Regression Estimated variance and Chi-square distribution

#1
Hi all

* If i have a data set and I model it using linear regression I estimate intercept, slope and variance Ymodel=a+bx and the corresponding MSE to estimate variance StdevModle^2 .

* The next NEW set of data come which follows same model as data in step 1. i calculate Sum of (YObservedNEW-YExpectedfrom Model)^2/StdModel^2, will this still follows chi-square as the denominator is not the population variance but estimated from data set of step1.


Thanks for your help!

Mod Note: Please only post a thread once. Thank you.
 
Last edited by a moderator: