Concretely speaking, "metaeducation" is just a handle I use to edit Wikipedia (though not so much these days). I use it for some email accounts too.
Yet over the years, this word has come to carry a somewhat unique meaning for me. Since this domain was blank, I thought I'd try and explain it as best I could.
I might be the only person who uses the words "meta" and "education" without putting a space or hyphen in-between. But I'm not the first to use them together! What instructors mean when they say this is "the portion of a learning program in which students are actively engaged with the reasons and theory being used to teach the material".
(Given this definition, Noam Chomsky would be more of a "meta-educator" than a traditional "educator".)
I have a slightly different take on "metaeducation". For me, this is a specific property you can measure in any kind of data. It is the information included by the creator intended for those who encounter a work out of context, to help them re-assemble that context.
After that description, the first example which jumps to mind for most people is watermarking. But that's not really what I mean, and I'll explain via this example from Wikipedia:
Obviously someone who makes a simple copy will also copy the watermark. But I don't think it counts for much. Pretty much anyone can write a program that stamps a line of text onto the dead center of an image—whether they took the picture or not! Plus, transparency alterations of this kind are actually rather easy to undo.
Simply put: the photograph didn't anticipate the watermark, and the watermark didn't embrace this specific picture. That means there's no conscious interplay between the photograph and the text. Under these conditions, there is nothing special which might awaken someone to what watermarks actually tell us about the identity of the photographer (or the middlemen who handled the images!)
Continuing with something else that might look like "metaeducation" but isn't really...let's study this image I found making the rounds on the Internet:

Here are examples of some questions one might have:
We tend to automatically answer these questions based on the context in which we view the work. I don't know about you—but I'd assume the three pictures were found with an image search— and were taken by three different individuals. A fourth person collaged them together, and added the caption as a joke...referring to the use of Adobe Photoshop.
I say this because there's a trend right now to mock debates over whether a photo has been digitally manipulated. Such alterations are legitimate to be concerned about—especially in light of cases like the famous faked Reuters photos. But there are those who challenge every image they come across, claiming expertise in detecting fakes which they don't actually have. (The full version of the joke is "This looks shopped, I can tell from some of the pixels and from seeing quite a number of shops in my time.")
Yet by studying this picture in isolation, it's not feasible to prove my assumptions are correct...it might not even be a joke! I'll acknowledge that it's very improbable that anyone could pass that off as a real photo to a human (even a child). But spam is a huge problem, and it is easy to flood a system with images like that are generated by algorithms. It gets considerably harder to filter them out by automated means.
Luckily, we can feel pretty certain that the images "on top" were arranged with awareness of the layers beneath. That's because the alignment of the horizon for the boat image is correct, and the positioning of the guitarist is on the shoreline. So that's a little information about the context of creation, but not going as far as the demands I have of "metaeducation".
Good "metaeducation" is difficult to remove because it is threaded into the work. I think this video set to the song of Hot Chip's "Over and Over" is a pretty good example. It pretty clearly lays out how to film a video in a green-screen process:
It would be tough for someone to pass that video to someone without also giving them a clue about the tools that were used to make it!
Revealing the tools for making a work inside of that work is a hallmark of metaeducation. In the Hot Chip video we see cameras, video monitors, motion-capture technology, and 3D wireframing! We can be quite certain that whoever made this video meant to show this—even if we could not communicate with them directly.
Also notable is that the name of the band and the title appear inside the work. It would be difficult for someone with amateur video-editing software to alter this. Not only because is the writing at an angle and in motion, but one of the band members bounces in front of it and covers it only partially!
This redundancy would go even further if the song lyrics were literally describing the actions in the video. Yet metaeducation is the information which travels with a work to teach you how to understand the context of its creation. It is supposed to leave plenty of room for the work's "education"— the message it was originally supposed to have under the creator's assumed context.
I've defined metaeducation as something a creator puts in to help viewers correctly connect to the context of creation. But a lot of times, we can get clues out of failures and errors which were not intended. An example everyone can probably relate to is when microphones accidentally make it into a shot in a movie.
Mistakes and glitches do give clues to how a work was created. But if they were not included on purpose, they don't fit my definition of metaeducation. If you meet the creator of a work with a mistake in it they can say "I meant to do that, it was to make a point that you should be looking for mistakes!"
As a curious case, consider Aimee Mann's recently-released video for "Freeway":
It starts out strange for a video from an industry professional—just by virtue of its simplicity and goofiness, and leaving timecode on the tape. But by 1 minute in she loses the lip sync—and I can't read lips so I don't know what she is actually saying. The whole thing descends into a maddening series of glitches—which unlike the timecode, would not be put on by any video hardware or software I've seen ("INSERT COIN TO CONTINUE"? "RUNTIME ERROR #H-0703"?)
Aimee is visibly frustrated, and we assume she doesn't regularly forget the lyrics to her songs. If she does, she probably pauses instead of continuing and just saying something else. So I think we can reasonably believe the context of creation of this video is that the footage was shot with anticipation of the layers to come.
It becomes less clear in cases like this video by "The Medic Droid":
The glitching here is over-the-top to the point of making the video unwatchable. It's not easy to see any references between the layers—for all I know, the footage of a woman walking around Tokyo could be from some Japanese movie. The glitches in this case are alien—non-informational.
metaeducation already has some epistemological limits. Certainly, it can't stop information from being deleted or filtered. And with the right tools, it can always be removed!
Plus you are always running a risk with any "added nonsense" in your work. Context is usually expected to be understood, and if every music video took place in a green screen like Hot Chip's—we'd get bored of it quickly. Beyond just boring people, you can actually scare them! (I think The Medic Droid video is kind of frightening...)
Still, something has struck me as being very important about this concept. So I have been thinking about techniques for putting metaeducation in my work. I don't have it all sorted out yet, and this is probably not the last I'll say on the subject. So I do invite your feedback!
(Note: I sometimes write metaeducation as metæducation. Though it visually looks like an a-and-e have been put into a single character, it is not typically pronounced the same way. It does have the nice property that I can tell when the word has passed through a gateway that isn't supporting non-ASCII characters!)