However, the posttest-only model has a built-in weakness. Without pretests,
the intervention and comparison groups could have been different in
relevant ways from the start, thus variations noted in posttests may simply
reflect initial differences in groups rather than effects of the program. This
problem may be less pronounced in the Community Gardens evaluation
because gardeners did not enter the program with a clear pattern of need.
Regardless, we strengthened the overall design bV including multiple
methods, multiple samples, multiple data sources with triangulation, multiple
measures, and multiple analyses in the overall evaluation plan.
The findings that appear in the following chapters reflect these varied
research strategies, and will demonstrate some unintended, as well as some
intended outcomes for Community Gardens.