Cluster Estimated Standard Errors in R (ceser)

Abstract

This paper presents an implementation in R of the Cluster Estimated Standard Error (CESE) proposed by Jackson (2019). The method estimates the covariance matrix of the estimated coefficients of linear models in grouped data sets with correlation among observations within groups. Cluster Estimated Standard Errors (CESE) is an alternative solution for the classical Cluster Robust Standard Error (CRSE) (Greene 2012; Eicker 1967; White 1980; Liang and Zeger 1986; MacKinnon and Webb 2017), which underestimates the standard errors in most of the situations encountered in practice (Esarey and Menger 2018).

Publication
Journal of Statistical Software (UNDER REVIEW)