Social Preferences and Well-being: Theory and Evidence

The paper investigates whether individuals with prosocial social preferences are happier than those with proself preferences. Given that laws and policies often encourage prosocial actions and may shape preferences, understanding the welfare consequences is important for social welfare analysis. Social preferences are defined as preferences over others’ payoffs; prosociality is the extent to which one cares about others’ outcomes. Prior research has largely examined prosocial behavior (observable) and happiness, but not the link between underlying social preferences (less observable) and well-being. The study addresses two gaps: (1) the absence of a formal theoretical model explicitly linking social preferences to subjective well-being, and (2) limited empirical evidence on this relationship. It develops a simple mathematical framework in which an individual’s well-being depends on own outcomes, others’ outcomes, and a prosociality parameter, and posits the hypothesis that higher prosociality is associated with higher subjective well-being. It then tests this using U.S. survey data.

The study connects three literatures: (1) social preferences/social value orientation (SVO), (2) subjective well-being (SWB), and (3) heterogeneous preferences. SVO research, originating from decomposed games, classifies individuals as prosocial, individualistic, competitive, or altruistic, and predicts behaviors like volunteering and donating (e.g., McClintock & Allison, Van Lange et al.). While SVO predicts prosocial behaviors across contexts, its relationship with well-being remains underexplored. SWB research documents determinants such as parenthood, political preferences, income, and education; happiness comprises multiple dimensions—remembered (general, eudaimonic, hedonic, social) and experienced well-being—with complementary measurement approaches (e.g., Diener et al.; Ryff; PANAS; Keyes). The heterogeneous preferences literature shows policy effects can depend on preference distributions (e.g., Fehr & Schmidt; Ziegler). The present work extends Decancq et al. (2017) by explicitly incorporating social preferences into a formal well-being model and empirically examining how prosociality correlates with multiple SWB domains.

Theory: Extends Decancq et al. (2017) to allow individual well-being WB(L, a_i) to depend on the full outcome matrix L (own and others’ outcomes) and on a preference vector a_i that includes a prosociality parameter. Prosocial preferences are defined such that improvements in others’ outcomes increase one’s well-being, ceteris paribus. Hypothesis: higher prosociality correlates with higher subjective well-being. Measurement of social preferences: Uses Social Value Orientation (SVO) Slider Measure (Murphy et al., 2011). Respondents make six allocation choices between self and another; mean own and other payoffs are used to compute an SVO angle (continuous measure of prosociality). Traditional categories (altruistic, prosocial, individualistic, competitive) can also be derived. Measurement of well-being: Uses the Pemberton Happiness Index (Hervás & Vázquez, 2013), covering remembered well-being (11 items spanning general, eudaimonic, hedonic, and social subdomains; 0–10 scale) and experienced well-being (10 yes/no daily experiences scored 0–10). Total well-being is the mean of remembered and experienced components. Participants and procedure: U.S. adults recruited via Amazon Mechanical Turk in March 2016; N=212 (power analysis target=200 for 80% power with small–medium effect). Procedure order: SVO Slider, Pemberton Index, demographics. Average completion time ~3.5 minutes; compensation $0.50. Demographics: 45.3% women; 65.2% aged <40; 51.4% bachelor’s or higher; 79.7% employed/self-employed; 43.8% household income ≥$50k; political: 50% Democrat, 20.8% Republican; 33.5% married; 41.5% with children. Econometric strategy: OLS regressions of seven dependent variables (total, remembered, general, eudaimonic, hedonic, social, experienced well-being) on SVO. Main specification uses continuous SVO score; robustness uses categorical SVO (prosocial vs individualistic). Controls include parenthood (binary), political preference (Republican, Independent, Other; reference Democrat), household income (ordinal categories; reference <$30k), education (categories; reference high school), and additional controls (gender, age, employment, marital status). Robust standard errors reported.

