Eating sugar will turn you into a criminal, and restricting carbs will make you fat. Those are the conclusions drawn from a couple of recent studies, at least as they were reported in the media.
Dangit, now what am I supposed to do … eat sugar and end up robbing a bank, or avoid the stuff and end up becoming an obese but law-abiding citizen? I suppose I could spend a lot of time lifting weights in the prison yard and get really buffed up …
Naturally, both of these studies are of the observational variety, which means they found correlations … and that’s it. As I’ve said before, there’s a strong correlation between gray hair and heart attacks, but nobody believes gray hair causes heart attacks. A statistical link doesn’t prove cause and effect. Unfortunately, too many reporters (and far too many researchers) can’t seem to grasp that concept.
Observational studies are iffy for a very simple reason: people are different. We’re all walking bundles of interrelated traits, many of which are largely genetic: intelligence, affability, athleticism, laziness, discipline, focus, a sense of humor, likes and dislikes, a predisposition to be fat or thin, etc. Those traits exert a powerful influence on our choices and behaviors … but people interpreting or reporting on observational studies often get the equation backwards.
In the first study, reported in Time Magazine, British researchers found a strong correlation between eating sugary treats during childhood and becoming a criminal later in life. Here is the opening paragraph from the Time story:
What parent hasn’t used candy to pacify a cranky child or head off a brewing tantrum? When reasoning, threats and time-outs fail, a sugary treat often does the trick. But while that chocolate-covered balm may be highly effective in the short term, say British scientists, it may be setting youngsters up for problem behavior later. According to a new study, kids who eat too many treats at a young age risk becoming violent in adulthood.
So what exactly prompted the writer to conclude that sugary treats lead to violent behavior years later? It was this finding:
Moore plumbed the data for information on kids’ diet and their later behavior: at age 10, the children were asked how much candy they consumed, and at age 34, they were questioned about whether they had been convicted of a crime. Moore’s analysis suggests a correlation: 69% of people who had been convicted of a violent act by age 34 reported eating candy almost every day as youngsters; 42% of people who had not been arrested for violent behavior reported the same.
Well, that’s it then … sugar must screw up your brain and make you decide it’s okay to mug people. Or perhaps — and this is the more likely explanation — we’re just witnessing the natural relationship between traits and behaviors. In other words, the kind of parents who end up raising criminals are also more likely to let their kids eat candy bars for breakfast.
Maybe we should talk to some teenage criminals and find out how many of their mothers spend a lot of time worrying about nutrition. I doubt many of them leave the house in the morning hearing “Johnny, if you rob a Walgreen’s this week, would you mind picking up some whey protein powder and a bottle of CoQ10? Oh, and your father likes magnesium supplements, so pick up some of those in case he ever shows up again.”
The same principle applies to positive behaviors and traits as well. For years, we were told that parents could raise more intelligent children by reading to them, limiting their TV time, and keeping a lot of books in the house. Those are, after all, common behaviors among the parents of intelligent kids. So read to little Johnny, and he’ll do well in school.
But as it turns out, the theory doesn’t hold up to actual research. Intelligent people who enjoy words and language and learning tend to read a lot of books. (You should see the size of Mike Eades’ library.) They’re also likely to produce intelligent kids who enjoy words and language and end up learning more easily in school. The kids inherited a trait — verbal intelligence — that tends to be exhibited as a related behavior — reading.
Thankfully, the lead researcher in the British study did seem to understand the concept (not that you’d know from the headline and lead paragraph of the article):
One of those questions is whether sweets themselves contain compounds that promote antisocial and aggressive behavior, or whether the excessive eating of sweets represents a lack of discipline in childhood that translates to poor impulse control in adulthood. Moore is leaning toward the latter… It’s also possible that children who are poorly behaved from the start tend to get more candy.
The second article, warning us that restricting carbs could make us fat, definitely gets the relationship between traits and behaviors backwards. Here are some choice quotes:
Low-carb eaters could be setting themselves up for obesity, suggests a new study from this month’s issue of the Journal of the American Dietetic Association. Researchers analyzed data collected from the Canadian Community Health Survey, which collected health data from a sample of over 35,000 Canadians. They narrowed the sample down to 4,451 people who had submitted information on their diets, including how much and what type of food they’d eaten on the day of the assessment. They found that people with higher intakes of carbohydrates ate fewer calories but more protein, fat, and fiber than low-carb eaters consumed.
Okay, wait a second … I’m writing late at night, after a play rehearsal, so maybe I’m missing something here, but please re-read that last sentence and answer this question for me: how can the people who eat more carbohydrates also consume more fat, more protein, and more fiber, but still end up consuming fewer calories? Did Monsanto create a new macronutrient I don’t know about? Anyway …
In fact, the incidence of overweight and obesity in the lowest-carb-intake group was 65 percent, while it was just 51 percent in the highest-carb group, and the risk for becoming overweight or obese was 40 percent lower in the highest-carb-intake groups.
This makes about as much sense as the observation that fat people are more likely to drink diet sodas, so diet sodas must make you fat. If you recruit a large group of people and tease out the data on those who restrict their carbs, you’re most likely looking at the dieters in the group. You’d probably also tease out dieters if you looked at who counts fat grams.
Now … what kind of people go on diets? Fat people, that’s who — those of us who tend to gain weight easily. Once again, the trait — predisposed to gain weight — produces the behavior — dieting. I restrict my carbohydrates, but I’m fatter than my son, who lives on them. He doesn’t watch his carbohydrates because he doesn’t have to. It’s a case of selection bias, not cause and effect.
Selection bias is also the reason that vegetarians tend to be leaner than the population as a whole. Yes, I’ve known some skinny vegetarians. And pretty much every one of them has been skinny since birth. They give up meat, they don’t gain weight, so they stick with it. They’re a self-selected group.
But I also know plenty of people — myself included — who tried a vegetarian diet and gained weight. So we became ex-vegetarians and selected ourselves out of the group. Once again, a trait — gains weight easily — produced a behavior — gave up the vegetarian diet. In other words, we’re not fat because we avoid sugar and starch; we avoid sugar and starch because we’re fat.
And also, of course, because we don’t want to become criminals.