November 21, 2007

Conservapedia, Homosexuality, and pranked statistics

The "Conservapedia" Homosexuality statistics story being echoed, where the right-wing Wikipedia-style site "Conservapedia" allegedly has nine out of its top ten most popular pages being against homosexuality, cannot be correct. That is, whether by accident or design, the alleged statistics don't pass the sanity test (I know, I know ...). The site's "Most viewed pages" statistics supposedly are:

1. Main Page [1,902,822]
2. Homosexuality [1,542,919]
3. Homosexuality and Hepatitis [516,745]
4. Homosexuality and Promiscuity [420,172]
5. Homosexuality and Parasites [387,876]
6. Homosexuality and Domestic Violence [351,556]
7. Gay Bowel Syndrome [343,273]
8. Homosexuality and Gonorrhea [331,090]
9. Homosexuality and Mental Health [277,119]
10. Homosexuality and Syphilis [265,042]

Except this makes no sense. While the "Homosexuality" page itself might be highly ranked, the "Homosexuality and Hepatitis" page is short and has been in existence only since October 17. There's no way something like that would a legitimate third-most popular page, even for raving homophobes.

And the top ten doesn't have "Bible"? Or "Jesus Christ"? [update - better: any other controversial topic?]. Those are supposedly less popular than "Gay Bowel Syndrome"?? That's ridiculous (I know, I know ...). Either a spider has run amok or someone is deliberately inflating the pageviews.

Of course, this post will have near-zero effect on a story "too good to check". Let's hear it for the self-deluding nature of the bogosphere and the futility of trying to be heard :-(.

By Seth Finkelstein | posted in wikipedia | on November 21, 2007 09:28 AM (Infothought permalink)
Seth Finkelstein's Infothought blog (Wikipedia, Google, censorware, and an inside view of net-politics) - Syndicate site (subscribe, RSS)

Subscribe with Bloglines      Subscribe in NewsGator Online  Google Reader or Homepage


Some of us at RationalWiki have manipulated page stats in the past, this was not a concerted effort and no one has claimed responsibility but occum's razor dictates that is probably someone from our camp.

However, your assumption that Jesus or Bible should be popular is wrong. CP is an "anti" site and things like abortion, evolution and homosexuality receive all the attention and promotion.

Posted by: Tmtoulouse at November 21, 2007 09:57 AM

This is what their fromt page said, below. Though you're right, I should have made my argument more general. As in, it's absurd no other hot-button topic would be in the top-ten, rather than the obscure Homosexuality-related pages.

Popular Articles at Conservapedia

* Examples of Bias in Wikipedia
* Theory of Evolution
* Homosexuality
* George W. Bush
* Jesus Christ
* Global Warming
* Barack Obama

* Bible
* Atheism
* Conservative
* Dinosaur
* Liberal Bias
* Creationism
* Big Bang Theory

Posted by: Seth Finkelstein at November 21, 2007 10:38 AM

Really the "bread and butter" of their site is their anti-evolution diatribe. The biggest problem in parsing CP in regards to stats is that sooo much of their traffic is of the "point and laugh" variety. Hence why dinosaur and kangaroo and unicorn, etc. were so popular. Most of our page bumping attempts were washed after we did them. But if someone is bumping homosexuality related articles they won't be washed. The main whitewasher is completely obsessed with homosexuality. He is the creator of the over a dozen articles and has edited homosexuality related articles over a thousand times in a others at rationalwiki have said even if you are skeptical of the "homophobe=repressed homosexual" argument his obsession strains credibility about his sexual orientation.

Posted by: tmtoulouse at November 21, 2007 11:28 AM

Problem resolved

Posted by: Shii at November 21, 2007 03:21 PM

Sorry, the Argument from Incredulity holds little favor among rational thinkers.

Got any evidence? That usually works.

Posted by: melior at November 21, 2007 07:21 PM

Shii: That's too unusual a search to account for the high numbers.

melior: Who has the burden of proof? The page above is merely a list of statistics. One hypothesis is that it represents the popularity of topics among Conservapedia readers. Another hypothesis is that it represents a bot-prankster forcing the stats into a parody of what's popular among Conservapedia readers. Given the unlikeliness of that list containing no other topics, versus the claim that similar pranks have been done to the site before by rivals, which hypothesis seems more reasonable?

Posted by: Seth Finkelstein at November 21, 2007 10:45 PM

Argument from incredulity works perfectly well for UFOs. What's the problem, again?

Posted by: Blah at November 21, 2007 10:50 PM

I for one am posting a correction. I expect as the word gets out, so will others.

