Monday, January 24, 2005

Lies, Damned lies, and Statistics

In response to a post by the venerable Jeff Jarvis I am creating this blog in order to debunk the statistics quoted so frequently by journalists. However I have limited time and resources to go searching for stupidity. Mostly I will look at and the NYtimes. If there is another article you think I should look at please, E-mail it to me.

Today I'm going to start slow with this article from on, of all things, the number of laptops and cellphones lost in cabs this past year. Which though stupid, does not say anything wrong that I can see.

"AMSTERDAM, The Netherlands (Reuters) -- An estimated 11,300 laptop computers, 31,400 handheld computers and 200,000 mobile telephones were left in taxis around the world during the last six months, a survey found on Monday."

Firstly, as completely uninteresting as these numbers are, who cares? More interesting to me would be the percentage of passengers leaving things behind. 1 in 5, 1 in 10? These raw numbers tell us basically nothing about people's propensity to leave things behind. And apparently these numbers are higher than the 2001 numbers, but there is no mention of an increase in the population of laptops/cellphones etc. Though there is an interesting anecdote about Hugh Grants girlfriend.

Secondly, these numbers are based on a SURVEY of Taxi Drivers. Which means they are if not completely wrong, they are at least grossly innaccurate. An individual's ability to recall specific instances of anything is not great and when it happens a lot, people stop counting. They might get within 5, but the reliability of a survey that asks "how many times" did this happen is well, low at best. So to extrapolate the data to the rest of the cabs in the city means, well, that the results are not good. Which doesn't really concern me since no one is drawing any conclusiosn from this study.

I tried to find the orgiginal survey etc to get a good look at their methodoligy and see if the surveyors actually had anything interesting to say. But I had no luck, if anyone finds it, please send it my way.

So basically the biggest problem with this article is that the study it is reporting on is, well, boring. Though interesting possibly to cab companies the results might be useful if setting up a lost item tracking program or something. Fortunately none of the articles I have seen on this study have drawn any outlandish conclusions like "people are getting more absent minded" or "cab drivers love returning lost items" so that's all I have to say.

Remember kids, Correlation is not the same as Causation.

No comments: