13.18
Fisher's Exact Test
Fisher's exact test is a statistical significance test widely used to analyze 2x2 contingency tables, particularly in situations where sample sizes ar…
Fisher's exact test determines the significance of a nonrandom relationship between two categorical variables in a two-by-two contingency table.
Unlike the chi-square test, which approximates the probability of observed outcomes, Fisher's test yields an exact P-value.
It helps analyze unequally distributed data with small sample sizes, especially for expected frequency values of less than five.
Despite its computational demands, Fisher's exact test ensures precision and integrity in result interpretation.
Researchers apply this test in various fields, including medicine, to compare the effectiveness or safety of treatments.
For example, it is used to compare the efficacy of drugs A and B, where the accurate P-value determines whether the differences in the success rate between the drugs are statistically significant.
The small P-value calculated implies that the difference between drug efficacy for the drugs is statistically significant.
View the full transcript and gain access to JoVE Core videos
Q1: When should you use Fisher's exact test instead of the chi-square test?
Use Fisher's exact test when analyzing a 2x2 contingency table with small sample sizes or when expected cell frequencies are less than five. Unlike the chi-square test, which approximates P-values and requires minimum expected frequencies of at least five, Fisher's exact test calculates the exact probability of observing your data under the null hypothesis, ensuring accurate results even with sparse data. This makes it part of introduction to nonparametric statistics methods.
Q2: What does a small P-value from Fisher's exact test indicate?
A small P-value indicates that the observed relationship between two categorical variables is statistically significant and unlikely to occur by chance. For example, when comparing drug efficacy, a small P-value suggests the difference in success rates between treatments is statistically significant, supporting the conclusion that the treatments differ meaningfully in their effectiveness.
Q3: Why is Fisher's exact test considered non-parametric?
Fisher's exact test is non-parametric because it does not assume the data follows any specific probability distribution. Instead of relying on distributional assumptions like parametric tests do, it calculates exact probabilities based directly on the observed contingency table, making it robust and reliable regardless of how the underlying data is distributed.
Q4: What types of data can Fisher's exact test analyze?
Fisher's exact test analyzes categorical data represented in a 2x2 contingency table where both variables are categorical. It is widely used in medicine to evaluate treatment effects in small clinical trials, in biology to analyze genetic associations or experimental outcomes, and in social sciences to examine relationships between categorical variables.
Q5: How does Fisher's exact test handle unequally distributed data?
Fisher's exact test calculates the exact P-value by computing all possible contingency tables with the same marginal totals as the observed data. This approach effectively handles unequally distributed data and small sample sizes by determining the precise probability of observing results as extreme or more extreme than those observed, ensuring precision regardless of data distribution patterns.
Q6: What are the main advantages of Fisher's exact test over approximation methods?
Fisher's exact test yields an exact P-value rather than an approximation, ensuring precision and integrity in result interpretation. Its exactness makes it especially valuable when sample sizes are small or expected cell counts are low, where approximation methods like chi-square may fail or provide unreliable outcomes, making it a robust alternative for researchers working with limited datasets.
Q7: In what research fields is Fisher's exact test commonly applied?
Fisher's exact test is applied across medicine, biology, and social sciences. In medical research, it evaluates treatment effectiveness and safety in small clinical trials. In biology, it analyzes genetic associations and experimental outcomes. In social sciences, it examines relationships between categorical variables, making it a versatile tool for assessing statistical significance when traditional approaches are unsuitable.