EViews.com

Hi,everyone.I have a big problem please help me out!

I run a logit model with 30 independent variables and want to check whether the first 4 variables are jointly significant or not. I do the Wald test with C(2)=C(3)=C(4)=C(5)=0. The result show that the p-value of F statistic is 0.15 which means the first 4 variables are jointly insignficant. However, when I took a look the logit model itself, I realized that the fourth variable which refers to C(5) has the p-value of 0.005. So the individual variable is significant. How come?

My question: Is it possible that individual variable is significant but all the variables as a whole are insignificant??
Thx a lot!

This is perfectly possible. The single coefficient test is more powerful, in this particular case.

This is perfectly possible. The single coefficient test is more powerful, in this particular case.

Would you please tell me why that happened? What's the reason behind that?
I suppose that if the result of joint significant test is insignificant, it means every individual variable is insignificant plus all variables as a whole are insignificant.

Loosely speaking, individual confidence intervals form rectangles and joint confidence regions form ellipses. Some areas are in one confidence interval but not the other. This can lead to statistically correct, but logically inconsistent, results. Perhaps it helps to remember that the tests don't conclude that a hypothesis is true or false, but only that there is a certain probability of the hypothesis being true or false.

Thanks a lot! Startz