Controlling For Variables

When investigating the efficacy of something, it is important to control for variables which may affect outcomes, or at least to be aware of them. In physical medicine, for example, for some treatments it is important to be aware of a person's blood type, or whether they have certain medical conditions like say a heart or liver problem.

Now, when trialling treatments and remedies beyond things like pharmaceuticals there is another problem. If we give a group of patients in a trial a 5mg pill of Olanzapine each day, we can be sure that all those 5mg pills are essentially identical. If we are trialling the efficacy of Mindfulness, Meditation, Taiji or Yoga, we no longer have that. Now two Taiji teachers are the same. No two have the same understanding. No two students understand what their teacher is telling them in an identical way. And so on. As such, if two people are getting, say, 'Mindfulness based therapy', we can't take for granted that, by dint of going to a 'registered Mindfulness based therapy practitioner', each participant in a trial is 'getting the same thing'. Yet, or at least I am worried that, in an attempt to give an 'evidence based medicine' facade to proceedings, trials are conducted under the implicit assumption that two people being sent to, say, a Taiji class, are getting the same thing as each other, and the same thing as anybody else being sent to a Taiji class, as if Taiji was something that came in 5mg pills which you received once per class.

In reality, there are myriad variables affecting the outcome of somebody taking classes in Taiji or Meditation. Crucially, there are far more relevant variables than is feasible to control for them. Yet I suspect this problem is silently ignored when analysing outcomes of sending someone to a Yoga class for 6 weeks.