R/read_hayes_plot_data_R.R

In daaawit/ABOHiwi: ABO Hiwi Funktionen

#' Read Hayes data as .txt file into R
#' 
#' @import stringr
#' @importFrom utils read.csv
#' @importFrom dplyr filter
#' 
#' @description The process Hayes Macro is quite nice for a lot of different analysis types. However,
#' the R implementation is not quite there yet in terms of usability, which is where this function comes in. You
#' can capture the output of the R implementation of Process Hayes using the capture.output() function from 
#' the utils package to save the output of the process function. You can then pass the captured output into 
#' this function to get data that is somewhat more usable, e.g. for plotting interactions or accessing certain
#' coefficients quickly.
#' 
#' To extract the data from the Hayes output, this function uses the text in specific lines of the Hayes module. 
#' It might be possible that these change between versions. The current version this function works with Hayes
#' version 4.0.
#' 
#' @param data The output of a process function call in R, captured using capture.output() from utils (or something else with an
#' equivalent function).
#' 
#' @returns A list of class HayesData consisting of the plot_data for visualizing the interaction as well as the various significance 
#' measures from the Hayes SPSS output as well as the model description.
#' 
#' @export

read_hayes_plot_data_R <- function(data){
  ##### Model description #####
  model_description_rows <- which(data[colnames(data)] == "*********************************************************************** ")
  model_description <- data[(model_description_rows[1]+2) : (model_description_rows[2] - 1), ]
  model_description <- gsub("\\s+", "", str_trim(model_description)) # Replace all spaces between variables with one ";"
  model_description <- read.csv(text = model_description, sep = ":", header = F)
  colnames(model_description) <- c("Var", "Val")
  
  ##### Model summary #####
  model_summary_row_index <- which(data[colnames(data)] == "Model Summary: ")
  model_summary <- data[(model_summary_row_index+1):(model_summary_row_index+2),]
  model_summary <- gsub("\\s+", ";", str_trim(model_summary)) # Replace all spaces between variables with one ";"
  model_summary <- read.csv(text = model_summary, sep = ";")
  
  
  ##### Plot Data #####
  
  plot_data_start_index <- which(data[colnames(data)] == "Data for visualizing the conditional effect of the focal predictor:") + 1
  plot_data_end_index <- which(data[colnames(data)] == "******************** ANALYSIS NOTES AND ERRORS ************************ ") - 1
  plot_data <- data[plot_data_start_index:plot_data_end_index, ]
  plot_data <- gsub("\\s+", ";", str_trim(plot_data)) 
  plot_data <- gsub(",", ".", plot_data) # Replace , with . for conversion to numeric
  plot_data <- read.csv(text = plot_data, sep = ";") # Convert to df 
  plot_data <- data.frame(sapply(plot_data, function(x) as.numeric(x)))
  
  ##### Significance data ##### 
  
  signif_start_index <- which(data[colnames(data)] == "Model: ") + 1
  signif_end_index <- which(data[colnames(data)] == "Product terms key:") - 2
  signif_data <- data[signif_start_index:signif_end_index,]
  signif_data <-  gsub("\\s+", ";", str_trim(signif_data)) # Remove whitespaces
  signif_data[1] <- paste("Var", signif_data[1], sep = ";") # Variables have no label in Hayes output. Add label "var"
  signif_data <- read.csv(text = signif_data, sep = ";") # Convert to dataframe
  
  ##### Interaction data ######
  interaction_start_index <- which(data[colnames(data)] == "Test(s) of highest order unconditional interaction(s):")+1
  interaction_end_index <- which(data[colnames(data)] == "----------")-1
  interaction_data <- data[interaction_start_index:interaction_end_index,]
  interaction_data <-  gsub("\\s+", ";", str_trim(interaction_data)) # Remove whitespaces
  interaction_data[1] <- paste("Var", interaction_data[1], sep = ";") # Variables have no label in Hayes output. Add label "var"
  interaction_data <- read.csv(text = interaction_data, sep = ";") # Convert to dataframe
  
  hayes_data <- list(plot_data = plot_data, 
                     model_description = model_description, 
                     model_summary = model_summary, 
                     signif_data = signif_data, 
                     interaction_data = interaction_data)
  class(hayes_data) <- "HayesData"
  
  return(hayes_data)
}