Wikipedia vs Britannica; continued

A while ago, Nature did a study comparing wikipedia to Britannica (you can read my take on it here - oh, just look at the title I used :-).

Now it seems that Britannica weren't very happy about the results, and have responded: We discovered in Nature's work a pattern of sloppiness, indifference to basic scholarly standards, and flagrant errors so numerous they completely invalidated the results. And so on.

First off, they whinge about contrary to the usual practice of making all data freely available in order to facilitate a study's replication by others, Nature declined our repeated requests to make the full reports available. This is twaddle. Nature provided a list of all the errors; its what wiki used to correct their mistakes. EB had all it needed.

There is a pile of stuff about Nature not having used the "real" Britannica. I don't know about that. It may have been related to trying to find pairs of articles from wiki and Britannica that matched, roughly, length-for-length. Which is difficult.

Next, Nature's review of the Britannica article "Pythagorean theorem" claimed the Britannica misspelled an Italian town that the reviewer said should be spelled "Crotona." Well, this is hard to judge. The wiki page for Crotone notes that it is Crotona in latin (and Crotona redirects to it). In EB, the latin is claimed to be Croton. Whereas Bartleby says Crotone, see Crotona. I can't read the full EB text, but it may well depend on exactly *when* the town is referred to.

The reviewer of the article on Paul Dirac objected that certain areas of Dirac's work were not covered. Our coverage, however, was appropriate for a general-reference encyclopedia. By design, the 825-word article explained, for the lay reader, Dirac's most significant contributions, not all of them. Well, this is a wiki-EB comparison problem, cos wiki Paul Dirac is clearly longer than 825 words. And I don't think that Nature published the error rate, just the error totals (which will be bigger for larger articles, as wikis were). Wiki has the magnetic monopole, and the large number, stuff that Nature criticises EB for omitting. OTOH it laso lacks the "least action" bit (oh, and the large-number bit was added *after* Nature review).

And EB complains about being marked down for not giving an equation in the Haber-Bosch process: The article's verbal description of the process was clear and sufficient for the general reader. Sounds dubious. Wiki has the equation.

But the main fault in the EB defense is the assumption that mistakes were only made in assessing EB entries, and that therefore the comparison should be tilted towads EB. That seems unlikely. Also I would guess that most reviewers guessed which of the pairs were EB and which wiki (the writing style is hard to disguise) so it wasn't a proper blind experiment; and this would likely have lead to a bias towards an easier review of the EB half of the pair.

Appendix B goes into the stuff in tedious detail. Sadly there is no global warming stuff there. Their insistence on retaining "evening" as against "night" sky is bizarre, and makes me suspect that much of their other responses are just stubborness, too.

The Kin selection bit is possibly interesting (PZ?): EB defends itself with the altruistic behavior is favored by natural selection because relatives share (in fractions depending on the degree of relatedness) all their genes. Errm, I can't see how you can share a fraction of all your genes. If you share them all, the fraction is 1.

And so on, quibbling into the far distance. The fun thing, of course, is not what they say but that they feel obliged to say it at all. Lofty distain is clearly out of the window; we're down to bare-knuckle fighting.

[Thanks to Brian J, who points out that Nature has responded: Interestingly, this clarifies one point: why stuff was taken from yearbooks not EB on occaision: The company also objects to the fact that in some cases we took material from Britannica's Book of the Year and its Student Encyclopedia. This was done in a few cases when the Britannica website provided articles from these sources when queried on the pre-determined topics; as we said, the survey compared the content of the websites. Which sounds fair enough, and is a point EB should have thought of. Nature also notes the obvious point that EB missed and I said: We realised that in some cases our reviewers' criticisms would be open to debate, and in some cases might be wrong. But this applied as much to criticisms of Wikipedia as of Encyclopaedia Britannica.

Another boring update: wiki does have least-action under Dirac, but cunningly disguised as path-intergral formulation, so searching for "action" doesn't work :-)



More like this

OTOH it laso lacks the "least action" bit

It's the principle of *extreme* action, not "least". The action is not always minimized, but it's always at the extreme point - i.e., it's variation is zero.

[OK, it lacks that too :-) -W]

By Roman Werpachowski (not verified) on 23 Mar 2006 #permalink

We do not accept this criticism, we do not accept this criticism, we do not... Maybe not a constructive approach? :)

Yes, the kin selection bit was weird. All members of a species should in most cases share all their genes, and they share alleles _by common descent_ "in fractions depending on the degree of relatedness". I wonder what Francisco Ayala was thinking, he is a geneticist after all?

The 'Acheulean Industry' bit sounds fishy as well, but hard to say without seeing the original:
Reviewer comment: "I would not use the term 'early Homo sapiens'. Instead, use Homo heidelbergensis."
They answer that anthropologists don't agree on the names, and that "In a short article like this we have decided not to distract the lay reader by raising such complex issues". Blah. Shouldn't they at least use 'archaic Hs', 'early' could mean anything?