As I post this (August 29, 2019), the page has the "No Real Life Examples, Please" category - not from having the "noreallife" template and an accompanying reason, but a raw category with no reason given. The page also has a non-trivial number of real-life examples.
What do we do? Delete the category? Delete the real life examples? Some unholy mix of both?
I recommend deleting any examples of non-notable people at the very least, especially if it seems like it was written in a fanfiction style. I am not against the idea of talking about real life abstractions of the Trope itself however.
I similarly wouldn't object, provided we address those specific examples in an appropriately tactful manner.
Let's not bandy about "notable" as a qualification, because There Is No Notability. That said, I would like some indication why some of these examples are worth cataloging -- some entries just give a name and their relationship, and I have no idea who they are or why/if they really merit an entry. Others -- like Mary Kay Letourneau -- are famous for this trope and I feel are ideal examples to retain.
Honestly I think the "no real life" warning was intended more to prevent a plethora "this teacher from my high school married this girl I knew as soon as she graduated" and "rumor was my professor boinked his female students" entries than to prevent any examples from non-fictional contexts.
Yeah, I concur on making sure any IRL examples are ones that at least make the news. Keeping random rumors off the page is entirely sensible, but IRL things that have objective news citable proof should definitely stand.
This keeps things from going into the gutter and avoids potential lawsuits for unsourced allegations.
To avoid this turning into Portal:juicy_tabloids, maybe limit to just the cases (whether proven or alleged) that inspired work(s) of fiction listed on the page (such as Letourneau). And perhaps para-scientific studies can be grandfathered in.
I like that idea - it's very much in keeping with the wiki's goal to look at how fiction is put together.
TBeholder's suggestion is probably the most agreeable consensus that we'll get.