- Joined
- Feb 26, 2024
This is true but doesn't take into account that the people who download this extension are a segment of YT users who care about like/dislike ratios enough to go ahead and do that, which is not at all reflective of the overall YT userbase. Hence selection bias is still a valid concern and the real stats are likely still off. We just don't know by how much.The selection bias argument is kind of a moot point. Hundreds of thousands of people have downloaded that dislike extension. People assume sample sizes need to be massive. In reality it's probably a fairly accurate representation of the dislikes.
Just an example to demonstrate how small sample sizes need to be, but if you wanted an accurate poll of the 4m people that watched that video, you'd only need to a sample size of 385 people.
View attachment 7241021
I am sure Idubbbz's screenshot is deceitful, there is no way he has a 85+% likes on that video. The truth is probably somewhere between what he posted, and what the extension is telling us.