teacher-report; there were a total of 776 measures across 622 studies.

3.5 Conceptualizations of prosocial behavior

The ways prosocial behavior was operationalized varied considerably across the 776 measures we extracted (see Table 4). For a plurality of measures (30%), researchers did not specify any particular domains of prosocial behavior, using general, nonspecific terms such as “prosocial behavior” or “prosocial tendencies” to describe their measures. Other researchers mentioned using measures that assessed specific subtypes of prosocial behavior, with the most common being helping behaviors or volunteering (17%). Other operationalizations of prosocial behaviors mentioned by researchers included cooperation, collaboration, or teamwork (14%); generosity, charitable giving, or donations (12%); caring, empathic, or compassionate behaviors (8%); altruistic behaviors (7%); sharing behaviors (6%); and trust or reciprocity (2%). A small subset of measures (5%) used by researchers included conceptualizations of prosocial behavior that did not fit into these categories, such as moral courage or fairness. See Supplementary Table 8 for additional information on how operationalization categories were coded.

Table 4. Conceptualizations of prosocial behavior.

The total number of conceptualizations across all 776 measures was 901.

^a Other conceptualizations included fairness, behavioral compliance, social reciprocity, moral courage, peer inclusion, etc.

3.6 Informant-report measures of prosocial behaviors: psychometric properties and intervention-associated change

Of the 776 measures extracted, informant-reports were the most common type (n = 196; 25%). Informant-report measures, which assess prosocial behavior as rated by someone other than the participant, were categorized as (1) teacher/caregiver reports, (2) peer reports, or (3) other reports. Of these, teacher and caregiver ratings were most frequent (n = 176) and included measures of adults' perceptions of students' and children's behavior. The Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) was the most widely used informant-report measure that relied on teachers' or caregivers' perceptions of prosocial behavior (n = 75). This informant-reported SDQ then had a relatively equal split between teacher reports (n = 38) and caregiver reports (n = 35). Other frequently used teacher/caregiver informant-report measures included various editions of the Social Skills Improvement System (SSIS) Rating Scales (n = 12) (Gresham and Elliot, 1990, 2008), the Teacher Observation of Classroom Adaptation—Checklist (n = 12) (Koth et al., 2009; Werthamer-Larsson et al., 1991), the Social Competence Scale (n = 9) (Conduct Problems Prevention Research Group, 1990) and the Children's/Preschool Social Behavior Scale (n = 5) (Crick, 1996; Crick et al., 1997). See Table 5 for a summary of the top teacher and caregiver informant-report measures.

Table 5. Top informant-report measures (teacher, caregiver) and psychometric properties.

Percentages based on a total count of 176 informant-report measures used by teachers and/or caregivers; α, Cronbach's alpha; PSB, Prosocial Behavior.

The reporting of psychometric properties across all teacher- and caregiver-report prosocial behavior measures varied, though some trends were noted. Of all teacher and caregiver informant-report measures (n = 176; 23% of the 776 included measures), reliability was reported in most cases (n = 142). Specifically, internal consistency was reported by researchers the majority of the time (n = 136), with test-retest (n = 11) and inter-rater (n = 3) reliability being reported less often. Validity evidence was reported far less consistently: construct validity was reported for 20 measures, criterion validity for 16, content validity for four, and face validity for one. For 36 of these measures, validity information was not specified, often because the authors cited another study for validation without indicating the type of validity. Overall, 122 of the 176 (69%) teacher- and caregiver-reported measures showed significant improvement in prosocial behavior.

