Social Scientists Need Better Training

I read A Matter of Taste: How Names, Fashions and Culture Change last week. It was very disappointing. The topic fascinates me and I was expecting a great book, but I was instead forcefully reminded how poorly trained social scientists are.

Even though the author is a professor at Harvard (ouch!), he makes many amateurish blunders. In an attempt to determine whether celebrities influence naming patterns, for example, he looked for naming spikes after celebrities' rise to prominence. If lots of people started naming their sons Charlton, Marlon and Elvis shortly after Heston, Brando and Presley become famous, we can assume that their celebrity influenced the naming spike. But in the dozens of famous entertainers considered, there was either no spike in the celebrity's name or the spike began before they were famous, suggesting that the entertainer wasn't the force behind the spike. The only spike that fit the hypothesis was the name Gary with the rise of Gary Cooper.

So, where does this information lead a Harvard sociology professor? To this probing question: "Why, for example, did the appearance of Gary Cooper lead to Gary becoming an enormously popular name, whereas Humphrey Bogart, Clark Gable, Bing Crosby and Cary Grant (other major stars of about the same period) had minimal effect?"

By this point in the book (page 185) I'd become so frustrated with the author, and by extension the editors, social scientists, Yale University Press and Harvard, that I felt like yelling. Gary DID NOT become popular because of Gary Cooper! You just disproved your hypothesis, you nincompoop! If naming spikes happen independent of the rise of celebrities, one should expect coincidental correlations nonetheless. That's what controls are for. Gary just happened to spike as Gary Cooper became famous. Aaargh.

It's as though he disproved the significance of Friday the 13th, then scratches his head trying to figure out why someone in Reno was hit by a train on Friday the 13th.