I've been fixing tag wiki plagiarism lately. Recently I've been getting a lot of "reject" reviews (though only a very small number of actual rejections--accepts are still largely cancelling them out). I think it's time for me to explain what I've been doing and get input from the community.
We have lots of plagiarized, or otherwise improperly cited tag wikis. By lots, I mean at least hundreds, and probably more.
I've been fixing some of them lately. How I have been doing it depends on the specifics of the problem. I've seen three main subtypes:
Uncited copied content in full tag wiki.
Coped content in excerpt.
Excerpts are usually seen without the full wiki accompanying them, so citations in the full wiki don't cover the excerpts. Excerpts with copied content that don't contain a citation in the excerpt are plagiarism. Excerpts also don't support formatting and are supposed to provide a very quick introduction to what the tag is about; for these reasons, copied material, quoted and cited, is undesirable too (though potentially acceptable).
Copied content in full tag wiki that has enough of a citation that it's not represented as our work (so not actually plagiarism), but where the citation is wrong (e.g., points to the wrong site or the wrong page on the right site) or worthless (e.g., consists just of the name of a site or company but is not a link at all).
All subtypes are common.
My goals in submitting tag wiki edits to fix plagiarism are:
- Reduce the amount of plagiarism on the site.
- Efficiently fix plagiarism, by focusing on it specifically; for example, I often fix a not-very-good plagiarized tag wiki so it is a not-very-good, not-plagiarized tag wiki. My goal is not to make every tag wiki beautiful, but just to get rid of the plagairism.
- Find and submit edits to a significant (not necessarily majority) percentage of all plagiarized tag wikis on Ask Ubuntu.
So far my general workflow has been:
- Go through the list of tag wikis I reviewed, and find instances where I rejected for plagiarism, or accepted a tag wiki that I shouldn't have because, on reflection, it was plagiarism. For each of these, check if it was ultimately accepted; if so, fix it.
- When I find someone who submitted a plagiarized tag wiki suggestion, go through their other suggestions and look for plagiarism.
- When I find someone who has frequently approved plagiarized tag wiki edits, go through their list of reviews to see what else they approved. If it's plagiarism, fix it, then return to step 2.
This is (roughly) an algorithm. It is not a moral analysis.
- I've found that this seems to yield plagiarized tag wiki edits extremely quickly and efficiently, especially compared to looking and searching through tags themselves.
- It has the result that my own tag wiki edit suggestions (submitted to fix plagiarism) come in blocks where I edit posts last edited by the same user for a while, before moving on to another user.
- While the revisions I'm editing belong to the same user for an extended time, I have not actually seen evidence yet that only a small group of users are responsible. At this point, I simply don't know. I'm not sure if I'll continue working on this problem long enough to find out.
- To sum up: I focus on users because users are the glue that connects plagiarized tag wikis to other plagiarized tag wikis. This is not about the users. It's about fixing plagiarism in tag wikis. As long as we have plagiarized tag wikis, the Ask Ubuntu community as a whole is offering and promulgating plagiarized content.
So far my specific workflow, at the level of an individual tag wiki, has been:
- If I suspect plagiarism, I paste two or three short-to-medium sections of text into two search engines. This usually reveals if it's plagiarism, and what the original source is likely to have been. It at least reveals reasonable sources. I make sure to keep in mind that it is possible for other sources on the Internet to copy material from an Ask Ubuntu tag wiki; I don't assume that duplicated content necessarily means we're the offending party.
- If citation is appropriate, I add it (or if there is a citation but it's wrong or not up to par, I improve it).
- If citation is not appropriate (as in tag wiki excerpts), and I am familiar with the topic, I rewrite.
- If citation is not appropriate and I am not very familiar with the topic, I research it until I feel comfortable rewriting, then do so.
When I rewrite material, my goal is to create something of comparable usefulness that is not plagiarized. My goal is not to create the most wonderful tag wiki ever. If the original tag wiki does not comply with tag wiki guidelines, in ways other than being plagiarized, I often don't fix that.
When I started doing this, my suggestions were accepted. The initial group of reviewers consisted largely but by no means entirely of moderators on the site.
Subsequently, a slightly overlapping but largely different group of reviewers started revewing them, and now they get lots of "reject" votes with reasons like "This edit does not follow any of our tag wiki guidelines..." and occasionally a custom reject reason about how the previous revision was better.
So: What should I be doing?
The reason tag wiki plagiarism exists is that it's quick and convenient and existing review mechanisms are insufficient to catch it much of the time. If I replace every plagiarized tag wiki with a well-written, excellent, totally helpful resource, I'll replace a very small number of tag wikis--possibly smaller than the number of plagiarized tag wikis that get submitted and accepted during the same time. Even if no new plagiarized tag wikis were accepted, I would still fail to make a perceptible dent on the plagiarism we already have.
Besides the plagiarism, sometimes my revision isn't better. Sometimes, what I write about a topic is not as good as what's copied improperly from the work of someone deeply familiar with the topic who writes about it frequently. This usually does not deter me--it seems to me that it's worth replacing plagiarism with something that approaches the value of the original, plagiarized version, even if it falls short in some ways.
It seems to me that there are three ways to deal with existing plagiarized tag wikis:
- Fix them efficiently. (As I'm attempting to do.)
- Leave them. Slowly replace some of them with very high quality tag wikis. (I see this as a bad option; as a community, we would be continuing perpetually to publish works we know are plagiarized.)
- Make a list of them so moderators can nuke them, removing the offending tag wikis entirely. (This would result in hundreds of tag wiki excerpts disappearing overnight, and might be considered undesirable.)
But I should not assume my perspective on this is shared by everyone in the community. So what do you think? Should I continue what I've been doing? Stop entirely? Continue but with some changes?