Intelligence of Patients in Subnormality Hospitals: A Survey of Admissions in 1961

¹ Lecturer in Social Psychology, University College, Swansea
² Lecturer in Psychology, Birkbeck College, University of London

Our survey suggests that in practice the categories of Severe Subnormality and Subnormality introduced by the 1959 Mental Health Act are being applied in some cases to patients whose intelligence level does not warrant such classification. A substantial proportion—up to one quarter—of patients classified as Subnormal had I.Q.s in the dull-normal and average ranges and thus cannot be regarded as showing "subnormality of intelligence", while of a selected sample of those classified Severely Subnormal, at least a half had I.Q.s of a level which suggested that they had the potential ultimately to leave hospital and lead a more or less independent life. These findings, and those concerning the use of intelligence tests, lead us to make the following suggestions.

(i) For the purpose of the Mental Health Act, the term "subnormality of intelligence" ordinarily should be taken to refer to scores of more than two standard deviation units below the mean on a standard intelligence test. It should never be applied to scores of less than 1.4 standard deviation units below the mean.

(ii) The classification of Severe Subnormality should be applied only to patients having scores of more than three standard deviation units below the mean on a standard intelligence scale.

(iii) There is an urgent need both for the restandardization of the Binet and Wechsler Scales on a British population and for the development of assessment techniques based on more recent research into cognitive processes. Meanwhile, in the training of all test-users likely to work with subnormal adults, particular emphasis should be laid upon the hazards and uncertainties in the use of tests (for example the Binet) which have not been standardized on an adult, British population.

Even if the borderlines we propose are not acceptable we would urge that some attempt be made to reach agreement on such borderlines, because at present there is too much variation in practice. Greater uniformity would improve not only official statistics; it would also assist treatment, training and the collection of data for research.