In an age of hyper-partisanship, Americans increasingly get their news from sites that align with their political beliefs. But more separates those right and left-leaning sides of the web than their opposite ideologies. According to a new study, the right end of the fractured online news industry also tracks its audience far more aggressively than the left does.
In a study published last week, researchers from King’s College London, the privacy-focused browser firm Brave, and the research arm of Spanish telecom firm Telefonica compared the surveillance practices of left- and right-leaning news sites across the web. They found that sites classified as rightwing plant on average about 10 percent more cookies—bits of data that allow sites to identify the user and their previous browsing history—than their leftwing counterparts: 65 cookies for the average rightwing site versus 58 for the average leftwing one.
When it comes to partisan news outlets that rank among the 10,000 most popular sites online the difference was even more stark. Popular right-leaning outlets analyzed by the researchers placed 227 cookies in a user’s browser versus 131 for the median popular left-leaning counterparts. When the researchers ordered popular sites by how many cookies they placed, the contrast at the top end of surveillance-happy sites was even more pronounced: The top 25 percent of conservative sites in terms tracking planted well over 300 cookies in browsers, versus less than half that number for that same top 25 percent slice of liberal sites.
“Basically, ad tech is more evolved in right-leaning websites than in left-leaning websites,” says Nishanth Sastry, a senior lecturer in computer science at King’s College London, who along with the other researchers will present the study at the Web Conference in Taiwan in April.
To carry out their study, the researchers started with a list of “partisan” news sites they took from an earlier analysis of the political news spectrum by Buzzfeed news. For that survey published in 2017, journalists at Buzzfeed manually categorized more than 500 sites by examining their About pages and Facebook pages for explicit mentions of their liberal or conservative leaning, and in some cases inferred those political leanings from story content, too. Right-leaning sites ranged from Dailycaller.com to Realclearpolitics.com to TheGatewayPundit.com, while left-leaning ones ranged from Salon.com to Rawstory.com to Alternet.com.
The researchers then crawled both the right- and left-leaning lists of political websites Buzzfeed had defined with a set of web-browsing “personas,” essentially bots designed to impersonate real users whose browsers had previously visited sites that marked them as certain demographics. Male personas were prepped by visiting sites like MensHealth.com and GQ.com, for instance, while female personas were pre-loaded with cookies from Cosmopolitan.com or Womansday.com.
The researchers found that female personas generally attracted more cookies from all the sites than male ones, and older personas received more cookies than young ones. That targeting of women and seniors fits with assumptions in the advertising industry that both groups respond well to targeting, says Abel Buko, online advertising analyst and consultant for ad firm Bannerboy. But less expected was the researchers’ other finding: that conservative sites placed far more cookies regardless of demographics. FoxNews.com, the most popular right-leaning site, placed around 4 percent more cookies in women’s browsers than MSNBC.com, the most popular left-leaning site as categorized by Buzzfeed. It also placed 34 percent more cookies in men’s browsers, 26 percent more cookies in young people’s browsers, and 30 percent more cookies in seniors’ browsers.
So why do right-leaning sites track users more than left-leaning ones? The researchers’ explanation is simple, if not altogether satisfying: Advertisers are willing to pay more to get their ads in front of conservative audiences. The researchers came to that conclusion when they measured the cost of the ads their personas were shown, using a methodology some of the same research group created for a 2017 study that intercepts ad prices in web traffic sent to browsers. (In cases where those prices were sent in an encrypted form, they used a machine learning technique to attempt to determine them.)