Effective sample size computation for Bayesian Estimation of Network Autocorrelation Model
The aim of the present study is to estimate the effective sample size for Bayesian Estimation of Network Autocorrelation Model (NAM), where the observations are correlated with each other because of the Network Effect. NAM is used to analyze/quantify/identify the network effect in a given network for a given specific theory of interpersonal influence while controlling for other possible factors. The other way around, NAM is also used to control for network effect to estimate other factors’ effect on an output variable. Thus, as network effect and density increase we expect higher correlation among the observations which implies lower level of ESS.
To estimate Effective Sample Size for Bayesian NAM, we used the ESS model proposed by Faes et. al (2012), where they estimated the ESS to devise a new denominator degrees-of-freedom for fixed-effects testing. We implemented that model into Bayesian Estimation of NAM model.