Peer-report measures were less common (n = 16; 2% of the 776 included measures) and assessed prosocial behaviors as reported by classmates or other peers. Nine studies (e.g., Schonert-Reichl et al., 2012; Von Salisch and Voltmer, 2023) used sociometric procedures in which classmates rated the extent to which their peers exhibited certain prosocial behaviors. Other examples included Condello et al. (2021), who assessed peer ratings of prosocial behavior using the Multisource Assessment of Social Competence Scale (Junttila et al., 2006; Magotsiou et al., 2006), and Cirimele et al. (2022), who had peers rate one another using three items from the Prosociality Scale (Caprara and Pastorelli, 1993). For nine of the 16 peer-report measures, no reliability was reported. For six of the peer-based measures, internal consistency was reported; in two studies, test-retest reliability was reported, and in one study, inter-rater reliability was reported. Regarding validity, in the vast majority of peer-report measures, no validity type was reported (n = 13). Only two studies stated unspecified validity, and one study reported construct validity. Overall, eight of the 16 (50%) peer-report measures reported significant improvements in prosocial behavior.

Four of the informant-report measures (< 1% of the 776 included measures) did not include teachers, caregivers, or peers as raters. For example, Gibbons et al. (2017) had mentors rate children's prosocial behavior using the Child Behavior Checklist (Ladd and Profilet, 1996). In a workplace context, Hafenbrack et al. (2020) included a measure in which supervisors rated employees' helping behavior toward others. Of these other informant-report measures, two studies reported internal consistency, and one reported construct validity. Additionally, only two of the four measures (50%) detected a significant improvement in prosocial behavior.

3.7 Self-report questionnaires of prosocial behavior: psychometric properties and intervention-associated change

Self-report questionnaires were the second-most used type of measure that emerged in our analysis (n = 169; 22% of the 776 included measures). These measures generally require participants to reflect on their own dispositional prosocial behavior through multi-item self-report questionnaires employing Likert-type rating scales. As seen with the informant-report measures, the SDQ (Goodman, 1997) was by far the most commonly used self-report questionnaire as well (n = 37). Other frequently used instruments included the African Youth Psychosocial Assessment Instrument (n = 7) (Betancourt et al., 2009; Bolton et al., 2007), the Prosocial Behavior Scale (n = 5) (Caprara et al., 2005), and the Self-Report Altruism Scale (n = 4) (Philippe Rushton et al., 1981). The most commonly used self-report questionnaires and their psychometric properties are displayed in Table 6. Additionally, several measures were used in only three studies, including the Alabama Parenting Questionnaire (Frick, 1991; Shelton et al., 1996), the Compassion Scale (Pommier et al., 2020), the Prosocial and Antisocial Behavior in Sports Scale (PABSS) (Kavussanu and Boardley, 2009), the Prosocial Tendencies Scale (Carlo et al., 2003), the Social Skills Improvement System (Gresham and Elliot, 1990), and the Prosocial Behavior Toward Different Targets Scale (Padilla-Walker and Christensen, 2011; Peterson and Seligman, 2004). In total, 92 of the 169 self-report questionnaires measuring prosocial behavior were used two or fewer times in the extracted articles, with 82 self-report measures being used just once.

Table 6. Top self-report questionnaires and psychometric properties.

SR, Self Report; PSB, Prosocial Behavior; α, Cronbach's alpha; r, Pearson's correlation coefficient.

The reporting of psychometric properties for self-report measures had some notable patterns. In terms of reliability, internal consistency was the most frequently cited, appearing in reports for 122 of the self-report questionnaires. Test-retest reliability was less commonly reported (n = 10). Although self-report measures typically do not report inter-rater reliability, one study by O'Callaghan et al. (2014) had caregivers rate children's behaviors using an adapted version of the African Youth Psychosocial Assessment Instrument (Betancourt et al., 2009; Bolton et al., 2007), which the youth themselves also completed, and report inter-rater reliability. However, in the current study, we define inter-rater reliability as the degree of agreement between two or more coders independently rating a participant (Hallgren, 2012). Therefore, the inter-informant agreement as reported by O'Callaghan et al. (2014) was not considered to be a form of inter-rater reliability in our analyses and was not included as such in our reporting. Information on at least one type of validity was included for approximately 62 self-report measures. Of these, construct validity was most often reported (n = 31), and unspecific validity claims were made for 23 of the measures. Other types of validity (i.e., criterion, content, face) were reported in eight or fewer uses of self-report measures. Of the 169 times a self-report measure was used, 125 (74%) detected a significant improvement in prosocial behavior.

