Wikipedia:Wikipediology/library/essays/R.fiend-1/Reviews
This is an extremely important essay, and I'm very glad that someone finally put together the time to put together such an analysis. This kind of analysis is far more meaningful than the standard {{NUMBEROFARTICLES}}. I'm especially pleased by the choice of metrics you chose to measure. This might better be categorised as a report than as an essay, since its focus is mainly on statistics and methodology, rather than on ideas or philosophy, so I have few other comments along those lines.
I hope to see more surveys like this published in the future. Perhaps the Wikipediology Institute could organise a sort of semiannual report, divying the research in 500-article blocks. We could set up a list of questions to be analysed, pool our findings, and give approximate accuracy measurements in our conclusion (see below for potential metrics). Such a survey could potentially be very important, so we should probably advertise the completion of the survey on the Signpost.
Some thoughts of metrics to be measured in the future:
- Flesch and Kincaid reading ease and grade level
- Length of article
- Number of images
- Number of non-free images
- Number of articles with talkpages (might be interesting)
- Age of articles
- Number of edits
- Number of distinct editors
- Number of distinct anonymous editors
- Number of anonymous edits
- Number of sources of various types
- Number of articles in given category (as you suggested)
Naturally, such an survey would be much larger an endeavour than the one you did, but I think it would be well worth our while to make such a report.
Error
[edit]With regards to error, I'm not entirely certain how to calculate the probability of error, but it is proportional to the square root of the sample size (this is refered to as the rule). Again, I can't calculate the exact probability of error, but I can tell you the accuracy triples each time the sample size is increased by an order of magnitude. Your Rambot metric is good way to measure your accuracy. Some potential techniques: number of articles marked by {{disambiguation}} or {{cleanup}}, number of various stub-types, and number of Featured Aricles.
Great essay. I'd been wondering about doing this kind of big sample random articles test - but I'm also happy that someone else did the work and I can just enjoy the result :)
I agree with your carefully considered conclusions. Sure, we have a tremendous coverage of popular culture - but we've got everything else down pretty well too. The same can be seen from Wikipedia:WikiProject_Missing_encyclopedia_articles - we're well on our way towards covering every subject traditional encyclopedias have got.
The quality issue is, of course, a much more difficult question - but the coverage issue is hardly even debatable anymore.