The difference in population proportions, written p₁ - p₂, is the parameter comparing the true proportion of successes in two distinct populations; in AP Stats you estimate it with a two-sample z-interval for a difference between population proportions (Topic 6.8).
The difference in population proportions, p₁ - p₂, is what you get when you subtract the true proportion of one group from the true proportion of another. It answers questions like "Is the proportion of households with internet access higher in County 1 than in County 2, and by how much?" Since you almost never know the true proportions, you estimate this parameter using the difference in sample proportions, p̂₁ - p̂₂, and build a confidence interval around it.
The procedure the CED names for this is the two-sample z-interval for a difference between population proportions. The interval is (p̂₁ - p̂₂) ± z*√(p̂₁(1-p̂₁)/n₁ + p̂₂(1-p̂₂)/n₂). Before you calculate anything, you have to verify conditions. The data must come from two independent random samples or a randomized experiment, each sample must be at most 10% of its population if sampling without replacement, and all four counts (n₁p̂₁, n₁(1-p̂₁), n₂p̂₂, and n₂(1-p̂₂)) must be large enough for the sampling distribution of p̂₁ - p̂₂ to be approximately normal. The big interpretive payoff is zero. If the entire interval is above 0, you have convincing evidence p₁ > p₂. If it's below 0, evidence points the other way. If the interval contains 0, you can't rule out that the two proportions are equal.
This term lives in Unit 6 (Inference for Categorical Data: Proportions), specifically Topic 6.8, and it carries four learning objectives. You have to identify the right procedure (AP Stats 6.8.A), verify the conditions (AP Stats 6.8.B), calculate the interval (AP Stats 6.8.C), and turn it into an interval estimate with real units in context (AP Stats 6.8.D). It's also the bridge from one-sample to two-sample thinking. Everything you learned about a single proportion (sample proportion, margin of error, critical value) gets repackaged here for a comparison question, which is exactly the kind of question real studies ask. Does the treatment work better than the placebo? Do two cities exercise at different rates? One nice break: the formula isn't on the formula sheet, but you don't need to memorize it. You can build it from the general estimate ± (critical value)(standard error) structure and the standard error formulas that are provided.
Keep studying AP Statistics Unit 6
Sampling Distribution of p̂₁ - p̂₂ (Unit 5)
The whole two-sample interval rests on Unit 5's result that the difference of two sample proportions has its own sampling distribution. The normality check in 6.8.B (all four success/failure counts big enough) is just you confirming that distribution is approximately normal before you trust the z* multiplier.
Confidence Interval for One Proportion (Unit 6)
Topic 6.8 is Topic 6.2 with a subtraction sign. Same skeleton of statistic ± (critical value)(standard error), but now the statistic is p̂₁ - p̂₂ and the standard error adds the variability from both samples under one square root. Variances add even when you're subtracting.
Independence Condition and Independent Samples (Unit 6)
Two-sample procedures only work when the two groups don't influence each other. That's why the CED requires two independent random samples or random assignment in an experiment, plus the 10% condition for each sample separately. If the samples are paired or overlapping, this procedure is the wrong tool.
Two-Sample Inference for Means (Unit 7)
Unit 7 runs this exact play again with quantitative data. The difference of two means, x̄₁ - x̄₂, gets the same logic of conditions, a standard error that combines both samples, and a check for whether 0 is in the interval. Master the proportions version and the means version feels familiar.
Multiple-choice questions love three angles here. First, computation setups where you're given counts like 240 of 300 households in one county and 280 of 400 in another and asked for the interval or its standard error. Second, interpretation, like being handed a 95% interval of (0.05, 0.15) for pₘ - p_f and asked for the most appropriate conclusion (the whole interval is positive, so there's convincing evidence pₘ > p_f). Third, conditions, where the wrong answers are intervals built without checking independence or the large counts condition. On the free-response side, two-sample proportion inference is a classic full-inference question. You name the procedure (two-sample z-interval for a difference between population proportions), define both parameters in context, check every condition explicitly, compute, and interpret the interval in context with a sentence about what the difference means for the actual populations. Forgetting to define p₁ and p₂ or skipping a condition is where points die.
p₁ - p₂ is the parameter, the fixed but unknown true difference between two populations. p̂₁ - p̂₂ is the statistic, the difference you actually compute from your two samples, and it changes from sample to sample. The confidence interval is built FROM p̂₁ - p̂₂ to estimate p₁ - p₂. On the exam, interpreting an interval as capturing the sample difference (which you already know exactly) instead of the population difference is a classic credit-killer.
The difference in population proportions, p₁ - p₂, is the parameter you estimate when comparing the true proportion of successes in two distinct groups.
The correct procedure is a two-sample z-interval for a difference between population proportions, computed as (p̂₁ - p̂₂) ± z*√(p̂₁(1-p̂₁)/n₁ + p̂₂(1-p̂₂)/n₂).
Before calculating, verify independence (two independent random samples or a randomized experiment, plus the 10% condition for each sample) and normality (all four counts n₁p̂₁, n₁(1-p̂₁), n₂p̂₂, and n₂(1-p̂₂) are large enough).
Zero is the magic number for interpretation, because an interval entirely above or below 0 gives convincing evidence the two population proportions differ, while an interval containing 0 does not.
You don't need to memorize the formula since it follows from estimate ± (critical value)(standard error) using the standard error formulas on the provided formula sheet.
Always interpret the interval in context with units, saying you are confident the interval captures the true difference between the two population proportions.
It's the parameter p₁ - p₂, the true difference between the proportion of successes in two populations. In Topic 6.8 you estimate it with a two-sample z-interval built from the sample proportions p̂₁ and p̂₂.
No. An interval containing 0 means you don't have convincing evidence of a difference, which is not the same as proving the proportions are equal. A confidence interval can never prove two parameters are exactly the same.
p₁ - p₂ is the unknown population parameter you want to learn about, while p̂₁ - p̂₂ is the statistic you compute from your data. The interval uses p̂₁ - p̂₂ as its center to estimate p₁ - p₂, and your interpretation must be about the population difference, not the sample one.
No. The CED explicitly says interval formulas don't appear on the formula sheet but don't need to be memorized, because you can build them from the general estimate ± (critical value)(standard error) structure plus the standard error formulas that are provided.
You need two independent random samples or a randomized experiment, the 10% condition for each sample when sampling without replacement (n₁ ≤ 10%N₁ and n₂ ≤ 10%N₂), and large counts, meaning n₁p̂₁, n₁(1-p̂₁), n₂p̂₂, and n₂(1-p̂₂) are all big enough for approximate normality.