3.8 Economic allocation games assessing prosocial behavior: psychometric properties and intervention-associated change

Several studies used incentivized economic games and tasks to measure prosocial behavior (n = 136; 18% of the 776 included measures), typically utilizing costly other-regarding choices with real or hypothetical stakes. For the purpose of this review, we grouped different iterations of similar economic games under common game names despite minor rule variations. The most frequently used economic allocation game was the dictator game and its variants (n = 45). Based on early work by Kahneman et al. (1986), the dictator game paradigm involves one participant (the “dictator”) being given a sum of money and deciding how much, if any, to give to a second participant who has no power to reject the offer (Forsythe et al., 1994). Another type of economic allocation game, the public goods game (n = 20), was used to assess prosocial behavior, in which participants were asked to contribute anonymously to a shared pot that was multiplied and then divided evenly, creating a tension between cooperative contribution and free-riding (Ledyard, 1995). The prisoner's dilemma was another commonly used economic allocation game (n = 12) to assess prosocial behavior. This task gives participants the option to cooperate or defect against the other player, with the threat that mutual defection leads to worse consequences for both players (Kelley and Stahelski, 1970). A small set of studies used generic allocation tasks (n = 10) that utilized an economic decision-making task focused on the tradeoff between self-benefit and others' welfare without naming the specific economic allocation game conducted, such as children being given the opportunity to allocate their own golden coins to another child who is ill (Lulciuc et al., 2020). The other most utilized economic allocation game category identified in the extracted articles was the trust game (n = 9). This type of measure assesses trust and trustworthiness by having an investor send an amount that is to be multiplied and given to a trustee, who may return some portion (Berg et al., 1995). Overall, there were 16 economic allocation games extracted from studies that did not fit into any specific category, either because the economic allocation game was named but only appeared once, or because the task was not named or described in a manner that made it easy to discern if it was similar enough to other economic allocation games. The most frequently used economic allocation games and their psychometric properties are displayed in Table 7.

Table 7. Top economic allocation game measures and psychometric properties.

PSB, Prosocial Behavior.

Although economic games are considered well-established tools for studying social behavior (Thielmann et al., 2021), reliability is not commonly reported because these games are typically behavioral tasks that yield a single observed score per round (Safra et al., 2022). Despite this, two notable exceptions emerge in our analysis: Laguna and colleagues (2023) reported a Guttman split-half coefficient for the decomposed game, and De Dreu et al. (2015) calculated Cronbach's alpha across five rounds of an intergroup prisoner's dilemma game. Overall, validity reporting in economic allocation games was rare, with 126 games not mentioning any type of validity. Only construct (n = 3) and face validity (n = 1) were reported for economic allocation games. Additionally, some articles made unspecified validity claims (n = 4). Overall, 110 (81%) of the 136 economic allocation games reported in the extracted studies detected significant improvement in prosocial behavior.

3.9 Experimental tasks assessing prosocial behavior: psychometric properties and intervention-associated change

Experimental tasks, which encompassed a broad range of paradigms in which participants completed a prosocial behavior/act in an experimental setting, accounted for approximately 15% (n = 119) of the 776 included measures. Some studies (e.g., Abrahams et al., 2018; Gino and Desai, 2012) had participants complete extra tasks or help the researcher with an optional survey. Other examples included participants helping pick up dropped pencils (Greitemeyer, 2009), collaboration in an online game (Heasman and Gillespie, 2019), and the Tangram Help/Hurt Task, in which participants assign puzzles of varying difficulties called “tangrams” to a supposed other participant (Saleem et al., 2015). Given the diversity in experimental tasks employed, reliability information was not applicable for this type of measure and is not reported. In terms of validity, construct validity was reported for just four of the 119 experimental task measures, and unspecified claims about a measure's validity were reported for three additional measures. All other types of validity were not reported. Overall, 93 (78%) of the 119 experimental tasks that were utilized in the extracted articles detected a significant improvement in prosocial behavior.

