Policy, Politics

111 – Tornado politics and abortion politics

In his book “The Honest Broker: Making Sense of Science in Policy and Politics”, Roger Pielke Jr. identifies that some political issues are amenable to solution through scientific input, but some are not.

Pielke gives two extreme example of political issues, for which the roles of science are quite different. Both of the examples are characteristically American, but they do illustrate an important point well. I’ve adapted them a bit in the following descriptions.

1. Tornado politics: The issue here is how to respond in the face of an approaching tornado. For example, you might choose to rush into the basement of the building you are in. In this case you would be reasonably safe, but not completely safe, if the tornado hits. Alternatively, you might choose to attempt to travel to a completely safe tornado shelter. Travelling in the open would be dangerous, and probably lethal, if you got caught by the tornado, but if you made it to the shelter, you would be better off than staying where you are. High quality scientific information about the likely course and timing of the tornado would extremely valuable in making the decision between these two options. The desirable outcome from the decision is clear and uncontroversial: maximum likelihood of personal survival. It is relatively easy for a scientist providing advice about the tornado to remain dispassionate and distant from the decision, and they would probably be more useful to the decision maker if they did so.

2. Abortion politics: The issue is whether a community should allow or ban non-essential abortions. This may be affected by scientific issues to some extent, but mainly by personal morality, emotions, religious positions, and so on. Information is relevant, but not just factual scientific information. For example, the decision of a community might be affected by personal anecdotes, or information about the views of others. In this case, scientific information is not the key driver of the decision. It might, in some circumstances, influence the decision, but most likely the decision will be dominated by moral and religious considerations. If a scientist does attempt to contribute to the debate by providing relevant factual scientific information, the information is likely to be picked up by one side and used to attempt to strengthen their entrenched position in the debate. The desirable outcomes is ambiguous and controversial. It may be difficult for some scientists to prevent their personal moral views from influencing the sort of information and/or advice that they give to the decision makers (i.e. they may behave as stealth issue advocates).

Pielke listed some characteristics of these stylised decision problems, as follows:

Tornado politicsAbortion politics
Information used for evaluation

Information used for rationalisation

Information used to help assess decision alternatives

Information used to help justify existing decisions

Comprehensive analysis desired

Selective analysis desired





The protagonists seek enlightenment

The protagonists seek power


He points out that neither type of politics is necessarily better. They are just inherently different, unavoidably reflecting differences between the natures of the issues.

Sometimes scientists who are trying to engage in an abortion-politics type of issue behave as if the issue were a case of tornado politics. I would nominate climate change as an obvious example.

The bottom line is that science can help resolve uncertainty but not conflicting values.

David Pannell, The University of Western Australia

Further Reading

Pielke, R.A. Jr. (2007). The Honest Broker: Making Sense of Science in Policy and Politics, Cambridge University Press, Cambridge.