I am trying to learn how to perform ordinal logistic regression, this time using SPSS (but it is not very important to my question).

I used the following tutorial:

Link to example

And either I don't understand it or they have a mistake in the interpretation of the results (first option is more likely).

The dependent variable has 3 ordinal levels: unlikely, somewhat likely and very likely.

There are two significant independent variables: One continuous (gpa) and one categorical (pared).

The coefficient estimate of the continuous variable is 0.616. The exponent is 1.85.

For the continuous variable, they say, and I quote: "For a one unit increase in gpa, the odds of the low and middle categories of apply versus the high category of apply are 1.85 times greater, given that the other variables in the model are held constant".

I have two problems with the above conclusion. To start with, as far as I can understand, the exponent of the coefficient is the odds ratio of being in a higher category, i.e., being very likely vs somewhat likely or somewhat likely vs unlikely. This is not what I understand from the conclusion they wrote. Secondly, the descriptive statistics clearly shows an opposite conclusion, more similar to what I wrote.

I wanted to ask you guys to explain to me how to interpret the coefficients and the exponent of the coefficient of the ordinal logistic regression properly. How do I interpret correctly the binary independent variable?

enter image description here

enter image description here

enter image description here


Browse other questions tagged or ask your own question.