Yates' Correction

Yates' correction is a small-sample adjustment for the chi-square test of independence in a 2x2 contingency table. In Honors Statistics, it lowers the test statistic when expected counts are small.

What is Yates' Correction?

Yates' correction is the adjustment you use with a 2x2 contingency table when the chi-square test of independence is being pushed past its comfort zone by small expected counts. Instead of using the raw difference between observed and expected frequencies, the correction subtracts 0.5 from the absolute difference in each cell before squaring and dividing by the expected value.

That extra step makes the chi-square statistic smaller, which makes the p-value larger. In plain terms, the test becomes more cautious about claiming there is an association between two categorical variables. This matters most when one or more expected cell frequencies are below 5, which is when the standard chi-square approximation can exaggerate how unusual the table looks.

A 2x2 table has two categories for one variable and two categories for the other. Think of a table like yes/no versus exposed/not exposed in a medical study, or pass/fail versus study method A/method B. If the sample is small, one or two cells can end up with tiny counts, and the chi-square test can overreact to those small differences.

Yates' correction is not a separate test. It is still part of the chi-square test of independence, just with a built-in adjustment for continuity. You will usually see it in the same unit as contingency tables, expected cell frequencies, and contingency analysis, because it comes up when you are checking whether two categorical variables look independent.

One easy way to think about it is this: the correction asks the test to be less jumpy about small gaps between observed and expected counts. That makes it a more conservative choice, especially in small-sample settings where a standard chi-square test might make an association look stronger than it really is.

Why Yates' Correction matters in Honors Statistics

Yates' correction shows up when Honors Statistics moves from counting data in a contingency table to making an inference about association. If you only know how to fill in a 2x2 table, you can list counts, but you still need a method for deciding whether the differences are meaningful or just random noise.

This term matters because it changes how you interpret the chi-square test result. A smaller test statistic usually means a larger p-value, so the evidence against the null hypothesis of independence has to be stronger before you reject it. That keeps you from overclaiming an association when the sample is small or the data are sparse.

It also connects to real data situations where rare outcomes matter. Medical studies, epidemiology examples, and other small-sample investigations often create tables with low expected counts. In those settings, using the corrected version shows you know not just how to compute a test, but also when the usual approximation needs help.

If you see a problem asking whether two categorical variables are independent, Yates' correction is part of the decision process when the table is 2x2 and the expected counts are low. That makes it a practical checkpoint in the whole contingency-table unit, not just a formula to memorize.

How Yates' Correction connects across the course

Contingency Table

Yates' correction is used on a contingency table, and it only makes sense when you are working with counts in rows and columns. A 2x2 table is the specific setup where the correction comes into play. If you cannot organize the data into a table first, you cannot decide whether the correction is needed.

Chi-Square Test of Independence

This is the test that Yates' correction adjusts. The correction changes the chi-square statistic so the test is less likely to overstate evidence against independence when expected counts are small. In a problem, you are still testing the same null hypothesis, but the calculation is slightly modified.

Expected Cell Frequencies

The decision to use Yates' correction depends on the expected counts, not just the observed counts. If expected cell frequencies are low, the standard chi-square approximation can be shaky. That is why teachers often have you check the expected counts before choosing the test formula.

Cell Frequencies

Observed cell frequencies are the raw counts you read directly from the table, and Yates' correction changes how those counts are compared to the expected ones. The correction shrinks the absolute difference in each cell by 0.5, which softens the chi-square contribution from each cell.

Is Yates' Correction on the Honors Statistics exam?

A quiz item or problem set question will usually give you a 2x2 contingency table and ask whether there is evidence of association between two categorical variables. Your job is to check the expected cell frequencies, decide whether the sample is small enough to make the standard chi-square test shaky, and then use the corrected test if needed. If the counts are low, you should recognize that Yates' correction makes the test more conservative, so the p-value may rise and the conclusion may change. In a worked solution, you may be asked to compare the uncorrected and corrected chi-square values, or to explain why the correction is being used before interpreting the result.

Key things to remember about Yates' Correction

• Yates' correction is a small-sample adjustment used with the chi-square test of independence in a 2x2 contingency table.

• It reduces the chi-square statistic by shrinking the observed minus expected difference by 0.5, which makes the p-value larger.

• You use it when expected cell frequencies are low, usually below 5, because the standard chi-square approximation can be too aggressive.

• The correction does not change the null hypothesis, it changes how cautious the test is about rejecting independence.

• If the table is larger than 2x2 or the expected counts are not small, you usually do not need Yates' correction.

Frequently asked questions about Yates' Correction

What is Yates' correction in Honors Statistics?

Yates' correction is an adjustment to the chi-square test of independence for a 2x2 contingency table. It makes the test more conservative when expected counts are small by reducing the chi-square statistic. That helps avoid overstating evidence of an association when the sample is limited.

When do you use Yates' correction?

You use it when you have a 2x2 table and the expected cell frequencies are small, often below 5. In that situation, the normal chi-square approximation can be less reliable. The correction is a safer choice when the data are sparse.

How does Yates' correction affect the p-value?

It usually increases the p-value because it lowers the chi-square test statistic. A larger p-value means weaker evidence against the null hypothesis of independence. That is why the correction is called conservative.

Is Yates' correction the same as the chi-square test?

It is part of the chi-square test, not a separate test. You are still testing independence between two categorical variables, but the calculation is adjusted for small counts. If the expected counts are large enough, the correction is usually unnecessary.