case the data collection procedure differed slightly. For the general

surgeon, once they scheduled the surgery, I mailed to the patients a

packet containing all the research instruments, the consent form (see

Appendix C), and a letter of explanation with a stamped self-addressed

envelope for return. A member of the office staff kept a list of

patients who had received the packets, the date of the surgery, the

hospital, and the type of surgery. I visited the patients on the third

postoperative day (approximately 86 hours after surgery) to assess their

pain, using the McGill Pain Index.

 Data from the hand surgery patients were collected in a slightly

different manner. The surgeon has two offices, one in Gainesville,

Florida, a university community classified as a standard metropolitan

statistical area, and the other 28 miles away in the small rural farming

and forestry community of Lake Butler. I divided my time between the

two offices, seeing patients in his conference room, explaining the

study, and asking them to participate. Because many of the patients had

accidental injuries to their hands, their surgery was done with

deliberate speed. When possible, these patients were seen before

their operation, or at least in the same day. Those who agreed to

participate and completed their questionnaires returned them in a few

days.

 Over an 8-month period 167 packets of research materials were

distributed. Of these, 103 were mailed from the general surgeon's

office and 40 were returned, a 39% return rate. Two of the 40 were

dropped because of incomplete data. Thirty-two of the remaining 38

subjects were gastric bypass patients. The remaining six were dropped