Thanks to the replies so far. I guess I should have explained the
goal of the analysis better, right?

In the end, I'm wanting to do a multilevel clustered observations
model with a recipient being nested with the doctor. Both have score
variables assessing the participant's ailments (ranging from 0 to
22). I haven't done the MLM yet as I was first working on the
directions given to me from my colleague that were based off of an
analysis done in a paper by Lyons, Zarit, Sayer, and Whitlatch.
(2002). Caregiving as a Dyadic Process: Perspectives from Caregiver
and Receiver. Journal of Gerontology: Psychological Sciences. 57b(3),
P195-P204.

I believe my directions were based on the last paragraph before
the "Results" section on P199. I was instructed to do the
following: "Physician and recipient responses need to be regressed
on -.5 and .5 respectively so that the intercept represents the mean
between self and physician ratings, the sign of the slope (+, -)
indicates who differs, and the magnitude of the slope represents the
degree of difference between the two."

Like I said, I haven't done the MLM analysis yet (and honestly need
to brush up on my class notes). Should I just temporarily ignore the
regression request?

Many, many thanks to the rockin' members of this group!
Sara