indices as unclassifiable, sexual and questionable responses, multiple themes and omissions, statistically significant differences have been established between a control group and subgroups which are (a) emotionally disturbed, (b) shy and reserved, (c) highly sexed. These results are based on IOI grammar-school boys in the fifth and sixth forms.

Tests AH 2 and AH 3 are group tests of reasoning with several features usually associated with individual tests of intelligence. In particular, they yield a three-point profile in addition to their total or general score, and their range of relevance extends from 10-year-olds to intelligent young adults. They have a certain flexibility in that they can be administered with either a long or a short time-limit, norms having been established on both sets of times. New techniques have been used in devising items and building-in checks to ensure that there is one and only one defensible answer for each question.In the course of standardizing the two tests, several widely accepted &dings have been confirmed, notably male superiority on Numerical, female superiority on Verbal and the greater variance found generally among males than females. Two further results are, however, worth noting: (a) the data do not support the suggestion, often made by teachers and other educationists, that girls tend to excel boys around 10-12 years of age owing to their earlier physical and mental maturationto be rejoined, or surpassed, by the boys in their early teens; (b) it seems well worth while incorporating pictorial items in a group test of reasoning, intended to cover a wide age range. This type of item enables the younger (and the less able) to show their strength in an area usually neglected in group tests and it proves to be a task in which they delight. AH 2 and AH 3 are two new, parallel, multiple-choice group tests of reasoning, suitable for literate subjects from 10 years to adulthood. This assertion of appropriateness is based on the fact that, whilst dull 10-year-olds do score above chance on the few questions they attempt, intelligent young adults such as students fail to gain 100 per cent. The percentage bar charts in Figs. 1 and 2 show the range of score and the changing level of mean with age.The wide range of relevance of the two tests has been attained in four ways. First, the test can be given with either a long (42 min.) or a short (28 min.) timelimit. In this way, younger (and also duller) subjects, using the long time-limit, produce quite a wide range of score (see Fig. 1); whereas effective differentiation obtains with older (and brighter) subjects, using the short time-limit. As in all group tests, it is important that subjects should be working throughout the time allowed, in order that comparison among individuals should be meaningful.Secondly, AH 2 and AH 3 (identical in format but having no items in common except the preliminary examples) both comprise three separately timed parts : 40Verbal questions (V), 40 Numerical questions (N) and 40 Perceptual questions (P).Thus a three-point 'profile' is obtained for each subject in addition to his total score. We had hoped by means of the three separate parts to achieve a group test with some of the qualities of an individual test. Some suggestion of the success of this aim is reflected in the high correlation found between AH 3 and the WISC, in a small group of mid-...

Summary. GCE O‐ and A‐level examinations of boys from a public school were correlated with the AH6 AG test scores obtained a year earlier. The results suggest that the predictive value of AH6 for success in individual subjects is almost as high as when the examinations and testing were taken within a few weeks of each other. This is equally true for estimates of overall success in O‐ and A‐level examinations where highly significant correlations are found.

One hundred and ninety-one first years and eighty-four fourth-years in a co-educational comprehensive school took either vocabulary test P or Q. The tests were divided into equated subtests, P1 and P2, Q1 and Q2, the subtests being given either in auditory or in visual form. The experimental design was counterbalanced, in order to verify subtest equivalence and absence of order effects. The hypothesis was confirmed that auditory presentation of stimulus words constitutes an easier task than visual presentation. The further hypothesis that this difference is greater for the younger and the less able subjects was partially confirmed.

No abstract