I have a question from my online statistics class.

Question

The study on calories burned per minute by walking had data from adult women only. If Shawna used the regression equation to extrapolate the calories burned per minute for a young boy, she would get a nonsensical result. Shawna used the regression equation to make a prediction __________ the range of the data.

Answer 1

What comes to mind is that Shawna used the regression equation to make a prediction "outside" the range of the sample data. This case is a little different because the subjects are different, one for adult women and one for young boys. 

 

One guideline with regression is that you should be cautious about extrapolating values (predicting Y or the response variable) outside the range of your independent variable (X) sample data.

 

For example, if the values of your sample data for the X independent variable (in your case minutes walking) ranges from, say, 25 to 250, you should restrict your predictions for Y to X values between 25 and 250. Of course, you could always predict outside this range, such as x=19 or x= 273, but the confidence interval for the y-value in your regression equation will be wider than one from an x-value between your sample data of 25 to 250.

 

Regression is a linear relationship between X and Y, and it's possible that the relationship is not linear beyond 250 or less than 25--it could be curved or quadratic. You don't know because your sample only contains values between 25 and 250.

 

OK, I hope this helped!