---
title: "Prediction Interval — AP Stats Definition & Exam Guide"
description: "A prediction interval estimates the response value for one individual at a given x. Learn how it differs from a confidence interval and how AP Stats tests it."
canonical: "https://fiveable.me/ap-stats/key-terms/prediction-interval"
type: "key-term"
subject: "AP Statistics"
unit: "Unit 5"
---

# Prediction Interval — AP Stats Definition & Exam Guide

## Definition

In AP Statistics, a prediction interval is an interval estimate for the response variable value of a single individual with a specific value of the explanatory variable, built around the predicted value ŷ from the regression equation (Topic 2.6).

## What It Is

A prediction interval answers the question "if I plug in one specific x, what [range](/ap-stats/key-terms/range "fv-autolink") of y-values should I expect for one individual?" The [regression line](/ap-stats/key-terms/regression-line "fv-autolink") gives you a single point prediction, ŷ = a + bx (that's EK DAT-1.D.2). But real individuals scatter around the line, so a single number is almost certainly a little off. A prediction interval wraps a margin of error around ŷ to capture where that one individual's actual y-value is likely to land.

The basic structure is ŷ ± ([critical value](/ap-stats/key-terms/critical-value "fv-autolink"))(standard error of prediction). The standard error measures how much actual points typically miss the line. Here's the part that trips people up. A prediction interval is for *one individual*, not for the average of all individuals at that x. Predicting one specific house's price is harder than predicting the average price of all 2000-square-foot houses, so prediction intervals are always wider than the corresponding confidence interval for the mean response.

## Why It Matters

This term lives in **[Unit 2](/ap-stats/unit-2 "fv-autolink"): Exploring Two-Variable Data**, specifically **Topic 2.6 Linear Regression Models**, supporting learning objective **2.6.A** (calculate a [predicted response value](/ap-stats/key-terms/predicted-value "fv-autolink") using a linear regression model). The prediction interval is the honest version of a prediction. Instead of pretending ŷ is exact, it admits uncertainty and gives a range. It also connects directly to **DAT-1.D.3** on extrapolation, because the reliability of any prediction (point or interval) falls apart when you plug in an x-value outside the range of your data. Understanding prediction intervals in Unit 2 sets you up for the inference mindset of Units 6-9, where everything becomes "estimate ± margin of error."

## Connections

### Predicted value ŷ (Unit 2)

The predicted value ŷ = a + bx is the [center](/ap-stats/key-terms/center "fv-autolink") of every prediction interval. The interval is just ŷ with an honest margin of error attached, acknowledging that real points scatter around the regression line.

### Extrapolation (Unit 2)

A prediction interval is only trustworthy for x-values inside the range of the original data. Extrapolate beyond that range and the [interval](/ap-stats/unit-1/representing-quantitative-variable-with-graphs/study-guide/VWtyLVDvjzEgtbAi6v6j "fv-autolink") becomes unreliable no matter how nicely you calculated it, per EK DAT-1.D.3.

### Confidence intervals for inference (Units 6-9)

[Prediction](/ap-stats/key-terms/prediction "fv-autolink") intervals share the same DNA as confidence intervals, estimate ± (critical value)(standard error). The difference is the target. Confidence intervals chase a population parameter like a mean or a slope, while a prediction interval chases one individual's value, which is why it's wider.

### Inference for slopes (Unit 9)

Unit 9 builds confidence intervals for the slope of the regression line. Don't mix them up with prediction intervals. A slope interval estimates the relationship itself (b), while a prediction interval estimates one individual's y-value at a chosen x.

## On the AP Exam

Multiple-choice questions typically hand you a regression equation and a standard error, then ask you to build or interpret the interval. For example, given ŷ = 32 - 2.5x for used car prices with a standard error of 1.8, you'd find ŷ at x = 6 and stretch out roughly two standard errors on each side for a 95% interval. Same idea with house prices, ŷ = 85 + 0.15(2000) = 385 thousand, then ± about 2(25). The 2026 FRQ Q6 used regression on baseball data, and predictions from regression models show up routinely in FRQ Question 6, the investigative task. The key skills are computing ŷ correctly, interpreting the interval in context ("we expect this individual 6-year-old car's price to fall between..."), and flagging extrapolation when the given x is outside the data range.

## prediction interval vs Confidence interval for the mean response

Both intervals are centered at the same ŷ, but they estimate different things. A confidence interval for the mean response estimates the *average* y-value for ALL individuals with that x. A prediction interval estimates the y-value for ONE individual with that x. Individuals vary more than averages do, so the prediction interval is always wider. If an exam question asks about "a randomly selected" or "a particular" car, house, or person, you want the prediction interval.

## Key Takeaways

- A prediction interval estimates the response value for a single individual at a specific x-value, not the average for all individuals at that x.
- Every prediction interval is centered at the predicted value ŷ = a + bx and extends out by a margin of error based on the standard error of prediction.
- Prediction intervals are always wider than confidence intervals for the mean response, because predicting one individual is harder than predicting an average.
- A prediction interval is only reliable when the x-value is inside the range of data used to build the regression line; extrapolating makes any prediction less trustworthy.
- When interpreting, name the individual and the context, for example "we expect the price of this particular 6-year-old car to be between $13,400 and $20,600."

