This would take a lot of teamwork, to share the research load.

It might possibly bring about a re-examination of Section 230, and it certainly would change a few minds about Wikipedia.

Some of you have seen Kato's and my recent "technique" of looking at vandalized edits on articles, assessing the length of time the vandalism stayed in place, and then estimating (using the Henrik-o-Meter) how many page views the vandalized edit received before it was corrected.

I propose that we take two months -- January/February 2008 -- and carefully quote verbatim, tally the incidence, and assay the duration of any vandalistic edits to the following articles:

435 U.S. Representatives
100 U.S. Senators
9 Members of the U.S. Supreme Court
1 Attorney General of the U.S.

We would compile this into a master password-protected spreadsheet (Google Spreadsheets?), then send hard-copy letters to each of the 545 individuals itemized above, indicating their own particular case study's results, and pointing them to the spreadsheet online for the full assessment of the 545.

Some of these edits won't be embarrassing, but just wrong, such as Senator Christopher Dodd living in Iowa, rather than in his Connecticut home; or a report of a fictional car accident. Other edits will be a bit disconcerting. Still others might possibly be true, but are worded in an embarrassing way.

Others articles, like [[Rick Renzi]] and [[Trent Franks]] will have nothing vandalistic at all over a two month period (congratulations, Wikipedia!). But, in these cases, we could do the sensationalist thing and find a good doozie somewhere in the article's history and just post that as an "historic example", but lying outside the boundaries of our more scientific "study".

Do we think that this is worth doing? It's probably going to take a good 30 minutes per article to review edits and copy/paste the doozies to the common spreadsheet. However, I think that this effort will garner the Wikipedia Review a lot of credibility, and it will certainly bring unwanted media and governmental attention to Wikipedia and its BLP mismanagement. If we had 10 volunteers sharing the load, that would be less than 30 hours' time per person -- easily managed within a month, right?

Let's discuss the relative merit of this plan before we solicit a volunteer team.

Greg

I think that this is an extremely practical idea which would clearly illustrate why law-makers need to examine Wikipedia more closely. I'm in.

Good idea.

Public Radio's On the Media had a story segment last week that touches on this issue. Brooke Gladstone interviewed Eve Fairbanks, Associate Editor of The New Republic about how editors on Wikipedia are shaping articles about candidates in the 2008 Presidential Election.


The audio segment can be accessed from On the Media's web site or downloaded as an MP3 file.

The original article by Eve Fairbanks in The New Republic is entitled Wiki Woman and includes reader comments.

The selection of the 545 is of obvious advantage - in the interest of demonstrating there's a problem of wider concern, how about doubling the sample size to include a group of US citizens, and relating them (as far as possible) to constituencies?

It makes for more work, but it would spotlight the potential for abuse of pretty much anyone, and would be a bullet-proof vest against the defence that "political BLP's" are special case (especially in the heat of an election year).

Very good idea. We could include mayors of major cities, governors, certain notable people (perhaps looking at who contributed to the person's campaign??), tieing things as much as possible into the issues that the political figure tries to defend. This would take some research, but it would be much more effective and if we split things up, it wouldn't be that much more work, really.

In the interest of getting 10 people dedicated to assisting this project, I suggest that we keep our "study" limited at first. If it goes over like a lead balloon on Capitol Hill, then we're going to be kicking ourselves for having researched 1,000 entities, rather than just 545.

In fact, this gives me an idea -- in the opposite direction.

We would build more "rolling interest" among the media and in government if we started with just the 100 U.S. Senators. With the release of those results, we could say "watch next month for the Wikipedia Review's report on vandalism to articles about U.S. Representatives". With that one's release, we'd have a preview of "the Wikipedia Review's report on vandalism to Capitol Hill's leading campaign donors". Then celebrities. Then attorneys and judges. Then state government officials. Then world leaders. Then business tycoons. Etc. Etc.

