I am dealing with an issue passing dichotomous data (1,0) where the '1' values have been filtered out to tbl_summary(). The type should be dichotomous, but I get an error message:
Summary type is "dichotomous" but no summary value has been assigned.
Reprex:
data <- data.frame(group = c("a", "b", "a", "a", "b"),
resp_1 = c(1, 1, 0, 1, 1),
resp_2 = c(0, 1, 0, 0, 1))
data %>% filter(group == "a") %>%
select(starts_with("resp")) %>%
tbl_summary(type = everything() ~ "dichotomous")
Not specifying type will give me a categorical response for values where there is no 1 values.
Results not as all dichotomous
I've searched for a solution but haven't found one that will allow "0%" as a dichotomous response. I'm sure there's a straightforward solution.
In the past, I have recoded the 1 and 0 values to "Y"/"N" or some other dichotomous values, but doing so would cause me to rewrite a lot of code. I am hoping to coerce tbl_summary to reporting 0% as a dichotomous result instead of a categorical one like in the attached screenshot.
I am using 'gtsummary' version 2.0.0 and R version 4.1.1.
@Edward is correct that this is a change in gtsummary v2.0. In the past, there has been an issue with unobserved values being incorrectly (e.g. via a typo) assigned as the dichotomous level to summarize. If you want an unobserved level to be summarized dichotomously, you need to explicitly make the level a factor level.
Below is an example of how I would do this:
library(gtsummary)
library(dplyr)
data <- tibble(group = c("a", "b", "a", "a", "b"),
resp_1 = c(1, 1, 0, 1, 1),
resp_2 = c(0, 1, 0, 0, 1))
data |>
filter(group == "a") |>
select(starts_with("resp")) |>
mutate(across(c(resp_1, resp_2), ~factor(., levels = c(0, 1)))) |>
tbl_summary(
type = everything() ~ "dichotomous",
value = c(resp_1, resp_2) ~ "1"
) |>
as_kable()
| Characteristic | N = 3 |
|---|---|
| resp_1 | 2 (67%) |
| resp_2 | 0 (0%) |
Created on 2024-08-20 with reprex v2.1.0