ABSTRACT
This chapter discusses approaches to sampling recipients or observers and the critical issue of how many program participants to include in the evaluation. Probability sampling helps to guarantee that the families chosen for study will be representative of the population of families served by the program. Probability sampling ensures that each family served by the program has an equal chance of being chosen for study. Three types of probability samples are most likely to be used in an evaluation of Family-Based Service: simple random samples, systematic samples, and proportional stratified samples. In simple random samples, each family served during the evaluation period is identified and given a number. There are basically four types of nonprobability sampling techniques that can be useful in evaluating family-based programs: convenience sampling, purposive sampling, quota sampling, and snowball sampling. The chapter concludes with a discussion of how to protect the rights of clients or other participants in the research process.
