RCT-traps and psychiatric epistemology.

While (Cochrane-oriented) evidence based medicine favours RCTs and RCT-based meta-analyses, psychiatry also highlights the values of individualised medicine, complexity sciences and medical epistemology. In a similar way, meta-theoretical considerations put RCTs into perspective: (i) randomisation does not necessarily generate analogous samples, (ii) differences between normal and actual distributions impact on statistical outcomes, (iii) Likert scales may not produce interval data, which are prerequisite for statistical processing, (iv) medical data may possess dynamic characteristics, and (v) standardised research designs have to be revisited from a correspondence-theoretical point of view. To overcome typical flaws, statistical correction, reliable ways of sample splitting and dynamic simulation models are in the ascendant.