I think part of the problem may be that a person isn't just one thing - a skinny person may also be anxious and poor, or something. There's a problem of correlation/causation and also general principles vs. individual variation. Scientific studies will mostly look at correlation and general principles. While amateur (and professional) online articles will try to present causation (while nobody except you can work out individual variation)... And in many cases nobody will tell you what is the balance of individual variation vs general principle... But I think it makes sense, if you can find that, to look at scientifically sound general principles and apply them. The Internet doesn't make this task easy, though... But literally copying an individual person's behaviour doesn't make sense to me, due to confounding factors (in scientific terms)...