Test affordances or test function? Did we get Messick's message right?

This paper follows a line of logical argumentation to claim that what Samuel Messick conceptualized about construct validation has probably been misunderstood by some educational policy makers, practicing educators, and classroom teachers. It argues that, while Messick's unified theory of test validation aimed at (a) warning educational practitioners and policy makers of the undesirable social consequences of test use and (b) entreating educators and test developers to think of a facet-driven item-banking-based construct-specific criterion-referenced common metric for any construct of interest in educational and other settings, his message has been misunderstood as a plea for alternative ways of evaluation and specifically a qualitative shift in educational assessment. The paper (a) draws on the conceptual differences between 'test function' and other construct-irrelevant peripheral 'affordances' to which any test can be put, (b) argues that the moment of truth for the qualitative camp has arrived, and (b) invites everyone to admit that even if qualitative assessment, 'thick' descriptions of achievement, and differentiated portraits and profiles of student performance might be much thicker than traditional norm-referenced psychometric tests, they are no match for any minimalistic facet-driven criterion-referenced common metric, nor is any of them an option. The paper suggests that the right way out is through an iron-clad criterion-referenced Occam's-razor-proof common metric for each construct of interest, perhaps the only option that is sure to transform the soft science of educational assessment into a hard science of educational measurement.