Wikipedia:Wikipedia Signpost/2012-10-22/Special report

How the German Wikipedia got on top of RfA
Unlike the long-running disputes that have characterised attempts to reform the RfA process on the English Wikipedia, the German Wikipedia's tradition of making decisions not by consensus but knife-edged 50% + 1 votes has led to a fundamentally different outcome. In 2009, the project managed to largely settle the RfA mode issue indirectly.

Historically, the question of how to make administrators accountable to the community in the absence of an English Wikipedia form of ArbCom has been highly controversial on the German site, leading to divisive debates and hard-fought votes until 2008. Due to the multifarious views within the community, administrators were largely unaccountable for their actions once elected (two-thirds majority support required). Personal conflicts in the aging community led to a rise in long-term tensions, fueled by the editorial community's unusually strong meetup-based culture, where face-to-face contact was sometimes confronting.

However, in early 2009 the overall administrative structure of the German Wikipedia beyond sysops themselves came under intense pressure. Due to problems with generational changes to the project's checkuser team under the narrow German interpretation of the WMF's privacy policy, and the need to bolster local oversight, the project embarked on a series of reform votes.
 * 2009 – the big change

Tackling the persistent lack of a majority view for ensuring proper checks and balances among functionary user-groups, and to address concerns over the oversight issue, the community approved a separation of powers among functionaries. Under the new arrangement, members of groups such as checkusers, oversights, and the local (electable) mediation committee (Schiedsgericht) have been unable to hold more than one of these functions at any given time since April 2009. Based on that decision, oversighters were introduced in May and all three affected user-groups were made accountable to the community in re-elections in September 2009.

After tackling the very small special-function groups, the project built on the momentum to finally take on the accountability of regular admins again, in October 2009. A decisive vote on RfA reform was based on the pre-existing voluntary system under which admins can choose to be open for recall, and introduced an obligatory recall page for every administrator. If 25 editors within three months, or 50 within six months, sign that they feel an admin should stand for re-election, the admin has the choice of either standing for re-election or standing down.

The reform has triggered a flood of re-election proceedings and calls for minor corrections to improve the handling of re-elections on the grounds of inactivity. To deal with permanent election-trolling and to give newly elected admins some space, a settling-in period was introduced for newly elected and newly re-elected administrators, to make it more difficult to pursue rolling elections as a strategy against an admin.
 * Aftermath – pros and cons

These 2009 reforms have never been significantly challenged, and are credited with bringing about a relatively low RfA barrier and better admin–community relations. When the numbers look tight, a candidate for election or re-election can use the option of voluntarily waiving the settle-in period to address concerns among the electorate. The result is that administrators generally appear to be regarded as accountable for their actions. Admins with a record of deciding controversial issues in an even-handed way, if forced into a re-election, are usually re-elected by the community.

Compared with the English Wikipedia, the Germans have a high number of RfAs relative to their community size. The English Wikipedia community – which is many times the size of the German – saw 121 successful RfAs in 2009, 75 in 2010, and 52 in 2011, and only 20 thus far in 2012; meanwhile, the German community had 67 in 2009 (including a flood of "inactivity" RfAs after the introduction of the new system), 43 in 2010, and 34 in 2011.

However, as the sponsor of the vote pointed out in a retrospective three years later, the system has been used to force through a new interpretation of the project's voting privileges. On the German Wikipedia, one has to do at least 50 main namespace edits per annum with an user account to be entitled to take part in the voting procedures, and parts of the electorate, by using the 2009 system, have insisted that admins who are no longer sufficiently active stand down.

Another issue allegedly caused by the system is the lack of containment of long-running controversies. While the English Wikipedia has ArbCom to take care of the most toxic and ideological disputes, the German project's mediation committee has insufficient powers to ensure respect for project guidelines in complex cases. Before 2009, administrators, free of accountability constraints, dealt with such issues pro-actively but have since risked being forced into re-election proceedings if they act decisively; therefore, the argument goes, they quite often ignore such complaints.

Key

{{legend|White|0 successful RFAs}} {{legend|#FBEC5D|1–5 successful RFAs}} {{legend|#FFDF00|6–10 successful RFAs}} {{legend|#FCC200|11–15 successful RFAs}} {{legend|Orange|16–20 successful RFAs}} {{legend|DarkOrange|21–25 successful RFAs}} {{legend|#E9692C|26–30 successful RFAs}} {{legend|#CC5500|31–35 successful RFAs}} {{legend|#b06030|36–40 successful RFAs}} {{legend|#804020|41–50 successful RFAs}} {{legend|#602000|51–60 successful RFAs}} {{legend|Black|More than 60 successful RFAs}}


 * Table courtesy of WereSpielChequers' "RfA by month" page.