Pooled Estimate

A pooled estimate is the combined sample proportion you use in Intro to Statistics when testing whether two population proportions are equal. It assumes the null hypothesis is true and gives one shared proportion for the standard error.

What is the Pooled Estimate?

A pooled estimate is the single proportion you calculate when you are testing two independent population proportions and your null hypothesis says the population proportions are equal. In Intro to Statistics, this usually shows up in a two-proportion z-test for a statement like H0: p1 = p2.

Instead of keeping the two sample proportions separate, you combine the successes from both groups and divide by the combined sample size. That gives you one shared estimate of the common population proportion, often written as p̂pooled. If group 1 has x1 successes out of n1 and group 2 has x2 successes out of n2, then the pooled estimate is (x1 + x2) / (n1 + n2).

Why pool at all? Because under the null hypothesis, you are acting as if the two populations have the same proportion. If that is the case, it makes sense to estimate that shared proportion from all of the data together rather than treating the samples as if they came from separate populations with different true proportions.

The pooled estimate is not the difference between the sample proportions, and it is not the final answer of the test. It is an input for the standard error in the z-test statistic. That means it affects how spread out the sampling distribution looks under H0, which then changes the z-score you compute.

A quick example: if one group has 18 successes out of 40 and the other has 12 successes out of 30, the pooled estimate is (18 + 12) / (40 + 30) = 30/70 ≈ 0.429. That number becomes the best combined estimate only for the null model where the proportions are assumed equal. If the null is not reasonable, pooling is the wrong move.

A common mistake is to average the two sample proportions without weighting by sample size. That can distort the estimate, especially when the groups are different sizes. The pooled estimate uses all observations, so bigger samples count more in the combined proportion.

Why the Pooled Estimate matters in Intro to Statistics

The pooled estimate is the piece that connects the null hypothesis to the actual calculation in a two-proportion z-test. In Intro to Statistics, you are not just comparing two sample percentages by eye. You are deciding whether the observed difference in p1 - p2 is large enough to suggest a real difference in the populations, or whether random sampling could easily explain it.

That decision depends on the standard error, and the standard error depends on the pooled estimate when H0 says the proportions are equal. So if you set up the test wrong, or forget to pool when the null requires it, your z-score will be off from the start.

It also teaches you something about model-building in stats: calculations change depending on the hypothesis. When you are estimating a proportion for a confidence interval, you usually do not pool. When you are testing equality of two proportions, you do. That distinction shows up often on problem sets because it checks whether you can match the formula to the goal.

The idea also makes interpretation cleaner. Pooling says, “If there is no difference, what single proportion best summarizes both groups together?” That one sentence is the logic behind the formula. Once you see that logic, the rest of the test, including the z statistic and p-value, makes more sense.

How the Pooled Estimate connects across the course

Population Proportion

The pooled estimate is built from sample data to estimate a population proportion under the assumption that two populations share the same p. If you do not know what a population proportion means, it is harder to see why combining the samples is justified in the null model.

Null Hypothesis

Pooling only happens because the null hypothesis says the two population proportions are equal. The null gives you the assumption that makes one shared estimate sensible, instead of separate estimates for each group.

Standard Error

The pooled estimate feeds directly into the standard error for the two-proportion z-test. That standard error tells you how much p1 - p2 would typically vary if the null hypothesis were true.

SE(p̂1 - p̂2)

This is the exact standard error formula used in the two-proportion z-test. It uses the pooled estimate, so it is the step where the pooled proportion turns into the spread measure for the test statistic.

Is the Pooled Estimate on the Intro to Statistics exam?

On a problem set or quiz, you will usually calculate the pooled estimate after writing the null hypothesis for two independent proportions. The task looks like this: find the number of successes in both groups, add them, divide by the total sample size, and then plug that value into the standard error formula for p1 - p2.

If a question asks whether two websites have different click-through rates or whether two classes have different pass rates, pooling tells you what proportion to use under H0. You may also need to explain why pooling is appropriate, which means saying that the null hypothesis treats the two population proportions as equal.

A common test question gives you a confidence interval setup and a hypothesis test setup side by side. If you are testing equality, you pool. If you are building an interval, you usually do not. That distinction is one of the easiest ways for intro stats questions to check whether you actually know the method or are just memorizing a formula.

The Pooled Estimate vs SE(p̂1 - p̂2)

The pooled estimate is the combined proportion you calculate first. SE(p̂1 - p̂2) is the standard error formula that uses that pooled estimate. One is the input, the other is the spread calculation for the test statistic.

Key things to remember about the Pooled Estimate

• The pooled estimate is the single combined proportion used in a two-proportion z-test when the null hypothesis says the population proportions are equal.

• You find it by adding the successes from both samples and dividing by the total sample size, not by averaging the two sample proportions equally.

• It is used to compute the standard error for the test statistic, so it changes the z-score and the p-value.

• Pooling makes sense only under the null model, where both groups are treated as if they come from the same population proportion.

• Do not use the pooled estimate in place of a confidence interval estimate, because pooling is tied to hypothesis testing, not interval estimation.

Frequently asked questions about the Pooled Estimate

What is pooled estimate in Intro to Statistics?

A pooled estimate is the combined proportion you calculate when testing whether two independent population proportions are equal. It uses successes from both groups together, then divides by the total sample size. In a two-proportion z-test, it stands in for the common proportion assumed by the null hypothesis.

How do you find the pooled estimate?

Add the number of successes in both samples, then divide by the total sample size across both groups. In symbols, p̂pooled = (x1 + x2) / (n1 + n2). A common mistake is to average p̂1 and p̂2 without weighting by sample size, which can give the wrong value.

Why do you pool proportions for a hypothesis test?

You pool because the null hypothesis says the two population proportions are equal, so one shared estimate makes sense. The pooled value gives the standard error the right baseline under H0. If you are making a confidence interval instead, you usually do not pool.

Is the pooled estimate the same as the sample proportion difference?

No. The pooled estimate is one combined proportion, while p1 - p2 is the difference between the two sample proportions. The pooled estimate is used inside the standard error formula, and the difference is what the z-test is measuring.