I am generating a cross table between two variables:

Sex (man and woman) and Pain_type (acute, severe, resolved, absent)

With expss package and custom tables feature.

The table contains cases made by tab_stat_cases and percentages made by per column tab_stat_cpct.

I am also doing a z test for significant differences between the percentages per column with the tab_last_sig_cpct function. However, I can’t find a way to generate additional columns with the z and p values ​​because the significance function only does the calculation and does not show me those values.

The code I am using is the following:

Data|>
  tab_cells("|"=unvr(Pain_type))|>
  tab_cols(total(label = "Total| |"), Sex)|>
  tab_stat_cases(label = "N")|>
  tab_last_add_sig_labels() |>
  tab_stat_cpct(label = "%")|>
  tab_last_sig_cpct()|>
  tab_pivot(stat_position = "inside_columns")

I have been trying to solve it with the tab_stat_fun_df function but it still doesn’t work for me.

Data|>
  tab_cells("|"=unvr(Pain_type))|>
  tab_cols(total(label = "Total| |"), Sex)|>
  tab_stat_cases(label = "N")|>
  tab_last_add_sig_labels() |>
  tab_stat_cpct(label = "%")|>
  tab_last_sig_cpct()|>
  tab_stat_fun_df(function(x){
     p_unified <- sum(data$Freq) / sum(data$Total)
     data$Z_value <- mapply(function(frec, total){
                     p1 <- frec / total
                     p2 <- data$Freq[1] / data$Total[1]
                     n1 <- total
                     n2 <- data$Total[1]
                     z <- (p1 - p2) / sqrt(p_unified * (1 - p_unified) * (1/n1 + 1/n2))
                     return(z)},
     data$Freq, data$Total)
                              }, label = "Z value")|>
  tab_pivot(stat_position = "inside_columns")

I hope to get a table like this

table of significant differences between percentages