It’s common for people on both sides of the political divide to regard the media as being inherently biased against them. While much of this debate has traditionally focused on the mainstream television and print media, this has lately spread into social media and search listings.
New research from Stanford University explores whether these accusations have merit. The researchers examined the first page of the Google search results for every single candidate running for office during the 2018 US election.
In total, around 4 million URLs were analyzed, and the analysis found that the results were largely free of any bias, with a centrist perspective the most commonly delivered angle.
“Our data suggest that Google’s search algorithm is not biased along political lines, but instead emphasizes authoritative sources,” the researchers say. “I think audits of large-scale algorithms that play such an important role in so many aspects of our lives are crucial. We need to be able to trust that these AI systems aren’t biased in important ways, and without audits, it’s difficult to assess these opaque algorithms.”
Fair coverage
In total, some 3,000 candidates who were collectively running for 225 seats across both the House of Representatives and Senate were assessed. The potential bias in the search listings was assessed by evaluating the news sources that appeared on the front page of each Google search result. They removed any social media platforms or personal campaign websites as they believed they were not relevant to any discussions around partisanship.
A scoring methodology was used that was originally created by Harvard University to assess mainstream media coverage during the 2016 presidential election. This provided a ‘partisan score’ to each source based upon the kind of people that shared that source on Twitter. Around 44,000 Twitter users were tagged as either conservative or liberal, and then their retweets were assessed to evaluate the partisanship. These were then applied to the sources that were appearing in the Google search listings.
After assessing the search listings, the researchers discovered that the results for the vast majority of candidates were derived from centrist news sources, with no noticeable biases for either party.
This should perhaps come as no surprise as Google’s search algorithm has evolved to prioritize highly reputable sources with a long track record of delivering high quality content over a prolonged period of time.
Real-world impact
But what does this mean for the actual elections themselves? To gauge the impact the search listings had on the way each election unfolded, the researchers compared the search results for a candidate with their incumbency status. They went down this route because incumbents are typically more moderate and centrist than challengers. The hypothesis that the search results would reflect this.
Lo and behold, that was precisely what emerged, with the search results for incumbents considerably less partisan than those for challenger candidates. They then compared the search results with the election results themselves, with winners typically having very different search results than losers, although this is largely because incumbents are statistically much more likely to win than challengers.
Overall however, the team believe their work opens the door for further research on the use of search results to provide real-time insight into elections.
