Reply to post:

Would you like to give concrete examples rather than making vague accusations of malpractice?

I work somewhere where we run very large computer models of physical systems, and I can tell you that people spend a lot of time worrying about the agreement or otherwise between the models and the data.

I think there is definitely a problem with some theoretical physics because the experiments to test theories have become essentially impossible to do -- we have theories that work essentially perfectly in all the cases we can construct experiments to test even though we know those theories can't be completely correct. So we're kind of stuck -- I suspect the best hope is either that a theory will come along which is compellingly good at explaining various astronomical observations or that someone will think up experiments which can be done without spending absurd amounts of money, or that something completely unexpected will turn up in existing experiments. Any of these would clearly be really good. But that problem is not the same problem.