Fri 9 Apr 2010
Taking tests helps you learn (maybe)
Posted by sciencegeekgirl under Educational Change , How People Learn1 Comment
I’ve written before on some interesting psychology studies on the benefits of retrieval for learning and memory. I recently heard a talk on the subject (by Sean Kang of UCSD) that spurred me to think about it again, and also generated some interesting discussion with a thoughtful colleague here at CU Boulder. Read on….
Many studies in the past have shown that testing is helpful, not just to see how much you learned, but to actually HELP you learn. A seminal study was that of Spitzer (waaay back in 1939) who found that subjects did better on a test on material 21 days later if they had had more tests on that material in the interim. There is a useful review on this field from 2006 here (Roediger, H.L. & Karpicke, J.D. (2006a). The power of testing memory: Basic research and implications for educational practice. Perspectives on Psychological Science, 1, 181-210. [PDF])
It’s unfortunate, said Kang, that these results haven’t translated to classroom practice. Teachers aren’t giving quizzes or tests as part of their instruction. But more on that in a moment — my colleague doesn’t think that this research should necessarily be applied in classroom practice.
First, let’s look at the research on testing, briefly. There are a few possible explanations for the effect that testing helps later retrieval.
- The test provides learned with one more presentation of the material
- The test gives practice for the next test
- The act of retrieving information generates connections in the brain, producing learning as part of that process.
Dr. Kang presented a series of studies that seemed to support the third hypothesis, since tests that required more effort (short answer instead of multiple choice, for example) appeared to produce more learning. Students also did better, not surprisingly, when they expected a harder test. Students regulate their studying by how tough they think the assessment will be.
Interesting results, and it makes sense to me. Testing is, in a way, a form of active engagement. The student is being required to recall information and write it down, which is a much more active brain process than, say, reading the information one more time. So, I’ve always liked these studies, as they demonstrate how learning is generated by active processing. These studies suggest effective ways for students to monitor their own learning and how it is that we can cement “facts” in our minds. I was seeing it as a piece of a broader picture (memorization as one useful but not sufficient piece of learning).
But my colleague Noah Podolefsky, a postdoctoral researcher at CU, questions the whole premise of these studies. On one of the original papers on retrieval and testing he wrote me,
But I have to say, they lost me at the first line, “Learning is often considered complete when a student can produce the correct answer to a question.”
I do think that in order to construct broader conceptual understanding, we do need to be able to regurgitate some facts. And Noah agrees that remembering facts is important. We can’t string facts together to make coherent meaning if we’re always beyond our cognitive load because we can’t remember any of the facts.
Noah doesn’t disagree with the validity of the results of these studies, but he thinks that overgeneralizing them could actually be dangerous to education.
I think this framing of what it means to learn is not just limited, but actually dangerous. I’m not convinced the advantages of being able to perform on these sorts of recall tests outweigh the consequences of students being subjected to this sort of instruction.
I don’t blame these researchers – in my opinion, the standard research methods of psychology and cognitive science drive researchers to ask just these sort of questions, and the sorts of results they produce are very appealing to people in the “hard” sciences because they are nice and clean. They look like physics experiments. Unfortunately, in order to reproduce these results in real classrooms, these classrooms have to look more and more like psych labs. To me, that is the wrong direction.
After all, even if you remember the facts, how do you use them? “The way you use facts when you are solving a complex problem may be fundamentally different from how you use facts when the task is simply to recall the fact,” says Noah.
Food for thought. Myself, I agree with Noah, but also find the research on testing and retention to be personally useful, so I hesitate to say that it’s not pedagogically useful as well. I’d be interested to hear what any teachers have to say on these ideas.
Got work? I love work. Drop me a line about your project in science education or communication, and read more about me at my client website .









No post about Halloween would be complete without a reference to the 



