It seems to me that most difficult problem here is how do we determine if an association between your weight and the weight of your friends implies that great weight cause great weight. Is it more likely that as you get older, your friends do also, age may imply weight. Or maybe some other shared trait, such as where you live, that you like the same foods and so.
In other words your being fat might indeed predict a greater chance that your friends will be also, but it doesn’t mean that if you stay thin so will your friends. There’s association, but not causation.
Statistical Modeling, Causal Inference, and Social Science