1
$\begingroup$

I have two datasets:

  1. I have an exam score (pretend is the GRE) for all students that took the exam from 2000-2005, although I do not have student's private information (names, id,etc) I several variables along with the score such as socioeconomic status, race, parent's education, type of school, etc.

  2. I have the average GRE score with standard deviation for all students that were accepted into a certain university for the same years 2000-2005. So I know that in med school the average GRE score was 60.05 with a standard deviation of 1.95.

Now, how do I manage to make inferences about students that were accepted into this university based on the GRE score in the first dataset? If I want to know what is the predicted probability of being accepted into Med School based on the characteristics of students in the first dataset (black, from public school) how do I calculate that?

I use STATA so I divided the GRE score into categories, and ran: ologit GREscore black publicschool

and then: prvalue, x(black=1 publicschool=1)

However, I am not sure if this is what I am looking for. Do I need to fit the whole model with all my variables or not? Is there another way that I can do this?

$\endgroup$

1 Answer 1

1
$\begingroup$

I don't see why you want to use ordinal logistic regression here (I don't use STATA, but I'm guessing that's what ologit is). You basically have 2 possibilities: accepted, & not accepted. Thus, you want to use regular logistic regression. Other than that, yes you want to fit the whole model & then solve for the predicted probability w/ the specific covariate values in question.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.