R/ttest_all.r

In samuelkaminsky/SKTools: SK's Tools

#' Conduct T-Tests at every nth percentile for list of IVs and DVs
#' @param df Dataframe with test data
#' @param ivs Names of independent variables to be inserted into dplyr::select()
#' @param dvs Names of dependent variables to be inserted into dplyr::select()
#' @param perc Nth percentile to conduct T-Test at
#' @return Data frame of tidy t.test results
#' @description Conduct T-Tests at every 5% interval for list of IVs and DVs
#' @export

ttest_all <-
  function(df, ivs, dvs, perc = .05) {
    dvs <- dplyr::enquo(dvs)
    ivs <- dplyr::enquo(ivs)
    IVs <-
      df %>%
      dplyr::select(!!ivs) %>%
      names() %>%
      purrr::set_names()
    DVs <-
      df %>%
      dplyr::select(!!dvs) %>%
      names() %>%
      purrr::set_names()

    IVs %>%
      purrr::map_dfr(function(x) {
        DVs %>%
          purrr::map_dfr(function(y) {
            y_quo <- rlang::enquo(y)

            stats::quantile(df[, x] %>% unlist(), seq(
              from = 0.05,
              to = .95,
              by = perc
            ), na.rm = TRUE) %>%
              as.list() %>%
              purrr::map_dfr(purrr::possibly(
                function(z) {
                  df$Grouped <-
                    dplyr::if_else(df[, x] >= z, 1, 0) %>%
                    as.factor()
                  cd <-
                    effsize::cohen.d(
                      stats::as.formula(paste0(y, " ~ Grouped")),
                      data = df
                    )
                  cd.df <-
                    tibble::tibble(
                      cd.est = cd$estimate %>% as.numeric(),
                      cd.mag = cd$magnitude %>% as.character()
                    )
                  stats::t.test(
                    stats::as.formula(paste0(y, " ~ Grouped")),
                    data = df
                  ) %>%
                    broom::tidy() %>%
                    cbind(
                      df %>%
                        dplyr::group_by(.data$Grouped, !!y_quo) %>%
                        dplyr::summarize(Count = dplyr::n()) %>%
                        dplyr::group_by(.data$Grouped) %>%
                        dplyr::summarize(Count = sum(.data$Count)) %>%
                        tidyr::spread(.data$Grouped, .data$Count),
                      Cutoff.Num = z,
                      cd.df
                    ) %>%
                    dplyr::mutate_if(is.factor, as.character)
                },
                otherwise = tibble::tibble(
                  estimate = NA_real_,
                  estimate1 = NA_real_,
                  estimate2 = NA_real_,
                  statistic = NA_real_,
                  p.value = NA_real_,
                  parameter = NA_real_,
                  conf.low = NA_real_,
                  conf.high = NA_real_,
                  method = NA_character_,
                  alternative = NA_character_,
                  `0` = NA_real_,
                  `1` = NA_real_,
                  Cutoff.Num = NA_real_,
                  Cutoff.Perc = NA_real_,
                  cd.est = NA_real_,
                  cd.mag = NA_character_
                )
              ),
              .id = "Cutoff.Perc"
              )
          },
          .id = "DV"
          )
      },
      .id = "IV"
      ) %>%
      dplyr::distinct() %>%
      tidyr::drop_na(.data$estimate) %>%
      dplyr::mutate(sig = dplyr::if_else(.data$p.value < .05, TRUE, FALSE)) %>%
      dplyr::mutate_at(
        dplyr::vars(.data$estimate:.data$conf.high, .data$Cutoff.Num),
        list(~ round(., 6))
      ) %>%
      dplyr::distinct(.data$IV, .data$DV, .data$estimate, .data$estimate1,
        .keep_all = TRUE
      ) %>%
      dplyr::select(
        .data$IV, .data$DV, .data$Cutoff.Perc, .data$Cutoff.Num,
        dplyr::everything()
      )
  }