Hi, Could I please have help understanding part i) of Qu8 in 2023 April Paper I understand the test is H0: Beta1 = 0 H1: Beta1 not equal to 0 I thought I could is the t distribution result at the bottom of p24 beta hat1 - 0 / se ~ t_n-2 however the solution says you can't use the t distribution , and was a common mistake in the paper. Please could I have explanation as to why the t distribution is not appropriate and why 2*[Z < test statistic] is? Thanks in advance - any help is much appreciated
This is a GLM question not a linear regression model, so the result at the bottom of p24 of the Tables is not appropriate, we consider the significance of the parameter from the estimator being asymptotically normal for large n, mean beta, variance from the CRLB, the standard error of beta, see Section 5.2 Chapter 13 for more detail.