Good News! Contradictory Research on Desirable Difficulties…

As we regularly emphasize here on the blog, attempts to recall information benefit learning.

That is: students might study by reviewing material. Or, they might study with practice tests. (Or flashcards. Perhaps Quizlet.)

Researchers call this technique “retrieval practice,” and we’ve got piles of research showing its effectiveness.

Learning Untested Material?

How far do the benefits of this technique go?

For instance, let’s imagine my students read a passage on famous author friendships during the Harlem Renaissance. Then they take a test on its key names, dates, and concepts.

We know that retrieval practice helps with facts (names and dates) and concepts.

But: does retrieval practice help with the names, dates, and concepts that didn’t appear on the practice test?

For instance, my practice test on Harlem Renaissance literature might include this question: “Zora Neale Hurston befriended which famous Harlem poet?”

That practice question will (probably) help my students do well on this test question: “Langston Hughes often corresponded with which well-known HR novelist?”

After all, the friendship between Hurston and Hughes was retrieved on the practice test, and therefore specifically recalled.

But: will that question help students remember that…say…Carl van Vechten took famous photos of poet and novelist Countee Cullen?

After all, that relationship was in the unit, but NOT specifically tested.

So, what are the limits of retrieval practice benefits?

Everything You Wanted to Know about Acid Reflux

Kevin Eva and Co. have explored this question, and found encouraging results.

In his study, Eva asked pharmacology students to study a PowerPoint deck about acid reflux and peptic ulcers: just the sort of information pharmacists need to know. In fact, this PPT deck would be taught later in the course – so students were getting a useful head start.

Half of them spent 30 minutes reviewing the deck. The other half spent 20 minutes reviewing, and 10 minutes taking a practice test.

Who remembered the information better 2 weeks later?

Sure enough: the students who took the practice test. And, crucially, they remembered more information on which they had been tested AND other information from the PPT that hadn’t been specifically tested.

That it, they would be likelier to remember information about Zora Hurston and Langton Hughes (the tested info) AND about van Vechten and Cullen (the UNtested info).

However, Eva’s tested students did NOT remember more general pharmacology info than their untested peers. In other words: retrieval practice helped with locally related information, but not with the entire discipline.

But Wait! There’s More! (Or, Less…)

About 2 month ago, I posted on the same topic – looking at a study by Cindy Nebel (nee Wooldridge) and Co.

You may recall that they reached the opposite finding. That is, in their research paradigm, retrieval practice helped students remember the information they retrieved, and only the information they retrieved.

Whereas retrieval practice helped students on later tests if the questions were basically the same, it didn’t have that effect if the questions were merely “topically related.”

For instance, a biology quiz question about “the fossil record” didn’t help students learn about “genetic differences,” even though both questions focus on the topic of “evolution.”

What Went Wrong?

If two psychology studies looked at (basically) the same question and got (basically) opposite answers, what went wrong?

Here’s a potentially surprising answer: nothing.

In science research, we often find contradictory results when we first start looking at questions.

We make progress in this field NOT by doing one study and concluding we know the answer, but by doing multiple (slightly different) studies and seeing what patterns emerge.

Only after we’ve got many data points can we draw strong conclusions.

In other words: the fact that we’ve got conflicting evidence isn’t bad news. It shows that the system is working as it should.

OK, but What Should Teachers Do?

Until we get those many data points, how should teachers use retrieval practice most effectively?

As is so often the case, we have to adapt research to our own teaching context.

If we want to ensure that our students learn a particular fact or concept or process, we should be sure to include it directly in our retrieval practice. In this case, we do have lots (and LOTS) of data points showing that this approach works. We can use this technique with great confidence.

Depending on how adventurous we feel, we might also use retrieval practice to enhance learning of topically related material. We’re on thinner ice here, so we shouldn’t do it with core content.

But, our own experiments  might lead us to useful conclusions. Perhaps we’ll find that…

Older students can use RP this way better than younger students, or

The technique works for factual learning better than procedural learning, or

Math yes, history no.

In brief: we can both follow retrieval practice research and contribute to it.

Leave a Reply

Your email address will not be published. Required fields are marked *