How to create a nested table in R

Question

I'm struggling to figure out how to create the following table;

using the following dataset

```{r}
df<-data.frame(ID=c(1,2,3,4,5,6),
       Treat_Cont = c("Treatment", "Treatment", "Treatment", "Control", "Control", "Control"),
       Q1 = c("Yes", "No", NA, "Yes", "No", NA),
       Q2 = c("Yes", "No", NA, "Yes", "No", NA)
       )
```

The main problem I have had using table(), the expss package or tables package is that they treat each level of the factor seperately, so the table will produce a count, for example, for; Treat_Control == Treatment & Q1 == Yes & Q2 == Yes .

Currently, I am at a stage where I'm unsure whether my issue is one of data structure, meaning I should reshape my dataset, or whether I'm missing a function or an argument to achieve this result.

Thanks,

@akrun Apologies, I'm not familiar with this syntax? Running the code, I'm unsure how it would help. Did you have something in particular in mind? — physicalcog, Oct 29 '18 at 18:01
Just checked your image. You may need `tables` i.e. `library(tables);tabular(Q1 + Q2 ~ Treat_Cont, df)` — akrun, Oct 29 '18 at 18:09

Gregory Demin · Accepted Answer · 2018-10-30T12:27:08.093

Solution with expss. Not very consise code from my opinion:

library(expss)
df = data.frame(ID=c(1,2,3,4,5,6),
               Treat_Cont = c("Treatment", "Treatment", "Treatment", "Control", "Control", "Control"),
               Q1 = c("Yes", "No", NA, "Yes", "No", NA),
               Q2 = c("Yes", "No", NA, "Yes", "No", NA)
)

df %>% 
    tab_total_row_position("none") %>% # suppress totals
    tab_rows("|" = Treat_Cont) %>%  # "|" suppress var. labels
    tab_cols(total(label = "|")) %>% # "|" suppress var. labels
    # if_na add values for NA
    tab_cells("|" = if_na(Q1, "<NA>")) %>% # "|" suppress var. labels
    tab_stat_cases(label = "Q1") %>% # calculate stats
    tab_cells("|" = if_na(Q1, "<NA>")) %>% # "|" suppress var. labels
    tab_stat_cases(label = "Q2") %>% # calculate stats
    tab_pivot(stat_position = "inside_columns") %>% # labels reposition
    tab_transpose() # transpose table

UPDATE: Shorter solution.

df %>% 
    calculate(
        cro(Treat_Cont %nest% if_na(Q1, "<NA>"), list("Q1"), total_row_position = "none") %merge%
            cro(Treat_Cont %nest% if_na(Q2, "<NA>"), list("Q2"), total_row_position = "none")
    ) %>% 
    tab_transpose()

Short solution with base R:

with(df, 
     rbind(
         "Q1" = table(Treat_Cont:addNA(Q1)),
         "Q2" = table(Treat_Cont:addNA(Q2))
     ))

How to create a nested table in R

1 Answers1