On this page
What this lesson covers
A regression prediction interval combines a fitted value with a critical t-value and forecast standard error to describe uncertainty about a new observation.
The lesson forecasts a new company's P/E and builds its prediction interval. It explains how fit, sample size, the forecast point and variation in the independent variable affect forecast precision.
How do you calculate a prediction interval?
Add and subtract the critical t-value times the forecast standard error from the predicted value, assuming the regression conditions, including homoskedastic errors, hold.
Forecast standard error = se × sqrt[1 + 1/n + (Xf − mean X)² / sum of squared X deviations]
- Standard error of the estimate
- Sample size
- Forecast value of X
- Sample mean of X
Prediction interval = predicted value ± critical t-value × forecast standard error
- Predicted dependent-variable value
- Standard error of the forecast
For example, assume the regression conditions hold with homoskedastic errors, n = 6, estimated line P/E = 8 + 1.5 × growth, mean growth 8 and squared growth deviations totaling 112. At forecast growth 9, predicted P/E = 21.5; with se = 4, sf = 4 × sqrt[1 + 1/6 + (9 − 8)²/112] ≈ 4.337. Using the 95% critical t-value 2.776 with 4 degrees of freedom gives 21.5 ± 12.04, approximately 9.46 to 33.54.
Common mistakes
- Omitting the intercept leaves a slope term that is not the predicted value.
- Using the standard error of the estimate in a prediction interval substitutes the wrong measure for the forecast standard error.
- Using the forecast standard error alone omits the critical t-value required for the interval.
Try a practice question
Choose an answer, then check it to read the explanation for every option.
Continue studying in Quizara
In Quizara the lesson comes with more practice questions on the topic, explanations for every option and an AI tutor for your follow-up questions.
Written by the Quizara Editorial Team under our editorial standards.