The reliability of results from national curriculum testing in England

• Published in: Educational research (Windsor) Vol. 51; no. 2; pp. 181 - 212
• Main Author: Newton, Paul E.
• Format: Journal Article
• Language: English
• Published: Routledge 01-06-2009
• Subjects: Curriculum; Educational Assessment; Educational Policy; Educational Testing; Elementary Secondary Education; England; Error; Evaluation Methods; Evidence; Examinations; Foreign Countries; High Stakes Tests; Internal Consistency; International Education; misclassification; Misclassification (Statistics); National Curriculum; Policy Analysis; Problems; Program Effectiveness; Program Evaluation; Reliability; Student Evaluation; Test Reliability; Test Results; testing; Tests; United Kingdom; United Kingdom
• ISSN: 0013-1881; 1469-5847
• DOI: 10.1080/00131880902891404

Background: National curriculum tests have been administered in England for well over a decade. Although reliability evidence has been published, critics have argued that there is not enough evidence (of the right kind) and that test results may be insufficiently reliable.
Purpose: This article collates and discusses evidence on the reliability of results from national curriculum testing in England. In particular, it examines the claim that at least 30% of students may be awarded incorrect national curriculum levels.
Sources of evidence: Empirical evidence is collated from a range of sources (mainly published) focusing primarily upon internal consistency, marking consistency and level-setting consistency.
Main argument and conclusion: There still remains some doubt over the extent of misclassification associated with national curriculum testing. However, there is no doubt that a substantial percentage of students would receive different levels were the testing process to be replicated within any particular year for any particular test. However, although results from national curriculum testing are somewhat unreliable, they are not necessarily too unreliable.