In AP Statistics, a Type I error occurs when the null hypothesis is actually true but you reject it anyway, claiming an effect that doesn't exist (a false positive). The probability of a Type I error equals the significance level α you chose for the test (EK 6.7.B).
A Type I error happens when the null hypothesis is true and your test rejects it anyway. In plain terms, you announce "there's a real effect!" when nothing is actually going on. That's why it's called a false positive. It doesn't mean you did the math wrong. Your sample data, just by random chance, looked unusual enough to push the p-value below α, even though the null was true the whole time.
Here's the part that makes the concept click. The probability of a Type I error IS the significance level α. When you set α = 0.05, you are literally saying "I'm willing to falsely reject a true null hypothesis 5% of the time." That's why choosing α isn't just a formality. If a Type I error has serious real-world consequences (like a quality engineer scrapping a perfectly good production line), you lower α to something like 0.01. The trade-off is that lowering α makes Type II errors more likely, so you can't make both risks disappear at once.
Type I error lives in Topic 6.7 (Potential Errors When Performing Tests) and is the direct target of four learning objectives. LO 6.7.A asks you to identify Type I vs. Type II errors, 6.7.B asks you to calculate their probabilities (P(Type I) = α), 6.7.C covers the factors that change error probabilities, and 6.7.D asks you to interpret errors in context and decide which one is more consequential. It also anchors Topic 7.1, where LO 7.1.A frames all of Unit 7 around the question "should I worry about error?" because random variation can produce errors in any inference procedure. Once you understand Type I error, every reject/fail-to-reject decision you make in Units 6 and 7, including two-sample t-tests in Topic 7.9, comes with a built-in awareness that "reject H₀" never means "H₀ is definitely false."
Keep studying AP Statistics Unit 9QyrTKQ1AuLyQwom
Type II Error (Unit 6)
These are mirror images. Type I rejects a true null; Type II fails to reject a false null. They're connected by a see-saw, since lowering α to reduce Type I risk raises the chance of a Type II error (EK 6.7.C), so you can't minimize both for free.
Significance Level (α) (Unit 6)
α isn't just a cutoff for p-values. It is literally the probability of a Type I error when the null is true. That's why CED 6.7.D says the consequences of a Type I error should drive your choice of α.
Null Hypothesis (Unit 6)
A Type I error only exists relative to H₀. You can't even define the error without first knowing what "no effect" claims, which is why writing hypotheses correctly comes before everything else in inference.
Two-Sample t-Tests for Means (Unit 7)
In Topic 7.9, the p-value is computed assuming μ₁ = μ₂ (LO 7.9.B). If you reject that null when the two population means really are equal, you've made a Type I error, like declaring one weight-loss program better when the programs are actually identical.
Multiple-choice questions love giving you a borderline p-value (like 0.049 against α = 0.05) and asking which error you risk, or asking what happens when a researcher changes α (moving from 0.05 to 0.01 cuts Type I risk but raises Type II risk). You're also expected to interpret a significant result skeptically, since a p-value of 0.022 in one study can balloon to 0.31 in a follow-up, hinting the first result may have been a false positive. On FRQs, error questions almost always come with context, like the 2025 FRQ about mean bedrooms in newly built houses. You'll need to (1) describe the error in context ("concluding the mean differs from 2.9 when it actually doesn't"), (2) describe a real-world consequence, and (3) sometimes argue which error is worse in that scenario. "Reject a true H₀" with no context earns nothing; the context is the point.
Type I error rejects a null that's true (you cry wolf when there's no wolf). Type II error fails to reject a null that's false (the wolf is there and you miss it). Quick memory hook: Type I = false Positive, Type II = false Negative, and the error types follow the same order as the words positive/negative. On the exam, identify which hypothesis is actually true in the scenario first, then ask what wrong decision was made. P(Type I) = α, while P(Type II) = 1 − power, so they're calculated completely differently too.
A Type I error means the null hypothesis is true but you rejected it, so you claimed an effect or difference that doesn't actually exist.
The probability of a Type I error equals the significance level α, so choosing α = 0.05 means accepting a 5% chance of a false positive when H₀ is true.
Lowering α reduces the chance of a Type I error but increases the chance of a Type II error, so the two risks trade off against each other.
On FRQs, you must describe a Type I error in the context of the problem and explain a real consequence; a generic textbook definition earns no credit.
Which error is more consequential depends on the situation, and that judgment is what should guide the choice of significance level (EK 6.7.D).
A statistically significant result is never proof; even a correct test procedure produces Type I errors purely from random variation in the sample.
A Type I error occurs when the null hypothesis is actually true but the test rejects it, a false positive. Per EK 6.7.B, its probability equals the significance level α of the test.
No. A Type I error happens even when every step is done correctly. Random variation in the sample produced data unusual enough to reject a true null, which CED LO 7.1.A flags as an unavoidable feature of inference.
Type I rejects a true null hypothesis (false positive); Type II fails to reject a false one (false negative). P(Type I) = α, while P(Type II) = 1 − power, so they're controlled in different ways and trade off against each other.
No, and this is a classic trap. The Type I error probability is α, the threshold you set before the test. The p-value is calculated from your data and changes with each sample. A p-value of 0.049 with α = 0.05 means you reject, with a 5% (not 4.9%) built-in false positive rate.
Lower the significance level, for example from α = 0.05 to α = 0.01. The catch, per EK 6.7.C, is that a smaller α makes the test less likely to reject anything, which raises the probability of a Type II error unless you also increase sample size.