Embedded Files
Science and policy are like apples and hay. They are not the same sort of thing. Making policy is about making decisions, and a decision can be made without using any knowledge or principles of science whatsoever. The policy related activity that is like science is called analytics, which encompasses a broad range of activities to gather data, investigate patterns, and give advice.
What's the difference between science and analytics? In a nutshell, science is an activity whose end goal is to advance knowledge, while the goal of analytics is to inform a decision. While both activities tend to use the same toolbox, they differ in the way they regard uncertainty. To illustrate what I mean, consider a situation where there are two parameters (P,Q) and two possible conclusions (A,B).
Top left) A separatrix separates regions where the evidence considers A and B to be equal (the black line); another line marks a point where the evidence strongly favors A (the red line). Top right) Scientific inference is epistemologically conservative, so it will fail to reject a null hypthesis (A) unless there is strong evidence favoring an alternative. Bottom left) Analytics need not accept this conservative bias; instead, it can offer both a statement about whether the evidence favors A or B and how strongly. Bottom right) The decision, however, must also consider the consequences of the uncertainty. The parameters P and Q affect both the weight of evidence and the likely outcome. The consequences of the outcome must be weighed through an objective function, a function that translates the information into a meaningful statement about the consequences, which can be non-linear. It may be that uncertainty favors one or the other, conveys indifference for most values of the parameters (blue); or it suggests that the uncertainty is meaningful and consequential (red).
Google Sites
Report abuse