FORMULATING THE PROBLEM
    Specifying Outcomes
    Measuring Experimental Contexts
    Specifying Interventions
    Specifying the Population
    New versus Existing Measures
    Operationalizing Success/Failure

    Often, the hardest part of evaluation research involves specifying exactly the intervention seeks to change.  Let's say we want to reduce spousal abuse.  What do we mean by that term?  Are we talking about physical abuse, sexual abuse, mental abuse, all of the above, two of the above, or what?  Are we interested only in married couples, or will we consider common law marriages?  How about gay and lesbian couples?  Whatever we choose, that's only the beginning, as each of the terms used above would need further specification.

    As you'll recall from the earlier discussions of conceptualization and operationalization, this process leads eventually to a specification of how we will measure the variables under consideration.  In the case of spousal abuse, we may use offical reports of abuse, or we may choose to seek data independently--through a survey or qualitative interviews, for example.

     As the subheadings above suggest, the main thrust here is to specify various aspects of the study: what the problem is, what solution will be tried, and what outcomes would reflect success or failure.