CODEWITHSUNDEEP
rchi-squared
Nate
Nate

Reputation: 577

Error in chisq.posthoc.test : All entries of 'x' must be nonnegative and finite?

I'm getting this error message even though I don't have any NA's, negative counts, etc.

I'm trying to do a post-hoc chi-square test on this data:

df <- structure(list(Zone = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 
2L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L), .Label = c("Crocodile", 
"Rankin", "West", "Whipray"), class = "factor"), Year = structure(c(1L, 
2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L), .Label = c("2016", 
"2017", "2018", "2019"), class = "factor"), empty_num_sum = c(0L, 
8L, 2L, 0L, 3L, 17L, 8L, 19L, 7L, 31L, 17L, 17L, 4L, 4L, 0L, 
3L)), row.names = c(NA, -16L), groups = structure(list(Zone = structure(1:4, .Label = c("Crocodile", 
"Rankin", "West", "Whipray"), class = "factor"), .rows = structure(list(
    1:4, 5:8, 9:12, 13:16), ptype = integer(0), class = c("vctrs_list_of", 
"vctrs_vctr", "list"))), row.names = c(NA, -4L), class = c("tbl_df", 
"tbl", "data.frame"), .drop = TRUE), class = c("grouped_df", 
"tbl_df", "tbl", "data.frame"))

But it keeps giving me this error...

chisq.posthoc.test(df, method = "bonferroni")

    Error in chisq.test(x, ...) : 
      all entries of 'x' must be nonnegative and finite

Upvotes: 1

Views: 705

Answers (1)

Allan Cameron
Allan Cameron

Reputation: 173928

You need to pass your counts as a matrix / contingency table. You can easily convert your data frame to this format using the xtabs function:

xtab <- xtabs(empty_num_sum ~ Zone + Year, data = df)

xtab
#>            Year
#> Zone        2016 2017 2018 2019
#>   Crocodile    0    8    2    0
#>   Rankin       3   17    8   19
#>   West         7   31   17   17
#>   Whipray      4    4    0    3

So now you can do:

library(chisq.posthoc.test)

chisq.posthoc.test(xtab)
#> Warning in chisq.test(x, ...): Chi-squared approximation may be incorrect
#>   Dimension     Value       2016        2017       2018        2019
#> 1 Crocodile Residuals -1.0938835  2.46306043  0.0594116 -2.03921942
#> 2 Crocodile  p values  1.0000000  0.22041100  1.0000000  0.66285000
#> 3    Rankin Residuals -1.0141479 -1.13659482 -0.4827683  2.35822986
#> 4    Rankin  p values  1.0000000  1.00000000  1.0000000  0.29379700
#> 5      West Residuals -0.1127333  0.04881493  1.3347770 -1.15317201
#> 6      West  p values  1.0000000  1.00000000  1.0000000  1.00000000
#> 7   Whipray Residuals  3.0363341 -0.45336940 -1.6889164 -0.04504248
#> 8   Whipray  p values  0.0383160  1.00000000  1.0000000  1.00000000

Created on 2022-02-07 by the reprex package (v2.0.1)

Upvotes: 1

Related Questions