I have collected a data from 200 customers on their brand preference of 5 international fast food chain shops. Each customer was asked to rank all the 5 brands from highest to lowest (5 for the most preferred and 1 for the least preferred brand). I have collected some IVs like customers age, income, gender, how often the customer takes fast food, mean service time of the shop and some other factors like this. I want to see how brand preference is affected by these IVs. I also want to find out how a customer usually ranks the brands. That is to predict the brand preference. I will also like to see if brand preference significantly differs from each other (multiple comparison? I am not sure!).

Although it looks to be a very simple statistical problem, but I am confused because the question related to the dependent variable (brand preference) contains 5 rankings for each customer. Because each customer gives a rank corresponding to each brand. What will be the dependent variable here then? What kind of regression should be performed?

Can someone give me suggestions on how do I perform the analysis?

  • $\begingroup$ One approach would be to simply use the top choice of each person as the dependent variable and then do multinomial logistic regression. $\endgroup$
    – Peter Flom
    Oct 7, 2012 at 12:27
  • $\begingroup$ Thank you Peter, I had thought about this, but looking for a better way, if exists. $\endgroup$
    – Blain Waan
    Oct 7, 2012 at 12:33
  • $\begingroup$ There are more complex ways but I only remember them vaguely; doubtless someone else will know more. $\endgroup$
    – Peter Flom
    Oct 7, 2012 at 12:34
  • $\begingroup$ Nice to know that there are ways, I really want to use these ranking information. Kindly let me know if you can remember the name of any of the methods. Thank you. $\endgroup$
    – Blain Waan
    Oct 7, 2012 at 12:48

1 Answer 1


We can apply binary logistic by modifying the dependent variable's rank order, like most preferred brand is ranked as 1 (code 1)and other considered as less preferred (code 0). Then, we can also try to find the correlation among customers age, income, gender and brand preference. One can also check the reference trend with the changing pattern of their income through graphical tools.


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.