3.10 Donation tasks assessing prosocial behavior: psychometric properties and intervention-associated change

A specific kind of measure that emerged from our review of studies was donation tasks (n = 79; 10% of the 776 included measures). Unlike more general experimental tasks, donation tasks specifically involve opportunities for participants to donate money (real or hypothetical) to a good cause. For example, many laboratory-based studies had participants choose whether to donate the money they earned for participating in the study to a predetermined or self-chosen charity (e.g., Alt and Gallier, 2022; Hahnel et al., 2015; Tear and Nielsen, 2014), and these donation decisions or amounts were used as measures of prosocial behavior. Drouvelis and Marx (2018), for instance, had participants complete a series of tasks and then gave each participant an opportunity to donate a portion of their earnings to a local charity. Similarly to experimental tasks, reliability was generally considered inapplicable to donation tasks, although reporting of validity was still expected. Across the 79 donation task measures, the reporting of psychometric properties was almost entirely absent: only one study (Laguna et al., 2023) reported internal consistency, and another (Alt and Gallier, 2022) mentioned face validity. In our review, we found that none of the studies using donation tasks clearly referenced the previous studies from which their procedures were derived. Overall, 53 (67%) of the 79 times that donation tasks were used detected a significant improvement in prosocial behavior.

3.11 Observational measures assessing prosocial behavior: psychometric properties and intervention-associated change

A smaller subset of researchers utilized observational measures of prosocial behavior (n = 77; 10% of the 776 included measures), which we sub-categorized as either structured observations (n = 66) or naturalistic observations (n = 11). Structured observations—in which trained raters (e.g., research assistants) coded video or live interactions for the quantity or quality of prosocial acts—were used in a variety of ways and occurred in both laboratory and real-world settings. For example, Li et al. (2016) recorded the frequency-specific prosocial behaviors (e.g., cooperation, sharing) of video recordings of children's interactions in a lab setting using an adapted version of the Play Observation Scale (Rubin, 1989). Anliak and Sahin (2010) had researchers directly observe children's prosocial behaviors in a real classroom using the Drexel Early Childhood Behavior Rating Scale (Shure, 2005). Of the studies using measures coded as structured observations, half (n = 33) reported at least some information on reliability; 20 included inter-rater reliability, 16 included internal consistency, and one included test-retest reliability. In terms of validity, no information was reported for 52 measures, and content validity was reported for five measures. Significant improvement in prosocial behavior was observed in 46 (70%) of the 66 instances in which structured observations were used.

Naturalistic observations—measures of prosocial behaviors observed or recorded in real-world contexts rather than by a trained rater—were less common than structured observations and included a range of measures (n = 11, 1% of 776 included measures). Blood donations were coded under this category and were reported in five studies. For example, Ou-Yang et al. (2020) measured blood donation rate—the percentage of participants returning to give blood within 12 months. Another example of naturalistic observations includes a large-scale field experiment conducted by Gallus (2017) in which they examined volunteer behavior by assessing their voluntary contributions to editing Wikipedia pages. Of the eleven naturalistic observation measures, inter-rater was the sole type reported (n = 2), and none contained information on validity. Significant improvement in prosocial behavior was detected in nine of the 11 instances in which naturalistic observation measures were used.

