Auditing Facebook Algorithms

In this paper, a proof-of-concept study is performed to validate the algorithmic auditing of Facebook News Feed. We tracked and documented public or otherwise known changes to the algorithms through Facebook public announcements, industry research, and information leaked to the press to parametrize a model that accounts for the variation in user engagement with Guardian news articles. To this end, we queried the Guardian API to collate a database of all Guardian articles published between 2010 and 2020 and subsequently queried the CrowdTangle API to retrieve Facebook engagement metrics for Guardian articles. We modeled this time series using time series analysis, including cross-correlation, anomaly detection, and granger causality tests to examine the relationship between changes to Facebook News Feed and engagement with Guardian articles over the past decade. Our results show that hard news items, particularly those classified in the section ‘News’ by the Guardian API, are significantly more likely to have been impacted by changes made to the News Feed Algorithm in the period. We conclude with a discussion on the asymmetric power exerted by social platforms on news organizations and the elapsed effects of algorithmic changes to website traffic, business models, and editorial decision-making in the news industry

Algorithm · Audit · Business · Guardian · Political science · Social media · World Wide Web · Complex Network Analysis Techniques · Computer Science · Digital Marketing and Social Media · Misinformation and Its Impacts · Accounting