Apparantly, it's
damn lies statistics time again in the Apple world, boys and girls. A few days ago, Net Applications published
OS market share figures which showed that the market share of the Mac OS
remained largely flat over the past year. However, today, the Switchtoamac website posted an article which looks at the
same set of figures in a different way, only to conclude that when you compare last year's figures to this year's, the Mac OS has risen 25% in marketshare. We all know the saying: there are lies... Damn lies... Funny detail: while the debate rages on about half percentage points for the Mac OS, Windows XP
increased its share by 8 percent the past 12 months. According to
these figures, of course.
Member since:
2006-02-07
Statistics can mean statistical analysis or information systematically collected. All data collections imply certain types of bias and problems of many kinds. But these data are actually systematically collected. They aren't unbiased, but it's likely that the bias persists over time hence an adequate analysis can draw conclusions from their analysis.
I'd say that the second analysis (the one adding macintel+macppc) is elemental, but hits on the nail. The data are from the same source, the comparison year-by-year helps address potential seasonality issues. It doesn't address what type of information this is, but at least compares the trends in the netapplications source with the trends in computer sales. It's not material for a statistics journal, but it's OK.
I'd be very interested, of course, in having better metrics and a better defined problem -- it's unlikely however that we'll ever have the chance to see a study like that in a public forum like this.