Exam pass rate analysis

#1
Hi, looking for some help with some analysis I am looking at on Minitab for exam pass marks.

I have a data set which shows global student pass/fail (attribute data) results for a specific exam entered for. I have a list of resources tools which the student either did or did not use (yes/no) to help them prepare for the exam. I also have demographics for each student such as age, country, gender, study method (online/face to face), years as a student.

I have performed a binary logistics regression test on minitab and have identified from p-value, coefficient and odds ratios which of the resource tools have an effect on whether student receives a pass mark. In short, the r-sq value for this is less than 5% and what I'm wondering is how best I could identify which other factors lead to an increase in the r-sq value and therefore are a predictor of whether student receives a pass mark.

Any help on which test to use and how to perform this on Minitab would be appreciated.

Thanks in advance!
 

hlsmith

Omega Contributor
#2
Traditionally the use of a pseudo-R^2 is frown upon when dealing with logistic regression. You can build models with multiple predictors in them. Models are typically examined using the coefficients like you did as well as looking at and testing differences in c-statistic (accuracy) along with performing 2loglikelihood tests on nested models.