NCJ Number
85277
Date Published
1981
Length
23 pages
Annotation
Simple methods are asserted to be adequate for many policy analyses, because the use of more complex measures does not change the results.
Abstract
In policy analysis, the main concern is choosing the best policy rather than in determining how much better the best policy is than the second best policy. In contrast, ordinary social science research usually focuses on exactly how much more important one independent variable is than another. However, policy analysis contains the following steps: determination of goals and available policies, ranking of the goals, identification of the targets of the policy, identification of the relationships between the goals and the policies, determining the causal relationships between policies and goals, and deciding which policies are best for achieving the goals under various circumstances. Simple dichotomies or trichotomies can often be meaningful in choosing among alternative policies. Although the relative value of the goals usually needs to be determined, the absolute value does not need to be known. Policy analysts can often work with small nonrandom samples and can relate policies to goals by surveying knowledgeable people or other methods rather than by using a complicated statistical regression analysis. In determining an optimum policy level where doing too much or too little is undesirable, analysis should try alternative numerical values for the policy until reaching the value that produces the lowest total costs, the highest total benefits, or the greatest difference between benefits and costs. An optimum mix of resource allocations across activities or places can be determined by allocating the total budget in proportion to nonlinear regression coefficients or to scores assigned to relative efficiencies on various dimensions. Notes, figures, and a list of 31 references are included.