There’s a lot of stupid at the New York Times, but an article today wins the stupid of the month award. Under the innocuous headline Hold the Phone: A Big-Data Conundrum, Times author Sendhil Mullainthan claims that “every time a new iPhone comes out, my existing iPhone seems to slow down,” and shows evidence that this is a huge conspiracy led by the evil masters of Cupertino.
You see, the author has a chart; based on research by a PhD student, who looked at Google Trends to see how many people searched for “iPhone slow,” the author concludes that somehow slows down your iPhone when a new model is released:
In Apple’s case, the company sells the device and makes the operating system. In principle, this creates the motive (to sell more devices) and the means (control over the operating systems) to slow down the old phone.
Yet this doesn’t happen with Samsung Galaxy phones; that curve is constantly ascending with little variation:
The author points out that “This data reveals only correlations, not conclusions.” Indeed. And the reason for this search – “iPhone slow” – is obvious, and the author of the article even explains it. But rather then ending with that explanation, he finishes the article with statement “And if those correlations allow conspiracy theorists to become that much more smug, that’s a small price to pay.”
But go up a few paragraphs to the point where the author tells why this happens:
Every major iPhone release coincides with a major new operating system release. Though Apple would not comment on the matter, one could speculate — and many have — that a new operating system, optimized for new phones, would slow down older phones. This could also explain the Samsung-iPhone difference: Because only 18 percent of Android users have the latest operating systems on their phones, whereas 90 percent of iPhone users do, any slowdown from a new operating system would be naturally bigger for iPhones.
Yea, that’s exactly what’s happening. iPhone users get free updates to iOS, whereas Android phones mostly do not get updates at all; or, if they do, they are staggered over a long period of time. And iOS has a high uptake rate, with more than 50% of users upgrading in the first week of the availability of a new version of iOS. So why dress this article up in conspiracy theories?
Yes, this is a good article for click bait, and all that “big data” ballyhoo is just a mistake in basic assumptions. Not only does correlation not equal causation, but one needs to know what one is looking for. In this case, the PhD student who played around with Google Trends clearly did not understand the issue.