Imo, it likely was/is due to the voting system — and, in a similar sense, awards. Redditors want to increase their Karma scores and seem to, at least subconsciously, view it as clout. So, they’ll create posts with the intent of farming these points — ie they post things that they know will get a specific response from the masses. What also doesn’t help, and is something that Lemmy similarly suffers from, is that there generally is no established consensus on how votes should be used. An upvote could mean agreement, or that a post is funny, or that it’s good quality, or that it’s on topic for a community, etc. A downvote could mean that the person disagrees with the post, or that they think that it isn’t relevant or they simply don’t like the OP. In reality, all that votes do, at the fundamental level, is tell the algorithm where it should place posts (a personalized recommendation algorithm changes this a bit, but the effect is essentially the same) — a post with a large upvote to downvote to ratio gets shown higher up and, by extension, more than one with a smaller ratio. This creates a sort of feedback loop where the posts that get farmed for upvotes get shown more. People don’t want their post to be buried, so they’ll only post what they think will get upvotes. And since upvotes are usually used for things that illicit an “agreement” response, only posts that people agree with will be shown.

The solution to these issues, imo, is to create an obvious standard for how votes are used and change how they’re interpreted by the algorithm. Imo, Facebook was on the right track with how they were using emojis as the voting method. People generally react to posts with emotion, and an emoji is a good representation of that. You could potentially still have an up/down form of vote (alongside the emotional voting options), but it would be standardized to only be used as a metric for relevance/importance/correctness. This could be enforced by moderation, if votes were publicly viewable, by allowing moderators to remove people that are vote brigading (not including emotional votes). Emotional votes probably shouldn’t be considered by the algorithm so that emotional bias can be avoided. Or, at the very least, there should be different algorithms that take these voting types into account I’m different ways (eg if you only want funny posts, you could sort to primarily get posts with a laughing reaction). In addition to this, also removing the gamification aspect (not showing (at least not publicly) total scores on profiles).