Systematic bias

Revision as of 20:36, 25 December 2003 by 142.177.103.52 (talk) (what it is, why it's not systemic bias, how it's been addressed)

A systematic bias is a bias arising from how participants in some joint process work together. It is related to, and may arise in part from, a systemic bias. However, one deals with the two in quite different ways - while a systemic bias can often be dealt with simply by disrupting the cliques and inviting in new participants, a systematic bias can only be addressed by what is known in legal terms as "due process" - a dragging-out and delaying of decisions so that persistent factual claims can be evaluated, and decisions are not made until a deliberation has occurred on the credibility of each one.

For ordinary distinctions such as, e.g. "simple vandalism" on a large public wiki, such a process may be overkill. However, for deep distinctions that involve the sysop power structure, such as sysop vandalism, there is going to be a "due process" or nothing, since, the sysops themselves have the technological power to prevent their own activities from being investigated.

The concept of the w:carceral state may be relevant: This was Michel Foucault's generalization of Jeremy Bentham's "Panopticon" or prison-like society where authority could see and control everything, but its own secret investigations and deliberations were entirely invisible to those it controlled.

Groupthink and the echo chamber are the natural result of an unaddressed systematic bias. While a group may have a systemic bias to accepting the outputs of such phenomena as "real", it takes a lack of process for them to accept those outputs as being viable consensus decision-making. Thus only a disciplined notion of such decision making can alleviate the systematic bias.

This is such a serious problem that some very respected thinkers have concluded that only a separation of w:epistemic community by value system can be effective in creating "objective" categories. The most famous examples being the burning of "pagan" texts by Christians, the development by Marxists and some syndicalists of a "working class" specific education system, various social and alternate histories from feminism, and recently and less famously the w:Islamization of knowledge proposed by the Palestinian Ismail al-Faruqi. In all these cases, people sharing a value system would be expected to share ethical principles as well, and those would guide their categorizations - this seems obvious for ethics and law, but is it also true for all facts?

There are over one thousand librarian-created categorization schemes in the largest collections of these. This suggests strongly that there is no single hierarchy of classification of knowledge, and that each means of agreeing on a single hierarchy is probably itself evidence of yet another systematic bias.