3.12 Overall psychometric properties of identified measures

Reporting of psychometric properties was inconsistent across the identified measures. Of the 776 measures we extracted, fewer than half (n = 356) provided any data regarding reliability or validity (see Table 8 for summary of psychometric reporting). Reliability was the most frequently reported psychometric index overall. However, because reliability reporting is often considered non-applicable or unconventional for certain paradigms (e.g., economic allocation games, experimental tasks, donation tasks), these measures were not included in the analysis of the overall psychometric properties. We emphasize, however, that although internal consistency may be less applicable for these types of measures, other forms of reliability may still be applicable. For example, researchers may report test-retest reliability when repeated trials are used (e.g., Safra et al., 2022) or inter-rater reliability when tasks are coded (e.g., Shoshani et al., 2022). However, we found that these methods were rarely used. Of the 445 measures considered to have the “opportunity” to report reliability, internal consistency (e.g., Cronbach's α) was the primary metric, reported for 283 of them. Inter-rater reliability was seldom reported (n = 26), and test-retest reliability was least common (n = 24). Notably, 30% (n = 134) of measures for which reliability was applicable failed to report any such data. Overall, 69% (n = 307) of eligible measures reported on reliability.

Table 8. Summary of psychometric properties of prosocial behavior measures.

Categories are not mutually exclusive, as a single measure may report multiple reliability/validity indices; percentages may sum to more than 100%.

^a Percentages for reliability types were calculated out of 445 measures only, as assessing reliability was not applicable for economic allocation games, experimental tasks, and donation tasks.

^b Percentage for validity was calculated out of the total of 776 measures.

^c Construct validity includes measures that report confirmatory factorial analysis and measurement invariance.

^d Criterion validity includes measures that report concurrent and predictive validity.

Compared to reliability, validity was less often reported. Of the measures that reported validity information, construct validity was the most commonly reported (n = 63). Other types of validity reported included criterion (n = 20), content (n = 17), and face (n = 11). In 74 of the 776 measures, authors made unspecified, generic claims (e.g., “previously validated”) about the measure's validity. Overall, 79% (n = 613) of the extracted measures did not report any information on validity. The dataset of the included articles, study information, interventions, and measures can be found in Supplementary Data Sheet 1.

4 Discussion

We conducted this scoping review to (1) identify extant measures of prosocial behavior used in contemporary intervention research articles published between 2008 and 2023, and (2) examine both the psychometric robustness of these measures and whether studies reported statistically significant intervention-associated change. In total, we extracted 776 measures used to assess various dimensions of prosocial behavior from 531 journal articles spanning more than 60 countries. We found a notable upward trend in the number of publications—a nearly 400% increase from 2008 to 2023—highlighting a growing scholarly interest in interventions measuring prosocial behavior.

Interventions with outcome measures of prosocial behavior varied widely in terms of intervention type, site, and target population, though several patterns emerged. When school-based SEL and non-SEL programs were combined, 24% of the included studies reported school-based interventions for pre-kindergarten through 12th-grade students. The popularity of school-based interventions designed to promote prosocial behavior reflects the understanding that prosocial involvement has a positive influence on children and adolescents as individuals (Hirani et al., 2022b; Greenberg, 2023; Miles et al., 2022; Schonert-Reichl et al., 2015), as well as in their relationships and the world around them (Leeuwen and Zagefka, 2017; Taylor, 2020). Due to the ongoing maturation of brain regions involved in social processing, childhood and adolescence in particular may be a sensitive period for developing prosocial intentions (Crone and Achterberg, 2022; Mills et al., 2014), further supporting the rationale for school programs that foster students' positive social behaviors. Lab-based experimental tasks conducted in university labs with university student participants were also frequently used interventions. The continued use of laboratory experiments, often conducted in university settings, provides valuable insights into the cognitive processes underlying prosocial behavior and the factors that incentivize these positive actions.

One notable finding from our scoping review was the considerable heterogeneity in how prosocial behavior was conceptualized across studies. The variability ranged from broad, non-specific terms to others that targeted distinct behavioral subtypes, such as helping, cooperation, or sharing. This breadth may reflect the inherently multidimensional nature of prosocial behavior; however, it also complicates synthesis and comparison across studies and interventions. In terms of conceptualizations, we found that in a substantial proportion of measures (30%), researchers used general, non-specific terms for prosocial behavior. This finding points to a lack of conceptual standardization. For the field to move forward, researchers must clearly define which aspect of prosocial behavior they are targeting (e.g., helping, sharing, cooperation) and select measures that align precisely with that definition.

