Methodological validity is therefore the major challenge to meditation research, and the chief problems within this broad category are first, the use of appropriate control strategies, second, the need for randomisation and other strategies to exclude bias and third, a definition of meditation that allows inter-trial comparability and remains consistent with the traditional ideas of meditation as a state of non-thought.
Plausible control groups are critical in behaviour therapy research because of the need to exclude the significant confounding effects of non-specific factors (such as placebo, therapeutic contact and researcher expectancy). The significance of this issue is borne out by the fact that even comparative trials of behaviour therapies often end up demonstrating equivalence of effect255. Such non-specific factors are also significant in meditation research. Expectancy alone, for instance, has been shown in a number of studies to positively influence the apparent effect of meditation4.
The essential criteria for a control strategy in meditation trials should therefore be first, convincing plausibility as an active intervention in its own right and second, a process that involves relaxation and reduction of somatic arousal (since this is the nearest conventionally understood phenomenon that resembles meditation).
Randomisation and other strategies to exclude bias
There are a large number of controlled meditation trials using dissimilar cohorts in non-randomised trials. The need for randomisation to exclude selection bias is obvious, yet as previously pointed out, less than 4% of the total number of peer-reviewed publications used random allocation of participants.