Watched another great lecture from Stanford’s HCI group by Dan Schwartz from the School of Education, who argues that choice is a better outcome to assess for than knowledge. The choices we make about our learning trajectory are better predictors of how well we will learn and do in the long term, because these decisions imply how we will deal with things we don’t know– which is only bound to happen in life. Traditionally, choices are thought of as classroom inputs, but not ever as outputs to test for.
- The test assessment model has significantly influenced classroom teaching models; because of the importance of tests in serving as a proxy for performance, teaching tends to be dictated and pressured by beating the test
- Students are taught to believe that learning is about memorizing and recall of concepts– and about ‘getting things correct’
- In fact, this is not very helpful for students to cope with real life for two reasons. 1) Tests are administered in ‘sequestered’ situations, i.e. students have to complete them independently. In real life, what we want them to do is the opposite: to get help and collaborate when they don’t know something.
2) It teaches them that being smart is about getting things right absolutely– and that you are smarter if you can do something without much help at all, when in fact– being smart is not about getting things right at first try but about knowing how to fix them when they are not.
- Knowledge is not always a good outcome to test for because what “knowledge” means can change with time and situation– static ‘facts’ are not as useful as knowing how to learn
- The reason why parents or educators want students to do well, to ‘have knowledge’ is often not very much about the knowledge itself at all (really, do you think anybody cares about the details of photosynthesis still?) but that it enables a path to more choices of the life they want to lead
- So ultimately, everybody cares about the same thing– which is that students can make good choices now, which will help them to make good choices later in life– so why not just test for that now?
I like the idea a lot; it is one of those really out-of-the-box things that challenges a fundamental paradigm that is so given that we rarely realize is a paradigm at all. Still, I do have a lot of questions surrounding the practical adoption of this ideal– I think it is more practical at the moment as a new way of thinking than a real model for change.
One of the key questions I have is how would you then assess/rate performance of the students? Are we giving better scores to students who make better choices for learning, or are we letting them make the better choices and then still ultimately testing for the post-results, i.e. the knowledge that they have learned? It seems very unlikely the traditional higher education institutions or companies will stop ranking people on some assessment of performance– so how would that play out? If you grade people based on choices:
- Does knowledge performance then not matter? How do you account for students who make good decisions about learning but who just cannot learn the material?
- It seems like a matter of time before there would be students trying to beat the test: learning the optimal strategies rather than truly making the choices because that’s what they inherently want to do
- But that might not be a bad thing– it could be a strange Pavlovian training to condition students into making the right choices– even if they are not quite instinctual at first
- Choices are extremely normative. Could schools really get away with having a right/wrong assessment that dictates one type of learning style is better than another?
It seems to me that there is a philosophical paradox somewhere in this model, even though the idea itself is very intriguing. I do think ultimately this emphasis on choice falls somewhere in between– i.e. it is important to make students cognizant of their learning choices and styles (and some sort of normative guidance on what is better), but probably quite problematic if you test for choice itself and not knowledge at all. It seems difficult to conceive a scenario in which you don’t have to test for knowledge; wouldn’t you have to test for it as a result to see if the intermediary choice was indeed superior for the individual?
Still, I do think it’s onto something pretty exciting.