In reviewing the measures used to assess prosocial behaviors in intervention research, we found that researchers employed multiple types of measures. Although informant-report measures were the most common type, we found that self-reports and economic allocation games were also frequently used across studies. Observational measures, both structured and naturalistic, were relatively rare. This pattern reflects methodological diversity in the measurement of prosocial behavior, yet it suggests that current intervention research is dominated by rater-based instruments (e.g., informant-reports, self-reports) and lab tasks, both of which are vulnerable to shared-method variance, demand characteristics, and limited ecological validity (Roe and Just, 2009). Although we found that naturalistic observations and economic allocation games more often detected significant improvements in prosocial behavior, this finding should be interpreted cautiously. These results may reflect differences in study design, sample size, intervention characteristics, and publication bias rather than the inherent sensitivity of particular measure types. Therefore, these findings should be used to describe patterns in our reviewed studies rather than to directly compare measure types in terms of sensitivity to change.

One of the clearest findings to emerge from this review is the relative pervasiveness of the Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) as a measure of prosocial behavior in intervention studies. The SDQ was not only the most frequently used instrument overall, but also the most common measure in the two largest categories: informant-reports (n = 75) and self-reports (n = 37). Originally developed by Goodman (1997), this questionnaire provides a comprehensive evaluation of emotional and behavioral functioning in young people aged 2–17 years and has been used in over 8000 publications across 117 countries (Youth in Mind, 2022). The 5-item prosocial behavior subscale of the SDQ assesses positive social behaviors that reflect concern for others and the capacity for helpful and cooperative interactions. It includes items that are rated on a 3-point scale ranging from not at all true to certainly true: “Considerate of other people's feelings,” “Shares readily with other children (treats, toys, pencils, etc.),” “Helpful if someone is hurt, upset or feeling ill,” “Kind to younger children,” and “Often volunteers to help others (parents, teachers, other children)” (Goodman, 1997). Although the SDQ was the most frequently used measure in the reviewed studies, its frequency of use should not be interpreted as evidence that it is necessarily the most robust or appropriate measure for all intervention contexts. Rather, its widespread use is likely due to several factors. First, the SDQ is available in multiple formats (teacher-, caregiver-, and self-reports), offering researchers a versatile tool for gathering data using multiple informants. Second, its brevity (i.e., 5 items per subscale) is a significant benefit, given that shorter measures are more easily implemented in implementation studies. Despite its brevity, the SDQ demonstrates relatively robust psychometric properties comparable to, and in some cases stronger than, long-form measures such as the Child Behavior Checklist (Goodman and Scott, 1999; Klasen et al., 2000; Stone et al., 2010). Furthermore, the SDQ is available free of cost for non-commercial purposes, translated into more than 80 languages (Youth in Mind, 2022), and has demonstrated cross-cultural validity, further contributing to its widespread adoption.

Another significant finding from our review was a pervasive gap in the reporting of the psychometric properties of measures of prosocial behavior used in intervention research. Consistent with broader concerns raised by other scholars (e.g., Flake and Fried, 2020; Hedge et al., 2018; Parsons et al., 2019), we found that researchers seldom reported the psychometric properties (e.g., reliability and validity) of their measures of prosocial behavior. Unsurprisingly, studies that used self-report and informant-report instruments provided more psychometric information than those that used behavioral tasks (e.g., experimental tasks, economic allocation games). Overall, reliability was reported more often than validity, with internal consistency the most frequently cited index; other types of reliability and all forms of validity were rarely addressed. Taken together, these patterns matter because they limit what we can conclude from the existing literature on prosocial behavior. When core psychometric information is missing or underreported, observed effects are harder to interpret, between-study comparisons become difficult or impossible, and meta-analytic analyses risk being biased by measurement error (Flake and Fried, 2020). In applied contexts involving interventions, underreported psychometrics make it unclear whether the findings reflect true effects or are due to imprecise measurement.

