Calculate the Yates-corrected chi-squared statistic for a 2×2 contingency table from four observed cell counts and expected frequencies.
Customize This Calculator
Build your own version. Describe what you want changed, added, or compared.
- All Statistics Calculators
- Expected Frequency Calculator
- Cramer’s V Calculator
- Likelihood Ratio Calculator
- Standard Error of Proportion Calculator
- Z-Score Calculator
Yates Correction Formula
Yates’ correction adjusts the standard chi-square test for a 2×2 contingency table by applying a continuity correction. It is most often used when sample sizes are small and you want a more conservative test of whether two categorical variables are independent.
For calculator inputs labeled A, B, C, and D, the same statistic can be written in shortcut form:
The correction subtracts 0.5 from the absolute observed-minus-expected gap before squaring, which usually reduces the test statistic compared with the ordinary Pearson chi-square test.
Calculator Cell Layout
| Column 1 | Column 2 | |
|---|---|---|
| Row 1 | Cell A | Cell B |
| Row 2 | Cell C | Cell D |
Enter the observed counts exactly as they appear in your 2×2 table. The calculator then computes row totals, column totals, expected frequencies, and the Yates-corrected chi-square statistic automatically.
Expected Frequency Formulas
The expected count in each cell comes from its row total multiplied by its column total, divided by the grand total.
| Cell | Expected Count |
|---|---|
| A | E_A = ((A + B)(A + C)) / (N) |
| B | E_B = ((A + B)(B + D)) / (N) |
| C | E_C = ((C + D)(A + C)) / (N) |
| D | E_D = ((C + D)(B + D)) / (N) |
How to Use the Yates Correction Calculator
- Enter the four observed frequencies into cells A, B, C, and D.
- Make sure the values are counts, not percentages, rates, means, or proportions.
- Click calculate to generate the Yates-corrected chi-square statistic.
- Compare the result to a chi-square critical value for 1 degree of freedom, or use it to obtain a p-value in your statistical workflow.
How to Interpret the Result
For a 2×2 table, the test has 1 degree of freedom. Larger values indicate stronger evidence against the null hypothesis of independence.
| Significance Level | Critical Value | Interpretation |
|---|---|---|
| 10% | 2.706 | If the statistic is at least 2.706, the association is significant at the 0.10 level. |
| 5% | 3.841 | If the statistic is at least 3.841, the association is significant at the 0.05 level. |
| 1% | 6.635 | If the statistic is at least 6.635, the association is significant at the 0.01 level. |
If the statistic is below your chosen threshold, there is not enough evidence to reject independence. If it exceeds the threshold, the table suggests a statistically significant association between the two categorical variables.
Example
| Column 1 | Column 2 | Row Total | |
|---|---|---|---|
| Row 1 | 12 | 5 | 17 |
| Row 2 | 3 | 10 | 13 |
| Column Total | 15 | 15 | 30 |
Because 4.887 > 3.841, this table is significant at the 5% level. The correction still indicates an association, but it is more conservative than an uncorrected chi-square calculation would be.
When Yates Correction Is Appropriate
- You have exactly two rows and two columns.
- Your data are frequency counts from mutually exclusive categories.
- The expected counts are small enough that a continuity correction is desirable.
- You want a conservative approximation instead of the ordinary Pearson chi-square statistic.
If expected counts are very small, many analysts prefer Fisher’s exact test because it does not rely on a chi-square approximation.
Assumptions and Input Rules
| Requirement | Why It Matters |
|---|---|
| Independent observations | Each subject or item should contribute to one cell only. |
| Raw counts only | Chi-square methods are based on frequencies, not percentages or summary statistics. |
| Nonnegative values | Cell counts cannot be negative. |
| No empty margins | If an entire row or column total is zero, expected counts cannot be formed correctly. |
| 2×2 structure | Yates correction is designed specifically for two-category-by-two-category tables. |
Common Mistakes
- Entering percentages instead of observed frequencies.
- Placing counts in the wrong cells and reversing the row/column structure.
- Using the calculator for tables larger than 2×2.
- Applying the test to paired or repeated-measures data, which violates independence.
- Interpreting statistical significance as proof of a large or practically important effect.
Used correctly, the calculator gives a fast way to evaluate whether a small-sample 2×2 table shows evidence of association while accounting for continuity correction.
