AP Statistics · Topic 5.3 · Unit 5

AP Stats 5.3: Linear Regression Models

By Jude Wallis · Published July 28, 2026

Topic 5.3 in the Fall 2026 AP Statistics course is linear regression models. When the form is linear you predict the response with y-hat = a + bx, where a is the intercept, b the slope, and x the explanatory value. Predicting inside the data range is interpolation; outside it is extrapolation.

AP Statistics: Unit 5 (topics 5.3). Topic 5.3 (Linear Regression Models) is part of Unit 5 in the Fall 2026 CED. It focuses on using the model y-hat = a + bx to predict a response and on telling interpolation from extrapolation.

The linear regression model

Topic 5.3 turns a linear pattern into a prediction tool. When the form of the relationship between the explanatory variable $x$ and the response variable $y$ looks linear, you can approximate it with a linear regression model, a linear equation that uses $x$ to predict $y$ . The provided formula sheet writes it as

\hat{y} = a + bx

Here $\hat{y}$ (say "y-hat") is the predicted response, $a$ is the y-intercept, $b$ is the slope, and $x$ is the value of the explanatory variable. The hat is the whole point, because the model returns an estimated average response for a given $x$ , not the exact value any single individual will show. Topic 5.5 gives $a$ and $b$ their contextual meaning; topic 5.3 focuses on using the model to predict.

Making a prediction

To predict, substitute a value of $x$ into the model and compute $\hat{y}$ . If a model for temperature from cricket chirps is $\hat{y} = 40 + 0.25x$ , then at $x = 120$ chirps per minute the predicted temperature is $40 + 0.25 \times 120 = 70$ degrees Fahrenheit.

The arithmetic is easy; the judgment is knowing when to trust the answer. That depends entirely on where $x$ sits relative to the data used to build the line, which is the difference between interpolation and extrapolation covered next. Keep the units of the response variable with every prediction, so the cricket model returns 70 degrees Fahrenheit rather than a bare 70.

Interpolation versus extrapolation

Interpolation is predicting for an $x$ -value inside the interval of x-values used to determine the line. Those predictions are the trustworthy ones, because data supports the pattern there.

Extrapolation is predicting for an $x$ -value beyond that interval. The AP course states plainly that a predicted value becomes less reliable the further you extrapolate, since nothing confirms the linear trend continues past the observed data. A cricket-chirp line built from rates of 80 to 160 chirps per minute says little about 300 chirps, and a decreasing line pushed far enough can even return an impossible value such as a negative price. Always check that your $x$ falls within the observed range before trusting a prediction.

A prediction is an average, not a promise

One habit keeps your answers honest. The value $\hat{y}$ is a predicted, or average, response for a given $x$ , not the exact value any single individual will show. Two crickets chirping at the same rate can sit in slightly different temperatures, and the model reports the center of that spread.

For that reason, describe results with hedged language such as the model predicts or on average, and round only at the final step to avoid piling up rounding error. A prediction stated with its units and a note about its reliability is a complete answer; a bare number is not.

Predict, then classify interpolation or extrapolation

A linear regression model predicting temperature (degrees Fahrenheit) from cricket chirps per minute is $\hat{y} = 40 + 0.25x$ . The data used to fit it had chirp rates from 80 to 160 per minute. Predict the temperature at 120 chirps per minute and at 200 chirps per minute, and state whether each prediction is interpolation or extrapolation.

Predict at $x = 120$ . $\hat{y} = 40 + 0.25 \times 120 = 40 + 30 = 70$ degrees Fahrenheit.
Classify $x = 120$ . Because 120 falls between 80 and 160, this prediction is interpolation and is reliable.
Predict at $x = 200$ . $\hat{y} = 40 + 0.25 \times 200 = 40 + 50 = 90$ degrees Fahrenheit.
Classify $x = 200$ . Because 200 is beyond the largest observed rate of 160, this prediction is extrapolation and is less reliable.

At 120 chirps per minute the model predicts 70 degrees Fahrenheit (interpolation, reliable). At 200 chirps per minute it predicts 90 degrees Fahrenheit (extrapolation, less reliable because 200 is outside the 80-to-160 range the line was built from).

Frequently asked questions

What does the hat on y-hat mean?

The hat marks a predicted value. Plain $y$ is an observed data value, while $\hat{y}$ is what the regression model estimates for a given $x$ . The gap between them, $y - \hat{y}$ , is the residual covered in topic 5.4.

Is extrapolation always wrong?

It is not forbidden, but it is risky. A prediction grows less reliable the further past the observed x-values you go, because no data confirms the trend holds there. Report extrapolated predictions with caution, and note that they assume the linear pattern continues.