If you haven’t already played with the NY Times dialect map/questionnaire game, I recommend trying it out. It picks out a random group of questions from one of the big American dialect surveys (the Harvard Dialect Survey, in which a Harvard prof surveyed students and linked the results to their hometowns, as opposed to any of the Labov telephone survey ones) and then maps the answers, trying to give you two or three cities or areas that your answers are most like. Users have variously reported it to be eerily accurate down to the county, and eerily inaccurate to the point of madness.
For me, it was quite accurate. However, Dayton has been the subject of a fair amount of interested study as a place of dialectal collision and mixing, so no big surprise. I also got Omaha and another Western city close to it, but the forces of dialectal collision are quite similar out there.