NotesFAQContact Us
Search Tips
Peer reviewed Peer reviewed
Direct linkDirect link
ERIC Number: EJ1002207
Record Type: Journal
Publication Date: 2013
Pages: 5
Abstractor: ERIC
Reference Count: 8
ISSN: ISSN-1536-6367
How Is Testing Supposed to Improve Schooling? Some Reflections
Wiliam, Dylan
Measurement: Interdisciplinary Research and Perspectives, v11 n1-2 p55-59 2013
In "How Is Testing Supposed to Improve Schooling?" Edward Haertel has proposed a framework for thinking about the mechanisms by which testing might improve the various educational processes undertaken in schools. The framework seems to the author to be quite general (he uses the word "general" here in its mathematical sense of including all cases) since he cannot think of any mechanisms that could not be easily accommodated within the framework. The distinction between direct and indirect purposes or mechanisms is also useful, although the author is not sure that these are the right terms to use. He can see that, from a psychometric standpoint, it may seem that mechanisms that require a test score might seem more direct, whereas the fact that students prepare for a test and, thus, learn more may well be an indirect benefit. However, from a lay perspective, a case could easily be made that the labels should be precisely the other way round. Any benefit that accrues to the test, whether a score is generated or not, could be seen as far more direct than a mechanism that needs to work via a test score. Indeed, it seems at least possible, if not likely, that the indirect effects are greater in magnitude. For the remainder of this response, however, the author wants to focus on the first of Haertel's mechanisms--use of assessment for instructional guidance. As a result of the work in which he and Paul Black have been engaged over the last two decades, the author is now firmly convinced that the use of assessment for instructional guidance offers one of the most powerful ways for improving schooling. However, during this time, he has become increasingly certain that traditional forms of testing have little to contribute here, for three reasons. The first relates to what one is learning about the ability of even well-designed tests to support inferences about multiple aspects of individual performance. The second relates to the "grain size" of the constructs on which the assessment is focused, and the third relates to the practical problems of using such test data in real classroom settings.
Psychology Press. Available from: Taylor & Francis, Ltd. 325 Chestnut Street Suite 800, Philadelphia, PA 19106. Tel: 800-354-1420; Fax: 215-625-2940; Web site:
Publication Type: Journal Articles; Opinion Papers
Education Level: Elementary Secondary Education
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A