SODA With Peers or SODA With Friends? An integration of Spatial Principles with Social Networks to Detect and Model Social Outcome Dependence or Autocorrelation—SODA
Description
Data prepared to replicate all the analytic procedures discussed in the paper: SODA With Peers or SODA With Friends? An integration of Spatial Principles with Social Networks to Detect and Model Social Outcome Dependence or Autocorrelation—SODA Manuel S. González Canché Abstract The goal of this study is to comprehensively describe an analytic framework designed to estimate peer or spillover effects in any social network. Conceptually, examples illustrate advantages and limitations of relying on classmates as sources of influence, compared to establishing arguably more meaningful connections among peers that transcend classroom settings (i.e., friendships). Methodologically, we guide readers to detect and address Social Outcome Dependence or Autocorrelation (SODA) by merging social networks analysis procedures with spatial econometric techniques. That is, although Moran’s I is used to measure SODA in spatial settings, we extrapolate this technique to social settings to measure peer influence or spillover effects. The analytic framework is applied to two-mode (different units) and one-mode (same units) networks to assess baseline equivalence in experimental settings (i.e., like clustered randomized control trials) and estimate and address SODA in experimental and observational studies. All replication resources (data and code) are available here https://cutt.ly/5wyTLBlw.
Files
Steps to reproduce
Open a working version of R with internet access and execute the code that can be accessed here https://cutt.ly/5wyTLBlw. The steps reproduce Tables 7, 8, and 9.
Institutions
- University of Pennsylvania