The expected value of the ordered residuals
under normality is displayed in the column titled: E(Resid). To find the expected value of the ordered
residuals under normality, we utilize the following basic assumptions:
1.
The
expected value of the error terms for simple linear regression is zeros.
2.
The
standard deviation of the error terms is estimated by square root of the MSE
(RMSE).
Therefore, a good approximation of the expected value of
the ith smallest observation in a
random sample of n is:
RMSE*[Z((I-0.375)/(n+0.25))]
Where z(A) denotes the (A)100 percentile of the standard
normal distribution.