Help with logistic regression assumptions

[dimitrios]

Hello,
I would be grateful to anyone who could help with this:

1. I have created a logistic regression model with dependent variable being of a dichotomous and variables are categorical except one which is continuous. Do I have to check for Linearity assumption since it is only one continuous variable? I did so and the interaction age* ln age comes back significant. What can I do to solve this?

2. What is the range that eigenvalues are considered very large?

Thank you in advance,

Dimitris

[Bruce Weaver]

Dimitris, Nabble shows that your post still has not been accepted by the mailing list after several hours.  Did you actually subscribe to the list, or just post via Nabble without subscribing?  

I'm responding to just one part of your first point below.  I don't understand why you're making a distinction between models with only one continuous explanatory variable and models with two or more continuous variables.  Why would having only one continuous variable make it any more acceptable to force a linear fit if the data suggest some non-linear functional relationship between X and the log-odds of Y being equal to 1?  

When my reply appears on the list, perhaps some others will jump in to respond to your other questions.

dimitrios wrote

Hello,
I would be grateful to anyone who could help with this:

1. I have created a logistic regression model with dependent variable being of a dichotomous and variables are categorical except one which is continuous. Do I have to check for Linearity assumption since it is only one continuous variable? I did so and the interaction age* ln age comes back significant. What can I do to solve this?

2. What is the range that eigenvalues are considered very large?

Thank you in advance,

Dimitris