There is an old joke physicists like to tell to their graduate students...
It seems a policeman walking his beat at about three in the morning notices a slightly disheveled drunk on his hands and knees searching for something under a streetlight. The drunk is very engaged and doesn’t take notice that the officer is watching him for about ten minutes. He goes around and around carefully examining each bit of asphalt and then returning and starting over after the sweep is done. Finally the policeman approaches the drunk.
“Excuse me sir, but what are you looking for?”
The drunk is startled, but not enough to stand up. Getting eye contact he replies
“Hello officer. I dropped my wallet and have been looking for it for at least an hour now.”
The policeman scans the area and sees nothing like a wallet.
“Are you certain you dropped it here?”
“Oh. I don’t think so officer. I probably dropped it across the street.”
The policeman is now confused and a bit amused..
“Why aren’t you looking over there?”
“Can’t you see?” the drunk replied “the light over there is terrible..”
_____
Today I read an analysis of some smartphone usage patterns. It was completely off base given the technique they were using. Their technique is well known and easy to use, but completely across the street - there is no way they could find what they claimed to be looking for with their measurements and analysis and they pushed a null result.
Getting at the question they posed would be difficult, but not impossible. It would require a different experimental model and they would need to measure different information. It was possible, but they opted to do something non-appropriate and got a result that didn’t make sense. It was too easy for them to search in an area they knew.
You see this sort of thing a lot in pseudoscience and every once and awhile elsewhere. If you are looking for something make sure you are looking in the right place. If there isn’t a good streetlight, build one of your own or work out something else that is just as good - perhaps a lawn rake would have worked for the drunk.
_____
Looking at results in clever ways can make an enormous difference. Sometimes seemingly complex patterns have simple underpinnings. I was reminded of that by this beautiful youtube video Jeri forwarded
(oddly enough I know three people with similar sounding names: Jeri, Jerry, and Jheri. All of them happen to be very interesting people)
What was the study, and what was the conclusion?
Posted by: Howard Greenstein | 10/04/2011 at 08:18 PM