• Windows 7 Wallpapers
• Vista Wallpapers
• Windows Wallpapers

I was going to let the ChangeWave data pass by. The results published on March 30, and it looked like maybe nobody would notice. But then yesterday Computerworld posted a story, and now I feel obligated to debunk ChangeWave's finding.

I hadn't planned on this being the week for challenging everybody. Yesterday it was Microsoft's ridiculous assertion that 96 percent of netbooks run Windows and overzealous iPhone rumormongers. Now it's a Windows satisfaction survey. Don't worry, I don't feel smug. I get my due from Microsoft Watch commenters, who are plenty quick to tell me where to go. You know, the repeated trips to hell and back are boring.

Between Feb. 9 and 17, ChangeWave surveyed 68 Windows 7 beta testers, 44 percent of whom reported being "very satisfied" with the operating system. The timing meant they had used the software for as much as one month. That's a fairly high percentage for "very satisfied," which suggests over 50 percent were satisfied to very satisfied. Again, that's a high percentage.

But the sample is too small. Thirty people is too small a number to call anybody really satisfied. For example, what's the mix of IT professionals to teenagers? The answer could make the finding useless. If I were a Microsoft product manager, channel partner or IT customer, I wouldn't make any Windows 7 deployment estimates based on 30 "very satisfied" beta testers.

Here's what I suspect and can't check because it's the middle of the night throughout much of the United States: The 68 is a subset of 2,000. "In the same survey we also asked 2,000 corporate IT buyers about their company upgrade plans," according to the ChangeWave report. That says to me that 68 of the 2,000 survey participants were testing Windows 7. Wow, so few?

The finding from the larger survey group conceptually is more compelling and believable based on size: 53 percent will skip Vista altogether. Sounds pretty good for Microsoft, right? Wrong. ChangeWave's question skews the data: "Will your company skip a Vista upgrade altogether to wait for Windows 7, or are you going ahead with a Vista rollout?" You've got to frakking be kidding me. You can't ask a question that way and expect good data. That's how you end up with 32 percent "other," as ChangeWave did.