R/centerESM.R

In PeterVerboon/lagnetw: The package builds a network for ESM data

#' Center variables in ESM data 
#' 
#' Group centering or grand-mean centering the specified variables and optionally add them to the data 
#'
#' @param data data frame 
#' @param subjnr variable, indicating subjects
#' @param varnames vector with the names of the variables which are centered.
#' @param center vector of same length as varnames that indicates for each variable which centering must be,
#'               applied, either "person" or "grand_mean" or a number. If "person" or "grand_mean" is specified once (default), 
#'               all variables are person or grand-mean centered, respectively. 
#'               If center is NULL no variables are centered and a warning is issued.
#' @param addmeans logical indicating whether a vector of person means should be added to the data 
#'               (works only if person centering is requested for that variable).             
#'
#' @return data frame with centered variables, and if requested, person means with suffix "_means".
#' @export
#'
#' @examples
#' data("news")
#' res <- centerESM(data = news, subjnr="subjnr",
#'        addmeans = TRUE, 
#'        varnames = c("Fearful", "Relaxed", "Anxious"), 
#'        center = c("person", "person", "person"))

centerESM <- function(data, subjnr = NULL, addmeans = TRUE, varnames = NULL, center = NULL) {
  
if (is.null(varnames)) return(cat("Argument varnames is not specified in function centerESM","\n"))
if (is.null(subjnr)) return(cat("Argument subjnr is not specified in function centerESM","\n"))
  
if (is.null(center))  return(cat("Warning: no centering is requested","\n"))
if (length(center) == 1 && center == "person")  center <- rep("person", length(varnames))
if (length(center) == 1 && center == "grand_mean")  center <- rep("grand_mean", length(varnames))
   
if (!is.null(center))
   if (length(varnames) != length(center))  
     return(cat("Number of variables in varnames does not equal length of center","\n")) 

  if(!subjnr %in% names(data)) 
    return(cat("Name of subjnr is not correctly specified in function centerESM","\n")) 

dat1 <- data



# centering of variables

  for (i in seq_along(varnames)) 
  {
    xx <- varnames[i]
    if (is.numeric(dat1[,xx])) {
      
      if (center[i] == "person") {
        group <- as.factor(dat1[,subjnr])
        pmean <- stats::ave(dat1[,xx], group, FUN = function(x) mean(x, na.rm=T)); 
        dat1[,xx] <- dat1[,xx] - pmean
        if (addmeans == TRUE) dat1[,paste0(varnames[i],"_means")] <- pmean
      }
      if (center[i] == "grand_mean") dat1[,xx] <- dat1[,xx] - base::mean(dat1[,xx], na.rm=TRUE)
      if (is.numeric(center[i])) dat1[,xx] <- dat1[,xx] - center[i]
    }
  }

 return(dat1)

}   # end function