|[1:00 PM] Apple Near Top of PC Data Rankings Despite Absence Of Mac Users From Sample
by Kyle D'Addario
Have you ever wondered how PC Data generates the numbers and Top 10 lists that we have been posting for so long? We did, and contacted PC Data to find out. As it turns out, much like any type of survey, PC Data uses a representative demographic sample of computer users, and monitors their Web surfing habits. PC Data representative Jim Carey (no, not THAT Jim Carey) told us that PC Data uses, "120,000 people who install a software application on their machine." PC Data then tracks those users, and "strings" together their surfing patterns by downloading data in 15 minute chunks. While 120,000 people is certainly not the entire Internet population, by selective random sampling and careful demographic consideration, computer usage can be generalized to the entire population.
For those unfamiliar with survey research techniques, this is no different than nearly every published survey in the United States today. Often, political surveys attempting to access who is going to win or lose and election are made up of far fewer responses than PC Data's 120,000. In fact, statistical methods, with careful demographic representation, allow an error rate of +/- less than 3% with a sample of 1000-1200.
The key to survey sampling is having a representative sample, and PC Data examines the Internet-using sample "..four times a year." According to Carey, "We want to make sure we are not over or under representing certain factions of the Internet-using population." While much of what PC Data does is sound, there is one monumental flaw in their design.
They do not survey Mac users.
At this time, the PC Data monitoring client is a Windows only product. In light of these methods obviously eliminating the millions of Mac users from their sample, not to mention Linux and other "alternative" OSs, the numbers that PC Data generates become even more interesting. Apple.com has been consistently in the top three of the PC Data Hardware Sites for months, even though their numbers include virtually no Macintosh users. Granted, statistical extrapolation can account for some of the results, but with literally no Macintosh representation, Apple's place near the front of the Top 10 list is even more extraordinary.
PC Data is, apparently, aware of the problem. Mr. Carey assured us that, "...a Macintosh client is being considered." Until that time, it will remain PC users generating top honors for the Apple site.
The Mac Observer Spin: Mr. Carey actually seemed somewhat surprised at the results generated by the PC Data collection methods regarding the popularity of the Apple web site. Since there are literally no Mac users, or more accurately actual Mac computers, making up the sample the success of the Apple web site is even more remarkable. As seen in postings in many forums around the web, there are many PC users following the progress of Apple and the Macintosh. PC Data has provided us with some numbers to back that claim up.
We can only assume that with a TRULY representative sample, Mac users included, that the Apple web site would rank even higher. Regardless, this is good news for Apple and the Apple.com web site.
Apple - PC Data