A social network approach to measuring social interaction related to c
hange of health behavior required 460 participants in a health promoti
on program to nominate up to five network members and to rate the supp
ortiveness of each person in changing health behavior during the previ
ous year. Exploratory factor analysis of intercorrelations of the eigh
t items suggested two factors representing supportive and negative int
eraction with internal consistency reliability of .89 and .61, respect
ively.