Binding moral values gain importance in the presence of close others

The study investigates whether and how social context—specifically, being with close versus distant others—shapes the importance people ascribe to different moral values. Moral values are theorized to serve social regulation, with binding values (loyalty, authority, purity) guiding group cohesion and obligations, and individualizing values (care, fairness) promoting universal rights and harm avoidance. Because cognitive schemas are activated when contextually useful, the authors hypothesize that the presence of close others selectively increases the importance of binding values, more so than individualizing values. Understanding this context sensitivity can clarify how morality supports cooperation, why people sometimes act contrary to stated principles across situations, and how relational factors trigger different moral priorities.

The paper builds on moral foundations theory distinguishing binding and individualizing values and on relationship-regulation accounts of morality. Prior work shows contextual cues activate concepts influencing attitudes and behavior, and that social presence affects egalitarianism, punishment, and cooperation. Close others elicit stronger moral emotions and prosocial behaviors (empathy, generosity, leniency, respect) and promote in-group cooperation. Psychological distance and abstract thinking differentially influence preference for individualizing versus binding values, suggesting binding values may be more context-sensitive. Research on whistleblowing illustrates trade-offs between loyalty (binding) and fairness (individualizing). Moral identity literature shows situational activation of moral self-concepts increases prosociality and may moderate context effects. Collectively, this background suggests relational closeness should preferentially heighten binding moral concerns.

Six studies examined how social context and closeness relate to moral value importance.

• Study 1 (experience sampling): European adults using a smartphone app (N = 1166; Mage = 35.7) received random prompts over months. At each prompt, participants reported who they were with (14 categories, later grouped into alone, family, friends, professional, other), rated current importance (0–100) of five moral values (care, fairness, loyalty, authority, purity). Binding importance = mean(loyalty, authority, purity); individualizing importance = mean(care, fairness). Affective state (happiness) was collected at ~75% of timepoints. Mixed-effects models controlled for age, gender, time of day, day of week with participant random intercepts. A separate MTurk sample (N = 94) provided estimated social closeness (0–100) for each relationship category; “alone” coded 0.
• Study 2 (preregistered large online survey): MTurk via TurkPrime (N = 2016; Mage = 36.9). Participants indicated current social company (17 categories including additions: boss, separate colleague/classmate, other), then completed two moral measures: (a) “moral behavior” items (same five as Study 1) and (b) “moral value” items (rating the abstract importance of Fairness, Loyalty, Respect, Purity/Sanctity, Care/Kindness). They reported political orientation, mood, social desirability pressures, demographics, and self-rated closeness to each present relationship. A separate MTurk sample (N = 110) provided estimated closeness ratings for the 17 relationships (excluding alone/other). Multilevel regressions with participant as random factor controlled for age, gender, political orientation, mood, and non-moral evaluation.
• Study 3 (laboratory experiment): University of Pennsylvania undergraduates (final N = 390; Mage ≈ 19.8) randomly assigned to “alone” vs “partner” condition. A “Getting To Know You” task was used to foster interaction and potential closeness in the partner condition. Participants completed the MFQ (30 items) to assess five moral foundations plus non-moral controls, and reported affect, political ideology, perceived social desirability pressure/judgment, and partner closeness (1–5). Primary analyses regressed moral importance on condition and value type, controlling for demographics, affect, non-moral evaluations, and political ideology. Exploratory analyses categorized partner closeness (“distant” vs “close” ≥ somewhat close) to test interactions with value type.
• Study 4A (preregistered field survey with manipulation): MTurk via TurkPrime (final N = 2031; participants currently with others). Random assignment to write about ways they felt close to or distant from those they were with (≥50 characters, ≥30 s). Manipulation check: current closeness; mood; then rated importance (0–100) of authority, loyalty, purity (binding) and harm, fairness (individualizing). Primary preregistered regression tested effect of condition on binding importance controlling for mood.
• Study 4B (preregistered imagined context): Prolific (N = 580). Randomly assigned to imagine sitting next to a stranger (distant) or a romantic partner (close), then rated the five values (0–100). Mood and demographics collected; suspicious respondents identified, analyses robust to their exclusion. Primary tests: effect of condition on binding vs individualizing importance and their interaction.
• Study 4C (preregistered imagined closeness generalized): Prolific (N = 752). Randomly assigned to imagine sitting next to someone they feel close to vs don’t feel close to; then rated the five values (0–100). Mood unaffected by condition. Primary tests mirrored Study 4B to generalize beyond specific targets. Across studies, analyses emphasized interactions between value type (binding vs individualizing) and social context/closeness, with extensive controls and mixed-effects or ANOVA models as appropriate. Code and materials were preregistered where noted and analyses conducted in R with hierarchical models (lmerTest).

