# Regression model that predicts the changes in the category of dependent variables?

I have some dependent variables that an event might changes their categories. Assume I have the following dataset

  Students   Category
A          Poor
B          Good
C          Excellent


I have an event (as my predictor variables) in a binary format ( $1$ indicates that they are exposed to the event and $0$ otherwise) and I wish to measure the effect of events on the changes of category. So, after events we will have the following scenario:

  Students    Event   Category
A           0       Poor
B           1       Good
C           1       Good


What is the appropriate regression model in this scenario? I know about Ordinal Logistic Regression that somehow can be used by transforming the data but does anyone has other suggestions?

• Do you have information about the variable in your data both before and after the event? Or are these separate data sets that can't be merged? And how many response categories (poor, good, excellent, etc.) do you have in your variable (Category)? Nov 3, 2016 at 15:08
• Hi Tahir, yes, we a have all the information before and after events. All the ids ( I in this example, students) have a unique identifier and two data set can be merge together. In total we have 9 response categories.
– MFR
Nov 3, 2016 at 21:54
• Hi, it might sound rather careless and I understand there is a categorical variable, but if response categories are ordered, what about assigning them numbers? How would be the distribution of this variable? Nov 3, 2016 at 22:31