Posted by: Josh Jasper at November 22, 2007 10:12 AM

Yep, Seth is right, it's definitely pranked.

I actually started writing a blog comment here saying maybe it was plausible, and then I thought to go check the Internet Archive. Turns out we have snapshots of the Conservapedia Statistics page dating back to December 2006. I grabbed all the data it had, typed it into a spreadsheet, and looked at the chart.

The giveaway is the "Homosexuality" page itself. That page was getting 30,000 pageviews a month from March 15 to July 15. That's consistent with the 40,000/month rate it had from March 6 to March 15, shortly after it was created.

Then from July 15 to today, Nov. 22, it got 350,000 pageviews/month, an increase of 11 times. That's especially not believable considering that the Main Page on Conservapedia got only 125,000 pageviews/month during that same period (which is just a slight increase over its March-July rate).

The "Homosexuality" page from March to July, the most recent period available, got 13.9% as many pageviews as "Main Page." But from July to November it got 281% as many.

And as you point out, Seth, since "Homosexuality and Hepatitis" was created on Oct. 17, we know its 517,753 pageviews arrived in those 37 days, or 420,000 pageviews/month. It's not believable that since Oct. 17 that one small page has been 338% as popular as the Conservapedia Main Page.

The Internet Archive stats also let us calibrate the "Homosexuality" page against the "George W. Bush" page from 3/6/07 to 7/15/07. It had 21% more pageviews than GWB in March, and 5.7% more in July. This strongly indicates that no tampering was done with the "Homosexuality" pageview count as recently as July.

Extrapolating the March-July pageview rate to today, the most popular "real" page on the site, besides the Main Page, is probably indeed "Homosexuality." But not by much (a few percent). The 3rd, 4th, 5th and 6th pages are probably "Examples of Bias in Wikipedia," "Unicorn" (!?), "George W. Bush," and "Theory of Evolution."*/

Posted by: Jamie McCarthy at November 22, 2007 11:33 AM

Wouldn't a big spike in a few isolated but related pages be entirely expected if there was some other larger site which posted links to those pages? Ie, say slashdot, or more likely wikipedia or something like that linked directly to those pages.

That said some of these seem like pretty unlikely pages to which to link directly.

Posted by: Gerg at November 23, 2007 04:40 PM

Ok guys, yes it's botted, I am one the guys that has helped to do it, although my contribution is only relatively minor, and I have no connection with the other guy(s) that did it.

I don't think anyone expected this to be picked up by the blogs in such a major way, or that so many people would for a moment think it was real. And yes, it was funnier a week or so ago when the order of popular pages was a bit more subtle.

People have been bumping other pages on CP for a couple of months now, but CP would eventually reset the statistics for them (quite easily, by deleting and recreating the page). After a break, the Homosexuality pages were targeted, I think to see if the guys at CP would actually be embarassed at having all the Homosex pages at the top of their stats. From the looks of things it seems they're not, and that they're quite liking the spike on the Alexa graph that all this coverage has given them.

Posted by: Anonymous Coward at November 24, 2007 08:21 AM

Kieran's update:
"Update: As emerges in the discussion below, this Top 10 is a little too good to be true, and probably reflects efforts to game the system either by critics or other participants in the Conservapedia world rather than the true degree of readership for these particular pages on the site."

In other words: "Mistakes were made. Guns were jumped. Facts got checked. Investigations were undertaken. Sources came forward. Hypotheses got validated. An update was issued. Blogging goes on."

Posted by: Jon Garfunkel at November 25, 2007 11:42 AM

It matters less whether it's true or not than whether it's funny.

Me, I reckon it's pretty funny.

Posted by: stavrosthewonderchicken at November 26, 2007 01:47 AM

Despite the spoofed numbers, those pages probably do represent some of the most visited and edited. As TMT commented above, the are very much an "anti-" site, not a "pro-" site.
For example, they had no article charity and other basic christian values, until relatively recently when an "outsider" started them. They had no page on Apologetics until an outsider started it.
Basically, it isn't really a "christian" wiki, it is a hate wiki.

Posted by: PalMD at November 28, 2007 12:02 PM

That's hilarious, I guess that shows what the religious right has on their mind.

I made my own spoof site

Posted by: Chris at November 29, 2007 12:57 PM

It was probably someone at RW using a bot-since the homosexuality page on CP is FAR more viewed than the second most viewed wikipedia article. ED!

Posted by: Jeb Berkeley at December 1, 2007 03:23 PM