A 2nd-order factor-analysis of the reliability and validity of the 11 plus examination in Northern Ireland

Abstract

In Northern Ireland the transition from primary to secondary education is a critical process since the type of secondary school attended is a powerful determinant of the life-chances of individuals. In Northern Ireland children enter the secondary sector of education at about age eleven and the allocation process for non-fee payers is decided by performance in two parallel tests each lasting for 50 minutes. The examination takes place early in the last year at primary school with an interval of about six weeks between the two tests. These tests, commonly referred to as the 11 plus, are of profound importance to the children, parents and the allocating agency. This paper attempts to assess the reliability and validity of this allocation mechanism. The two test papers for three consecutive years namely 1982, 1983 and 1984 (six papers in all) were completed by the children (n = 47) over a period of weeks. The analysis of the participants' marks was based on a second order factor analysis using both the structural and measurement model in LISREL. This permits the variance to be partitioned into three components: (a) a common component, (b) a unique component, i.e., that which is specific to any particular test, and (c) the measurement error variance. The validity of a test is measured by (a), the invalidity by (b) and the unreliability by (c). The reliability and validity of this restricted sample, suggests that should such results be generalised, a large number of children would have been incorrectly classified by these test papers.