Like it or not, science is politics. And despite one party's attempt to brand itself as the voice for all science, the reality is fans from either side of the ideological spectrum kneel before the altar of data. Or rather, their respective altars. Because Liberals and Conservatives tend to prefer to read about different types of science.
The proof is in the perusing. A new study published today in Nature Human Behavior looked at which science books and which political books people shopping at Amazon.com and BarnesAndNoble.com typically bought together. The huge consumer dataset revealed some pretty clear trends: Red-tinged readers prefer applied science, like criminology or medicine, while lefties alight upon books that explore science for science's sake, like zoology, or abstract physics. And while it's great that the reasoned examination of facts appeals to everyone, the study seems to suggest that---unsurprisingly, but depressing nonetheless---people seek out the stuff that supports their worldview.
"The political divisions within this country seem to be splitting the country into mutually hostile camps that increasingly don't like each other, and don't engage intellectually," says Michael Macy, computational social scientist at Cornell University, and co-author for the study. "We were interested to find out if science might be able to serve as a bridge between these camps." He and his co-authors figured that books might be a good way of assessing peoples' actual interest in science. And, rather than fuss over low response rates and other survey design flaws, they figured they could just scrape sales trends from online bookstores instead.
If you've been to Amazon, you're probably familiar with one of their most successful sales tactics: Recommending books other customers have bought alongside the one you are currently browsing. The data underlying these suggestions is right there in the open, in Amazon's API.
The authors started with a two "seed" books: Barack Obama's Dreams of My Father, and Mitt Romney's No Apology. For each, they scraped the top 100 results from the "Customers Who Bought This Item Also" section. Then they repeated this process with every book on those lists---looking for other books customers bought alongside---and again with the results from those books. They repeated this cycle again and again until they had a complete library of nearly 1.5 million books. Then they began winnowing out all the political and science-based books, based on Amazon.com's classification system---many political books are labeled for conservative or liberal readers. In order to delineate subcategories of science, they used the Dewey Decimal and Library of Congress systems of categorization.
From that thicket of data, they parsed. "There are two important general differences between the two ideologies," says Macy. "Liberals tend to be more interested in basic science that is motivated by intellectual puzzles, empirical exercises, philosophical musings, and conservatives are looking for solutions, problem solving, and applied research." A liberal might be more likely to buy a bundle of books featuring Al Franken and Carl Sagan; while conservative shopping carts would be full of Star Parker and Mary Roach.
The second broad trend is a little more nuanced. Liberals tend to purchase science books that are interesting to anyone who is interested in science, regardless of whether they read political books. And conservatives are more cloistered, preferring science books that are only of interest to people who buy conservative political books. For instance, a liberal reader of books on environmental science is more likely to read something with broad popular appeal, like Andrea Wulf's The Invention of Nature: Alexander von Humbolt's New World, whereas a conservative reader would go for something more niche, and mostly of interest to other conservative readers, such as Lukewarming: The New Climate Science that Changes Everything.
So where's the common ground? Dinosaurs, mostly. "And perhaps that not too surprising, given that it's not an area of research that's particularly politically-charged," says Macy. Which seems counterintuitive, given the whole creationism thing, but that's not what the data says. Physical science topics more generally were the least partisan, followed by life sciences---biology, environmental science, zoology---and finally the social sciences, like psychology, which might as well have trenches.
The Amazon data has its limitations. "The biggest is that we don't have individual-level purchase data," says Macy. They had to draw their conclusions from broad aggregate trends, which could mean they are missing nuances in Amazon's algorithm that could be skewing the data one way or anther. To account for this, they repeated the entire experiment on Barnes & Noble's online store. Interestingly, the two websites did not share a high number of people making the exact same co-purchases of this political book to that science book. However, the high level correlation between political ideology and scientific discipline held.
Science isn't the only topic where Macy and his co-authors are investigating partisan social divides. They maintain a website called www.lifestyle-politics.com where they rate things like sports teams, professional wrestlers, and TV shows based on users' Twitter feeds. "What we've found is there's a strong correlation between ideology and cultural preferences that have seemingly nothing to do with ideology." For instance, if enough people who follow ideologically conservative accounts like @realDonaldTrump and @FoxNews also follow @ChickfilA and @BigBangTheory, that fast food restaurant and that TV show also get grouped as conservative cultural touchstones.
To Macy, these cultural trends seem to match what's happening in science. "We don't know for sure, but we speculate that lots of interest in science is politically-motivated, and people are interested in reading about science that supports their political views," he says. If scientists want to do a better job of making their research more accessible---which they probably should if they don't want their line of work targeted by the same kind of ideological philocide currently being perpetuated against climate science---they should try to preach beyond their own choir.
