Also, lack of rigour in study design can introduces sources of bias, confounding and cointervention, all of which can lead to positive results for reasons other than a real difference between placebo and study drug.Originally Posted by Ken G
It's a big problem for mainstream medicine generally, and the journal editors and their referees should be weeding out studies with poor design, so that they don't reach the literature. (In fact, medical research ethics committees are also taking a role, by refusing to allow patients to be enrolled in studies that are badly designed.) However, this doesn't happen perfectly, by any means: and there are particular issues with the scientific skills of both the researchers and the editors of journals dealing with "alternative medicine" themes.
So that's another reason that anyone who wishes to make a superficial case for homeopathy can find plenty of statistically significant results to point at.