research

Working Papers

Social Influence in Online Reviews: Evidence from the Steam Store (link)

• Winner of the Best Paper award of the Business and Industry Section of the Royal Statistical Society at NIE Conference 2024

How does social influence affect consumer ratings? Using a dataset from the popular Steam gaming platform I investigate how quality judgements depend on pre-existing consumer assessments. In 2019, Steam introduced a new review system which decreased the exposure of users to previous ratings. Firstly, I find that user ratings are dependent on average ratings. A 10% increase in average rating increases the probability a review is positive by 5.4% before the policy change, but only by 2.8% after. The result is not due to selection, and is robust to a wide range of alternative specifications. Secondly, the effect is heavily asymmetric: individual reviewers are more negative when exposed to a lower average rating, but do not respond to a higher one. This negativity compounds and inflates the gap between lower rated and higher rated games. Overall, these social influence effects are driven by less experienced users on the platform. Finally, using estimates of owner data, I run a structural model of game choice. A 1% increase in rating is equivalent to a 2.5 dollar price reduction. This suggests social influence has large implications for buyers and sellers.

Work in Progress

Can Hate Speech be Banned Online? Evidence from Reddit with Lily Shevchenko

Is deplatforming effective in reducing toxicity on social media? To answer this question we study a change in content policy on a large social network Reddit in June 2020. The change led to a simultaneous ban of approximately 2000 forums containing hateful content, but not the users of these forums. We use the complete history of user activity on Reddit to examine the impact of the ban on the behaviour of affected users. Preliminary results suggest that, although on aggregate users of banned forums become less active, the effects are highly heterogeneous, with least active users posting 10% less after the ban, but the most active ones posting 7% more. Increased activity of “core users” is not associated with higher toxicity: instead, most active users of banned forums substitute away from ideologically similar non-banned communities, become more positive, and reduce their use of hate speech.

Selection in Online Reviews: Evidence from the Steam Store

How good are reviews as signals of product quality for consumers? Using a data-set derived from the popular ‘Steam’ gaming platform I investigate the self-selection of reviewers. A policy reform on Steam in 2019 lowered the transaction cost of reviewing, with this randomly occurring within a game and reviewer’s life cycle. I find that the new individuals elicited to review by the policy change are 4% more likely to rate any game positively, leave 20% shorter reviews and are less experienced both within and across games. This selection is heterogeneous across games, greatly affecting their rank order. This suggests that product ratings are not even ordinally robust in the presence of user selection. The policy reduced selection bias and improved the consumer’s information set, but new reviewers were rated as less helpful by their peers implying a trade-off for platforms in who leaves a review.