| Abstract [eng] |
This paper examines whether Facebook’s algorithms enclose users in ideologically one-sided echo chambers thereby fostering polarization. It draws on two principal theoretical explanations for echo chambers — whether they arise from the inherent properties of algorithmic operations or from users’ social inclination to select content that confirms their own views. In an experimental study, four Facebook accounts were created—one pro-Ukraine, one skeptical of Ukraine, one mixed, and one control—which for 30 days posted preselected keywords twice daily. The daily news feeds captured (N = 1,981) were coded by ideology (0 = pro-Ukraine … 1 = pro-Russia) and by source diversity. Using quantitative content analysis alongside OLS and mixed-effects models, it was found that account type explained less than 4 % of the variance in ideological slant, and source diversity remained high—each account encountered at least 26 different sources. The results suggest that merely passively sharing content does not quickly trigger algorithmic “closure” into echo chambers—likely social factors, such as interacting with posts and a network of friends, are also needed. The study’s hypotheses were not confirmed: no link was found between account type and observed ideological narrowing of content. In its conclusions, the study recommends that social-media regulation address not only technological dimensions but also social ones, since the findings indicate that algorithms likely operate in a complex manner and require a broader social perspective for effective oversight. |
