Myths and (Urban) Legends Example: Skim Milk Causes Fat Kids

A couple of days ago I wrote about health-related myths and urban legends.  It was a fun post to research, and a fun one to write as well.  I enjoyed watching the video by Tom Naughton, linked at the end of my post, quite a bit; not only did I learn a lot about different types of studies and different code words to look for in evaluating the study results, but I also found his sense of humor to be engaging.

Didn't you?

"Okay, I admit, I don't follow links from blogs very often," I can hear you saying.  That's okay; I don't either.  But please do, in this case--go back to the blog post I linked above, and down at the bottom follow the link to Naughton's video.  It's 46 minutes of goodness.  I'll wait; heck, I'll probably go enjoy it again myself in the meantime.

Okay, now that you're back, he makes some good points, doesn't he?  Butter, for example, might be relatively solid (though squishy, in my experience) at room temperature, but it's perfectly runny at artery temperature.  Bananas, meanwhile, are rather solid at artery temp.  Bananas bad, butter good.  See how cool science is when you really get into it?

And then there's that word "associated."  I hadn't thought much about it, but he's right.  Using the word "associated" to describe a statistically significant correlation (in other words, enough members of the population who have one thing also have another to satisfy statistical tests) is one thing, while using it to then suggest causation is another.  After all, having red hair is associated with sunburns, since people with red hair are also often fair-skinned, and people who are fair-skinned are more likely to get sunburned, but red hair neither causes nor is caused by sunburns.

So all that said, I was a little surprised to come across an example presented this morning, so soon after my own blog post.  And I wasn't even looking for it; rather, I was checking online to see the caloric difference between whole milk and 2% milk.

Here's a link to the blog over at NPR.  It suggests that, surprisingly enough, giving kids milk containing less fat makes them fatter.  The quoted researchers seem sincere enough in expressing their own surprise at the results.

Giggle, giggle, snort, snort.

I mean, c'mon, we all know that consuming things lower in fat are better for the weight conscious folks than consuming things higher in fat, right?  Or, um, do we?  Hopefully you paid attention to Naughton's video and glared at me just a little when I said that.  After all, basic nutritional science tell us that a human body needs to consume a certain--albeit low, for most of us hard-charging desk-jockies--amount of fat regularly in order to operate.  The more active, the more calories burnt, and thus the more fat needed.  The extremely active Eskimo people consume a great deal of fat, after all, and they're just fine.  So why would removing fat from an active kid's diet be a good thing?

Bottom line: are you really gonna tell me that skim milk causes more fat kids than whole or 2% milk does?  With a straight face, even?

I guess the two doctors from University of Virginia are.  Their study, or at least its abstract (a term used by academicians for "making a long story shorter but no more readable") is linked from the blog post.  Turns out they did a longitudinal study (a study in which the same population is checked over time) and found that the kids fed skim milk had more weight problems (defined as being overweight or obese) than those fed normal milk.  Yeah, there are a lot of numbers in the abstract, and as I've already pointed out on this blog, Americans in general don't know math.  But generally speaking, "p < .01" means that the probability of the hypothesis being true is less than 1%, which in turn means that the researcher is more than 99% certain the hypothesis (generally that there's no relationship between the variables) is false, which means they're 99% certain that there's a relationship between, in this case, kids who drink skim milk and kids who tend to be overweight.  Similarly, "p < .0001" means the researchers are 99.99% certain of their results.

See how easy that is?  The researchers are at least 99% certain that there is a statistical correlation between consumption of skim milk in the studied population (2-4 years old, diverse demographics) and weight issues.

So skim milk causes fat kids?  Really?

Let's think about it.  I'd love to see a study done on how many kids despise skim milk.  Hell, I do, and I'm not even a kid.  Skim milk is like watered down plaster, no offense intended to the skim cows.  But why on earth would parents, in the aggregate at least, want to feed their kid something that's fairly icky unless the kid is trending toward being overweight anyway?

In other words, sure, I'll buy that being overweight might cause skim milk consumption.  But I won't buy it going the opposite direction.  Sorry.

- TOSK

P.S. (caveats):

1. The researchers in question did, in fact, put in their conclusion what I said in my next to last paragraph.  But people didn't read to the conclusion in school, nor do we do so at work, and so seldom do we do so in the other parts of our lives either.  
2. I admit, I did not read the entire report of study.  I'm too cheap to pay for the download.  I'll bet most agencies reporting on the story are the same way.
3. No, skim cows are not used to make skim milk; centrifuges are.  Was fun to say, regardless.