Saturday, June 13, 2009

The crud factor

Many thanks to Vassyana for opening a page within user space to brainstorm for solutions to Wikipedia's chronic conflicts. This post is related to his excellent drive.

One thing to be alert for within arbitration is the crud factor: disruptive editors fill the case pages with crud. The principle is simple: when the evidence is against you and you don't have a useful rebuttal, clutter the case and make it harder to read. Crud appears at Wikipedia arbitrations across many unrelated topics because crudism is an intuitive strategy. Crud is the noise in the signal-to-noise factor of Wikipedia arbitration. This post describes crudism and how to defend against it.

First, what exactly is the crud factor? A simple example of pure crudism occurred at the Gundagai arbitration of 2006. All of the participants except one substantiate their assertions with diffs, while a single individual makes a long series of aggressive claims backed by no evidence at all. See if you can spot the crudite at this page.

Most of the cases the Arbitration Committee gets in 2009 are more complex than that because the community has gotten more effective at dealing with simple crud at the community level. One well-known example of recent crudism was the John254/Kristen Eriksen sockpuppet team that played both sides of the fence at the Scientology arbitration workshop until the sockteam was identified and banned by the community through independent action.

Community action doesn't stop crud from occurring, though, because John/Kristen was a rare type of crudite. Most crudites are either directly involved in the dispute being arbitrated or else strategically aligned with one or more partisans. Most crudites pursue three goals:
  • Protection of one or more allied partisans against arbitration sanction.
  • Aggressive sanction against one or more opposing partisans.
  • Establishment of arbitration principles that can be leveraged to the crudites' advantage in future disputes.
These days, when a case reaches arbitration it often comes with multiple crudites acting in tandem and subtle forms of crud. Instead of long rants with no diffs they provide cherry picked or irrelevant diffs. Crud expands and migrates across case pages the longer the case remains in evidence phase.

The really damaging thing about subtle crud is that it resembles evidence; other editors may consider themselves compelled to rebut it. If they do, more crud follows. If they don't, exhausted arbitrators might suppose no defense was attempted. The result of this double bind is a negative feedback loop during which the case grows exponentially.

The Solution to Crud
Arbitrators can put an effective cap on crud by putting up proposed decisions sooner. Note the time frame of the Gundagai arbitration mentioned earlier:

Case Opened on 21:50, 18 October 2006 (UTC)

Case Closed on 18:02, 4 November 2006 (UTC)

In most arbitrations the useful evidence gets posted within about three weeks of opening the case. Occasions when useful evidence gets delayed are usually identifiable because someone has asked for more time due to health, work, or other tangible reasons. By contrast, crud accumulates at a steady trickle according to whenever the crudites have enough free time to generate more crud.

The key defense, from the arbitrators' standpoint, is not to rush the voting upon a proposed decision but to initiate the proposed decision sooner. Once a case moves to voting, crud naturally migrates to the proposed decision talk page--which slows the growth of crud at the evidence and workshop. It is more useful to distract hardened crudites off those pages and to leave the arbitrators somewhat at leisure to sort out the existing mess.

Fred Bauder was brilliant at that while he served on the Committee. Many of the old Bauder proposed decisions are superb demonstrations of crud management.

Remember, the most important response to crud is to identify it, separate it from useful input, and redirect the crudites' attention where they cause less damage until the case closes. For the most part, only arbitrators can achieve this.

No comments: