The main risk in a study like this is that (1) with inclusion criteria but (2) without controls, you’re susceptible to reversion to the mean. Imagine that your

The main risk in a study like this is that (1) with inclusion criteria but (2) without controls, you’re susceptible to reversion to the mean. Imagine that your weight fluctuates by about 3% every few days. The risk is that by imposing a cutoff for inclusion, you are more likely to identify people at the top of their usual range. So, the natural decline as they oscillate around their normal weight will look like a treatment effect. Adding controls without intervention would be helpful to identify (and control for) this problem. Unrelated, but also, I note that 6 people were excluded because they already eat for fewer than 10h/day.