Assessing pupils’ skills: implications for research in education

Abstract

The subject of this article is the structure and evolution of skills developed by pupils at primary level. Starting from an analysis of the panel data provided by the French Ministry of Education, the main object of this paper is an original measurement of skills using structural models. The findings of this research raise two complementary issues. The first issue concerns the cross‐curricular nature of skills suggesting a logic of disciplinary interdependence. The acquisition of a skill may be strongly correlated to the acquisition of other skills belonging to the same subject area or to other subject areas. The notion of a set of skills is used to account for the connections between the different aspects of acquired skills. The second issue concerns the evolution of skills over time. Here the aim will be to identify the kinds of skills that are the most predictive of subsequent educational success. This issue is highly relevant for educational policy‐making. It is hoped that the results presented in this paper will improve one’s understanding of the ways in which schools might provide pupils with the best chances of success in the earliest stages of their educational career.

Keywords:

• structure and evolution of skills
• structural models
• key skills

Notes

1. See http://www.hce.education.fr.

2. See http://www.education.gouv.fr/cid1180/direction-evaluation-prospective-performance.html.

3. In the 1970s, a group of researchers (Jöreskog and Van Tillo 1973) conceived and developed a software application known as LISREL (Linear Structural Relations) that enables the analysis of structural models in the social sciences and which subsequently generated other applications such as AMOS, EQS, LISCOMP, MPLUS, etc. LISREL compares two series of values in variables: the theoretical values implied by the model with which it is provided and the observed values. Calculations using LISREL involve two distinct phases: the use of PRELIS (a pre‐processor of LISREL) that calculates a matrix of observed correlations providing input for the second phase of calculation where it is compared with the matrix of theoretical correlations.

4. The ‘correlations’ column indicates relations between items; for example, ‘F67/F62’ translates the correlation between items 67 and 62 in French.