A 2 x 2 table comparing an exposure with a rare disease outcome has one cell with an expected frequency of 2. Which test should replace the usual chi-square test?
- A Yates correction chi-square
- B Fisher exact test ✓
- C McNemar test
- D Student t-test
Correct answer: B. Fisher exact test
Explanation
The chi-square approximation breaks down when any expected cell frequency falls below 5 in a 2 x 2 table, because the sampling distribution of the statistic is then far from chi-square. Fisher exact test calculates the exact probability of the observed table from the hypergeometric distribution and is valid at any cell size. Yates correction only improves the approximation modestly and is not adequate when an expected value is as low as 2.
Reference: Mahajan's Methods in Biostatistics for Medical Students and Research Workers, 9th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.