R/ueq_scales.R

In gitc23/ueqr: Analyzes user experience questionnaire data

#' Summarize results for the six UEQ scales
#'
#' Creates a dataframe with scale name, mean, variance, standard deviation,
#' N, confidence, left and right confidence interval, alpha, lambda^2, and
#' the benchmark comparison.
#'
#' @param df dataframe with 26 columns in the order specified in the UEQ documentation
#' @param is.clean Boolean whether the passed dataframe is already cleaned
#' @param ueq_range the range of columns where the 26 UEQ Items are to be found
#'
#' @return dataframe with 11 columns and one row for each of the six scales
#'
#' @export
#'
ueq_scales <- function(df, is.clean = FALSE, ueq_range = c(1:26)) {

  scales <- list(attractiveness = c(1, 12, 14, 16, 24, 25),
                 perspicuity = c(2, 4, 13, 21),
                 efficiency = c(9, 20, 22, 23),
                 dependability = c(8, 11, 17, 19),
                 stimulation = c(5, 6, 7, 18),
                 novelty = c(3, 10, 15, 26))

  if (!is.clean) {
    cleaned_df <- clean_ueq(df[ueq_range])
  } else {
    cleaned_df <- df
  }

  scales_list <- list()
  for (i in 1:length(scales)) {
    scale <- case_when(
      i == 1 ~ "attractiveness",
      i == 2 ~ "perspicuity",
      i == 3 ~ "efficiency",
      i == 4 ~ "dependability",
      i == 5 ~ "stimulation",
      i == 6 ~ "novelty")
    means <- mean(rowMeans(cleaned_df[unlist(scales[i])]))
    vars <- var(rowMeans(cleaned_df[unlist(scales[i])]))
    sd <- sd(rowMeans(cleaned_df[unlist(scales[i])]))
    n <- nrow(cleaned_df[unlist(scales[i])])
    confidence <- qnorm(0.975)*sd/sqrt(n)
    ci_left <-  means - confidence
    ci_right <-  means + confidence
    alpha <- psych::alpha(cleaned_df[unlist(scales[i])])[["total"]][["std.alpha"]]
    lambda2 <- psych::alpha(cleaned_df[unlist(scales[i])])[["total"]][["G6(smc)"]]

    instance <- tibble(scale, means, vars, sd, n, confidence, ci_left, ci_right, alpha, lambda2)
    scales_list[[i]] <- instance
  }

  scales_results <- bind_rows(scales_list)

  return(cbind(scales_results, benchmark = ueq_benchmarks(scales_results)))
}


ueq_benchmarks <- function(scales_results) {

  bm_attrac <- function(scale_mean) {
    bm <- case_when(
      scale_mean > 1.75 ~ "Excellent",
      scale_mean > 1.52 ~ "Good",
      scale_mean > 1.17 ~ "Above Average",
      scale_mean > 0.7 ~ "Below Average",
      scale_mean <= 0.7 ~ "Bad"
    )
    return(bm)
  }
  bm_pers <- function(scale_mean) {
    bm <- case_when(
      scale_mean > 1.9 ~ "Excellent",
      scale_mean > 1.56 ~ "Good",
      scale_mean > 1.08 ~ "Above Average",
      scale_mean > 0.64 ~ "Below Average",
      scale_mean <= 0.64 ~ "Bad"
    )
    return(bm)
  }
  bm_eff <- function(scale_mean) {
    bm <- case_when(
      scale_mean > 1.78 ~ "Excellent",
      scale_mean > 1.47 ~ "Good",
      scale_mean > 0.98 ~ "Above Average",
      scale_mean > 0.54 ~ "Below Average",
      scale_mean <= 0.54 ~ "Bad"
    )
    return(bm)
  }
  bm_dep <- function(scale_mean) {
    bm <- case_when(
      scale_mean > 1.65 ~ "Excellent",
      scale_mean > 1.48 ~ "Good",
      scale_mean > 1.14 ~ "Above Average",
      scale_mean > 0.78 ~ "Below Average",
      scale_mean <= 0.78 ~ "Bad"
    )
    return(bm)
  }
  bm_stim <- function(scale_mean) {
    bm <- case_when(
      scale_mean > 1.55 ~ "Excellent",
      scale_mean > 1.31 ~ "Good",
      scale_mean > 0.99 ~ "Above Average",
      scale_mean > 0.5 ~ "Below Average",
      scale_mean <= 0.5 ~ "Bad"
    )
    return(bm)
  }
  bm_nov <- function(scale_mean) {
    bm <- case_when(
      scale_mean > 1.4 ~ "Excellent",
      scale_mean > 1.05 ~ "Good",
      scale_mean > 0.71 ~ "Above Average",
      scale_mean > 0.3 ~ "Below Average",
      scale_mean <= 0.3 ~ "Bad"
    )
    return(bm)
  }

  benchmark <- rbind(
    bm_attrac(scales_results$means[1]),
    bm_pers(scales_results$means[2]),
    bm_eff(scales_results$means[3]),
    bm_dep(scales_results$means[4]),
    bm_stim(scales_results$means[5]),
    bm_nov(scales_results$means[6]))

  return(benchmark)
}