When Politics and Maths Collide: Politics Wins, We Lose
Do you think that if we are better at mathematics we are more apt at ignoring our biases and interpreting scientific data effectively?
If you did (like me), then you’d probably be wrong (like me).
A recent study has shown that not only are people significantly impaired when looking at politically charged data, but it seems that those who are more mathematically literate on average seem to fare worse than those who are less numerate.
The study by David Kahn and his colleagues had a really ingenious design. They selected over a thousand participants, tested their mathematical abilities and then asked them to assess a fake scientific study. There were four variations of the study in which the data stayed identical but the results changed. There was one neutral topic (effects of a skin cream) and one politically charged topic (law banning concealed handguns). This resulted in four outcomes: skin cream positive results, skin cream negative results, gun ban positive, gun ban negative.
As you can see there is a big gap between the neutral and politically charged groups getting the answer right. It also seemed that those who performed well in numeracy (about 7) performed worse in the politically charged question than those less numerate. This effect only decreased (but was still high) in the most numerate people.
The most interesting part of the study is when we start to look at possible mechanisms, however, these are less surprising than the initial results. It seems to all about our skills of rationalisation (to make the data fit our preconceived ideas) and our faith in our own ability to be rational (when in fact we are just using motivated reasoning).
In addressing the facts of the world we should be choosing science over belief, but to do that it seems we are going to need to reduce our beliefs instead of fighting them, lest we fall in to the trap of motivated reasoning. The less things we choose to believe, the less obstructions we have and the better we can impartially look at the data.
[Image from Creative Commons licensed Flickr photo by mstewartphotography]