Chi-square or Fisher's exact test: how to choose
Both test whether two categorical variables are associated. The choice depends on the expected counts in the table, not the observed ones.
Rule of thumb
- Expected count for a cell = row total × column total ÷ grand total.
- If expected counts are small (in a 2 × 2 table, any expected count below 5), report Fisher's exact test.
- Otherwise Pearson's chi-square test is appropriate.
Reporting
χ²(df, N = n) = x.xx, p = .xxx, Cramér's V = .xx. For 2 × 2 tables, an odds ratio with its confidence interval is often more informative than V.
Frequently asked questions
Does Fisher's test only work for 2 × 2 tables?
It is most often used for 2 × 2 tables; extensions exist for larger tables but are computationally heavier.
Try it with your data See a sample dashboard