CODEWITHSUNDEEP
rpurrr
Sajjad khani
Sajjad khani

Reputation: 21

How to perform pairwise.t.test in nested data frame

I would like to perform pairwise.t.test in the nested data frame. Is it possible?

library(dplyr)
library(broom)
library(tidyr)
library(purrr)
a <- mtcars%>%
  dplyr::select(disp, gear,am)%>%
  nest(-am)%>%
  mutate(t_test = map(data, ~ pairwise.t.test(disp, gear, p.adjust = "bonferroni")))

# Error in factor(g) : object 'gear' not found

Upvotes: 1

Views: 1276

Answers (2)

Jan Kislinger
Jan Kislinger

Reputation: 1563

You have to specify the data frames inside the nested column using .x.

a <- mtcars %>%
  select(disp, gear,am) %>%
  nest(data = c(-am)) %>%
  mutate(t_test = map(data, ~pairwise.t.test(.x$disp, .x$gear, p.adjust = "bonferroni")))

If you don't need the (nested) column data you can do it with just group_by and summarise

a <- mtcars %>%
  select(disp, gear,am) %>%
  group_by(am) %>%
  summarise(t_test = list(pairwise.t.test(disp, gear, p.adjust = "bonferroni")))

In both cases, you can get the results into a table form using broom (i guess you wanted that since you load that package).

a %>% 
  mutate(t_test = map(t_test, broom::tidy)) %>% 
  unnest(t_test)

Upvotes: 3

Jeremy K.
Jeremy K.

Reputation: 1792

Does this give the output you were looking for?

library(dplyr)
library(broom)
library(tidyr)
library(purrr)

mtcars_sub <- mtcars %>% 
  dplyr::select(disp, gear, am) %>% 
  group_by(am) %>% 
  nest()  


t_tests <- function(df){
  pairwise.t.test(df$disp, df$gear, p.adjust = "bonferroni")
}

map(mtcars_sub$data, t_tests)

#> [[1]]
#> 
#>  Pairwise comparisons using t tests with pooled SD 
#> 
#> data:  df$disp and df$gear 
#> 
#>   4    
#> 5 0.048
#> 
#> P value adjustment method: bonferroni 
#> 
#> [[2]]
#> 
#>  Pairwise comparisons using t tests with pooled SD 
#> 
#> data:  df$disp and df$gear 
#> 
#>   3     
#> 4 0.0027
#> 
#> P value adjustment method: bonferroni

Let me know if it's not the right output you need.

Credit to @ronak-shah

https://stackoverflow.com/a/62408521/8742237

Upvotes: 0

Related Questions