Is it always true that in the simple linear regression model the width of the prediction interval corresponding to a new observation x=xo increases linearly with xo?
Thanks in advance
Is it always true that in the simple linear regression model the width of the prediction interval corresponding to a new observation x=xo increases linearly with xo?
Thanks in advance
The width of prediction intervals does not increase linearly; the uncertainty in parameters involving constant and slope induces a spreading out that increases a little more slowly than linearly as you move further away from the mean of the predictor variable; the limits of the prediction interval will asymptote to a pair of lines that cross at $(\bar{x},\bar{y})$.
There's an intuitive illustration of the way the confidence interval behaves here: Shape of confidence interval for predicted values in linear regression
The prediction interval has similar behavior but includes an additional term that has a constant spread.