Conservapedia:An illustrated guide
What is going on?
|Best of Conservapedia|
Differences with Wikipedia
CP in the News
More about CP
|Conservapedia claims to be a conservative, family-friendly Wiki encyclopedia, founded by Andrew Schlafly in 2006. According to its statistics, in June, 2009, there were ~31,000 content pages, 30,000 registered users and fewer than 200 active users.||Notes: Conservapedia had troubles with its logo from the very beginning: The first one had a hidden message, and the next one (pictured) had an embarrassing mistake. (See: Conservapedia: Timeline March 2007)|
|Active users are defined at Conservapedia as Users who have performed an action in the last 7 days. Other wikis use different periods of time, e.g., at the English language Wikipedia (henceforth referred to as Wikipedia), active users are those who have performed an action in the last 30 days. ~156,000 editors contributed to Wikipedia over the last month, while less than 600 did so on Conservapedia.||Notes: The data used for this graph differs slightly from the definition of active user'. In statistics generated by the mediawiki software all users who performed an action, i.e., made an edit, marked an edit as watched, blocked another user, etc., are counted, while for the graph only those are taken into account who actually made an edit in any namespace. The difference seems to be negligible.|
 Page views
|Conservapedia brags about the number of page views on their main page. This number is manually updated several times a week by the owner of the site. An update of the software at CP resulted in the loss of a couple of million page views in June, 2008.||Notes: 355 data points, each representing a single update by Andrew Schlafly from June 27th, 2007 until June 17th, 2009 (687 days), an average of 3.6 updates a week.
While there were 69.4 million page views on June 16th, 2008, the next entry (June 29th, 2008) features only 50.4 million page views, stating "surpass 50 million page views even with the narrower definition of a page view in the upgraded software" Another upgrade of the mediawiki software lead to the complete loss of data for a whole week in January 2009.
|The accuracy of this data is disputed. It almost certainly includes the actions of bots — like the one which gathered the data for this article.||Notes: Andrew Schlafly's page views linearly interpolated and transformed into page views per week. The week of the upgrade in 2008 is left out. The data loss in 2009 is visible, too: while in 2008, only the page views were lost, in 2009 a whole week of edits, account creations, page views etc. vanished.
Most traffic occurred in November 2007 and May/June 2008, perhaps due to reports about CP in other media. The Lenski affair aroused some interest in August/September 2008.
 The basics
|Conservapedia has a strange tendency to prevent as many potential editors as possible from editing. More than half of the accounts created at Conservapedia are blocked today.||Notes: The data is taken from Conservapedia's user creation log. An account is counted as blocked when it is blocked at the time of data gathering. As less than 0.6 % of the current blocks in Conservapedia's block log are shorter than a year, there is a shift from unblocked accounts to blocked accounts in newer versions of this graph. In 2009 on average 7-8 new accounts were created a day and not blocked subsequently, in contrast to 10-11 accounts a day which got blocked after creation.|
|Simply editing at CP puts you in danger of doing something wrong — therefore, less than half of the active editors (see above) are currently not blocked.||Notes: Looking only at unblocked editors, these numbers are comparable to RationalWiki's own statistics — though RationalWiki has only 10% of Conservapedia's registered editors.|
|A block at Conservapedia tends to be long: most editors are blocked for five years — or infinity.||Notes: As the data was taken from Conservapedia's block log, all blocks are counted, including those that have expired. As there is a drop-down menu for the block length when an entitled editor banhammers someone, certain periods are preferred. Other periods (9 years, 9 months, 9 weeks and 9 days or 1000 days) are capricious — and very rarely used. To exclude such blocks, only periods which were used at least 5 times were counted. Therefore blocks measured in minutes are generally shorter than those measured in hours, blocks in hours shorter than blocks in days, etc.
Over 75% of the blocks at Conservapedia last for at least five years, short blocks for minor infringements are hard to find.
|And the blocks are becoming longer on average.||Notes: A boxplot of the monthly blocks for editors at CP. The numbers are taken from CP's block log. Above the image you find the number of blocks for editors per month. The percentage states the quota of indefinite blocks. In the boxplot itself only finite blocks are included. Since 2009 virtually no block is shorter than five years!|
|Even a year is an eternity in Internet time, mind you. A block of one day is often enough to turn an editor away.||Notes: Here the date from the block log entry was used — and for each editor the list of his or her contributions. An editor is counted as pardoned if they made an edit during the block period (i.e. they must have been unblocked before their block expired). The editor returned if they made an edit only after the block run out. While there is a tendency to wait out a two hour block, other blocks often were appealed, but less than 10% of the blocks were amended — a number which dropped dramatically since the end of 2009.
