i have to analyse two binary categorical data by using chi square test. but one of the cell in 2/2 table showed cell count less than 5. i am in doubt whether i would continue my analysis with chi square or there is some way. thank you in advanced for your suggestions.
In SPSS there is a general rule that the expected count cannot be below a certain value, so long as this is not affected you could continue with the test. Or run fishers exact test. You could also use the Kruskal Wallis test.
In this situation; For 2x2 tables: Fisher's exact test must be used.
For other tables: Chi-square test can be used if no more than 20% of the expected frequencies are less than 5 and none is less than 1.
It is possible to 'pool' or 'collapse' categories into fewer, but this must only be done if it is meaningful to group the data in this way.
R, for example, can perform a Fisher's exact on tables larger than 2 x 2.
during google search i found that if cell count is less than 5, either we can use Yates correction or Fisher exact test. what should be my choice then? is it according to sample size or something else? please make my doubt clear.
For small samples, you are better off with Fisher exact test, as long as you have the computing power to complete the test.
By the way, the "less than 5" criterion is an approximation. This page gives a reasonably good discussion of small sample sizes with chi-square tests.
http://www.biostathandbook.com/small.html
please use Fisher exact Test. it will be of help using SPSS. Yates correction is a correction simply subtraction of .5 in the standard formula for chi square.
If you have expected counts below 5, Fisher's Exact Test is an alternative to a chi-square test.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis