Great idea for a study. A couple random thoughts:
-- Substitution. Teams can not only call a play or change defenses
during timeouts, they can also change who is on the court. Of course,
trying to separate out the two effects might be a real drag.
-- Losing teams. As has been pointed out, teams often call timeout
when they are faced with a run by the opponent. This strongly
suggests, although I have no hard date to back this up, that losing
teams will use more timeouts than winning teams, and could
potentially bias the data if they used enough more.
Best of luck with it and keep us posted....