Second, we understood from our qual

Second, we understood from our qualitative findings that some individual characteristics across all
program categories would be important variables to match. For example, we learned during
interviews that cultural background played a role in the motivation to garden. We also knew that
we would have to match Participants by age level and gender because we would be measuring
several age- and sex-sensitive behaviors, such as daily exercise and diet. Thus we sought
matching pairs of participants by ethnic group, age, and gender. 7
Third, we understood from the literature that comparison participants should be involved in
programs, or receiving some kind of common service (as long as the service was not that which
was provided by the program being evaluated). Posovac and Carey (1989) argue that the attention
alone that program clients receive may account for changes in their behaviors over time. Thus we
sought comparison groups that were linked by common services.
With these three principles in mind, we selected several comparison group sites. The sites are
outlined below, with descriptions of the aggregate and individual traits we sought to match with the
program groups.