The aim of the present study was the construction of a sensitive objective instrument for the evaluation of changes in social adjustment of chronic psychotic patients.
The patients were subjected to an intensive multidimensional therapeutic programme. 
In assessing social adjustment emphasis is laid on interpersonal behavioural patterns, rather than on the various external, and often arbitrarily chosen 'qualities' that are considered in the usual symptom rating scales.
The instrument had to be sensitive enough to register small differences in the interaction between judge and subject with a reasonably high degree of reliability (interjudge agreement). The (un)reliability of the psychiatric judgements is discussed. Four aspects of the psychiatric judgement receive special attention:
1 The expertness of the judge with regard to the judgement he is required to give. (E.g. it is generally known that in certain respects the expertness of the nurses in judging social adjustment is greater than that of the psychiatrists).
2 The fact that the judgement on adjustment is of value ij formed in the particular social situation on which it bears.
3 The fact that in the course of forming a judgement the situation conditioning the judgement changes. 4 The distinctive power of a given rating instrument. The descriptions used should be unambiguous and so chosen as to ensure the greatest possible interjudge agreement, while retaining maximum significance for the purpose they serve (i.e. judging social adjustment).
The validity of judgements in psychiatry is discussed.
The existence of any measurable and valid external criterion is denied, considering that the decisions of the psychiatrist in the mental hospital are inevitably influenced and conditioned by a host of still intangible factors.
h is one of the primary functions of a rating instrument, to help make the factors determining the psychiatrists' decisions explicit.
Apart from this practical value of an instrument, reliable in terras of interjudge agreement, it is important, from a theoretical point of view, to assess its construct validity.
The construction of the PUK rating scales is discussed. An exposition is given of the specially devised method by which the building-stones of the rating scales were obtained. Six undergraduate students, working as parttime nurse aids in the above mentioned hospital were instructed to write down in detail their remarks (as participants) on the interaction patterns immediately following ten minutes' observation periods, which each individual student had with every patient over a period of approximately one year. They were told that neither symptoms nor any other characteristic qualities of the patients were of significance of that they had to confine their attention to the interaction patterns on the ward, while keeping in view social adjustment and social acceptibility.
Those observations which recurred in a majority of the individual reports on one and the same patient, were assumed to constitute reliable judgements. It were those judgements which formed the basis for the construction of our rating, scales.
Twenty-one scales consisting of 5, 6 or 7 ordinally arranged descriptions were subsequently tested for their interrater reliability.
Both in the Second Psychiatric Hospital of Utrecht University as well as in another psychiatric hospita! (Zon en Schild, Amersfoort), the results proved to be highly satisfactory (each of the 21 scales used by 4-8 raters, in three different trials, showed a reliability of about 90). As a preliminary indication of the validity of the PUK rating scales it may be mentioned that the median test showed a highly significant differentiation of two populations in each scale.
(P < 0.001). Investigations on the construct validity are continued.