Read I’ve Been Studying All Wrong: Highlighting and Rereading if you haven’t already.
Retrieval Practice (AKA Practice Testing)
Retrieval practice, also known as practice testing, has consistently shown to be one of the most effective and efficient learning strategies.
What is retrieval practice?
Retrieval practice is any sort of low-stakes (or no-stakes) testing or quizzing where students retrieve course material from memory. This includes not only instructor-supplied practice exams but also flash cards and student-generated practice questions. I prefer the term retrieval practice over practice testing because students may have a negative association with the term testing, and the retrieval from memory is the critical component. In their review of learning strategies, Dunlosky et al. (2013) concluded:
“…we rate practice testing as having high utility. Testing effects have been demonstrated across an impressive range of practice-test formats, kinds of material, learner ages, outcome measures, and retention intervals. Thus, practice testing has broad applicability. Practice testing is not particularly time intensive relative to other techniques, and it can be implemented with minimal training.”
*Important Note
For retrieval practice to be most effective, it is critical that a student actually retrieves (or attempts to retrieve) the information or concept from memory. For example, if a student doesn’t know an answer to a question, it is more beneficial if they make an attempt (even if it is incorrect) before looking up the correct answer! This is actually referred to as a desirable difficulty. When students struggle to retrieve information from memory, subsequent retrieval will be easier.
Why should students use retrieval practice?
Retrieval practice leads to a number of direct and indirect effects that facilitate learning (see Roediger et al. 2011 for a detailed review).
Direct Effects
When done correctly, retrieval practice forces a student to retrieve (or attempt to retrieve) knowledge from memory. If the student correctly recalls an answer, they will be able to more easily recall it in the future, and it may improve students’ mental organization of the information. Even if they answer incorrectly, if they are given (or look up) the correct answer, they will be more likely able to retrieve the information in the future. The key is the retrieval. For example, if a student is reviewing a flash card but flips to the solution without attempting an answer first, this is no more effective than rereading.
Indirect Effects
Frequent quizzing or testing, even low-stakes, by an instructor may motivate students to study more or to space out their studying over a longer time frame, rather than cramming before the midterm or final. (See Part 2 for more on why this is important.) This can also give feedback on students’ understanding so the instructor can adapt accordingly and students can identify gaps in their knowledge in order to focus their studying.
If retrieval practice is so effective, why don’t students use it more?
Many students approach testing, even practice testing, with trepidation, and may therefore avoid it. It is also uncomfortable to not know the answer to a question. Again this is actually a desirable difficulty that aids learning, though it may just feel like a difficulty to students. This is in contrast to an approach like rereading, where it may seem to a student that they understand everything since they aren’t forced to actually assess what they can recall or understand. Another barrier to retrieval practice is that, in contrast to highlighting or rereading, it requires some preparation. Either an instructor has to administer or supply practice tests or quizzes, or students need to create questions or flash cards on their own.
How I facilitate retrieval practice for my students
Explaining the Technique
In my first lecture in a course, I actually present some learning strategies to students, including suggestions on how to incorporate them into their studying. For example, writing their own questions from the material and exchanging them with a friend. I do however recognize that a portion of a lecture spent talking about learning strategies isn’t going to upend how students have been studying for the past 13 years, so I also facilitate retrieval practice through the use of clickers and low-stakes quizzes.
Clickers
In the last course I taught, I began every lecture with 3 or 4 questions that students would answer with clickers. The questions were primarily on material from the previous lecture. The clicker questions were only worth 5% of the final grade, and I scored them generously (e.g., ≥ 50% correct -> full points for the day, ≥ 80% of clicker points over the semester -> full 5%). The low-stakes was to balance incentive for attendance and participation against the motivation for misconduct.
This was helpful to students, but I was missing a key component of retrieval practice. I allowed students to discuss the question with their neighbors and look up the answers in their notes. I now know that both of those factors reduced the effectiveness of the technique. In my next course, I will explain the logic of retrieval practice and propose a deal to students. Their clicker scores will be based solely on participation if they are willing to answer the questions without notes or help from a neighbor, because my goal is for them to recall (or attempt to recall) the information from memory, regardless of whether they know the correct answer. The beauty of a clicker is that I can then immediately share the correct answer.
Low-Stakes Online Quizzes
I also posted a weekly online review quiz on Canvas, covering that week’s lectures. Students could take it any time between Friday’s and Monday’s lectures. Retrieval practice is most effective if students aren’t referencing their text or notes, but there is some evidence of an effect even with open-book quizzes. In my last course I would have preferred the quizzes to be closed-book, but, because the temptation to look up an answer would be high, I allowed the use of notes and books. The quizzes were cumulatively worth 10% of the final grade, enough to motivate students to do them, but hopefully not so much to incentivize misconduct (students working together on quizzes). In order to be able to post the quiz answers immediately after the quiz closed, I didn’t allow makeup quizzes, but I allowed each student to drop their two lowest scores (so it was not an issue if they missed a quiz due to illness or other obligations).
In order to improve the effectiveness of the retrieval practice, I’m currently trying to figure out a way to encourage students to attempt to answer a question without looking it up. I can’t simply make the quizzes closed-book, because the temptation to look up an answer would be too high. My inelegant workaround at the moment is to give students two attempts at each quiz, only counting the score of the second attempt. I will limit each attempt to 10 minutes (unfortunately the time limit for each attempt must be the same in Canvas). My hope is that 10 minutes on the first attempt will be just enough time for the students to answer (or attempt to answer) each question, but not enough time to look up the answers. The trick is that I can allow students to see the graded quiz after the initial submission. For incorrect answers, they won’t see the correct answer, but they now have a chance (with potentially unlimited time) to use their text and notes to figure out the correct answer for each question prior to starting their second attempt.
This is a bit convoluted, but my hope is that by explaining the science of retrieval practice beforehand, I can get buy-in from the students on this approach. The two-stage quiz is a bit of extra effort, but they are rewarded with the fairly good chance (with some effort on their part) of getting all the question correct on the second, graded attempt.
Recommended Reading
Dr. Pooja Agarwal and colleagues have a great website, www.retrievalpractice.org, devoted to retrieval practice. Their site includes a helpful guide, How to Use Retrieval Practice to Improve Learning.
Next: How I Should Have Studied. Part 2: Distributed (Spaced) Practice
References