First - Scientific method has nothing to do with statistics. It is to do with formation of hypothesis, using that hypothesis to postulate results that are not in the current known data, and then the discovering the predictions were true (not 99% true, ACTUALLY true). For example Quantum mechanics has made a large number of weird and wonderful predictions of things no-one has ever seen before at the level of small scale things. Then, with the improvements in technology over they years we have observed all those things were actually true which means it evolves to a well respected theory. If something, anything was discovered that didn't fit, then this is a problem with the theory, not a statistical anomoly (QM doesn't work for very large scale things, this tells us that the theory is not a perfect picture of everything, and points to where to look to make a better one) You seem to have confused the fundamental principles of science with the mathematical technicalities of analysing results.
Second - Physics deal with statistics at many levels; it is their bread and butter. Quantum mechanics is largely the study of probability distributions (take a look at Quantum Statistical Thermodynamics, which is takled at the undergraduate level). To infer that (real, qualified) physicists have a fuzzy idea of dealing with statistics compared to psychologists is absurd. Its like a shop assistant saying that mathematicians dont know anything about maths because they havn't seen them work the till. If you *genuinely* believe this, then please point out a few statistical flaws in published physics papers.