We could become like the "J.D. Powers" reports, rolling out on a monthly basis, such that everyone in the media is salivating in eager anticipation of the results.

I hereby propose that we simply begin with the 100 U.S. Senators, see how that goes, then undertake the next waves of defamation, one by one.

To get more data points, maybe we expand the assessment period to January through March, 2008.

Greg

That makes great sense: it's best to spread it out to get better impact.

...plus it's less daunting a task.

Three months sounds like a good start.

Sounds like a good idea. It's particularly timely because this is an election year, and the political biographies will inspire more input, with more consequences for the subject of that biography.

My guess is that Seigenthaler would probably approve of this project, based on what I know about his views from a couple dozen telephone conversations over the last two years. He might be able to help get it noticed in Congress and the media.

I think we should select a sample, and keep it quiet until we get the last three months history downloaded for each bio in the sample. If there's any funny business by oversighters, we'll be able to expose them.

Also, some data on the extent to which defamatory material was propagated around the web, on scraper sites and/or blogs, might be interesting. Scroogle could help out by searching for exact sentences, in a somewhat automated fashion (like I did with the plagiarism report).

I can also write programs to download three months of history for a given bio.

I'm a bit confused. Removing defamatory versions from article histories is a bad thing? Oh, so that's why you were so strongly opposed to the deletion of the history of the article about you, and so supportive of JoshuaZ's attempts to stop it getting deleted.

You have that 180 degrees backwards, or maybe you are trying to be sarcastic. I was supportive of Doc_glasgow's efforts to bury the history so that it could no longer be found by the simple procedure of turning off the redirect from Daniel_Brandt to Public_Information_Research, and then clicking on history at Daniel_Brandt, and then reading all of the past versions. This would have invited every Tom, Dick, and Harry to splice items from my former bio into other articles, just to show me who's boss.

Removing defamatory versions is a good thing, but not while research is underway to document the scope of the problem.

I was trying to be sarcastic, wasn't it obvious? I actually tried to help get it deleted, by the way - it was my idea to preserve the history listing in text form to satisfy the claims of GFDL requirements (you may remember that at the time, east718 mentioned a discussion on IRC with an "unnamed admin" - I didn't really think my involvement there was a secret, I created the listing page myself)



fair enough. The other issue is, the 545-article set Greg Kohs suggested is fairly well-known.

Random832, you are showing advanced signs of wikipedia-speak, and wikipedia-think. Just about everything gets twisted, the messenger gets shot, and it's the same tired old business as usual, with the same tired old miscreant methods.

You admins at Wikipedia should be down on your hands and knees thanking Brandt for doing something about BLPs. He is doing a great service, and WP is beginning to take heed. And if WP don't take heed, their number will be up very shortly. All it will take is one little court case to crack the egg, then all-comers will stampede in.

That was why I was confused - having a problem with _removal_ of defamatory versions (since all oversight is capable of is removing things) seemed like such a blatant reversal that I was honestly surprised, until it was explained to me.

Probably best to get this underway PDQ, before anyone has time to pick through with their oversighting buttons ready?

Good idea, I could probably do a few local senators and congresspeople as well and get them in the loop. I'll probably take a look at Jane Harman's article tomorrow.

See tobacco-related lawsuits for an example. If people don't like you and are jealous of your power and influence, you can argue till you're blue in the face about "fairness," and it won't get you anywhere.

Which it comes Jimbo's time to do the sacrifice dance, he'll no doubt do a great "innocent persecuted lamb of God" imitation. As though nobody gave him any warning about how he was headed to get stomped on, for his nasty BLP policies. But the jury will get stuff like what we've been talking about for years, and stuff those Marsden chat transcripts where he shows he knows the problem, to show that he knew exactly what he was doing. So it's gunna cost him, is my guess.

He will have blown the funds by then.

Okay, I have enough volunteers to launch Phase I of the project. I just hope we're faster than the Oversight team.

We will need more volunteers in the future if this first wave is a media success.

Greg