Well, think about that.


A bad trial will not have some of the following essential features:

- appropriate placebo.
The placebo must account for all effects other than that due solely to the treatment. It must look, smell and taste the same and come in identical containers.
- random allocation (of patients to treatment and placebo groups).
This is to ensure that the patients in each group is identical with respect to all possible confounding factors such as age, sex, state of health, degree of illness etc.
- double-blinding (of patients and practitioners).
Neither the patient nor the practitioner must kow whether the patient is receiving placebo or remedy. The person who codes the bottles must have no contact with either patients or practitioners.

I think cross-over trials are not appropriate for homoeopathy for reasons stated in other posts.

So, the better the trial, the smaller the effect, because you are eliminating all other possible causes of the patient feeling better apart from the effect of the remedy.


What do you think?