Shapiro has developed a method of constructing a questionnaire for an individual psychiatric patient, which gives a relatively coarse scaling of fluctuations in a fairly large number of his own symptoms, expressed in his own words, and calibrated by himself. It has been pointed out that this is only one member of a general class of such techniques: the present paper describes another, suitable for obtaining relatively precise (interval) scalings of a fairly small number of symptoms.

For each symptom (i) statements are derived, in consultation with the patient, representing graded levels of its possible intensity and (ii) optionally, a preliminary scaling of this set of statements may be obtained by the patient's making paired comparison judgements of the differences between these levels. Then (iii) on each occasion when a scaling of the current level of the symptom is required, the patient makes paired comparison judgements of the difference between it and each of the statements in turn (an estimate of the level on each such occasion being given by the average of these judgements). After all the data have been collected the significance of changes in the level of the symptom is assessed by means of a final analysis of variance, which also checks the preliminary scaling of statements and tests for systematic inconsistencies in the patient's responses.

ALGOL 60 programs to check, analyse and graph the results are reported, and a clinical example presented, together with supplementary statistical analyses to answer further questions of clinical relevance which typically arise.