How to avoid bias in likes?

Community Building

AI-generated summary

The discussion revolves around avoiding bias in likes on a forum. Himanshu_Singh is rewarding users for high-quality posts based on likes, but wants to prevent users from colluding to inflate likes. He considers tracking the percentage of likes given by a user to a single user and has implemented a daily likes limit.

Himanshu_Singh seeks a feature or resource to address this issue. CaptainZac suggests manually checking users’ likes in their profile, while merefield recommends using a Data Explorer query to analyze likes distribution. Himanshu_Singh decides to write a SQL query to check likes distribution and acknowledges the potential for genuine likes to be misinterpreted as biased, requiring real-world testing to refine the approach.

1

I am giving users rewards for posting high quality topics on the forum. The metric I am using for high quality posts is likes from the community. However, this leaves the window open for users to collude and like each other’s posts to get more rewards.

I want to avoid that by checking what percentage of likes given by a user were concentrated to just one user? I am also using the daily likes given limit to restrict mass likes.

People must have faced this issue in their forums but I could not find a feature or a topic that addresses this issue. Can someone please share your experience or point me to an existing resource if I missed it?

Thank you!

3 Likes
2

So there’s not a direct feature that you can see the percentage, at least I don’t think, but you are able to go into their likes column and see what posts they have liked. In their profile, you can go to the activities tab and go down to like. There it will show you what posts they have liked.

If you really do want to see the percentage, you can post in Marketplace and someone can make a plugin/theme for you (if you’re willing to pay for someone to do it.)

4 Likes
3

You can probably come up with some Data Explorer query too if you don’t need a UI.

You can see the query code I use to generate a connectivity score in Community Network Visualisation

5 Likes
4

Thanks Robert. The visualization looks great. I have decided to write a SQL that checks for likes_given distribution per user and use it to grant or withhold rewards.

The trick here is some people could come back and say - hey, I just follow these 5 people and 1 of them is a very active poster and I like this author’s work. I am not biased, I genuinely like the posts.

Lot of nuances here but only way to sort this out is by seeing how it actually works.

4 Likes