A summary of the paper ” The power of successive relearning: Improving performance on course exams and long-term retention.”
by Rawson, K. A., Dunlosky, J., & Sciartelli, S. M. (2013). Educational Psychology Review, 25(4), 523-548.
Did it ever happen to you that you crammed to a test, did well, and then forgot almost everything you learned in a matter of days ?
If you are a teacher, this becomes a crucial question– are we teaching our students to pass the tests or are we teaching them so they will have useful knowledge for the rest of their lives? Ideally, these two goals would be aligned, nevertheless, in practice we know that they not always are. However, as you may probably know, it is definitely possible to teach for long term results. Findings from the cognitive sciences converge with very effective (but not popular enough) teaching approaches.
Specifically, it is possible to talk about three main principles that promote effective long-term learning:
- Meaningful connections – explain every new concept in terms of concepts that learners already know.
- Retrieval practice – to maintain the information and make it accessible, learners should practice is by attempting to recall the information.
- Spaced Practice – to make the practice effortful and effective, distributing the practice repetitions over time is beneficial.
How effective would it be to combine these three well-established ideas into one powerful practice strategy? Rawson, Dunlosky, and Sciartelli from Kent University, designed such a procedure and tested the effects among several groups of undergraduates at their university. Their design and study outcomes provide a compelling learning strategy with high translational value to application across classrooms.
They termed their strategy “Successive relearning”, to emphasize the repeated effortful reconstruction of memory during practice sessions that followed the lectures. We will refer to their strategy as “Spaced Retrieval Practice” to reference to the strategies above.
All the participants in this study were undergraduate students that attended an introductory psychology course. Lectures in this course took place as usual, the experimental manipulation included additional practice sessions on lectures’ content.
The lecturer defined a list of concepts that were all covered during the lectures (e.g. “classical conditioning”, “confirmation bias”). The participating students were assigned to one of several experimental groups and learned half of the concepts in a method that was designed by the researchers. The other half of the concepts they learned on their own, in a way they saw fit as a preparation for course exams. At the end of the process the researchers compared the learning outcomes for the concepts that were learned using the designed strategy vs. the concepts that were learned in the “business-as-usual” mode*.
The learning strategies
The researchers developed several variations of practice routines; we will consider here the two major ones. The practice sessions were conducted in the laboratory, individually, in front of a computer screen, using a simulation of learning with flash-cards. The studied concepts were presented in practice sessions only after they were learned in class.
The two practice strategies were as follows:
1. Spaced Retrieval Practice– In this strategy the learning of each concept progressed in 3 steps:
- Retrieval – the concept was presented along with a free-text field, the participants were asked to type the definition to the best of their ability.
- Monitoring – The concept and the just-entered response were presented, along with the correct definition broken into its main ideas. Participants were requested to indicate for each main idea if it was included in the definition they have provided.
- Feedback – The concept appeared again along with the complete definition.
The practice sessions were self-paced and ended after each concept was correctly recalled 3 times (or after 60 minutes).
2. Restudy – each concept appeared along with its full definition in a self-paced manner. After the initial presentation, each concept appeared 5 more times (the number of presentations was chosen to be similar to the average number of presentations in the spaced retrieval condition).
Practice Sessions and Tests
Practice Sessions – all participants took part in a total of 6 practice sessions over 3 weeks, 2 sessions per week.
Course exam – took place 2 days after the last practice session, included multiple choice questions.
Exam 2 – took place 3 days after the course exam, included cued recall question (write definitions).
Exam 3 – took place 24 days after the course exam, included cued recall question (write definitions).
What was Measured?
The participating students were taking this course for credit and learned for the exam like any student would. They practiced in the lab as part of the experiment only half of the concepts that were covered during lectures, the other half they had to study by themselves and as they saw fit – “business as usual”.
In each of these exams, the researchers compared memory performance for the concepts the participants learned in the lab using the practice strategy (either spaced retrieval or restudy) with the concepts that they learned by themselves in the “business-as-usual” mode.
The design of the study is quite complex, nevertheless, it is worth our attention from two main reasons:
First, the experiment was conducted in natural conditions and schedules: it is a real course, students’ motivation is genuine, the schedule is reasonable and the exam schedule represents time points that are long-term and of great interest in educational settings.
Second, the performance in the experimental conditions is compared against the performance of the same students in “business as usual” conditions! That is, the comparison is performed within the group of participants rather than between different groups.
Additional reason to pay attention is the highly interesting findings of the study:
Performance in the course exam shows that concepts that were learned using the spaced retrieval practice (Orange) were better remembered than the concepts participants learned by themselves (business-as-usual, adjacent Grey). The difference was more than a full letter grade.
The restudy practice condition (Yellow) was not as effective when compared with “business as usual” concepts (adjacent Grey).
The figures were re-created using the data provided in Rawson, Dunlosky, Sciartelli, 2013
These results are impressive: Spaced retrieval practice is better than just spacing and better than whatever students are choosing to do on their own.
However, an inevitable and justified question would be:
“Does it worth the effort?” do six after-class study sessions worth one letter grade?
The answer becomes much clearer when looking at how performance unfolds over time: In the figure below you can see the performance in the first course exam (left, as above), followed by the performance in the exams, that took place 3 and 24 days later (middle and right):
The figures were created from the data provided in Rawson, Dunlosky, Sciartelli, 2013
First, it is clear that the two spaced practice methods (Orange and Yellow) are much more effective in supporting students’ long-term retention of the information than the “business as usual” approach (Grey). We can probably assume that “business as usual” meant cramming the day before the exam, and while it was enough for the course exam, what we see in the longer-term time points is the representation of the common phenomenon “I studied to the test, I did well, but forgot everything as soon as I walked out the door”. The spaced practice methods did not just yield better results on the exam, but prevented the post-exam rapid forgetting.
