R/romr.fnc.R

In LMERConvenienceFunctions: Model Selection and Post-Hoc Analysis for (G)LMER Models

romr.fnc <-
function(model,data, trim = 2.5){
	data$rstand = as.vector(scale(resid(model))) 
	row.names(data)=1:nrow(data)
	outliers=as.numeric(row.names(data[abs(data$rstand)>trim,]))
	data0=data
        if(length(outliers) > 0){
                data=data[-outliers,,drop=TRUE]
        }
	cat("n.removed =",(nrow(data0)-nrow(data)),"\n")
	cat("percent.removed =",(nrow(data0)-nrow(data))/nrow(data0)*100,"\n")
	return(list(data=data,data0=data0,n.removed=nrow(data0)-nrow(data),percent.removed=(nrow(data0)-nrow(data))/nrow(data0)*100))
}