Exploring the Role of Negative Peer Feedback on Social Media

What are the implications of negative peer feedback on social media posts, and how can content creators use this feedback to alter messaging? On this episode of Michigan Minds, Jessica Fong, PhD, discusses findings from a new study on the role of negative peer feedback on social media, how creators can use this information to enhance their presence on platforms, and the ways in which feedback encourages users to moderate their tone. Fong is an assistant professor at the University of Michigan Stephen M. Ross School of Business, where she researches matching markets, platform design, advertising, and behavioral economics.

Fong recently coauthored a study on the role of negative peer feedback on social media platforms. She explains the background experiences that led her and her colleagues from Michigan Ross to this area of research: they were interested in the topic of user-generated content (UGC) and what drives people to create content. For example, why does someone post a review for a restaurant or a comment on a YouTube video or a post on Reddit? 

“As we were looking into this topic, something that stood out to us was how YouTube had changed its policy about the ‘dislike’ button. Before 2021, YouTube users were able to dislike videos on YouTube and the number of dislikes for each video was public. In 2021, YouTube changed its policy where users can still dislike videos, but they can no longer see how many dislikes each video gets,” says Fong.

According to YouTube’s blog, the reason the platform made this change was to reduce dislike attacks, which is where people work together to drive up the number of dislikes for a creator’s video. Fong says that YouTube’s policy change is what inspired her team to think about how on some social media platforms, you can see some semblance of the number of likes and dislikes, like on Reddit, and on some you can’t.

“We became interested in how dislikes impact the creation of user-generated content. And we chose to study this on Reddit for a few reasons. First, Reddit is just a really popular website—it’s one of the most highly visited websites in the world, so there’s really not a better place to study UGC. We also needed a platform that had downvotes, and we as the researchers can see … approximately how many downvotes each piece of content receives. And the third reason is we wanted a setting where there’s no objective right answer—so that upvotes and downvotes generally reflect approval or disapproval of an opinion. And this is important if we want to study the implications on echo chambers and polarization.”

The study found that negative feedback, represented by downvotes on Reddit, actually encourages people to post more on Reddit. So when a user’s post on Reddit sees a decrease in its score, which is the net difference between upvotes and downvotes, they are subsequently more likely to make a new post on Reddit. So negative feedback, explains Fong, actually increases engagement on Reddit.

“Something that I didn’t really expect to see was that negative feedback increases engagement on average. Of course, there’s going to be some people in the population who will respond positively to negative feedback, right? Internet trolls would be an example, but I just didn’t expect that this effect would be large enough to be true on average.”

Jessica Fong on Michigan Minds

The study also looked at how downloads impact the intensity of language. If the initial post is more intense or if the opinion expressed is more extreme, Fong says, downvotes actually cause that user to moderate their language the next time they talk about the topic.

Fong shares a potential explanation for these findings: reputation management. Content creators can use negative feedback to enhance their presence on the platform and moderate their tone. Fong references existing research that shows people care about their online reputations even when they’re posting anonymously. Reddit has a reputation system called “karma,” where karma “points” are accrued when the user’s posts get more upvotes, and are lost when they get more downvotes. 

“Karma seems to be something that Redditors really care about. There’s even something called karma farming, which is when people repost popular content in different subreddits just to get more karma. So what we think is happening here is that a Reddit user makes a post that turns out to be unpopular—they get downvotes, so their karma takes a hit. To try to increase their karma back to where it was before in response to getting these downvotes, they create more posts. Hence why we see higher engagement in terms of moderating their tone. People seem to be learning from their past posts so they know what they said previously was unpopular. So they’re going to moderate their opinion next time to avoid those downvotes in the future.”

An important takeaway from this research, Fong says, is that the way digital platforms are designed can have an impact not only on user engagement, but on what people are saying and how they’re using the platform. She notes that the common perception is that negative feedback, criticism and downvotes are bad, not only for the platform’s engagement metrics, but also for real-world issues like increasing levels of polarization. But at least in the context of this study, there are benefits of negative feedback both for the platform itself and for reducing polarization.