The real bias in Wikipedia

[ Tue. Jun. 6. 2006 ]

More than a dozen years ago I was involved in a project to build an internet-delivered encyclopedic reference source. Those of us who worked on it were dazzled by the potential that seemed to be opening up before us. There was a worldwide communication network that anyone could use; here in hand was the most comprehensive and authoritative generalreference work in the English language; and in between us and the goal that grew more ambitious each day were only some technical challenges and the limits of our imaginations. It was a wonderful time to be an encyclopedia editor.

Well, things didn't work out just as we hoped, for reasons too numerous to mention here. I recall this episode mainly to make the point that I understand the enthusiasm, the evangelism, that Wikipedia evokes in many, many people. I wish I could share it with them now. But, as David Shariatmadari's openDemocracy article "The sultan and the glamour model" (25 May 2006) shows once again, Wikipedia's most eloquent advocates fail, or refuse, to acknowledge certain issues.

Bias and imbalance

Shariatmadari's article praises the work of a group calling itself by the unfortunately self-congratulatory labelWikiproject: Countering Systemic Bias and ends with a call for more such efforts to improve the coverage of the encyclopedia. Certainly such work is needed. I would suggest that it needs to begin with a clear distinction between "bias" and "imbalance", terms that Shariatmadari uses interchangeably but that to an editor mean quite different things. The Wikiproject seems to concern itself with topics that are treated in insufficient detail or not at all; to me, this is addressing imbalance. "Bias" denotes a lack of objectivity or fairness in the treatment of topics. Thus, when a writer called Joseph McCabe alleged in a widely distributed pamphlet that certain articles in theEncyclopedia Britannica had been unduly influenced by the Catholic church, he was charging bias. (That was in 1947, and he was quite wrong, by the way.)

Is imbalance in Wikipedia "systemic"? I should rather say that it results inevitably from a lack of system. Given the method by which Wikipedia articles are created, for there to be any semblance of balance in the overall coverage of subject-matter would be miraculous. Balance results from planning. As an example, the planning of the coverage of the fifteenth edition of Britannica took an in-house staff and dozens of advisers several years to complete. That was forty years ago; it would be harder now.

It is unremarkable that the topics covered at present in Wikipedia reflect the interests of those who contribute to it, and that these contributors represent a relatively narrow, self-selected segment of society. In the absence of planning and some degree of central direction, how else could it have been?

It is well to bear in mind also that imbalance is a judgment, not a fact, and that it cannot be reduced to numbers. To say that article A is longer than article B is not to show that B has not been given its due. Some subjects require more background, more context, more sheer wordage to convey a sense of understanding to the reader. Are 260 lines too much to devote to the Scots language? Clearly, someone does not think so. Someone else might well feel that there ought to be much more. Three lines for the language of the Yi is almost certainly too few, but what is the right number? Who – I'm asking for a showing of hands here – knows? What is lacking is not some numerical standard but editorial standards: a set of principles that define what constitutes adequate treatment of various kinds of topics for an intended audience.

Truth and openness

David Shariatmadari writes that the situation is "uncannily like free market economics applied to knowledge." This is quite inapt. I suppose it is meant to shock; what could be worse than, you know, capitalism? I'll just point out that another shocking word that might properly be applied to Wikipedia is "globalist." Sorry, but I calls 'em as I sees 'em.

More seriously, a better analogy might be a children's soccer team. It is notorious that, in the United States, at least, a game involving the youngest children will consist of a swarm of twenty or so players buzzing ineffectively about the ball. As the children grow older, however, they will develop individual skills and learn to play positions and to execute strategies. Just so, traditionally, have editors honed skills, learned appropriate methods and processes, and developed the synoptic view required by the job.

No complex project can be expected to yield satisfactory results without a clear vision of what the goal is – and here I mean what a worthy internet encyclopedia actually looks like – and a plan to reach that goal, which will include a careful inventory of the needed skills and knowledge and some meaningful measures of progress. To date, the "hive mind" of Wikipedia's "digital Maoism" (as Jaron Lanier'svigorous critique on edge.org calls it) displays none of these.

That vision of the goal must do something that Wikipedia and Wikipedians steadfastly decline to do today, and that is to consider seriously the user, the reader. What is the user meant to take away from the experience of consulting a Wikipedia article? The most candid defenders of the encyclopedia today confess that it cannot be trusted to impart correct information but can serve as a starting-point for research. By this they seem to mean that it supplies some links and some useful search terms to plug into Google. This is not much. It is a great shame that some excellent work – and there is some – is rendered suspect both by the ideologically required openness of the process and by association with much distinctly not excellent work that is accorded equal standing by that same ideology.

One simple fact that must be accepted as the basis for any intellectual work is that truth – whatever definition of that word you may subscribe to – is not democratically determined. And another is that talent, whether for soccer or for exposition, is not equally distributed across the population, while a robust confidence is one's own views apparently is. If there is a systemic bias in Wikipedia, it is to have ignored so far these inescapable facts.

People Mentions: