Dropping out of treatment is a wellknown fact. The meaning of this phenomenon and causal factors appear to be inadequately investigated. As a rule one supposes that a patient ends therapy prematurely, when sufficient improvement or satisfaction is achieved. Mostly the interpretation of the therapist is used for the classification in dropouts and non-dropouts. In this article, after a theoretical introduction, an investigation is presented making above mentioned assumptions questionable. The results point to the fact that therapists' ideas about ending therapy are not congruent with those of his patient and that this patient is certainly not in worse condition than one who does not dropout of treatment. Besides this it shows that a retrospective investigation is insufficient to produce sound results and recommendations are given for future investigation into this, as yet not well enough understood, subject.