Saturday, March 25, 2006

The battle between Britannica and Nature

Let me begin with one important point. This battle is not about Wikipedia. It's largely about the claims to authority and fairness of two (commercial) entities: Nature and Britannica. If their mud-slinging battle continues, it's their reputation -- and not Wikipedia's -- that will (continue to) suffer. [Given the publicly available information, Britannica appears to stand to lose more. More on this below.]

Let me get back to Wikipedia for a few more moments. For all its flaws -- there are many, and they are well documented -- it has one great virtue: its open culture. As David Weinberger notes:

Wikipedia has been a continuous state of self-criticism that newspapers would do well to emulate. It has discussion pages for every article. It has handled inaccuracies not defensively but with the humble understanding that of course Wikipedia articles will have mistakes, so let's get on with the unending task of improving them. Wikipedia's ambitions are immodest, but Wikipedia is not.

Indeed. I love that last sentence, so let me repeat it, with emphasis: Wikipedia's ambitions are immodest, but Wikipedia is not.

Ah, that made me feel better! With that out of the way, we can now turn to take a clear view of the battle between Britannica and Nature. From all the 'he said, she said', this is how I see the battle:

Here is the chronology:

First, Nature's report was published on 4 December 2005. It also released some additional information (doc) regarding its study on 22 December 2005.

Britannica issued its response (pdf) on 22 March 2006. And, Nature issued its 'rebuttal' (pdf) the following day.

In studies such as the one by Nature, there is a lot of methodological leeway: how the articles are chosen, how much of each article is chosen, who the experts are, how to tally their reports of 'inaccuracies' (defined by Nature as 'factual errors, omissions and misleading statements'). In other words, a lot of judgement is involved in pretty much every step. Nature has made a particular set of choices, and Britannica doesn't like it. Thus, its complaint is about (a) the methodology, (b) the headline used by Nature ('...Wikipedia comes close to Britannica in terms of the accuracy of its science entries ... '), and finally, (c) the list of 'inaccuracies' compiled by Nature.

Let me take the last one first: Out of some 123 inaccuracies compiled by Nature, Britannica has disputed nearly sixty (less than 50 percent). They are in 29 entries (out of 42). What's the point behind this exercise? Look at the best case scenario (for Britannica): it just shows that the experts hired by Nature weren't infallible. I would just shrug my shoulders, and say 'so what?'.

But, look at the worst case scenario: in complaining so loudly, Britannica has implicitly admitted that the remaining 60 were real errors: pure, simple, indisputable, and gold-standard (!). True, this brings its error rate from three down to 1.5 per entry. Does it entitle Britannica to claim victory? "Hey, I have only 1.5 wrinkles on my face. Not 3!" If this makes the folks at Britannica comfortable, well, ...

For its complaints about the study's methodology to hurt Nature, Britannica has to prove a bias. I find (at least, as of now) its case less than convincing: as Nature has responded, flaws -- if any -- in its methodology applied to both Wikipedia and Britannica. In other words, there is no bias. With whatever information we have right now, I would go with Nature on this one.

Britannica's other major complaint is misrepresentation of the study's results; its grouse is that even though an average Wikipedia entry, with four errors, had a third more inaccuracies than a Britannica entry (which has three), Nature chose to say in its headline: "Wikipedia comes close to Britannica in terms of the accuracy of its science entries".

Now, this whine can fill a lot of barrels! Let me just say that I fell off the chair while reading it. To Britannica, this was so important that its list of complaints leads off with this one!

It's one thing for Wikipedia enthusiasts -- like me! -- to celebrate its being compared to the old, venerable and mighty Britannica (even if the comparison is slightly adverse); it's an entirely different thing altogether, when Britannica chooses to shout from its own rooftop 'Hey, listen up. With 1.5 wrinkles on my face, I am actually more beautiful than that wiki lady. She has four.'

Britannica has done itself and its reputation a serious disservice by its 'response'. It should have left the Nature study alone.