When samples sizes are small, the significance from values from a chi-square distribution can become inaccurate. What test can be applied to reduce this effect?
A) Fisher's exact test
B) Loglinear analysis
C) Seymour's test of precision
D) Multiple t-tests

Question

Quizplus · Accepted Answer

Fisher's exact test is a suitable test to apply when sample sizes are small because it calculates the exact probability of an observed distribution of data under the null hypothesis. It does not rely on the chi-square distribution and is thus not affected by small sample sizes. Loglinear analysis and multiple t-tests are not appropriate tests for this situation. Seymour's test of precision is not a commonly used statistical test and is not a suitable option.

When Samples Sizes Are Small, the Significance from Values from a Chi-Square