Questions not applicable given a certain answer. How to analyze?

Asked Feb 27 '20 at 13:39

Active Feb 27 '20 at 15:01

Viewed 30 times

So I am dealing with a questionnaire where given an answer on a specific questions the next three following questions are not applicable to the person. I would like to include those variables in a regression as independent variables but I dont know how. For example:

Do you own a car?
What is the color of the car?
What is the brand of the car?
How old is the car?

So there will be a high correlation between question 1 and the questions 2,3,4. I could remove question 1 but multi collinearity between 2,3,4 remains because all people who do not own a car have a "not-applicable". Also I would lose the easily interpret-able effect of owning a car.

I also cant really impute those values as they are not missing at random.

Is there a way of solving it, or what would be search terms I could use to find an answer?

edited Feb 27 '20 at 15:01

asked Feb 27 '20 at 13:39

Janosch

Could you elaborate on what you mean by "dealing"? That is, exactly how do you intend to use this set of related results? What is the purpose of your analysis? – whuber Feb 27 '20 at 13:48
Oh sorry. I want to use those variable in a regression. I edited it – Janosch Feb 27 '20 at 14:05
How, exactly? As a response or as explanatory variables? – whuber Feb 27 '20 at 14:24
1

As predictors variables – Janosch Feb 27 '20 at 15:01
This might be helpful: https://stats.stackexchange.com/questions/372257/how-do-you-deal-with-nested-variables-in-a-regression-model/372258#372258 – kjetil b halvorsen Dec 03 '20 at 12:40

Questions not applicable given a certain answer. How to analyze?

0 Answers0