User:Angela/Vandalism study

Vandalism studies/Study1 found that almost all vandalism (97%) is made by unregistered users. Looking at vandalism on my own user page, I find a very different result. Almost half of the vandalism is made by registered users.

There were 624 edits to my user page between February 20, 2003 and March 29, 2007.

There were 29 edits in the first three months of 2007, 142 edits in 2006, 110 edits in 2005, 211 edits in 2004, and 132 edits in the last 10 months of 2003. I was more active editing in 2004 (before I started Wikia) which explains the higher rate of edits that year.

219 edits (35%) were made by me.

112 edits (18%) were made by users who were not logged in. 79 different IPs edited my user page. This is fewer than 79 people since a blocked user will often switch IP and try vandalise again.

512 edits (83%) were made by logged in users.

Excluding the edits made by me, 405 edits were made. 112 (29%) were made by users who were not logged in. 210 different registered users (excluding myself) edited my user page. Of the usernames who edited, many were made specifically to attack or impersonate me, often by the same user over a short period of time. For example, Aangela, 2Angela, AngelA, AnGela, Angels, Angely, Angela (sysop-vandal), Angela is a bitch, Angela is a troll, Angela is a virus, Angela is anti-wiki, Angela is fused, Angela's ex-boyfriend, and Anti-Angela.

Of the 624 edits, at least 203 of them were bad edits. That's 33% of all edits and 50% of edits not made by me. A few more were questionable - changing colours in a way I later reverted, other edits that weren't malicious but weren't useful. 2 people attempted to change my British spelling to American spelling. At least 182 of the edits were malicious. Quite a few people accidentally left a message for me here instead of on my talk page, so those are bad edits that aren't malicious. 187 edits were reverts. A few vandals reverted themselves.

Of the malicious edits, 96 (53%) were made by unregistered users and 86 (47%) were made by registered users. This is very different to the percentage found in the other vandalism study where they found 97% of vandalism to come from unregistered users. Only one revert (excluding self-reverts) was made by an unregistered user.

The vandalism included blanking, silly comments, negative statements about me or other Wikipedians, adding sexist comments and images, and one death threat. There were two instances of spam and six instances of page move vandalism.

There were at least 37 good edits made by others that were not reverts. That's less than 10% of the edits not made by me, but shows the benefit of having a user page openly editable. Those good edits include added categories, added interlanguage links, awarding barnstars, fixing links, fixing images, improving formatting, improving wording, spelling corrections, and making minor updates. Only two good edits were made by unregistered users. I suggest this does not imply my user page should be semi-protected since then vandals would be more likely to log in to vandalise this page, making the vandalism harder to detect.

Ideas for future studies:
 * Is vandalism by registered users more likely on user pages than on articles?
 * Are user pages of females subject to more sexist comments than those of males or users who don't openly state their gender on Wikipedia?
 * Is user page vandalism reverted more quickly than article vandalism?
 * Does the amount of user page vandalism correlate with the number of vandals that user has blocked?
 * Is vandalism by unregistered users reverted more quickly?

Other possible future studies ideas (added by other people).
 * Is the amount of vandalism on user pages well-correlated with number of edits made by that person?
 * Is the amount of vandalism on user pages correlated with popularity of that person's username in the general population? (I.e. does user:Fred get more vandalism than user:Szljybf92j, everything else being equal?)
 * Does having a picture on your user page increase vandalism of that page? (Will need a control group of users with very few edits, but pictures on their pages, which may be hard to find).