I think your argument is too strong: this is not because a test is not a measurement instrument that it follows that the test should not be used. The issue is the purposes they are used for, and, of course, the usefulness of using tests to solve specific problems.
In my view and in short, the point is not to develop pseudo-measurement instruments by focusing on psychometric properties (which only describe statistic variables but not tests--tests have no statistical properties at all), but to address the problems scores or response patterns enable one to solve. Maybe you could be interested in details provided in Afzali, Lacot and Vautier (2016). The basic idea is that assessment (in French, evaluation) does not requires measurement at all--assessment and measurement are not synonyms; assessment requires to be clear with the function of scoring techniques in the context of specific settings. It is striking to me that usual test validation papers tell virtually nothing about the purposes of the tests uses.
Lacot, E., Afzali, M. H., & Vautier, S. (2016). Test validation without measurement: Disentangling scientific explanation of item responses and justification of focused assessment policies based on test data. European Journal of Psychological Assessment, 32, 204-214.
-----Message d'origine-----
De : IDANET (Individual Differences and Assessment Network) [mailto:[log in to unmask]] De la part de Colin Cooper
Envoyé : dimanche 12 novembre 2017 21:34
À : [log in to unmask]
Objet : Re: Undergraduate readings?
Sorry about that! I mean that I basically agree with Michell (etc.). However students and practitioners don't have the luxury of saying "Michell is correct and I will no longer use tests as a methodod of assessing behaviour". For someone (a lecturer, licensing body or whoever) will have decreed that students, practitioners and whoever else reads the book will need to use and develop tests, evaluate their reliability, validity and perform all the other analyses which we enjoy so much. So the book covers the usual psychometric theory and techniques, whilst reminding readers that if what Michell said is indeed true, it is not clear whether these techniques are appropriate or whether the results are meaningful.
|