The analyst then works with experts to identify the decision alternatives (choices), uncertainties, costs, and benefits. In our simple example, the choices available to a decision maker include issuing a boil-water advisory now or ordering tests (of water and any sick individuals that can be found) and waiting for the results. The uncertainty in this example is whether the surveillance data indicate a Cryptosporidium outbreak or some other phenomenon. The costs and benefits associated with each possible outcome of the decision include the cost of a boil-water advisory and the benefit in reduction of the number of sick individuals, should the increase in sales of diarrhea remedies actually prove to be a result of an outbreak.
One of the techniques used in building decision models is the clarity test (Howard, 1988). For each element of the model (i.e., choices, uncertainties, costs, and benefits), the analyst checks whether it has been defined in an unambiguous manner. Ambiguous definitions of model elements will generally backfire at some modeling stage, especially when it comes to elicitation of numerical parameters. The definition of an outbreak, as discussed in Chapter 2, is an example of an ambiguous definition that would not pass a clarity test. The phrases higher levels than normal levels and localized increases are not precise. A decision analyst would ask the experts to provide an operational definition for the term outbreak were the experts to indicate that the model include a variable called outbreak.
Was this article helpful?