In AP Statistics, p-hat (p̂) is the sample proportion, calculated as the number of successes divided by the sample size n. It serves as the point estimate for the population proportion p and sits at the center of a one-sample z-interval, written p̂ ± z*√(p̂(1-p̂)/n).
p-hat (written p̂) is the proportion of "successes" in your sample. Count how many individuals have the characteristic you care about, divide by the sample size n, and that's p̂. If 64 out of 200 surveyed teens say they've skipped breakfast, p̂ = 64/200 = 0.32.
The "hat" is the whole point. In statistics notation, a hat means "this is an estimate computed from data." The plain letter p is the true population proportion, a parameter you almost never get to see. p̂ is a statistic, the thing you actually calculate, and it changes from sample to sample. That's why the CED treats p̂ as the point estimate for p. When you build a confidence interval for a population proportion (Topic 6.2), p̂ does double duty. It's the center of the interval, and it's also plugged into the standard error formula SE = √(p̂(1-p̂)/n), since you don't know the real p to use instead.
p̂ is the workhorse of Unit 6 (Inference for Categorical Data: Proportions) and shows up in every learning objective in Topic 6.2. You use p̂ to verify the large counts condition (LO 6.2.B), to compute the standard error and margin of error z*√(p̂(1-p̂)/n) (LO 6.2.C), and to build the full interval p̂ ± z*√(p̂(1-p̂)/n) (LO 6.2.D-E). It also reaches backward into Unit 5, where the sampling distribution of p̂ is what makes inference work in the first place, and forward into significance tests for proportions. If you can't keep p̂ and p straight, basically all of categorical inference falls apart.
Keep studying AP Statistics Unit 6
Population Proportion (Unit 6)
p̂ exists to estimate p. The population proportion is the unknown truth about everyone; p̂ is your best guess from a sample. Every confidence interval in Topic 6.2 is really a statement about how close p̂ probably is to p.
Sampling Distribution of p̂ (Unit 5)
Take many samples and each one gives a different p̂. Unit 5 shows that those p̂ values pile up in an approximately normal shape centered at p when the conditions hold. That normal shape is exactly why a z-interval is the right tool in Unit 6.
Margin of Error (Unit 6)
The margin of error is built from p̂ itself, since SE = √(p̂(1-p̂)/n). A p̂ near 0.5 gives the widest interval, which is why sample-size problems often assume p̂ = 0.5 as the worst case.
Confidence Interval (Unit 6)
Every interval estimate follows the pattern point estimate ± margin of error, and for one proportion the point estimate is p̂. The interval is literally centered on it.
No released FRQ asks you to define p̂ by itself, but you can't survive a proportion inference FRQ without using it correctly. On the free response, you're expected to compute p̂ from a two-way table or survey count, check the large counts condition using p̂ (both np̂ and n(1-p̂) at least 10), and write the interval p̂ ± z*√(p̂(1-p̂)/n) with correct notation. Multiple choice loves notation traps, like asking whether a number describes p or p̂, or whether the standard error uses p̂ (confidence intervals) or p₀ (significance tests). One quiet exam fact helps here. The interval formula isn't printed on the formula sheet as one piece, but the standard error formula is, so you build the interval from point estimate ± (critical value)(standard error).
p is the true proportion in the entire population, a parameter that is fixed but unknown. p̂ is the proportion in one sample, a statistic that varies from sample to sample. You calculate p̂; you infer about p. On FRQs, defining your parameter as "p̂ = the true proportion of all..." is a classic notation error that costs credit, because the true proportion is p, not p̂.
p̂ is the sample proportion, calculated as the number of successes divided by the sample size n.
p̂ is a statistic and a point estimate for the parameter p; the hat always signals a value computed from sample data.
The standard error of p̂ is √(p̂(1-p̂)/n), and the one-sample z-interval is p̂ ± z*√(p̂(1-p̂)/n).
The large counts condition uses p̂ in confidence intervals: check that np̂ ≥ 10 and n(1-p̂) ≥ 10 before building the interval.
Because p̂ varies from sample to sample, it has its own sampling distribution (Unit 5), which is approximately normal when the conditions are met.
Larger samples make p̂ a more precise estimate, since the standard error shrinks as n grows.
p-hat (p̂) is the sample proportion, found by dividing the number of successes by the sample size. If 45 of 150 people in a sample own a dog, p̂ = 45/150 = 0.30, and it serves as your point estimate for the population proportion p.
p is the true population proportion, a fixed but unknown parameter. p̂ is the proportion in your sample, a statistic that changes with every new sample. You use p̂ to estimate p; writing "p̂ = the true proportion" on an FRQ is a notation error.
No. p̂ is an observed proportion from data after sampling, not a theoretical probability. In a binomial setting, p is the probability of success; p̂ is just your sample-based estimate of it.
For a confidence interval, you use p̂, so SE = √(p̂(1-p̂)/n), because p is unknown. In a significance test you use the hypothesized value p₀ instead. Mixing these up is one of the most common Unit 6 mistakes.
Not as a single formula. The formula sheet gives the standard error of p̂, and you assemble the interval yourself as point estimate ± (critical value)(standard error), which becomes p̂ ± z*√(p̂(1-p̂)/n).
Connect this key term to the AP exam workflow: review the course, practice questions, and check related study tools.
Review units, study guides, and course resources.
Check this vocabulary in multiple-choice context.
Apply key concepts in written AP responses.
Estimate the exam score you are working toward.
Review the highest-yield facts before practice.
Put the full course together before test day.