Bridging self-control and prosocial behavior in early adolescents: a simulation-based node-perturbation analysis

Prosocial behaviors during early adolescence are closely related to self-control; however, they are not completely accounted for through variable-centered methods, as person-centered variations in these behaviors may co-exist independently. The present study used a person-centered, network-informed approach to explore self-control profiles latent within persons, the corresponding networks of prosocial behavior, along with responses to node perturbation.

A cross-sectional sample of 1,417 early adolescents (M = 13.57 years, SD = 1.10) was recruited from junior middle schools in Guangxi, China. Using self-reported multidimensional self-control and prosocial behavior, we conducted latent profile analysis (LPA), Ising network estimation, Gaussian Graphical Model estimation, and simulation-based node-perturbation analysis (Network Intervention Response Analysis; NIRA) to identify profile-specific association patterns and candidate node-level leverage points.

LPA identified four self-control subgroups: Lowest, Low-to-Moderate, Moderate-to-High, and Highest. Overall prosocial behavior tended to be higher in profiles with higher self-control, although item-level patterns were not uniformly monotonic. The Low-to-Moderate self-control subgroup showed the densest pattern of associations among prosocial nodes, a pattern that may reflect stronger behavioral coupling but should not be interpreted as necessarily adaptive. Hypothetical node perturbations suggested profile-specific sensitivities: lower self-control profiles were more responsive to emotionally reactive and request-based helping nodes (e.g., PB24: Compliant helping), whereas higher self-control profiles were more responsive to non-reciprocal and principle-oriented helping nodes (e.g., PB25: Pure altruism). These patterns indicate different association structures rather than moral superiority or confirmed intervention effects.

The present research offers person-first, network-informed evidence that self-control profiles correspond to both the extent of, and structure behind, prosocial behaviors during early adolescence. The node-perturbation results should be interpreted as hypothesis-generating evidence for potential targets requiring longitudinal and experimental validation. These findings suggest that school-based programs in the future should incorporate cautious, profile-sensitive measures instead of claims that simulating specific targets is a better method than universal target methods.