I should warn you: this entry is a rant. Not only that, but it is a rant about something so insignificant that it really doesn’t even deserve comment, but it is annoying me and I can’t sleep. So I’m commenting.
I have recently become a fan of the show House, M.D., which is full of more silly medspeak than E.R. ever was. I usually assume that most of it is near nonsense and enjoy the show. But tonight — oh tonight — the show really pushed my buttons.
There was a lecture scene where a supposedly eminent scientist (I’ll call him Lecture Dude) was speaking in an auditorium, and two of the show’s regular characters were sitting in the audience carrying on a side conversation. The shot switched back and forth from the lecture to the conversation, with the research-babble creating background noise for the conversation.
As part of his babble, Lecture Dude said “ANOVA analysis showed that….” The problem is that he pronounced ANOVA wrong. It should be pronounced “an ova” (as in an egg, with equal stress on both syllables). He pronounced it “ainova” (the closest I can come to reproducing this is “this ain’t ova” with the t left out and the stress on the first syllable).
Scientists pronounce science words in many ways — most of us just say the words with confidence and hope for the best, and no one really comments. In fact, some interesting social networks can be identified by how certain words are pronounced (because someone heard it said that way by their advisor who heard it from their advisor who heard it from their advisor and so on and so on and so on).
But mispronouncing ANOVA is different. ANOVA is short for Analysis Of Variance. If you know what ANOVA is (and it was clear from the context that the writers’ intention was to convey that the lecture dude did know his stuff), then there’s one and only one way to pronounce it.
A smaller, but in my opinion less irritating mistake, followed when Lecture Dude said “my results are significant with p equal to point zero zero one.” He even wrote “p=.001″ on the white board and underlined it for emphasis. He should have said “at the the p equals point zero zero one level” or — even better — “with p equal to less than point zero zero one.” This error is arguaby more egregious than the ANOVA mistake. However, I’m ok with it. It’s the kind of error I see and hear all the time, from scientists and non-scientists alike.
But mispronouncing ANOVA? That’s wrong. Wrong AND irritating.
(Also, since I’ve already taken this stroll down the path of obsessive pedanticness, I might as well start running. I’m not a medical researcher, but based on my limited knowledge, it appeared that while ANOVA was appropriate for the fake study Lecture Dude was describing (which sounded like a randomized controlled study of a new medication, based on what little bits of description I caught), it wasn’t the best choice. Lecture Dude’s sample size was large enough — he said it was an impossibly high 500 — that he should probably have used some other kind of analysis. Here my knowledge of medical research fails, because I can’t suggest what the other technique might have been. But in my own research, I definitely wouldn’t waste a sample size of 500 on ANOVA. Given the kind of outcome Lecture Dude was studying (getting a particular illness), I would have done a logistic regression, or if time-varying covariates were available, a survival analysis. But I know that ANOVA is very ingrained in certain research areas where sample sizes tend to be small. So maybe Lecture Dude felt that another technique would be so non-normative that it wouldn’t get past reviewers, even though it would be mathematically and substantively appropriate. Still, it would surprise me if someone used ANOVA in this situation in real life).
Oh, and I need I remind you? I did warn you ahead of time that this was a rant over something of the barest consequence. It’s not my fault if you read this far.