Only one user returned after the expiry of his one year block (cp:User:Gulik), but he was blocked for good shortly after.
|It doesn't matter how valuable you have been for the site or how many contributions you made: These kind of merits doesn't count in the meritocracy of Conservapedia...||Notes: The thousand editors with the most comments in the database, ordered by the number of their comments. Blue: not blocked, red: blocked. The black line indicates the percentage of blocked editors. For instance when the data was gathered, 476 editors made more than 100 comments on Conservapedia which are still in the database. 48.3% of these 476 editors are blocked currently.|
|But not only editors are blocked. Even though Conservapedia doesn't allow anonymous editing in the first place, not just single IPs (sometimes referred to as range \32) are blocked, but ranges of up to 216=65,536 IPs. As of June 2009, the absolute number of range \16 blocks is greater than the number of range \24 blocks — why use a scalpel if you can hit them with a sledgehammer?||Notes: Again data from the block log is used. Single IP blocks were used since the start of the project, but the possibility to block whole ranges was discovered only in August 2007. The first steps were quite clumsy: only the biggest range blocks possible (\16 range, i.e., 65,536 numbers) were made. In September 2007, lots of these blocks were refined into \24 range blocks (256 numbers). Somehow, the instrument of blocking ranges fell out of interest in March 2008, and virtually all blocks which weren't set for at least one year were allowed to run out in November 2008. In December 2009 a new blocking spree started: this time, \16 blocks were preferred to \24 blocks, again, so since April 2009, the absolute number of \16 blocks is bigger than that of \24 blocks — an idiocy par excellence...|
 Comparison with Wikipedia's blocklist
|Range \16 blocks account for roughly 90% of blocked IPs at Conservapedia.||Notes: ''Here, the number of blocked IPs is counted: 65,536 in each \16 block, 32,768 in each \17 block, 256 in each \24 block and a single one in each \32 block (an intro on range blocks at mediawiki). As every range covers two times as many IPs as the next smaller one, only the contributions of the first couple ranges can be seen.
The blue line indicates a special milestone which Conservapedia reached in April/May 2009: 224 = 16,777,216 blocked IPs, equivalent to an \8 block
|At Wikipedia, they constitute less than two thirds of the total number of blocked IPs.||Notes: At Wikipedia, the effect of other block ranges is visible, as smaller block ranges seem to be preferred over bigger ones. Furthermore the graph fluctuates more, indicating that the number of block is high, but their length is short.|
|The bizarre consequence is that the number of IPs blocked by Conservapedia surpassed the number of IPs blocked by the English Wikipedia at around January 2009, and as of June there are roughly twice as many IPs blocked at Conservapedia than at the English Wikipedia.||Notes: Even more stunning is a direct comparison of the curves in the same scale. Wikipedia is nowhere near the \8 range milestone.|
|Conservapedia not only blocks more numbers in larger groups than Wikipedia, it also blocks for longer time spans. Convervapedia frequently blocks for all eternity, and even when they don't, the terms are excessive by any sane standard.||Notes: The boxplot (maximum, .75 quantil, median, .25 quantil, minimum and outliers) illustrates the excessive blocking policy at Conservapedia in contrast to Wikipedia: the scale on the y-axis is logarithmic to cover the great span of different block times. In most ranges half of the not-infinite blocks at Wikipedia are shorter than a week or even a day, while at Conservapedia, in every range but one half of the non-infinity blocks are longer than six months. The degenerate boxes of some ranges at Conservapedia stem from a lack of diversity in the length of the blocks. For instance, a \24 block is either infinite, six or three month long — with a single exception of a block length of one day.|
|Which parts of the net are affected by these blocks? xkcd showed us the unique topography of the world of IPv4.||Notes: xkcd's map may be a little bit outdated, but the idea is lovely: The LANDER project used it to present the output of its ANT Censuses of the Internet Address Space, where they pinged every IP of the IPv4 space|
|At Wikipedia, there are numerous (~90,000) blocks of one IP each, distributed over the whole IPv4 address space, with some range blocks scattered in between.||Notes: The IPs are ordered in the same way as in the previous image. Different colors indicate the various Regional Internet Registries (RIR). There map differs slightly from xkcd's as some allocations have changed since 2006. The big squares represent a range of 216=16,777,216 IP addresses, each pixel represents an address range of 65,536 distinct IP addresses. Every white tick mark represents a selective block of one specific IP address out of these 65,536. Black squares indicate the entire range is blocked. Different shades of gray represent the intermediate states. Again this image shows that blocks are handed out individually at Wikipedia.
