Merge pull request #844 from cmu-delphi/fb-package-is-selected-missing

[fb survey] `is_selected` to map missing responses to `NA`

Author: krivard

facebook/delphiFacebook/R/variables.R

@@ -19,6 +19,9 @@ split_options <- function(column) {
 
 #' Test if a specific selection is selected
 #'
+#' Checking whether a specific selection is selected in either "" (empty
+#' string) or `NA` responses will produce `NA`s.
+#'
 #' @param vec A list whose entries are character vectors, such as c("14", "15").
 #' @param selection one string, such as "14"
 #' @return a logical vector; for each list entry, whether selection is contained
@@ -27,9 +30,11 @@ is_selected <- function(vec, selection) {
   selections <- sapply(
     vec,
     function(resp) {
-      if (length(resp) == 0) {
-        # All our selection items include "None of the above" or similar, so
-        # treat no selection the same as missingness.
+      if (length(resp) == 0 || all(is.na(resp))) {
+        # Qualtrics files code no selection as "" (empty string), which is
+        # parsed by `read_csv` as `NA` (missing) by default. Since all our
+        # selection items include "None of the above" or similar, treat both no
+        # selection ("") or missing (NA) as missing, for generality.
         NA
       } else {
         selection %in% resp

facebook/delphiFacebook/tests/testthat/test-variables.R

@@ -12,14 +12,32 @@ test_that("is_selected handles selections correctly", {
   expect_equal(is_selected(split_options(c("", "14", "4", "4,6,8")), "1"),
                c(NA, FALSE, FALSE, FALSE))
 
-  expect_equal(is_selected(split_options(c("1", "15", "14", NA)), "14"),
-               c(FALSE, FALSE, TRUE, FALSE))
+  expect_equal(is_selected(split_options(c("1", "15", "14", NA, "")), "14"),
+               c(FALSE, FALSE, TRUE, NA, NA))
 
   expect_equal(is_selected(split_options(c("4,54", "3,6,2,54", "5,4,45")),
                            "54"),
                c(TRUE, TRUE, FALSE))
 })
 
+test_that("activities items correctly coded", {
+  input_data <- data.frame(
+    C13 = c(NA, "1,2,4", "3", "", "6", "2,4")
+  )
+  
+  out <- code_activities(input_data)
+  
+  # expected result
+  input_data$a_work_outside_home_1d <- c(NA, TRUE, FALSE, NA, FALSE, FALSE)
+  input_data$a_shop_1d <- c(NA, TRUE, FALSE, NA, FALSE, TRUE)
+  input_data$a_restaurant_1d <- c(NA, FALSE, TRUE, NA, FALSE, FALSE)
+  input_data$a_spent_time_1d <- c(NA, TRUE, FALSE, NA, FALSE, TRUE)
+  input_data$a_large_event_1d <- c(NA, FALSE, FALSE, NA, FALSE, FALSE)
+  input_data$a_public_transit_1d <- c(NA, FALSE, FALSE, NA, TRUE, FALSE)
+
+  expect_equal(out, input_data)
+})
+
 test_that("mask items correctly coded", {
   input_data <- data.frame(
     C14 = c(NA, 1, 3, 6, 2, 4),