## FAQs

### What is a prediction interval in AP Stats?

It's an interval estimate for the response variable value of one individual with a specific explanatory variable value. It's built as ŷ ± (critical value)(standard error of prediction), centered at the predicted value from the regression equation in Topic 2.6.

### What's the difference between a prediction interval and a confidence interval?

A confidence interval estimates a population parameter, like the mean response for all individuals at a given x. A prediction interval estimates one individual's actual y-value at that x, so it's always wider because individuals scatter more than averages.

### Is a prediction interval always wider than a confidence interval?

Yes, at the same x-value and confidence level. The prediction interval has to account for both the uncertainty in the regression line and the natural scatter of individual points around it, so it's strictly wider.

### How do you calculate a prediction interval from a regression equation?

First compute ŷ = a + bx for the given x, then add and subtract the margin of error. For example, with ŷ = 85 + 0.15x and x = 2000, the point prediction is 385 (thousand dollars), and a 95% interval stretches roughly two standard errors of prediction (here 25) on each side.

### Can you use a prediction interval for an x-value outside your data?

You can compute one, but you shouldn't trust it. That's extrapolation (EK DAT-1.D.3), and predictions get less reliable the further the x-value sits outside the interval of x-values used to fit the regression line.

## Related Study Guides

- [5.3 Linear Regression Models](/ap-stats/unit-5/linear-regression-models/study-guide/PSt5cfDuvB5nu60DHulR)

## Structured Data

```json
{"@context":"https://schema.org","@graph":[{"@type":"LearningResource","@id":"https://fiveable.me/ap-stats/key-terms/prediction-interval#resource","name":"Prediction Interval — AP Stats Definition & Exam Guide","url":"https://fiveable.me/ap-stats/key-terms/prediction-interval","learningResourceType":"Concept explainer","educationalLevel":"AP® / High School","about":{"@id":"https://fiveable.me/ap-stats/key-terms/prediction-interval#term"},"audience":{"@type":"EducationalAudience","educationalRole":"student"},"dateModified":"2026-06-11T05:22:55.013Z","isPartOf":{"@type":"Collection","name":"AP Statistics Key Terms","url":"https://fiveable.me/ap-stats/key-terms"},"publisher":{"@type":"Organization","name":"Fiveable","url":"https://fiveable.me"}},{"@type":"DefinedTerm","@id":"https://fiveable.me/ap-stats/key-terms/prediction-interval#term","name":"prediction interval","description":"In AP Statistics, a prediction interval is an interval estimate for the response variable value of a single individual with a specific value of the explanatory variable, built around the predicted value ŷ from the regression equation (Topic 2.6).","url":"https://fiveable.me/ap-stats/key-terms/prediction-interval","inDefinedTermSet":{"@type":"DefinedTermSet","name":"AP Statistics Key Terms","url":"https://fiveable.me/ap-stats/key-terms"}},{"@type":"FAQPage","mainEntity":[{"@type":"Question","name":"What is a prediction interval in AP Stats?","acceptedAnswer":{"@type":"Answer","text":"It's an interval estimate for the response variable value of one individual with a specific explanatory variable value. It's built as ŷ ± (critical value)(standard error of prediction), centered at the predicted value from the regression equation in Topic 2.6."}},{"@type":"Question","name":"What's the difference between a prediction interval and a confidence interval?","acceptedAnswer":{"@type":"Answer","text":"A confidence interval estimates a population parameter, like the mean response for all individuals at a given x. A prediction interval estimates one individual's actual y-value at that x, so it's always wider because individuals scatter more than averages."}},{"@type":"Question","name":"Is a prediction interval always wider than a confidence interval?","acceptedAnswer":{"@type":"Answer","text":"Yes, at the same x-value and confidence level. The prediction interval has to account for both the uncertainty in the regression line and the natural scatter of individual points around it, so it's strictly wider."}},{"@type":"Question","name":"How do you calculate a prediction interval from a regression equation?","acceptedAnswer":{"@type":"Answer","text":"First compute ŷ = a + bx for the given x, then add and subtract the margin of error. For example, with ŷ = 85 + 0.15x and x = 2000, the point prediction is 385 (thousand dollars), and a 95% interval stretches roughly two standard errors of prediction (here 25) on each side."}},{"@type":"Question","name":"Can you use a prediction interval for an x-value outside your data?","acceptedAnswer":{"@type":"Answer","text":"You can compute one, but you shouldn't trust it. That's extrapolation (EK DAT-1.D.3), and predictions get less reliable the further the x-value sits outside the interval of x-values used to fit the regression line."}}]},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"AP Statistics","item":"https://fiveable.me/ap-stats"},{"@type":"ListItem","position":2,"name":"Key Terms","item":"https://fiveable.me/ap-stats/key-terms"},{"@type":"ListItem","position":3,"name":"Unit 5","item":"https://fiveable.me/ap-stats/unit-5"},{"@type":"ListItem","position":4,"name":"prediction interval"}]}]}
```