This indicates that practicing in a spaced schedule has qualitatively different consequences to the fate of memory. Massed repetitions are activating the learned information so it is easily retrieved only when the context of retrieval has not changed much. When the practice is distributed over longer periods of time, we are challenging ourselves to reconstruct the information, as well as the pathways leading to it in varying contexts and conditions – this effort establishes more meaningful and interconnected representation of the memory that will be accessible even when conditions change over time.
In this Image, we use the pyramids model of knowledge building (introduced here) to demonstrate how the learning strategy influence the fate of learned material over time. Without meaningful associations (upper left, incorrect placement of bricks), the performance is highly dependent on the context of learning (orange box), it is high shortly after learning but does not survive when the context fades over time (upper right). When the structure is meaningful (bottom left, correct placement of bricks), the knowledge structure survives even when the context of learning no longer exists (bottom right).
The other factor that plays a crucial role here is Retrieval: this study demonstrates that when the practice sessions are not only spaced (which happened both in the Spaced retrieval and in the Restudy conditions), but also include retrieval practice, the long-term performance is even better. You can see the difference at 24 days highlighted by the arrows in this figure: the Yellow arrow represents the contribution of spacing alone, while the Orange arrow represents the contribution of spaced retrieval.
The results of this study are powerful because they show us how a well-designed practice routine that combines two important principles of effective learning: spaced practice and retrieval practice, can help real students to do better on the exam. More importantly, it helps students retain the information even after considerable time has passed.
The effect of elaboration is not directly discussed in the paper, but it is included in the spaced retrieval procedure, as students are required to judge inclusion of core ideas in their initial answers. so the spaced retrieval conditions actually involve a combination of all three important components: elaboration, retrieval and spacing.
The results were replicated and extended to include several other interesting conditions.
By using several variations of the practice procedure the researchers replicated their results and demonstrated that spaced retrieval practice was effective also when:
- The criterion for completing each learning session was reduced from 3 to 1 correct response.
- The participants decided by themselves how to study each concept (by retrieval, judging correctness of previous definition or re-studying it).
- The participants completed the study sessions at home rather than at the lab.
In addition it was found that:
- Similar gains in performance were evident for both lower-performing and higher performing students (according to their performance in exam 1)
- The effect was seen in both lower-level memory based questions and higher-level comprehension-based questions.
Importantly, the time students spent practicing dramatically decreased with every subsequent study session!
This last point is crucial, as often teaches (and students) worry that retrieval practice is time-consuming. Another concern is that retrieval would be hard or impossible if too much time has passed – therefore, it is important to remember that as this study shows, repeated attempts are much easier and much faster. This finding may support the notion that the information is not “lost” but has became inaccessible – spaced retrieval is effective in restoring the neural access pathways.
This study helps us make a leap from theory to practice. It shows that by applying a well-spaced retrieval practice, that follows an initial learning phase, students can:
- Do better on exams
- Better retain the knowledge for future use, and as a basis for higher-level learning.
- Potentially also gain hands-on experience in applying effective learning strategies.
This study clearly demonstrates that practice is a crucial part of the learning process that should not be neglected. This aspect is sometimes hidden, as we only see students’ performance on short-term exams, which do not reveal the full picture (i.e. the fast decay that occurs after the exam). However, on the longer term we often witness the lack of effective practice; for example when starting to teach advanced material, implicitly assuming that students are familiar with the required background knowledge which was on last year’s curriculum, only to find out that they don’t.
To truly support students in integrating effective practice routines into their learning routine, the strategies should be integrated into lessons.
With persistence, students will hopefully acquire them as learning skills.
* A special note to those who are interested in field-research:.
The design of this study can inspire effective field- research in any subject. Two points are worth highlighting:
- Note that in this study, the comparison between the two learning conditions (with vs. without strategy) was performed within the group of participants rather than between two separate groups. This way to compare conditions may prove highly useful in natural conditions: First, all the participants are exposed to both strategies, you don’t need to chose “experimental” and “control” groups. In addition, this reduces the problem of possible differences between the groups.
- Long-term evaluation is critical! the results of this study clearly show, and not for the first time, that short-term performance is not an indication of long-term learning (as explained by prof. Robert Bjork). Long-term assessment is often harder to pursue, but it is highly valuable!
References and Resources:
Rawson, K. A., Dunlosky, J., & Sciartelli, S. M. (2013). The power of successive relearning: Improving performance on course exams and long-term retention. Educational Psychology Review, 25(4), 523-548.
In this recommended youtube lecture (51 min), prof. John Dunlosky, one of the researchers, presents this study, as well as some findings about spaced learning and retrieval practice, and some very useful classroom applications
More on the topic:
Another study proved the benefits of spaced retrieval practice in a simulated higher-education learning environment: in very short, students attended a lecture, were tested at the end of it and then again either 1 day or 8 days later. The differential timing of the review improved their performance on the reviewed materials, but also on material that was not reviewed, 35 days later.
Kapler, I. V., Weston, T., & Wiseheart, M. (2015). Spacing in a simulated undergraduate classroom: Long-term benefits for factual and higher-level learning. Learning and Instruction, 36, 38-45.