Hello. Your finalfit package is very nice. I very often use this package. Thank you!<b

Pass missing values to statistical tests about finalfit HOT 2 CLOSED

ShuntaroS commented on July 29, 2024

Pass missing values to statistical tests

from finalfit.

Comments (2)

ewenharrison commented on July 29, 2024

Many thanks for your comments. Help with this is much appreciated.

I can't replicate the na_to_p issue. This is working for me as expected. Can you please provided code illustrating the issue:

> library(finalfit)
> library(magrittr)
> explanatory = c("obstruct.factor")
> dependent = "perfor.factor"
> colon_s %>%
+   summary_factorlist(dependent, explanatory, p = TRUE, p_cat = "fisher",
+                      na_to_p = FALSE, na_include = TRUE)
       label    levels         No       Yes     p
 Obstruction        No 715 (79.3) 17 (63.0) 0.026
                   Yes 166 (18.4) 10 (37.0)      
             (Missing)   21 (2.3)                
> 
> colon_s %>%
+   summary_factorlist(dependent, explanatory, p = TRUE, p_cat = "fisher",
+                      na_to_p = TRUE, na_include = TRUE)
Explanatory variable(s) missing data included in hypothesis test (p-value).
       label    levels         No       Yes     p
 Obstruction        No 715 (79.3) 17 (63.0) 0.069
                   Yes 166 (18.4) 10 (37.0)      
             (Missing)   21 (2.3)                
> 
> colon_s %$% 
+   fisher.test(perfor.factor, obstruct.factor)

	Fisher's Exact Test for Count Data

data:  perfor.factor and obstruct.factor
p-value = 0.02574
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 1.015853 5.977041
sample estimates:
odds ratio 
  2.530265 

> 
> colon_s %>%  
+   dplyr::mutate(
+     obstruct.factor = forcats::fct_explicit_na(obstruct.factor)
+   ) %$% 
+   fisher.test(perfor.factor, obstruct.factor)

	Fisher's Exact Test for Count Data

data:  perfor.factor and obstruct.factor
p-value = 0.06915
alternative hypothesis: two.sided

For the second bit, I couldn't see what missing data was being excluded. Again, would be delighted to look at code and fix anything that is not working:

> explanatory = c("obstruct.factor")
> dependent = "perfor.factor"
> colon_s %>%
+   summary_factorlist(dependent, explanatory, p = TRUE, p_cat = "fisher",
+                      na_include = TRUE, 
+                      add_dependent_label = TRUE,
+                      dependent_label_prefix = "Test")
 TestPerforation                   No       Yes     p
     Obstruction        No 715 (79.3) 17 (63.0) 0.026
                       Yes 166 (18.4) 10 (37.0)      
                 (Missing)   21 (2.3)

from finalfit.

ShuntaroS commented on July 29, 2024

Thank you for your quick reply.
After restarting R, the problem could not be reproduced.
I'm very sorry to have answered carefully. Next, leave a record.

from finalfit.

Recommend Projects