An iPhone Survey from BankMyCell has bounced around the Apple blogosphere. It claims that iPhone retention is down 15.2% this year compared to last year. But The Macalope tells us why it’s flawed.
You can’t compare results for different demographies and declare them meaningful. You’re not controlling for anything…The only constant here is the gullibility (or culpability) of the technology press.
If BankMyCell were interested in meaningful results, it would have compared its own results over two years if it couldn’t get CIRP’s demographic breakdown. But it’s clearly not…If the methodology is crap, then you can’t trust the numbers.
That last line is the lesson we can learn from this: If the methodology is flawed, so are the results. You can’t p-hack your way around this one.