I have a data set containing n observations (rows) of mobile applications. The dependent variable is the number of installations -- the variable is categorical (with more than 2 categories). Therefore the econometric approach is an ordinal probit/logit model. Besides my dependent variable I have plenty variables which describe the apps characteristics (pricing, rating, reviews, description length, the category etc.).

I would like to find a theory which explains the customer demand for apps. I have limited prior knowledge, but I found that one could use either the theory of product space or product characteristics. To apply the former, I would need to find groups in which the apps "compete" in. This cannot be done sufficiently in my case.

The second approach (product characteristics) utilizes a discrete choice approach. One customer maximizes its utility by comparing the characteristics of all products and chooses the one which suits him/her best. However, all customers can choose only one product, this is not the case here. One customer could choose n applications to download and therefore maximize his/her utility.

Is there any theory which supports my observed data?

EDIT: An app can fall in one of 12 categories. They have a natural ordering, but different sizes (Google reports them this way). They look like this:

1-5 Downloads = Category 1
6-10 Downloads = Category 2
11-50 Downloads = C 3
51-100 Downloads = C 4
and so on.

  • $\begingroup$ Could you please state more clearly the definition of your depvar? Like which are the categories and the like? $\endgroup$ – user_newbie10 May 21 '16 at 17:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.