Mary, I'm a little confused how what you're looking for is different from the respondent-level WTPs you say you've created.
It sounds like from the utilities you're calculating the utility value of changes in quality and then converting those changes in utilities to WTP $ because you know how utility points convert to $, am I correct?
If so, then it it seems you've already captured the interesting information. I don't quite understand what you mean by a correlation coefficient for each individual. Since the WTP calculation is deterministic, I might imagine that the correlation at the respondent level would be 1.0, assuming utilities have been constrained to move in the right directions.