One easy mistake to make as a policy analyst is to get ahead of yourself.
When analyzing a public policy problem, policymakers are often already bringing diagnoses of the problem to the table before the analysis has even begun. This is why problem definition is such a crucial step in the policy analysis process: it helps us strip away assumptions we bring to the table and analyze a public policy on the merits of its core goal. This keeps our problem definitions squarely rooted in our goal of advancing social values.
In Eugene Bardach’s A Practical Guide for Policy Analysis, the political scientist makes a distinction between two ways of thinking about problem definitions that is helpful here: problem definitions can be both descriptive and diagnostic. The more descriptive a problem definition is, the closer we can tie it to empirical evidence. The more diagnostic it is, the more we can construct concrete policy options to address the problem.
One way to think about problem definition is a type of criteria selection: we are describing what public problems policymakers care about. When we define a problem such as “the quality of air in the state of Michigan is too low,” we are laying the groundwork for us to construct certain types of policy options. We are also already implying one criteria: the effectiveness of a policy option in improving air quality.
Bardach talks about the treacherous line policy analysts walk when they define a problem. He uses the following example to show what a more diagnostic problem definition looks like:
One of the problems in the air pollution area is that states have not been willing to force motorists to keep their engines tuned up and their exhaust systems in proper order.
He says that on the one hand, this more diagnostic problem definition is useful. By more narrowly defining the problem, we can come up with more practical policy options to address this problem. “How can the public sector force motorists to maintain their vehicles?” is much more concrete than “how can the public sector improve air quality?”
On the other hand, such a strongly diagnostic problem definition can (a) limit the range of policy options to address the underlying problem, and (b) detract focus from larger causal factors to small causal factors for the underlying problem.
Policy analysis is an iterative process, and remembering this is an important way to figure out how to navigate the treacherous territory of description versus diagnosis in problem definition. Your first pass at problem definition will look different from your approach to problem definition in the eighth step when you are telling your story.
For this reason, I provide the following advice: describe, then diagnose. In the earliest phase of your policy analysis, focus on description. Make your problem definition abstract and idealistic: strip what your client wants down to its most naked connection to public interest. “The quality of air in the state of Michigan is too low” is an excellent first-pass problem definition.
Later, particularly after the second step of “assembling evidence,” diagnose. Say why air quality is low. Compare air quality in Michigan to that of other states. Talk about causal mechanisms. Build a model.
At this point, your problem definition can change. And the changes should be driven by (a) empirical evidence, and (b) client needs. This allows your problem definition to be both truthful and helpful.
If your problem definition starts out rooted in social values, then it can be further specified with empirical evidence and client needs. If you don’t start with social values, however, it can be a lot harder to add those in later in the policy analysis process.