Teaching to the test displaces the education the test was meant to check

The argument

A second argument concerns what testing does to teaching. Proponents argue that once a score carries serious consequences for students, teachers and schools, the rational response is to optimise for the score, and that this reliably changes instruction for the worse. Time goes to the subjects and topics the test samples and away from those it does not, so music, art, physical education, practical subjects and anything unexamined are compressed. Within examined subjects, time goes to the question formats that appear on the paper, to drilling, to timed practice and to the techniques of answering rather than to the understanding the questions were designed to detect. Proponents argue that this corrupts the measurement as well as the teaching. A score that rises because students have been coached on the test tells nobody anything about whether they understand more, which means the instrument stops reporting the thing it is used to report at exactly the moment it starts being taken seriously. The familiar consequence is a cycle in which results improve, the improvement is found not to transfer to anything else, the test is revised, and the optimisation begins again against the new version. Proponents also point to effects on students that do not appear in the results. Where a single examination determines a path, the pressure attached to it is considerable, and it is applied to young people over months or years. This produces anxiety, sleep loss and a relationship to learning organised around avoiding failure rather than curiosity, and it teaches students that the purpose of understanding something is to be scored on it. On this account the costs are paid in the quality of education itself, which is why proponents argue for abolition rather than reform of the instrument.

Premises

Counter-arguments

Defenders argue that the distortion described is a consequence of attaching high stakes to a score rather than of measuring anything, and that the remedy is to loosen the stakes, sample more broadly or use tests for diagnosis rather than allocation. They also argue that teaching to a well-designed test is simply teaching, since the content of the paper is the content of the course, and that the complaint is really about badly designed tests. On the pressure point, defenders note that competition for scarce places produces anxiety under any selection method, and that assessment by continuous teacher judgement distributes pressure across the whole of a student's time at school rather than concentrating it where it can at least be prepared for and ended.

Rejecting the premises

[Rejecting P1] High-stakes use is a policy choice that can be loosened, so the distortion is not an inherent feature of measuring. [Rejecting P2] Teaching to a well-designed test that samples the course broadly is simply teaching the course. [Rejecting P3] Score inflation through coaching is detectable and can be addressed by rotating content and auditing results. [Rejecting C] The argument supports reforming the stakes and the design rather than abolishing assessment altogether.