Why does a good word (praise) help? The effect of positive feedback on contributions to online public goods
Good Question! The Effect of Positive Feedback on Contributions to Online Public Goods
HCI Today summarized the key points
- •This article reports a study that tested how a single praise indicator received on Stack Overflow changes people’s subsequent participation.
- •The researchers randomly attached anonymous recommendations to 22,856 questions and then observed changes in activity over the following four weeks.
- •People who received recommendations were 6.3% more likely to ask questions again, and 12.9% more likely to answer other people’s questions.
- •There was also an effect of making questions more visible, but the impact of asking questions was driven more by the praise itself, while answering was driven more by the visibility effect.
- •In other words, even small positive feedback increases participation in online communities—especially by helping people sustain helping behaviors over time.
This summary was generated by an AI editor based on HCI expert perspectives.
Why Read This from an HCI Perspective
This article experimentally shows whether very small feedback signals—like ‘likes’—can change people’s next actions. For HCI practitioners and researchers, the key point isn’t how smart the model is, but how users interpret that signal and decide to re-engage. In particular, it separates out the effects that arise when recommendation, ranking, and notifications operate together, which directly informs real service design.
CIT's Commentary
The core of this study is not the upvote itself, but the way an upvote bundles two things together: ‘human recognition’ and increased system exposure. In interaction design, this kind of bundling doesn’t always come with only benefits. Users may feel praised, but at the same time they come to expect more visibility and faster responses. So when you add a piece of feedback, you need to consider both what signal it sends and what failure modes it may create. Especially in Korean community services or AI-agent-style products, you need to design not just a simple reaction button, but also when users can intervene and whether, when a response arrives, it actually leads them back to participate—because that’s what sustains engagement. An interesting follow-up question is whether answers generated by LLMs can replace these social-reward effects. If the intervention comes from AI rather than a person, satisfaction might increase while community participation could decrease.
Questions to Consider While Reading
- Q.When you separate the direct effect of upvotes from the exposure effect caused by ranking, what combination of design choices in real products can increase participation with the fewest side effects?
- Q.In a Q&A environment where AI answers first, what interface is needed to preserve the ‘help users receive’ while not losing their re-engagement with the community?
- Q.If the effect of good feedback ends after just one instance, why wouldn’t a second round of praise or rewards add anything? What UX hypotheses could explain this nonlinearity?
This commentary was generated by an AI editor based on HCI expert perspectives.
Please refer to the original for accurate details.
Subscribe to Newsletter
Get the weekly HCI highlights delivered to your inbox every Friday.