Online purchase patterns show left-wingers and right-wingers read very different science books

While faith in politicians and journalists is in the doldrums, surveys in countries like Britain, Canada and the US suggest scientists are among the most respected professions, and citizens are appreciative of the contribution science makes to their lives. As the authors of a recent article in Nature Human Behaviour note, it seems that "we may disagree on emotionally charged social issues, but at least we can agree on science".  Sadly, their findings show that it's not so simple. Although politically diverse people may be fans of science, we're not fans of the same science.

The researchers, led by Feng Shi at the University of Chicago, explored the book-buying habits of tens of millions of Americans by analysing two huge datasets belonging to Amazon and Barnes & Noble, which encoded the purchase patterns that make possible recommendations like "customers who bought this item also bought…".

The researchers coded over a million books in the databases: about half a million covered topics like art, fiction and sports; there were over 400,000 science books, categorised into different domains like life sciences and further into disciplines like medicine; and finally, there were political books, from which the researchers identified nearly 600 emphatically liberal books and another 700 conservative ones. The key question was, what other books were buyers of these "blue" (liberal-left in US parlance) and "red" (conservative) books also buying?

Overall, science books were more likely to be read by people also reading political books, regardless of ideology. And many disciplines of science, like chemistry or economics, were of similar interest to both red and blue readers. So far, so good. But many other disciplines were more favoured by one tribe. Liberals were far more drawn to engineering, anthropology, and purer sciences like biology, astronomy or to a lesser extent physics. Conservatives were drawn more to applied disciplines such as medicine and law, and – in the highest association for the red tribe – climate science. (This may seem surprising, but we'll come back to it shortly.)

What's more, when you scrutinise the sciences that appear to be equally favoured, you find polarisation is still happening, just at a more fine-grained level. Digging into chemistry, for example, reveals its most applied branch (organic) is very strongly red whereas zoology is exceptionally blue. Even within subfields that have interest to both tribes, like criminology, conservatives and liberals are siloed into reading different authors (palaeontology was one of the rare exceptions to this rule).

The science books read by conservative readers also tended to be less varied, such as the same three climate books cropping up repeatedly, whereas liberals tended to read more widely across an area. The conservatives narrower choices also appeared to contain more fringe publications rather than core texts. This may explain why the analysis appeared to suggest that conservatives are more interested in climate change: they were dipping into the same minority of books sceptical of climate change, but generally they weren't reading on the topic more widely.

Anonymous purchasing patterns may provide access to true preferences, getting round the potential problem of people hiding their unpopular beliefs in surveys. That said, the new findings may give a somewhat distorted picture of people's true interests because the recommendation algorithm used by Amazon and Barnes & Noble pushes similar consumers towards the same choices (even so, the findings still provide another demonstration of the way filter bubbles may be shaping how we approach the world). Together with the filter effect, the authors also consider the possibility that "scientific puzzles appeal more to the left, while problem-solving appeals more to the right" as an explanation for the preferences for pure and applied science respectively.

Science is a powerful engine for finding truth. But truth pursued too selectively can cause problems, in this case via political cultures each with a partial picture of the world, each increasingly talking past the other. It's a big problem to tackle, but also a reminder that if we're willing to poke our heads above the parapet, there's an unseen world out there just waiting for our attention.