Opinion Dynamics via Search Engines (and Other Algorithmic Gatekeepers)

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractRanking algorithms are the information gatekeepers of the Internet era. We develop a stylized framework to study the effects of ranking algorithms on opinion dynamics. We consider rankings that depend on popularity and on personalization. We find that popularity driven rankings can enhance asymptotic learning while personalized ones can both inhibit or enhance it, depending on whether individuals have common or private value preferences. We also find that ranking algorithms can contribute towards the diffusion of misinformation (e.g., "fake news"), since lower ex-ante accuracy of content of minority websites can actually increase their overall traffic share. JEL-Codes: D830, L860.

The few-get-richer: a surprising consequence of popularity-based rankings?

“…Closer to our work, several papers have proposed models of the dynamics of interactions between individual searches and ranking algorithms, e.g., for understanding the feedback loop between ranking system and user queries [12], explaining the observed mitigation of search engines' popularity bias [16], or the competition of memes using limited attention [31]. The paper closest to ours is [17], which also obtains a few-get-richer effect in a model where individuals get multiple signals and where (news) items are ranked via a probabilistic popularity-based ranking. Besides being simpler, our model works with a discrete and deterministic ranking of the websites rather than a continuous and probabilistic one.…”

The Political Economy of New Media: Challenges and Opportunities Across Democracies and Autocracies

“…Accordingly, studying the effects of information platforms relying on ranking algorithms on the accuracy of individuals' beliefs requires a rather different approach compared to that used by theoretical models of media bias (e.g., Gentzkow and Shapiro, 2006;Mullainathan and Shleifer, 2005;Sobbrio, 2014a;Strömberg, 2004a). Germano and Sobbrio (2018) provide a first attempt to address this issue by proposing a theoretical model capturing the interaction of individual search behavior with some of the key components of automated ranking algorithms, namely popularity and personalized rankings. The insights of the model show that popularity-driven rankings have an overall positive effect on the probability of individuals reading a website reporting correct information.…”

“…Different individuals typically observe a different ranking of websites for the same search query or a different order of news posts by their Facebooks friends. Germano and Sobbrio (2018) point out that, while personalized rankings might be useful on private value issues (e.g., attributes of a commercial product), they may end up decreasing the probability of individuals reading correct information when it comes to common value issues (e.g., side effects of a vaccine).…”

Dr. Google, what can you tell me about homeopathy? Comparative study of the top10 websites in the United States, United Kingdom, France, Mexico and Spain

“…When searching a controversial health issue, whether at the scientific or social level, the type of content and the position on the topic of the best positioned websites have a determining potential to form the user's opinion (Pogacar et al, 2017;Germano;Sobbrio, 2017). The variety of points of view in the ranking seems to be related to the popularity of the topic in the public opinion (Gerhart, 2004).…”