Suppose you have 2 researchers interested in Cat Studies. Researcher A performs an experiment using 12 cats that shows a remarkable correlation between how much TV cats are exposed to and how much they meow. Researcher B performs a similar experiment using 60 cats, but he finds no correlation.
Which of these experiments will get written up in Cat Studies Journal? It’s not hard to guess that it’s Researcher A who will publish an article and then go on to do interviews for TV Guide and NPR.
Researcher B didn’t even bother to submit an article to the journal, because he knew it wouldn’t get published. Who wants to read about experiments whose conclusion is “Nothing to see here?” Instead, Researcher B takes his video tapes and R programs and shoves them in a drawer.
Which brings me to today’s new-to-me word: the file drawer problem, also known as publication bias. I first learned this term from an old Planet Money podcast about why so many experiments can’t be replicated. The term file drawer problem was first used in 1979 in a paper whose abstract begins this way:
For any given research area, one cannot tell how many studies have been conducted but never reported. The extreme view of the "file drawer problem" is that journals are filled with the 5% of the studies that show Type I errors, while the file drawers are filled with the 95% of the studies that show nonsignificant results.
The file drawer problem has important implications for how research is reported. As noted, the communication in the field is biased toward research that has positive results. Moreover, these positive results might be have come about through defective methodology, or maybe just as a statistical fluke. (The PsychFileDrawer site lets researchers share the results of attempting to replicate published studies—for example, there's a page that collects studies about the efficacy of "brain training" games.) Anyway, the next time you hear in the popular media about some study that shows an intriguing correlation between, dunno, playing Monopoly and success in the stock market, remember that there might be other studies in file drawers that fail to replicate that study.
Ok. What do you suppose bachelors have to do with vaccines? Cows. Well, maybe; it's not certain where the word bachelor came from, which entered English from French in medieval times. A possible origin is that in Latin there was a word baccalarius that referred to someone who worked on a baccalaria, a dairy farm; hence, a baccalarius could have been a cowhand. Bacca is a variant on vacca, a Latin word for cow (vaca in Spanish, vache in French). We also got vaccine and vaccination from this root, because when Edward Jenner did the groundbreaking work with smallpox vaccinations, he deliberately infected people with cowpox, a related but much milder disease.
In English, bachelor initially referred to a knight who was too poor or too young to have his own banner. It also referred to a junior member of a guild. By Chaucer's time, a bachelor was someone who earned the lowest (pre-master) degree at a university, or even a man who hadn't yet married—both of these senses appear in The Canterbury Tales (1386). While we contemplate the bachelor-cow connection, we can think about how to capitalize and punctuate a bachelor's degree. Fortunately, Grammar Girl is on the case.
Like this? Read all the Friday words.