Bug Report: requested continuous2 statistics not applied when N is low

[PaulC91]

Hey.

As you can see from the reprex below, the requested statistics are not applied when the sample size is reduced to 10. Perhaps this is intended behaviour but I can't find anything in the docs about it.

Thanks!

library(gtsummary)

trial %>%
  dplyr::slice_sample(n = 100) %>% 
  dplyr::select(age, marker) %>%
  tbl_summary(
    type = all_continuous() ~ "continuous2",
    statistic = all_continuous() ~ c("{median} ({p25}, {p75})", "{min}, {max}"),
    missing = "no"
  ) %>% 
  as_kable()

Characteristic	N = 100
Age
Median (IQR)	48 (39, 57)
Range	6, 83
Marker Level (ng/mL)
Median (IQR)	0.61 (0.18, 1.42)
Range	0.01, 2.77

trial %>%
  dplyr::slice_sample(n = 10) %>% 
  dplyr::select(age, marker) %>%
  tbl_summary(
    type = all_continuous() ~ "continuous2",
    statistic = all_continuous() ~ c("{median} ({p25}, {p75})", "{min}, {max}"),
    missing = "no"
  ) %>% 
  as_kable()

Characteristic	N = 10
Age
9	1 (11%)
34	1 (11%)
47	1 (11%)
50	1 (11%)
54	1 (11%)
57	1 (11%)
65	1 (11%)
67	1 (11%)
78	1 (11%)
Marker Level (ng/mL)
0.161	1 (11%)
0.169	1 (11%)
0.175	1 (11%)
0.238	1 (11%)
0.389	1 (11%)
1.107	1 (11%)
1.321	1 (11%)
1.328	1 (11%)
1.68	1 (11%)

^{Created on 2022-10-11 with reprex v2.0.2}

[ddsjoberg]

You need to use the type argument to tell tbl_summary() that these variables are indeed continuous2 (they default with so few unique levels is categorical). Because their default is categorical, you can't use all_continuous() to cast them as continuous2. https://www.danieldsjoberg.com/gtsummary/reference/tbl_summary.html#type-argument-1

[PaulC91]

Got it! thanks for the quick response.