It was classic Malcolm Tucker. Informed that scientific evidence had led the minister to question government policy, the Downing Street spin doctor fired back: "Yeah, but I've got an expert who will deny that."

Who was this authority, the minister inquired? "I have no idea, but I can get one by this afternoon."

As so often with The Thick of It, the exchange captured a truth.

Politicians of all parties love to portray themselves as champions of science who follow the evidence wherever it leads. But they rarely want all the evidence – just the bits they can use to justify decisions they're taking anyway. What politics really values isn't evidence-based policy. It's policy-based evidence.

When the Gordon Brown government decided to reclassify cannabis as a class B drug, home secretary Jacqui Smith ignored the misgivings of the Advisory Council on the Misuse of Drugs (ACMD). Instead, she sought more convenient evidence – already weighed and found wanting by her advisers – to support "a compelling case to act now rather than risk the future health of young people".

The coalition has been just as bad. Stung by the ACMD's insistence on banning "legal highs" only on a scientific basis, Smith's successor Theresa May abolished its statutory scientific membership. It can now be packed with the "right experts", as Tucker might put it. Caroline Spelman claimed her long-promised badger cull as "evidence-based" over the protests of the scientists behind the very research she cited. And to justify his NHS changes, Andrew Lansley embarked on a wholly misleading comparison of heart attack deaths in Britain and France.

Each of these policies emerged from political calculation, to appeal to supporters or public opinion. Yet none of the ministers responsible had the guts or the honesty to admit that.

In search of cover for populism or ideology, they cherry-picked data to create a veneer of scientific authority for an unscientific case. This is evidence abuse, and the ease with which politicians get away with it has two pernicious consequences. They feel less pressure to act on scientific advice to make policy that is properly fit for purpose. And it devalues evidence as a currency that voters can use to make up their minds. When every policy is presented as evidence-based, it becomes difficult to know which are genuinely founded on science.

This damaging culture of evidence abuse is partly born of the staggering under-representation of science in the House of Commons. Some 158 of our 650 MPs have a background in business. Another 90 had prior careers in politics, 86 in law, and 38 in the media. Yet just a solitary MP – Julian Huppert, the Lib Dem member for Cambridge – has worked as a scientist, and only two more have scientific PhDs.

This shortage denies politics a cadre of people who understand that evidence isn't a kind of magic dust you can sprinkle on a policy to give it credibility, but that you have to look at it all. We need more MPs whose instinct is to consider research in the round before they decide, not twist it afterwards.

For those who lack that instinct, however, policy-based evidence will always be catnip for as long as it carries no political cost. We need urgently to create one, so ministers who aren't ashamed of evidence abuse avoid it out of self-interest. This government has already established an independent Office for Budget Responsibility as a disincentive to fiddling economic figures. It should set up an Office for Scientific Responsibility as well.

Like the OBR, this body wouldn't limit the executive power of ministers, who would remain free to act without, or even in the face of evidence. But when science is cited to justify a policy, the OSR would audit it, naming and shaming those who bend it to their political advantage. The Smiths and Spelmans who reject scientific advice would thus have to admit to it, or face an official rebuke that their Malcolm Tuckers would struggle to spin. It's only by making evidence abuse politically painful that we can hope to put a stop to it.