This really should be a comment to an answer from a now deleted account, but I want to elaborate on a point, and also get a reality check from others. So an answer it is.
The truth is that 99% of math.stackexchange is just people repeating information they learned in some class or from reading a book or lecture notes. There is no actual difference between helping someone in algebraic topology or functional analysis and calc 1 or linear algebra besides ego. The gatekeeping of, "oh, is this question good enough for us to reply?" is nonsense. If it's below your standards, just ignore it.
The complaint seems to be that the bulk of janitorial work is directed at lower level questions (=sophomore or below) while advanced undergrad and graduate level homework questions in their opinion fall through the cracks.
Given that at least here I wanted to give more advanced questions some slack, I took this a bit personally in the sense that the accusation somewhat fits my tendencies. I thought about this, and want to share.
Yes, I have voted to put on hold at least ten times as many calculus / elementary-number-theory / combinatorics questions in comparison to higher level questions.
But this is not really such a heavy bias in my voting policies, nor those of the other CURED-activists.
- Sheer numbers. There are many more low quality questions in lower level tags. Simply because world wide there are more students at those levels.
- I'm on record for fighting duplicates with passion. The more advanced questions are much less likely to be duplicates. The number of questions at a particular level correlates with the number of people at that level in their studies, and this forms a triangle (or a pyramid) sitting on its low level base. The amount of mathematical knowledge, on the other hand, forms a tree-like structure, more like an inverted pyramid. These two factors imply that the duplication is disproportionately concentrated on the lower level tags. Given that I scan the close queue mostly for posts that PSQs as well as duplicates (particularly those likely to become fodder for FGITWs), this will make my "vote-to-close" stats look skewed towards lower level. My desire to "punish" FGITWs probably skews it further.
- If I run into a PSQ in stochastic math I will likely just click "Skip" simply because I lack the expertise to determine whether sufficient context has been given in the question. Likely many other follow similar (intellectually honest) policy, again with the net result that basic questions get more attention from the reviewers.
- But, I'm just a single reviewer. Judging from my observations the more systematic CURED-activists don't pay attention to the factors of the previous bullets. For them a PSQ is a PSQ irrespective of its level. Undoubtedly there are many other active reviewers all using their subjective criteria.
So it is not surprising at all that the net effect is that lower level questions are more likely to be put on hold by the reviewers. On the other hand, seeing this as a bias favoring deeper stuff seems misguided, when statistical explanations exist.