The construction and validation of a spatail test, using diagrammatic material based on projections and sections of solid objects

An important development in mental testing has been the construction of tests for measuring the ability to obtain,and utilise, visual spatial imagery. These spatial tests have proved valuable in predicting subsequent success in such spheres as engineering apprenticeships, technical drawing and woodwork. They have also been employed successfully i n selection for secondary schools. Although a limited number of such tests already existed, it was felt that a new test based on three-dimensionalmaterial, and particularly projections and sections of this material, could prove of value. Two hundred possible test items, of thirteen types, were therefore prepared and "tried-out" on a representative sample of school children. An item analysis of the resulting dataprovided indices of Facility and Discrimination which were used to select the one hundred most suitable items. The selected items were then organised into a revised draft which was inserted into the selection examination for the entire ten-year-old population of a city. This large- scale trial, as well as showing that the test and instructions are suitable for the age group, provided information from which tables for converting raw scores to standardised scores were constructed; enabled a second, confirmatory item analysis to be made; showing a significant difference in mean scores for boys and girls, a recognised property of spatial tests; and provided the following figures: Range: 1 - 99 Mean: 42.1 Standard Deviation: 20.745 Reliability: 0.9642 In an investigation with 85 boys the test correlated more highly with a recognised spatial test than with a verbal test, and was a better predictor of success in metalwork than was the verbal test. Extended validation, notably a four year follow-up of the 85 boys, and a factor-analysis are required before it can be certain that the test is truly "spatial" with the uses associated with such tests.