If in the final analysis the value of the projected treatment can only be considered relative to results obtained by therapy which it has motivated, would it not be possible as implied to assure from the very beginning its effects? The author's research based on statistical and clinical analysis of 233 files drawn up by the Policlinique psychiatrique de Lausanne, between 1967 and 1975, allows an affirmative reply. And if on the one hand this work goes as far as to relativise or even to negate the capacities of the therapist (a therapist considered mediocre obtaining almost identical results to one who is not seen as such), and on the other to show that all the inconstant elements studied are no longer likely to be mastered by the medical prerogatives alone, the latter is not put into doubt but simply becomes better adjusted to prevailing socio-cultural conditions.