1
$\begingroup$

I have the following 2X2 contingency table with observed counts to perform Chi-square test. But i've read about Chi-square test restriction:

"If any expected counts are less than 5, then some other test should be used (e.g., Fisher exact test for 2x2 contingency tables)"

I'm confused, since in my case the contingency table contains observed count not the expected ones. Is that restriction related to observed counts as well? Contigency table

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

I cannot see your table, there is seemingly some problem with your image. Even though the table shows the observed number of cases, nothing prevents you from computing the expected number under the null distribution (likely, independence or homogeneity). Then you can check if the condition is fulfilled.

Improve this answer
$\endgroup$
4
  • $\begingroup$ Thanks for your response! The table has attached now. $\endgroup$
    – Denis
    Commented Apr 19, 2020 at 20:12
  • $\begingroup$ I'm wondering, if there is any straightforward way (may be a special function in R) to count expected counts? $\endgroup$
    – Denis
    Commented Apr 20, 2020 at 21:11
  • $\begingroup$ I thought, the common approach is just applying the same restriction to the observed frequencies. Am i right? $\endgroup$
    – Denis
    Commented Apr 20, 2020 at 21:15
  • 1
    $\begingroup$ You can use for instance result <- loglin(table, margin=c(1,2), fit=TRUE). In the component result$fit you will find the expected counts. $\endgroup$
    – F. Tusell
    Commented Apr 21, 2020 at 9:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.