Scientists, said Winston Churchill, should always be on tap when ministers decide, but never on top. His meaning was that scientific advice is sometimes only one of many factors that elected representatives must take into account: in a democracy political considerations matter, too. It remains a maxim with great merit, but it is often abused to set expert evidence to one side.

Sound science has rarely been more important to the great issues of the day. Governments cannot address challenges such as global warming, stem-cell research, GM crops and nuclear power without access to the very best evidence. Yet while Britain has expertise in abundance, our political system is not always well placed to tap it.

Both the House of Commons and the Civil Service are dominated by graduates in the humanities, the social sciences and the law. Both the Cabinet and Shadow Cabinet each boast a solitary full member trained in the natural sciences. They exhibit a discomfort with science that is shared by most junior ministers, backbench MPs and officials. Most have not engaged with the subject since they were teenagers, and lack the skills with which to weigh and interpret evidence. This leaves them vulnerable to two very different errors. Some politicians see scientific advice as little more than pleading by another special interest. Others make the opposite mistake and expect to be told what to do, when the very nature of science is uncertainty. Both courses can be catastrophic. The first led ministers initially to seek a ban on human-animal hybrid embryos. The second informed the BSE debacle.

The Conservatives' plan to require new MPs to attend science literacy classes (see page 10) is thus timely and welcome. It will not turn them into experts on genetics or greenhouse gases, but has a more limited goal. That is to teach the rudiments of the scientific method, which is the best tool that exists for making sense of the world.

Science is not best understood as a body of facts. Rather, it is a process by which these are discovered, by forming and then testing hypotheses which stand or fall on the evidence amassed. It is about drawing conclusions from data, rather than cherry-picking convenient observations that seem to support a cherished idea.

This intellectual approach should be axiomatic to good government, and not only in the obvious departments. Education policy would benefit from trials that compare class sizes or teaching techniques. In criminal justice, alternative sentences could be properly evaluated to determine which best reduce recidivism. The aim should be to make evidence-based policy and not, as happens too often, to seek policy-based evidence.

Adam Afriyie, the Shadow Science Minister, is not the only politician who understands this. John Denham, the Innovation Secretary, is a chemistry graduate who spoke eloquently on this issue in January, and Phil Willis, a Liberal Democrat, chaired the select committee inquiry from which the Tory policy emerged. Both should encourage their parties to emulate it.

We cannot expect politicians to grasp the minutiae of every branch of science: even scientists cannot do so. Nor should we expect ideology always to take a back seat. But we should expect decision-makers to learn how to seek out and interrogate evidence. They must be informed consumers of what science has to say.