• Study 1 (experience sampling; 13,123 evaluations at 1951 timepoints; 58.7% in presence of others): Relationship type predicted overall moral importance (F(4,1608)=5.74, p<0.001), moderated by value type (F(4,2891)=2.68, p=0.030). Binding values varied more by social context than individualizing (F(4,3553)=10.56, p<0.001 vs F(4,3016)=5.93, p<0.001). Social closeness (from separate sample, N=94) positively predicted moral importance (B=0.03, SE=0.01, t(1750)=4.38, p<0.001); across relationship types, estimated moral importance correlated with average closeness r(14)=0.77, p=0.001. The closeness–importance link was stronger for binding (B=0.05, SE=0.01, t(3534)=4.59, p<0.001) than individualizing (B=0.02, SE=0.01, t(3531)=2.00, p=0.045); interaction B=0.027, SE=0.014, t(3561)=1.98, p=0.048.
• Study 2 (preregistered, N=2016): Value type × estimated closeness interactions significant for moral behavior (B≈−0.12, SE=0.01, t≈−16.10, p<0.001) and moral values (B≈−0.08, SE=0.01, t≈−12.56, p<0.001). Binding importance increased with closeness; individualizing showed negative associations. Using self-rated closeness, robust interactions held for moral behavior (β=−0.11, SE=0.01, t(3530)=−16.58, p<0.001) and moral values (β=−0.09, SE=0.01, t(3512)=−14.72, p<0.001); binding behaviors (β=0.07, SE=0.01, t=8.05, p<0.001) and values (β=0.06, SE=0.01, t=6.94, p<0.001) rose with closeness, while individualizing behaviors (β=−0.05, SE=0.01, t=−5.36, p<0.001) and values (β=−0.03, SE=0.01, t=−3.66, p<0.001) declined. Relationship type × value type interactions significant in two-way ANOVAs for both moral behavior (F(4,3510)=94.4, p<0.001) and moral values (F(4,3499)=57.43, p<0.001). Main effects collapsing across value types were mixed or null, underscoring the interaction.
• Study 3 (lab, N=390): No overall effect of partner vs alone on moral importance (B=0.02, SE=0.06, t(689)=0.38, p=0.701) nor on binding vs individualizing (B=−0.09, SE=0.08, t(346)=−1.12, p=0.262); no simple effect on binding (B=0.07, SE=0.06, t(683)=1.18, p=0.238). Manipulation yielded low closeness: only 32% felt at least somewhat close. Exploratory split (alone vs distant partner vs close partner) showed value type × context interaction (F(2,345)=4.02, p=0.019). Binding varied by context (F(2,340)=3.69, p=0.026): higher in close partner (M=5.03) vs alone (M=4.86), planned comparison t(107)=2.09, p=0.039; individualizing did not vary (F(2,340)=1.63, p=0.197).
• Study 4A (preregistered in-situ closeness writing, N≈2031): Manipulation check showed close > distant in felt closeness (e.g., M=82.2 vs 72.6 on 0–100), but preregistered effect of condition on binding importance was not significant (B=−1.15, SE=0.84, t≈−1.37, p=0.169). Exploratory analyses found value type × closeness interaction (B=−0.08, SE=0.02, t(1930)=−5.00, p<0.001), with stronger positive association for binding (B=0.08, SE=0.02, t=4.94, p<0.001) than individualizing (B=0.04, SE=0.01, t=3.48, p=0.001). Mood likely confounded the manipulation (close vs distant writing affected mood strongly).
• Study 4B (preregistered imagined stranger vs romantic partner, Prolific N=580): Condition increased binding importance (B=12.05, SE=1.76, t(540)=6.86, p<0.001) and also increased individualizing (B=3.75, SE=1.58, t(542)=2.36, p=0.018); effect was significantly stronger for binding (interaction B=−8.30, SE=1.62, t(540)=−5.13, p<0.001).
• Study 4C (preregistered imagined distant vs close other, Prolific N=752): Condition increased binding importance (B=11.08, SE=1.47, t(727)=7.55, p<0.001) and individualizing (B=5.15, SE=1.30, t(730)=3.96, p<0.001), with a stronger effect for binding (interaction B=−5.82, SE=1.33, t(730)=−4.36, p<0.001). Mood unaffected by condition (t(734)=1.26, p=0.20). Overall, across methods, being with or imagining close others reliably increased the importance of binding moral values, with weaker and inconsistent patterns for individualizing values.

Findings indicate that moral value priorities are context-sensitive: the presence (real or imagined) of close others increases the salience and importance of binding values, whereas individualizing values show little or inconsistent association with relational context. This supports the view that moral cognition functions to regulate social relationships and that different values serve different roles—binding values addressing obligations within relationships and groups, individualizing values applying more universally. The results help explain trade-offs in practical contexts (e.g., whistleblowing loyalty vs fairness) and align with research on psychological distance and moral identity. The selective activation of binding values with close others may guide adherence to relationship-specific duties and monitoring of others’ conduct. Variability across studies for individualizing values (including negative associations with closeness in Study 2) may reflect sample differences, measurement formats, or that true associations are near zero. The lab null in Study 3 and the preregistered null in Study 4A point to challenges in effectively manipulating closeness without mood confounds and suggest that imagined standardized contexts (Studies 4B/4C) can isolate the effect more cleanly. Overall, the mere presence of close others appears sufficient to shift moral emphasis toward binding concerns.

The paper demonstrates that whom people are with shapes what they morally value in the moment: close others increase the importance of binding values, with individualizing values showing no reliable change. This advances understanding of morality as relationship-regulation and reveals a mechanism for within-person variability in moral priorities across contexts. Future research should: (1) examine other taxonomies of moral motives and more granular value–context pairings; (2) identify which specific binding values are activated by which social relationships; (3) test moderating roles of moral identity and culture/media modalities; and (4) extend to implicit measures and behavioral outcomes to assess non-self-report manifestations of these context effects.

• Some preregistered tests did not confirm the hypothesis: Study 3 showed no main effect of partner presence; Study 4A failed to produce the expected effect, likely due to mood confounds and possible compensatory activation of binding values when distance is salient.
• Manipulation fidelity issues: Study 3’s partner closeness was low (only 32% somewhat close), limiting causal inference about closeness.
• Measures largely relied on self-reported importance ratings; implicit or behavioral measures were not used, potentially susceptible to demand effects despite controls.
• Mixed patterns for individualizing values (including negative associations with closeness in Study 2) raise questions about sample or method differences (European smartphone vs US MTurk/Prolific) and generalizability across populations and modalities.
• Imagined scenarios (Studies 4B/4C) enhance internal control but may limit ecological validity relative to real interactions.
• The binding vs individualizing dichotomy may obscure value-specific nuances; supplemental analyses suggest heterogeneity across specific values.