For loop in R: error in model.frame.default()

I'm quite new to loops so please be patient with me

I have calculated alpha indices (Observed, Shannon, InvSimpson, Evenness) for which I want to perform a Kruskal-Wallis statistical test with the variable Month of my table.

My table (df) looks something like this :

Before trying with a loop I performed the test, one indices at a time, like so :

obs <- df %>% kruskal_test(Observed ~ Month)
sha <- df %>% kruskal_test(Shannon ~ Month)
inv <- df %>% kruskal_test(InvSimpson ~ Month)
eve <- df %>% kruskal_test(Evenness ~ Month)
res.kruskal <- rbind(obs, sha, inv, eve)
res.kruskal

And it worked, that's the same result I want to get with the for loop :

# A tibble: 4 × 6
  .y.            n statistic    df       p method        
  <chr>      <int>     <dbl> <int>   <dbl> <chr>         
1 Observed      45      20.6     9 0.0144  Kruskal-Wallis
2 Shannon       45      24.0     9 0.00434 Kruskal-Wallis
3 InvSimpson    45      20.3     9 0.0159  Kruskal-Wallis
4 Evenness      45      22.0     9 0.00899 Kruskal-Wallis

However, when I try it with a for loop like so :

Indices <- c("Observed", "Shannon", "InvSimpson", "Evenness")
result.kruskal <- data_frame()

for (i in Indices) {
  kruskal <- df %>% kruskal_test(i ~ Month)
  result.kruskal <- rbind(result.kruskal, kruskal)
}

I get the following error :

Error in model.frame.default(formula = formula, data = data) : 
  variable length differ (found for 'Month')

From similar errors found on the forum, I don't think my problem comes from the Month variable as the error message says, I don't have NA in my table df either. Am I writing the for loop wrong?

I would be thankful for any insight you might have.

Sophie

Using the first rows of your dataset as example, both lapply() and apply() can be used to iterate over the columns. Then, with bind_rows() the results of single tests can be combined together as a data frame:

library(tidyverse)
library(rstatix)
Indices <- c("Observed", "Shannon", "InvSimpson", "Evenness")

using lapply

result.kruskal <- bind_rows(
               lapply(df[Indices], FUN = function(x)   kruskal_test(df, x ~ Month))
               , .id = "variable") %>% 
               select(-2) %>% as.data.frame()

result.kruskal

 variable        n statistic    df   p    method        
1 Observed       8      5.14     2 0.0766 Kruskal-Wallis
2 Shannon        8      2        2 0.368  Kruskal-Wallis
3 InvSimpson     8      3.22     2 0.2    Kruskal-Wallis
4 Evenness       8      1.44     2 0.486  Kruskal-Wallis

or with apply

result.kruskal <- bind_rows(
  apply(df[Indices], 2, FUN = function(x) kruskal_test(df, x ~ Month))
, .id = "variable") %>% select(-2) %>% as.data.frame()

result.kruskal

 variable        n statistic    df   p    method        
1 Observed       8      5.14     2 0.0766 Kruskal-Wallis
2 Shannon        8      2        2 0.368  Kruskal-Wallis
3 InvSimpson     8      3.22     2 0.2    Kruskal-Wallis
4 Evenness       8      1.44     2 0.486  Kruskal-Wallis

Example data

df <- read.table(text = "Observed	Shannon	InvSimpson	Evenness	Month
688	4.5538	23.365814	0.696963	February
749	4.3815	15.162467	0.661992	February
610	3.8291	11.178981	0.597054	February
665	4.2011	16.284009	0.646343	March
839	5.1855	43.198709	0.770260	March
516	3.2393	4.765211	0.518611	April
470	3.9677	11.614851	0.644873	April
539	4.2995	15.593572	0.683583	April", header=T)