Add chi square attribute to the datawizard_crosstab objects.

DESCRIPTION

@@ -85,4 +85,4 @@ Roxygen: list(markdown = TRUE)
 Config/testthat/edition: 3
 Config/testthat/parallel: true
 Config/Needs/website: easystats/easystatstemplate
-Config/roxygen2/version: 8.1.0
+Config/roxygen2/version: 8.1.1

R/data_tabulate.R

@@ -44,6 +44,8 @@
 #' any combination of `N` (frequencies including `NA`), `"raw"` (percentage
 #' including `NA` values), `"valid"` (percentage excluding `NA` values) and
 #' `"cumulative"` (percentage excluding `NA` values).
+#' @param chisq include chi square for two dimensional tables. This is always
+#' calculated with the `chisq.test()` `simulate.p.value = TRUE`.
 #' @param ... not used.
 #' @inheritParams extract_column_names
 #'
@@ -163,6 +165,7 @@ data_tabulate.default <- function(
   name = NULL,
   verbose = TRUE,
   metrics = c("N", "raw", "valid", "cumulative"),
+  chisq = FALSE,
   ...
 ) {
   # save label attribute, before it gets lost...
@@ -211,7 +214,8 @@ data_tabulate.default <- function(
       remove_na = remove_na,
       proportions = proportions,
       obj_name = obj_name,
-      group_variable = group_variable
+      group_variable = group_variable,
+      chisq = chisq
     ))
   }
   # frequency table
@@ -334,6 +338,7 @@ data_tabulate.default <- function(
     attr(out, "by") <- NULL
   } else {
     attr(out, "by") <- by_name
+    attr(out, "chi_square") <- chisq
   }
 
   class(out) <- c("datawizard_table", "data.frame")
@@ -357,6 +362,7 @@ data_tabulate.data.frame <- function(
   proportions = NULL,
   collapse = FALSE,
   verbose = TRUE,
+  chisq = FALSE,
   ...
 ) {
   if (!is.null(by)) {
@@ -391,6 +397,7 @@ data_tabulate.data.frame <- function(
       remove_na = remove_na,
       name = i,
       verbose = verbose,
+      chisq = chisq,
       ...
     )
   })
@@ -422,6 +429,7 @@ data_tabulate.grouped_df <- function(
   remove_na = FALSE,
   collapse = FALSE,
   verbose = TRUE,
+  chisq = FALSE,
   ...
 ) {
   grps <- attr(x, "groups", exact = TRUE)

R/data_xtabulate.R

@@ -7,8 +7,19 @@
   remove_na = FALSE,
   proportions = NULL,
   obj_name = NULL,
-  group_variable = NULL
+  group_variable = NULL,
+  chisq = FALSE
 ) {
+  if (is.null(group_variable) && is.null(attr(x, "groups"))) {
+    chi_square <- chisq.test(x, by, simulate.p.value = TRUE)
+  } else if (nlevels(x) > 1 && nlevels(x) > 1) {
+    chi_square <- chisq.test(x, by, simulate.p.value = TRUE)
+  } else {
+    chi_square <-
+      "Both 'by' and 'x' must have at least 2 unique levels\n
+    in order to calculate chi square."
+  }
+
   if (!is.null(proportions)) {
     proportions <- match.arg(proportions, c("row", "column", "full"))
   }
@@ -97,7 +108,24 @@
   if (!is.null(by_name)) {
     attr(out, "by") <- by_name
   }
-
+  attr(out, "chi_square") <- chi_square
+  if (chisq) {
+    if (is.list(chi_square)) {
+      attr(out, "chisq_caption") <- paste0(
+        chi_square$method,
+        "\n ",
+        "X-squared = ",
+        round(chi_square[["statistic"]], 3),
+        ", ",
+        "df = ",
+        chi_square[["parameter"]][["df"]],
+        ", p = ",
+        round(chi_square$p.value, 3)
+      )
+    } else if (is.character(chi_square)) {
+      attr(out, "chisq_caption") <- chi_square
+    }
+  }
   attr(out, "grouped_df") <- !is.null(group_variable)
   attr(out, "prop_table") <- .prop_table(out)
 
@@ -312,6 +340,16 @@ print_html.datawizard_crosstab <- function(x, big_mark = NULL, ...) {
     caption <- paste0("Grouped by ", x[["Group"]][1])
     x$Group <- NULL
   }
+  if (!is.null(attr(x, "chisq_caption"))) {
+    caption <- paste0(caption, attr(x, "chisq_caption"))
+  }
+  if (!is.null(attr(x, "chi_square_warning"))) {
+    paste0(
+      caption,
+      "\n",
+      insight::format_warning(attr(x, "chi_square_warning"))
+    )
+  }
 
   # this function is used by all four supported format, markdown, text, html
   # and tt (tinytable). For tt, we sometimes have format "html" and backend = "tt",

tests/testthat/_snaps/data_tabulate.md

@@ -270,6 +270,31 @@
       5     | 2002107 | 20.02 |   20.02 |       100.00
       <NA>  |       0 |  0.00 |    <NA> |         <NA>
 
+# data_tabulate default by with chisq
+
+    Code
+      data_tabulate(x = efc$e16sex, by = efc$c172code, chisq = TRUE)
+    Output
+      Pearson's Chi-squared test with simulated p-value
+      	 (based on 2000 replicates)
+       X-squared = 3.88, df = NA, p = 0.163
+
+      efc$e16sex | low level of education | intermediate level of education
+      -----------+------------------------+--------------------------------
+      1          |                      5 |                              32
+      2          |                      3 |                              34
+      <NA>       |                      0 |                               0
+      -----------+------------------------+--------------------------------
+      Total      |                      8 |                              66
+
+      efc$e16sex | high level of education | <NA> | Total
+      -----------+-------------------------+------+------
+      1          |                       4 |    5 |    46
+      2          |                      12 |    5 |    54
+      <NA>       |                       0 |    0 |     0
+      -----------+-------------------------+------+------
+      Total      |                      16 |   10 |   100
+
 # data_tabulate print multiple, collapse
 
     Code
@@ -861,6 +886,26 @@
       |e42dep   |     4| 32  (50.8%)| 28 (44.4%)| 3 (4.8%) |    63|
       |e42dep   |  (NA)|  1  (33.3%)|  2 (66.7%)| 0 (0.0%) |     3|
 
+---
+
+    Code
+      display(data_tabulate(efc, "c172code", by = "e16sex", proportions = "column",
+        remove_na = TRUE, weights = "weights", chisq = TRUE))
+    Output
+
+
+      Table: Pearson's Chi-squared test with simulated p-value
+      	 (based on 2000 replicates)
+       X-squared = 4.217, df = NA, p = 0.117
+
+      |c172code |       male|     female| Total|
+      |:--------|----------:|----------:|-----:|
+      |1        |  5 (12.5%)|  3  (6.5%)|     8|
+      |2        | 32 (80.0%)| 32 (69.6%)|    64|
+      |3        |  3  (7.5%)| 11 (23.9%)|    14|
+      |         |           |           |      |
+      |Total    |         40|         46|    86|
+
 # data_tabulate, correct 0% for proportions
 
     Code