Overall, our scoping review raises important concerns for the measurement of prosocial behavior in intervention research. The lack of measurement consistency has been implicated as one of several contributing factors to the replication crisis concerns raised by many scholars (Lilienfeld and Strother, 2020; Open Science Collaboration, 2015). When operationalizations vary and construct-measure alignment is weak, estimates become more sensitive to analytic choices and less likely to reproduce across labs, samples, and contexts (Flake and Fried, 2020). Moreover, limited psychometric reporting-especially sparse validity evidence-exacerbates attenuation and interpretive ambiguity, further contributing to issues in replication (Flake and Fried, 2020; Lilienfeld and Strother, 2020).

4.1 Strengths, limitations, implications, and future directions

This scoping review has several notable strengths. First, we conducted our scoping review across a broad time span, from 2008 to 2023, enabling an expansive examination of studies spanning more than a decade. Second, we used expansive search parameters when conducting our search, thus reviewing a wide breadth of literature on how prosocial behavior measures are used in interventions. Third, the review team followed best practices in conducting scoping reviews, including pre-registration, peer review of our search strings by an experienced librarian, and utilization of state-of-the-art software (i.e., Covidence) to support and streamline our screening process. Finally, by including a broad spectrum of intervention modalities, settings, and measurement types, we were able to provide a comprehensive map of the field.

This scoping review also has several limitations. We limited the articles included in this review to only those published in English, so we cannot conclude whether these findings hold in measures in other languages. In addition, we only examined peer-reviewed journal articles, meaning measures reported in other formats, such as dissertations or research reports, were excluded, which may introduce publication bias. Studies reporting nonsignificant or null findings are less likely to be published, potentially resulting in an overrepresentation of positive results in the studies included in this review. Consequently, this reliance on published studies may have inflated apparent intervention effects, particularly if measures of prosocial behavior demonstrated limited sensitivity that went unreported. Similarly, although our coding of whether an intervention detected a statistically significant improvement in prosocial behavior is potentially useful as a descriptive indicator, it should be interpreted cautiously and not as a direct estimate of psychometric sensitivity to change. Furthermore, our scoping review focused on quantitative measures of prosocial behavior, so the ways in which prosocial behavior is captured qualitatively remain uncertain. Finally, our review is limited to the psychometric information that authors chose to report. A lack of reported reliability or validity in a paper does not necessarily mean the measure is not psychometrically sound, as the information may have been omitted due to other reasons (e.g., journal article space constraints).

There is also a limitation to non-naturalistic studies, specifically when it comes to measures of prosocial behavior, as people who are more prosocial tend to comply more often with requests to participate in psychological research (Van Lange et al., 2011), so any study that actively recruits participants may already be slightly skewed toward a more prosocial sample. A further concern lies in the potential disconnect between how prosocial behavior is measured and how it is actually manifested in daily life. Self- and informant-report questionnaires capture perceived likelihoods or frequencies of prosocial behaviors, which may not accurately reflect individuals' behaviors in real-world contexts. This disconnect is evident in social desirability bias, wherein respondents are inclined to provide responses that align with perceived social norms and expectations. This may lead to inflated reports of socially desirable behaviors, including prosocial behavior (Fisher and Katz, 2000). Because individuals are motivated to present themselves in a favorable light and avoid embarrassment, social desirability bias may undermine the validity and interpretation of self-report questionnaires which assess prosocial behavior.

