What I find most amusing is that Michael Crichton holds up medical trials as the example for what all scientific studies should emulate. "I come from medicine, where the gold standard is the randomized double-blind study, ... For a person with a medical background, accustomed to this degree of rigor in research..." MichaelCrichton.com | Testimony before the United States Senate

The irony of course is that medecine has unique problems and frankly shares most of his complaints about scientific rigor with climate change.

In terms of uniqueness, unlike almost any other field, human experiments need to deal with the placebo effect on their experimental subject. Double-blind's are specifically designed to avoid this effect (among others). If a doctor knew what he was administering, humans are adept at reading body language and could cue in on this, which in turn could influence the answers via the placebo effect. No one has to worry that a fundamental particle or thermometer might read their body language and change the results.

As far as weaknesses, the primary ones would of course be:

1) Small sample size. The number of people involved is miniscule in most studies compared to the number of events in say particle physics. Even if results are nominally significant, limited numbers limit internal self-consistency checks for systematics.

2) Biased reporting (which Crichton complains about for climate change). Forget raw data, there's no requirement for any results or the existence of a study to be released which is a statistical abyss in terms of potential cherry-picking.