I like Stephens’s notion that tests are used as proxies (see the very first post of this week). This is particularly true when also the subject matter to be learned is just a proxy for the capabilities needed later.
In the past when the “Nuremberg funnel” mechanisms of stuffing content into students’ heads were not yet very sophisticated, the capability to learn a lot of subject matter was a good proxy for the capability to learn the real important stuff later. The chaotic teaching of professors who had no clue of how to mediate content, led us students to developing skills of self-directed learning. So the teaching of useless stuff was useful for learning how to learn. Learning of relevant competencies worked only indirectly, not by transmission but by induction. And assessing the memorized content was a fair proxy of assessing the induced capabilities that were really important later.
The problem is when we forget about the indirectness, and think that the content is the true value, and optimize and maximize the content and throughput through the funnel. Then measuring the memorized subject matter is no longer an indicator for expectable later success, no longer a proxy for valuable learning.
Jenny reminded us that “assessment has such an impact on so many people’s lives”. Therefore we need to be aware of when we are assessing proxies rather than the real learning that has an impact on people’s lives.