estimate peer effects

This study addresses the identification and estimation of social effects in linear-in-means models with endogenous network structures. By leveraging randomized assignment combined with observed network data, and invoking the assumption that initially randomized peers exert no direct social influence, the authors construct local heterogeneity–robust moment conditions applicable to the full sample. Building on ψ-dependent network structures, they develop a generalized method of moments (GMM) estimator that admits a closed-form solution, ensuring both theoretical rigor and computational tractability. Monte Carlo simulations demonstrate strong finite-sample performance. Empirically, the analysis reveals a significant positive spillover effect: among Hong Kong high school students, being assigned by teachers to sit next to a randomly allocated peer positively influences the mathematics achievement of that peer’s study partners.

This study addresses the challenge of identifying key individuals who exert significant peer influence on risk behaviors—such as smoking and marijuana use—from cross-sectional observational data in multilayer social networks. To tackle the endogeneity arising from homophily, the authors innovatively construct instrumental variables using observable characteristics of distal individuals within the multilayer network structure and integrate these with instrumental variable estimation and multilayer network modeling. Empirical analysis based on the Add Health dataset reveals robust positive peer effects from both friends and classmates on risk behaviors and successfully pinpoints influential individuals. The findings offer a novel methodological approach and empirical evidence for understanding behavioral contagion mechanisms in social networks.

Identifying heterogeneous peer effects in count-response data remains challenging due to unobserved heterogeneity and network endogeneity. Method: This paper develops a rational-expectations game-theoretic model under incomplete information to characterize individual decision-making under heterogeneous peer influence in social networks. It extends linear identification conditions to generalized nonlinear count models for the first time; proposes a novel network-exogeneity identification strategy leveraging “friends-of-friends” to jointly identify network endogeneity and heterogeneous peer effects; and introduces a nested pseudo-likelihood (NPL) estimator, implemented in the open-source R package CDatanet for empirical replication. Contribution/Results: Empirical analysis reveals significant same-gender peer effects on girls’ extracurricular participation, whereas boys’ participation is unaffected by same-gender peers—highlighting gender-differentiated social influence mechanisms in adolescent behavior.

This paper addresses causal inference in network experiments subject to interference. We propose a purely design-based, model-agnostic weighted least squares framework. Methodologically, we first establish the equivalence between the Hájek estimator and a specific inverse-probability-weighted regression coefficient. Second, we develop a bias-corrected network-robust covariance adjustment that ensures design-based validity of standard errors under arbitrary regression misspecification. Theoretically, our estimator is consistent and asymptotically normal. Simulations and empirical applications demonstrate stable confidence interval coverage exceeding 95%. Our approach balances practical implementability, flexible incorporation of covariates, and design-based robustness—offering a new paradigm for causal inference in network experiments that unifies theoretical rigor with empirical applicability.

Exposure contrasts—commonly used in causal inference to estimate treatment and spillover effects—may yield estimates with signs opposite to the true unit-level causal effects, particularly under interference. Method: We systematically characterize the causal interpretability boundary of exposure contrasts within a nonparametric framework, formalizing exposure mappings via causal graphs. We derive verifiable necessary and sufficient conditions for sign consistency, ensuring that exposure contrasts robustly reflect the direction of causal effects—even under arbitrary assignment mechanisms (e.g., cluster-randomized trials, network experiments, or observational data with peer selection) and arbitrary interference structures. Contribution/Results: This work establishes, for the first time, a function-form–free theory guaranteeing sign preservation of exposure contrasts. It provides a rigorous foundation for reliable identification of spillover effects in social network analysis and policy evaluation, bridging theoretical causality and empirical practice without restrictive modeling assumptions.

This study addresses the bias and potential spurious significance in peer effect estimation arising from measurement error in linear-in-means models of social networks. It demonstrates that the direction of this bias hinges on the interaction between individual attributes and network structure. Departing from conventional approaches that rely on external instrumental variables, the paper innovatively leverages intrinsic network topology for identification, thereby circumventing traditional paradigms for handling measurement error. Building upon generalized method of moments (GMM) and two-stage least squares (2SLS), the authors develop a class of consistent estimators that explicitly incorporate network structure. The validity and robustness of these estimators are confirmed through extensive Monte Carlo simulations.

This study addresses estimation bias arising from heterogeneous peer effects and endogenous network formation in social interactions by proposing a Selection-Corrected Heterogeneous Spatial Autoregressive (SCHSAR) model. This framework achieves, for the first time, credible identification of heterogeneous spillover effects under endogenous networks by jointly modeling the processes of network formation and outcome generation. It incorporates a finite mixture structure to capture individual heterogeneity in responses to peer influence and employs a fully Bayesian data augmentation approach to overcome computational and identification challenges posed by complex endogeneity. Application to U.S. firm innovation networks reveals significant and heterogeneous positive peer effects on R&D investment. The analysis further quantifies both direct and spillover effects of policy shocks, offering empirical foundations for designing targeted innovation policies.

This study addresses the challenge posed by network interference, which complicates the accurate estimation of both direct treatment effects and spillover effects in conventional experimental designs. To overcome this limitation, the authors propose a novel randomization scheme based on ego-clusters—clusters formed by each focal individual and their immediate neighbors—and develop a tailored clustering algorithm aimed at minimizing the asymptotic variance of the resulting estimators. Within a formal model framework, the approach integrates model-based estimation with asymptotic normality theory to simultaneously identify global average treatment effects and spillover effects. Simulation studies and empirical analyses demonstrate that the proposed method substantially improves estimation precision and inferential efficiency compared to existing network experiment designs.