Descriptives and reliability: Mean SVO=23.883 (SD=14.514), indicating average prosociality. Mean scores: total well-being 6.830; remembered 6.846; experienced 6.656. Subdomains: general 6.764; eudaimonic 7.068; hedonic 6.722; social 5.925 (lowest among subdomains). Cronbach’s alpha: remembered 0.940; total 0.942. Correlations (Pearson): SVO correlated with total (r=0.139, p<0.05), remembered (r=0.139, p<0.05), hedonic (r=0.189, p<0.01); weaker/non-significant with general and social well-being. SVO category distribution: altruistic 0.5%, prosocial 55.7%, individualistic 43.9%, competitive 0% (similar to Murphy et al. benchmarks). Regression results (continuous SVO): Positive, statistically significant associations between SVO and multiple well-being measures after controls. Unstandardized coefficients for SVO: total 0.020 (p<0.05), remembered 0.020 (p<0.05), eudaimonic 0.019 (p<0.05), hedonic 0.028 (p<0.01); general 0.016 (ns), social 0.015 (ns), experienced 0.021 (ns). Interpretation example: a 1-point increase in SVO associates with +0.028 in hedonic well-being. Standardized coefficients for SVO: total 0.159; remembered 0.157; eudaimonic 0.147; hedonic 0.187; general 0.115; social 0.096; experienced 0.123. Effect sizes for SVO on total and subdomain well-being are comparable to key determinants: parenthood (std ~0.158), highest income ≥$150k (std ~0.167), and doctoral/professional degree (std ~0.176). Other covariates: Parenthood positively associated (e.g., eudaimonic coef 0.672, p<0.1). Republicans (vs Democrats) show higher well-being (e.g., eudaimonic 0.973, p≤0.01). Higher income categories generally positive, with notable larger coefficients for $70–79k and ≥$150k. Education: doctoral/professional degree shows sizeable positive associations (e.g., total 1.500, p<0.01). Adjusted R² range: ~0.024–0.093 across models. Robustness (categorical SVO: prosocial=1 vs individualistic=0; altruist merged with prosocial): Positive, significant coefficients for total 0.579 (p<0.05), remembered 0.566 (p<0.05), eudaimonic 0.568 (p<0.05), hedonic 0.833 (p<0.01), experienced 0.725 (p<0.05); general and social ns. Standardized effects closely match continuous-specification results. Experienced well-being significance appears in categorical but not continuous models; authors caution this may be chance and warrants further research.

Findings support the central hypothesis: higher prosociality (as measured by SVO) is associated with higher subjective well-being. The relationship holds for overall well-being and key subdomains, especially hedonic (momentary affect) and eudaimonic (purpose, growth), suggesting that prosocial orientation relates not only to transient positive affect but also to more enduring aspects of functioning and self-realization. The effect size of prosociality is on par with established determinants such as parenthood, income, and education, highlighting prosociality as a substantive factor in individual well-being. From a policy perspective, since laws and interventions can influence prosocial preferences and behaviors, understanding that more prosocial individuals tend to report higher well-being suggests that preference-shaping policies need not impose a well-being cost and may align individual welfare with social goals. The non-significant results for general life satisfaction and social well-being imply that prosocial orientation may not translate into higher global life satisfaction or satisfaction with society per se, but rather into greater positive affect and eudaimonic fulfillment. The mixed evidence for experienced well-being (significant only with categorical SVO) indicates the need for cautious interpretation and additional study.

The paper contributes a formal model linking heterogeneous social preferences to individual well-being and provides empirical evidence from U.S. adults that prosociality correlates positively with total, hedonic, and eudaimonic well-being. Effect sizes are comparable to those of major well-being determinants, underscoring prosociality’s importance. The results inform policy debates about interventions that may encourage prosocial preferences, suggesting such policies can be compatible with individual welfare. Future research should: (1) clarify causal pathways (e.g., longitudinal or experimental designs), (2) examine external validity across cultures and populations beyond MTurk samples, (3) refine measurement of experienced well-being and probe why general/social well-being show weaker links, and (4) explore mechanisms (e.g., social connectedness, meaning, autonomy) mediating the prosociality–well-being relationship.

• Design identifies correlations, not causality; reverse causality and omitted variables cannot be ruled out. - Convenience sample from MTurk U.S. adults (N=212) in 2016 may limit generalizability to broader or non-U.S. populations. - SVO measures one facet of social preferences; other dimensions (e.g., social mindfulness) were not included. - Experienced well-being results are sensitive to specification (significant with categorical SVO but not continuous), indicating measurement or statistical uncertainty. - Self-reported well-being may be subject to recall or response biases; remembered and experienced well-being capture complementary but incomplete aspects of SWB. - Some categories (e.g., altruistic, competitive SVO) were rare or absent, limiting analyses across full SVO typologies.