Hi All,
I'm running an Ordinal regression on a continuous independent variable and (obviously) an ordinal dependent variable. I'm doing the same for a number of continuous variables. So, I have many models. Some of these went well, but others either the assumption of proportional odds was not met and/or the model didn't fit the data well. I wonder, can I still use the p-value from these variables to say 'x' variable had or did not have an effect o the response variable?
All the best,
Laura