I have data for 110 companies re. the price they have paid for a service annually between 2004 and 2011. Now I would like to find out if there is a statistically significant difference between the price they are charged pre and post 2008 . . . how should I deal with this, given that I have 8 observations per company (Can I pool them for pre and post 2008 and use the paired samples t test?)


It is possible to group pre-2008 and 2008 and later (the way you had it worded leaves 2008 out altogether, I am not sure if that is your intention). But there are probably better methods.

Are you interested ONLY in this comparison (before 2008 and after) or do you want to look at other possible effects of time?

Have you other information about the companies that might be important? (e.g. size, amount of the service that they used, etc)

If the answer to either of these is "yes" then I would look into multi-level models (aka mixed models and various other terms).

But, first, I would plot the data. Given that you have 110 companies, one plot I would do is to first sort the companies by the amount paid in 2004; divide them into 5 quantiles; then, for each quantile, make a plot with "year" on the x-axis and "price paid" on the y-axis and a line for each company (22 lines per graph). (You might divide them into 10 groups and have 11 lines per graph).

  • $\begingroup$ Thanks for your reply, I would rather stick to this comparison pre and post 2008 (with 08 included in pre as pointed out) ... what would you recommend? Thanks a lot. $\endgroup$ – user21100 Feb 21 '13 at 11:32

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.