Meditation research to date has been plagued by conceptual and methodological problems. One of the most significant difficulties involves developing control strategies involving interventions which blinded participants might consider plausible, that have no specific therapeutic effects. Randomisation and management of other sources of bias is another area of concern; a large number of controlled trials have used non-randomised, dissimilar cohorts.
Dr Ramesh Manocha further discusses the specific problems of defining meditation at his blog.