Wikipedia not as good as Britannica

Abi points to a study which finds that the average Wikipedia article contains four errors while the average Encyclopaedia Britannica article contains three. Abi thinks that the study vindicates Wikipedia, and if you think that Wikipedia is ridden with errors, then indeed it does. But if you, like me, expected Wikipedia to be better than EB, then it must be a bit of a disappointment.

So here is a question. There are two ways to measure the quality of Wikipedia. One is to take a sample of articles in Encyclopaedia Britannica, take their corresponding articles in Wikipedia, and compare the two on quality. The other is to take a sample of articles in WP, take their corresponding articles in EB and compare the two in quality.

Is there a difference between the two methods? If so, why and which is the better way?

6 thoughts on “Wikipedia not as good as Britannica”

  1. 1.This analysis, either way, would assume the errors are comparable. That is “quite” an assumption.
    2. And, picking something up from Wiki at random would obviously be a better first choice. That would give Wiki the right weightage for covering aspects that might not be covered in Britanica – my assumption being, Wiki would have more entries.
    3. The rate of correction would be another statistic worth looking at.
    4. Each article’s outgoing links would also be a good way to measure the two up. That would be a race in which I believe Wiki wins pants down.

  2. EB *pays* some ‘experts’ to write its entries, while the Wikipedia gets its content from ‘experts’ as well as interested ‘amateurs’. While keeping in mind that there is still some more distance to go, what’s wrong in saying aloud ‘Hey, we are on course’? When WP has had a rough fortnight, celebration is certainly in order.

    The study covered 42 corresponding entries in WP and EB. They were all in the sciences. Trust me, if there was a BIG goof-up in WP, Nature would have gone to town with it.Instead of such BIG ones, the reviewers found four serious errors (such as ‘misinterpretations of important concepts’) each in WP and EB. Most of the other errors were considered minor.

  3. I agree with you. I am second to none in my admiration of WP and I believe quite a lot in the Wiki process. But don’t you find it ironical that all the criticism that WP has faced has ended up lowering the bar that it has had to cross?

    I mean, it is like someone criticising an extremely bright student in very harsh terms and making it sound like he is a hopeless duffer. Then if he comes 5th in class, it will look like a vindication. On the other hand, if the criticism had recognised that he is in fact a very intelligent student, then it would have added value and perhaps have helped him rank first.

    So I guess my point is that we should accept that WP is a success and then constructively criticise it from that viewpoint.

    Nilu, I think that you got the answer to your first point – the articles were handed out randomly and they in fact classified the errors into mild and severe.

  4. I’d say, comparing a 4 yr old work to a century old mammoth itself is a great thing. As it is said, Wikipedia is a collective effort. Indeed it needs wikification on the whole. I’d say it is in the process of unexpected growth phase. Although some changes are need, things will settle down.

  5. Ravi,

    Both tests you propose will suffer an inherent selection bias. You should just pick the same number of articles at random from either source and then calculate the average number of errors present in each encyclopaedia.

Comments are closed.