Assessing Development, Not “Static Stuff”
3.19.09 | Too often in today’s assessment system we can’t tell a lack of knowledge from “errors” that reveal deep knowledge. Many tests examine what a learner has done through weeks and sometimes years of instruction by a test given on one day and in a format not remotely like the instruction being assessed. The model of learning behind this practice is the idea that instruction has poured static “stuff” into peoples heads and the test can tell us whether it is stored there or not.
But learning is a form of development, and in any developmental process we want to know where in a trajectory or course of development a person is. And, of course, to know that, we have to know how different courses of development work in different domains of learning. Today we often assess (in the sense of “test”) while not knowing either how development works in a given domain or where a learner is in a course of development in that domain.
Think about what would happen if we gave little children language tests as they were learning their native languages. A child who says “goed” instead of “went” would lose points—perhaps be sent for some remediation. But in the normal course of language development, children first say “went” (because they have memorized past tense forms and do not see present and past forms as related by a general rule) and then say “goed,” showing that they have discovered a major, deep property of human grammar, namely that it is not a list or a set of memorized facts, but a rule-governed or pattern-governed domain. The little child would get a “bad” grade just at the point he or she had accomplished a major step forward in language development. And what showed the child had made such a major accomplishment? A so-called “mistake” or “error.” In terms of development, “goed” is not, in fact, a mistake. It is not an indication of a lack of knowledge, but, rather, an indication of knowledge.
We have known for years that in many areas of human development, development is U-shaped: learners get better, then look like they are getting worse as they reorganize their knowledge and find deeper patterns, then get better again. Today we have an assessment system that, in many cases, cant tell a real failure from the bottom of a fruitful U-shaped course of development.
Of course, we do not do anything this stupid to little children learning their native languages, but we do when we assess with no due regard for courses or trajectories of development (and keep in mind we do not even really know them for many important domains taught in school—but we test away nonetheless).
Leave a comment
Comments are moderated to ensure topic relevance and generally will be posted quickly.