Additionally, economic allocation games constitute another measurement approach that may not adequately reflect prosocial behavior as it is enacted in everyday life. Economic allocation games often occur in controlled laboratory settings that are strikingly sparse in context—stripped of the social cues, relationships, and rewards or costs that ordinarily shape prosocial decisions—making it difficult to generalize findings to everyday behavior. Indeed, there is conflicting evidence on whether outcomes from economic allocation games relate meaningfully to prosocial behavior in the real world (e.g., Franzen and Pointner, 2013; Galizzi and Navarro-Martinez, 2019). Evidence from dictator game studies, for instance, illustrates this well: participants' generosity is highly sensitive to the format of the task and subject to experimental demand characteristics, such that even minor variations in procedure can significantly alter giving behavior (Bardsley, 2008; List, 2007). Although concerns remain regarding the extent to which laboratory tasks reflect real-world behavior, such tasks allow experimenters to tightly control extraneous variables, thereby enhancing the internal validity of a study. In contrast, naturalistic experiments are subject to a myriad of uncontrolled contextual influences that are difficult to account for in both design and analyses; however, they may better capture behavior as it occurs in real-world settings. Laboratory and naturalistic experiments always involve a trade-off between external and internal validity (Roe and Just, 2009), a strength and limitation of nearly every study included. Conflicting evidence may always arise from comparisons between laboratory and real-world behaviors, and studies using both approaches are needed.

Our scoping review has several important practical implications for researchers and practitioners involved in intervention research aimed at promoting prosocial behaviors. Our results highlight the importance of researchers clearly defining the specific aspect of prosocial behavior their intervention is intended to promote prior to selecting a measure and ensuring that chosen measure directly aligns with their intervention's focus. For example, if an intervention is designed to increase broad prosocial tendencies, a global questionnaire (e.g., SDQ; Goodman, 1997) may be appropriate; however, if an intervention targets specific behaviors such as helping others, cooperating in groups, or engaging in charitable giving, measures should be selected that directly capture those prosocial behaviors. A second practical implication is that researchers should ensure transparent reporting of the measures they use, including item content, administration and scoring procedures, and available reliability and validity evidence. This information is especially important when researchers adapt existing measures or develop new measures, as incomplete reporting limits the ability to evaluate, compare, and replicate findings across intervention studies. Additionally, researchers should consider the developmental stage of participants (e.g., young children, adolescents, adults), reporter perspective, and the social context of their interventions. When possible, we suggest combining measurement approaches, such as using self-reports in addition to behavioral tasks or naturalistic observations, to provide a more comprehensive assessment of prosocial behavior.

The findings from this scoping review highlight several avenues for future work. Future research should compare different prosocial behavior instruments (e.g., self-report, informant-report, behavioral observations) within the same intervention and compare results. This would help establish convergent validity and clarify which types of measures are most sensitive to change for different populations and contexts. Our finding that psychometric data are rarely reported clearly underscores a need for more validation studies for existing measures of prosocial behavior. This is particularly true for diverse cultural and linguistic contexts to ensure measures are culturally appropriate for different populations. To reduce publication bias, future researchers are advised to incorporate diverse sources of evidence, such as dissertations, theses, and unpublished or preregistered studies, when conducting similar reviews of prosocial behavior measures. Lastly, to address issues in psychological measurement (Elson et al., 2023), a practical next step would be the creation of an open-access online repository of prosocial behavior measures. This resource could serve as a hub for psychometrically validated instruments and promote the use of similar instruments across studies.

5 Conclusion

There is a vibrant, burgeoning literature on interventions that include measures of prosocial behavior. Across 531 publications and 776 measures of prosocial behavior, our results illustrate a diverse range of measurement techniques. Yet, our findings also highlight a striking lack of consistency in measures used and a paucity of psychometric evidence supporting their reliability and validity. Ultimately, without a more unified and psychometrically robust approach to measurement, the field's capacity to build cumulative evidence and offer clear, actionable recommendations for cultivating prosocial behavior remains significantly stymied. Moving the field forward will require sharper construct specification, multi-method, multi-informant assessment strategies, and transparent reporting of psychometric properties. We contend that advancing these practices will yield more interpretable findings about when, how, and for whom interventions foster prosocial behavior and will better support educators, clinicians, and policymakers seeking to promote kindness, cooperation, and other prosocial behaviors across diverse contexts.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frsps.2026.1804116/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Alessandro Geraci, University of Palermo, Italy

Reviewed by: Lucie Rose, Université de Paris, France

Linwei Li, Universiti Teknologi Malaysia, Malaysia

Footnotes

https://osf.io/hzb4a/.