"What we learned from the experiment:
Those in the experiment could still see and use the public dislike button, but because the count was not visible to them, we found that they were less likely to target a video’s dislike button to drive up the count. In short, our experiment data showed a reduction in dislike attacking behavior 1. We’ve also heard directly from smaller creators, and those just getting started with their YouTube channel, that they are unfairly targeted by dislike attacks. Our experiment data confirmed that this behavior does occur at a higher proportion on smaller channels."
It's not about appeasing Disney, it's about discouraging shitty behavior.
> because the count was not visible to them, we found that they were less likely to target a video’s dislike button to drive up the count
How is this any different than users targeting a video with lots of likes to drive up its count, though?
If YouTube were hiding both the public like and dislike counts, I wouldn't give this a second thought. But to say "seeing counts influences people" as a reason for hiding dislikes, and not extending that same argument to likes as well just seems a bit shallow.
Public counts are useful because they provide a gauge of perceived quality. If I see a video with lots of likes and no dislikes, it's fair to assume I'll like that video. If I see a video with lots of likes and I can't see the dislikes, I have absolutely no idea whether I'll like it or not; maybe it's a really good video and everyone liked it, or maybe the majority of people disliked it but I just can't tell. If you're removing my ability to gauge quality anyway, then why not just remove the like count as well?
This change is essentially shifting from "three out of five dentists prefer Trident" to "three dentists prefer Trident". The loss of information is so substantial that the remaining information is essentially useless.
I don't think the ratio of likes to dislikes is a reliable indicator of quality. It probably will be for certain categories of video, but not for anything polarizing like a review of an Apple product or anything even remotely politics-adjacent.
Totally agreed, that's why I wanted to call out that it's a measure of perceived quality.
If there's really great documentary about the International Space Station, I'd expect flat-earth believers to think that it's rubbish, that's just part of being human.
Conversely, if I saw a documentary about abortion (or some other very polarized topic) which had a 9:1 like-to-dislike ratio, the difference between that ratio and my expected 1:1 ratio would tell me that lots of people thought that the video was high-quality, even those who didn't share the same ideological views.
"Dislike attack behavior" could just be a video that is widely disliked. It doesn't necessarily mean there's any sort of coordinated attack on it, so your "shitty behavior" comment doesn't really fly.
So people watch it without the prejudgement and bias of knowing what the wider group thinks and dislike on their own judgement. That seems like a simple solution to coaching people into good curation patterns and behaviour.
It will likely so give some good results in the background for YT and media partners but it is allowed to be both.
If nobody can see the dislike number, what is the point of disliking a video? It doesn't make the video less likely to appear on feeds. Having the number makes disliking a video feel like you've actually done something rather than just piss in the ocean.
This seems like the wrong target to me. Why not go after the user?
If there's a user that is going around disliking a ton of videos, especially on a single channel, and/or without watching some minimum time or % of that video, it seems likely that user is being abusive; so perhaps shadow-ban that user's votes?
Similar with new users.. disallow use of dislike completely until they've been around a while and built up some reputation/history (by watching, upvoting, etc). If the user is possibly a bot (I assume there's algorithms to determine confidence of bot vs real user), require more verification.
I'm sure there's a reason they're not taking this approach, but I find it hard to believe they're choosing they way they are based on the reason they're giving here.
"What we learned from the experiment: Those in the experiment could still see and use the public dislike button, but because the count was not visible to them, we found that they were less likely to target a video’s dislike button to drive up the count. In short, our experiment data showed a reduction in dislike attacking behavior 1. We’ve also heard directly from smaller creators, and those just getting started with their YouTube channel, that they are unfairly targeted by dislike attacks. Our experiment data confirmed that this behavior does occur at a higher proportion on smaller channels."
It's not about appeasing Disney, it's about discouraging shitty behavior.