About This Fisher Exact Test Tool
A Fisher exact test is useful when a two by two table has small counts. It checks whether two categorical variables are associated. The method does not rely on a large sample approximation. It uses the exact probability of the observed table, given fixed row and column totals. That makes it helpful for clinical studies, surveys, laboratory checks, quality reviews, and classroom statistics.
Why Exact Testing Matters
A chi square test can be weak when expected counts are small. Fisher's method avoids that concern by listing every possible table that keeps the same margins. It then compares the probability of each table with the observed one. The calculator reports one sided and two sided evidence, so the result can match the research question.
Reading The Output
The p value is the main decision measure. A small value means the observed arrangement is unusual under independence. The odds ratio shows the direction and strength of association. A value above one favors the first row and first column pairing. A value below one favors the opposite pattern. Confidence limits give a practical range for the odds ratio. When any cell is zero, the corrected interval prevents impossible division.
Advanced Use Cases
This calculator also shows expected counts, risks, risk difference, and risk ratio. These extra fields help explain the practical meaning behind the exact test. A statistically significant result may still be small in real terms. A nonsignificant result may still need a larger sample. Always review the table, the p value, and the effect measures together.
Good Data Practice
Enter raw counts, not percentages. Make sure each observation belongs to one row and one column only. Avoid mixing repeated measures with independent counts unless the study design supports it. Name rows and columns clearly before exporting. The CSV file is best for spreadsheets. The PDF file is useful for quick reporting. Keep the calculation settings with the result, especially the alternative hypothesis and confidence level.
Interpreting Results Responsibly
The test measures evidence, not cause. Causation depends on design, controls, timing, and sampling. Report the counts beside the p value. This helps readers verify assumptions and understand the actual comparison clearly before taking action with confidence.