The data for this image is taken from the wp:Special:BlockList — instead of the block log. This list holds all block which are currently in place. At Wikipedia — as at Conservapedia — some IPs are blocked several times: perhaps, different editors handed out blocks of the same range of IP addresses, or such ranges overlap. While in the previous images all blocks are counted independently, here this effect is taken into account: every blocked IP is counted only once and thus the overall number is a little bit smaller.
|Conservapedia presents a noticeably different picture: whole areas are blocked. Not all that dramatic at first glance, but clearly problematic at closer examination: range blocks at Conservapedia completely prevent users with IP addresses in the range in question from editing. At Wikipedia, registered users may still be able to edit anyway after logging in (anonymous only blocks or "school blocks").||Notes: On a first glance this image may look alright compared to the previous one. But the scatter shot approach of blocking at Conservapedia leads to many dark gray pixels: sometimes whole Internet Service Providers (ISPs) are blocked because one of their customers displeased the powers that be at Conservapedia.|
|Wherever there is connectivity, there is the occasional prankster, abusing the service and bringing a temporary block down on his IP. If you want to contribute to Wikipedia, be prepared to have to ask your provider for a new IP address every other decade...||Notes: A map of the Internet is nice — a real map is nicer. The problem: Though it is easy to match an IP to the country from which it originated (the RIRs provide this data), finer matching is more difficult. Luckily MaxMind provides a free service which matches IP addresses and geographical positions with some accuracy (and with even more accuracy if you are willing to pay for it). With this software most of the blocked IP ranges (~95%) could be mapped and visualized. The area of the circles represents the number of IPs in the blocked range — and bigger ranges have brighter colors. The circles are translucent so that overlapping blocks can be identified. Again the map shows an extensive use of blocks of single IPs — not a surprise as Wikipedia allows anonymous edits. Range blocks are relatively sparse|
|...but if you want to contribute to Conservapedia, try not to live in Europe, or go to college in the US.||Notes: Conservapedia sees enemies everywhere — and in great hordes.|
 Edits and editors
|And still, edits happen at Conservapedia...|
|As common editors have to tread very carefully, it isn't surprising that most of the comments are made by the thirty-five sysops — or the score of other editors (sysops light) which have earned extended rights (foremost, the right to block their fellow editors) ...|
|... though roughly half of the not blocked commentators are editors without special privileges.|
|Most of the edits are changes to existing articles — including many reverts of proposed improvements (or, indeed, parody). The creation of new articles seems to have become less and less important. Andrew Schlafly himself said: Article count beyond a certain critical mass (say 20,000 entries) doesn't have any particular significance. People who are interested can look at the statistics. But fewer than 20,000 entries, and perhaps fewer than 10,000 entries, are of any value to 99% of visitors.|
|That may be the case, but most encyclopedias take pride in the number of articles. Perhaps this comment is just sour grapes, as the number of articles created in a year is roughly half of the number of new articles in the previous year...|
|While Conservapedia started strong, Wikipedia took some time to take off, though even in the first — and slowest — year, there were more articles created per day than at Conservapedia in its strongest year, 2007.|
|Therefore, Wikipedia has on the order of 2,900,000 actual articles, i. e. pages containing encyclopedic content (as opposed to pages discussing procedure), while there are ~31,000 such articles at Conservapedia ...|
|...and the respective rates at which Wikipedia and Conservapedia gain new articles makes it highly unlikely that this gap will ever be closed.|
|Conservapedia trails behind Wikipedia in every aspect save one: blocking editors and blocking IP ranges. That's Conservapedia's single strength. So, build on it!|
|The next goal should be to prevent every European from editing at Conservapedia....|
 Data sources
Data: the underlying data was gathered by getting the logs of CP and WP via MediaWiki API. For the creation of articles at CP, all articles were looked up — while at WP, a sample of size 50,000 was taken. Images: the images were created using R. For the geographical location of IPs, the free version of MaxMind was used. For the map of the Internet images, nested or overlapping range blocks were taken into account, while for the other images, all blocks were counted independently, resulting in a higher number of blocked IPs.