Wednesday, September 30, 2009

The Plural of "Anecdote"...

My status message for the last few days has been:
 
The plural of "anecdote" is not "data".
 
As a few people have asked me to clarify, I thought I'd just post my response.
 
The status came as a result of an article by The SciGuy over at the Houston Chronicle. This specific article was about a study that found that children who were spanked a certain number of times per week lost a point to 3 points off of their IQ over the course of 3 years. In the comments section, there were a number of people saying variations on the theme: "I was spanked as a child and I have an IQ of 130! Imagine if I hadn't been spanked! I'd be a genius!"
 
I was getting frustrated because these people seemed to think that:
 
1. If their IQ was high, that meant it hadn't been lowered.
2. If their IQ was 3 points higher, that would be a noticeable change to them.
3. Their personal assumption that their IQ hadn't been lowered meant that the research was false.
 
#3 is where the status update stems from. The study evaluated over 1000 children. Of those 1000+, there were some who did not follow the pattern of the majority, that doesn't mean that the pattern is wrong, it just means that there is variation. Some children lost much more than 3 IQ points. Some may have gained. Some of those children may have had extenuating circumstances not captured in the research. Your one personal story does not mean that the research is false.
 
My dad said it like this:
 
I could sit in McDonald's and watch an old guy come in and say, "MacDonald's is where old people go to eat." I could see another old person come in and say, "That confirms it! All old people must eat at MacDonald's!" But that anecdotal evidence does not mean that it's good research. Until I study a large, random group of old people, I can make no conclusive statements about the eating habits of old people. The plural of "anecdote" is not "data".

No comments:

Post a Comment