Route 2(A)

I’ve noticed lately that people pluralize Latin and Greek second-declension words like “memorandum” and “encomium” like “memorandums” and “encomiums” and not “memoranda” and “encomia”. When did this become acceptable? What’s next? “criterions?” “alumnuses?”

I was going to make a proper post about a week ago. I still intend to make a real post. This isn’t it. I just wanted to vent.

I was casually paging through my Google Reader’s new posts today when I saw this short piece shared by one of the people I’m following from the Atlantic’s Daily Dish. It’s a blurb titled “Alcohol and Exercise” by Jonah Lehrer.

In the past I’ve usually liked things I read by Mr. Lehrer. One article that comes to mind specifically was an article he wrote about self-control for the New Yorker. (And if you like reading it, he had a follow up Q&A as well, both of these I found via kottke.)

This bit was not really an exception. He links to an article from Reuters that reports about a study in which it was found that drinking regularly is positively correlated with spending more time doing exercise. Most of the Reuters article is pretty basic and puts the paper’s interesting points out there but towards the end it started to bother me a little.

From that article:

Moderate drinking, on the other hand, has been linked to potential health benefits, including a decreased risk of heart disease. While part of that might be attributed to moderate drinkers’ overall lifestyle — which, based on this study, includes higher exercise levels — research also suggests that alcohol has some direct benefits, like elevated levels of “good” HDL cholesterol.

To gain those potential benefits while reducing the chances of harm, experts generally recommend that women have no more than one drink per day and men no more than two.

I don’t like the phrase “has been linked to” and moreover the phrasing “To gain those potential benefits” suggests that it is indeed the act of drinking that conveys these health benefits. That strikes me as misleading and is doing nothing to help people reading differentiate correlation from causation.

It’s at times like these that I am reminded of a conversation I had with Ani a while ago in which he suggested that it would be a great thing if newspapers in their online versions would actually provide citations for things they say. In print there are obvious limitations to printing out all that extra mostly unused information but online the cost presumably much closer to trivial.

At this point I got a bit annoyed with the phrasing in Lehrer’s article too.

alcohol – a drink that seems to protect the heart, prevent dementia, raise levels of good HDL cholesterol and makes us go jogging

I’m probably being too sensitive about this and I also agree with the main point he’s making in the paragraph I’m quoting about how it is unfortunate that we seem to have no problem taxing alcohol but can’t bring ourselves to stop people from sucking down their sweet sugary soda.

The article linked to in the NYTimes about preventing dementia I think understates the simple idea that people who can and do drink moderately are the same people who are doing a lot of other things right in terms of living healthy. I still suspect that total abstinence is probably correlated with something else wrong with you.</snark>

Anyway, thoughts?

There’s a great xkcd about correlation too, if you haven’t seen it.