There was an article in the Washington Post yesterday, Diet soda is fine, and 3 other food truths it’s time you believed, by Tamar Haspel.
I believe it’s the first and only time I’ve seen the phrase “observational study” in a popular press article. I was so impressed I wrote a lengthy comment.
Which, because I have nothing better to blog about today, I’m reproducing below. Obviously, you should at least skim the article if you want to make sense of the comment.
…
On your first point, this is also the reason poor people eat a poor diet. Try planning a month’s worth of meals at the current SNAP limit of $194 a month. You — like poor people everywhere — will find yourself loading up on starch, sugar, and fat, and skipping the fruits and vegetables. Rice at $0.60/lb provides about twenty times as many calories per dollar as apples at $2/lb.
(Highest calories/dollar among grocery-store items? Vegetable oil. Fried food, anybody?)
Second, bless you for using the phrase “observational study” in a news article. I was a health economist by trade, and if there were one little bit of understanding that I wish I could spread, it’s that not all “science” is created equal. Randomized controlled trials sit at the top of the heap, in terms of their strength of inference. Observational studies sit at the bottom. (“Natural experiments” of various sorts sit in-between).
Whenever you see the results of a study, the first thing to ask is whether or not it was a randomized trial. Hint: Almost no studies of diet are randomized trials. And if not, then is there a plausible alternative explanation of the facts, e.g., fat people drink diet soda, instead of diet soda makes you fat?
Finally, I note the absolutely toxic interaction between the frequently false and counterintuitive “findings” of observational studies, and the modern media’s thirst for click-bait. This virtually guarantees that every oddball and counterintuitive (and wrong) conclusion by every half-baked academic researcher will be hyped. And that any actual science — which by-and-large tends to show boring things, e.g., weight loss is all about restricting calorie intake — gets buried under an avalanche of pseudoscientific nonsense.
