如何使用 dplyr 和 rstatix 包使用来自 ANOVA 的数据表的列执行计算？

Question

I have a data frame and have done an ANOVA between the data. After the ANOVA I want to use one of the resulting columns to do a calculation and create a new column with the mutate() function. However, an error appears indicating that this operation cannot be done on an anova class object:

Error: `x` must be a vector, not a <anova_test/data.frame/rstatix_test> object.

Can someone help me perform calculations ( F + 1 ) with the F column of the ANOVA result?

library(dplyr)
library(rstatix)

Temperature <- factor(c(rep("cold", times = 4),
                        rep("hot", times = 4)),
                      levels = c("cold", "hot"))

Light <- factor(rep(c(rep("blue", times = 2),
                      rep("yellow", times = 2)),
                    times = 2),
                levels = c("blue", "yellow"))

Result <- c(90.40, 85.20, 21.70, 25.30,
            75.12, 77.36, 6.11, 10.8)

Data <- data.frame(Temperature, Light, Result)

NewColumn <- Data %>%
  anova_test(formula = Result ~ Temperature*Light) %>%
  mutate(New= `F` + 1) #<-------- Not working

Answer 1

As mentioned by JKupzig in the comments, this is a known issue in dplyr as documented here: https://github.com/tidyverse/dplyr/issues/5286 .

The issue is caused by anova_test() creating an output data frame with classes anova_test , data.frame and rstatix_test , in that order, while mutate() from dplyr seems to get hung up if the last element in the class vector is not data.frame . You can verify the classes of the output of the anova as follows:

Data %>% anova_test(formula = Result ~ Temperature*Light) %>% class()

[1] "anova_test"   "data.frame"   "rstatix_test"

As a workaround, you can add as_tibble() to your dplyr pipe after anova_test() . This resets the classes to tbl_df , tbl , and data.frame , in that order.

Data %>% anova_test(formula = Result ~ Temperature*Light) %>% as_tibble() %>% class()

[1] "tbl_df"     "tbl"        "data.frame"

Data %>%
    anova_test(formula = Result ~ Temperature*Light) %>% 
    as_tibble() %>%
    mutate(New= `F` + 1)

# A tibble: 3 x 8
  Effect              DFn   DFd        F         p `p<.05`   ges     New
  <chr>             <dbl> <dbl>    <dbl>     <dbl> <chr>   <dbl>   <dbl>
1 Temperature           1     4   42.2   0.003     "*"     0.914   43.2 
2 Light                 1     4 1041.    0.0000055 "*"     0.996 1042.  
3 Temperature:Light     1     4    0.725 0.442     ""      0.153    1.72

Note, that this action removes the classes anova_test and rstatix_test . If these classes are important down the line, use a different workaround with set_class() from the magrittr package ( magrittr is a dependency of dplyr , so no need to install it separately).

Data %>%
    anova_test(formula = Result ~ Temperature*Light) %>%
    magrittr::set_class(c("anova_test", "rstatix_test", "data.frame")) %>% 
    class()

[1] "anova_test"   "rstatix_test" "data.frame" 

Data %>%
   anova_test(formula = Result ~ Temperature*Light) %>%
   magrittr::set_class(c("anova_test", "rstatix_test", "data.frame")) %>% 
   mutate(New = `F` + 1)

ANOVA Table (type II tests)

             Effect DFn DFd        F        p p<.05   ges      New
1       Temperature   1   4   42.250 3.00e-03     * 0.914   43.250
2             Light   1   4 1041.366 5.50e-06     * 0.996 1042.366
3 Temperature:Light   1   4    0.725 4.42e-01       0.153    1.725