[Use and abuse of rating scales and statistics in psychopharmacology]

Scientific method in clinical work and research is still fully accepted in th world of medicine. There is still much misunderstanding of the principles and errors in its application. Rating scales are particularly useful in the evaluation of treatments, though they have other applications. But they are not absolutely necessary for this purpose and there are simpler methods available, though they are less efficient. Scales provide information in a standard form which permits the making of many different kinds of comparisons. They are not a substitute for sound clinical Judgement applied to properly designed clinical trials. Uncontrolled trials cannot distinguish between good and effective treatments and useless or even harmful ones. Statistical analysis is necessary because all patients are different. Unfortunately, incorrect or inappropriate statistical analysis is still not uncommon. Some examples are given.