[Stata] - Logit Results: 2 variables only have coeff output no p value???

#1
I am using a binary dependent variable with multiple explanatory dummy variables. I used the logit function to get my statistical model.

However 2 of the variables only show the coeff value and not the p-value to show it's significance. What are some reasons why this output may have happened?

I tried to embed the screen shot in the thread but it isn't working. I uploaded it to flickr to better illustrate what is going on: http://www.flickr.com/photos/35531322@N02/4952706657/

Stata 10
 
Last edited:
#2
I am using a binary dependent variable with multiple explanatory dummy variables. I used the logit function to get my statistical model.

However 2 of the variables only show the coeff value and not the p-value to show it's significance. What are some reasons why this output may have happened?

I tried to embed the screen shot in the thread but it isn't working. I uploaded it to flickr to better illustrate what is going on: http://www.flickr.com/photos/35531322@N02/4952706657/

Stata 10
this might be too late but i would be interested to know too...

my guess is that it's due to collinearity... your model specification seems to be very poor judging by the output you have posted... i think if you run a specification test you'll actually get a negative chi-squared...
 

Link

Ninja say what!?!
#3
You have a huge number of covariates and only 76 observations. What's happening is that there is only one (or very few) observations where R2=1 and Tech2=1.

Another possible explanation is that you've put everyone into the model and so there's no control group. Lets say for example that I have a variable called "test" that has only 3 possible values (1,2,3). If I make dummy variables test1, test2, and test3 and throw them all into my model, there won't be a baseline group to compare to. In that situation, funky things would start happening.

HTH