Carousel: A Dive Into The Evidence (Part 1: Retrieval Practice)
Recent years have seen teachers becoming more discerning when it comes to their classroom activities. Perhaps more than any time in the history of teaching, teachers are demanding to see the evidence that stands behind certain ideas and practices. Gone are the days when consultants, senior leaders or teacher trainers could expect to be able to tell teachers what “best practice” is without challenge. Teachers today have higher standards, and expect to be provided with the evidence that such practices work.
We welcome the same scrutiny at Carousel. If we want teachers to use Carousel, we need to be prepared to make an evidence-based case in our favour. Since we started, we have tried to follow the evidence as much as possible, and in this article I am going to examine some of the evidence that has guided us as we built Carousel. The two main areas of examination will be retrieval practice and homework. Entire books could be written about these areas, so I have tried to be selective without cherry-picking. I have also included a section at the end with things I am less sure about, as well as a number of reasonable evidence-based challenges to adoption of Carousel.
What is retrieval practice?
Once something has been taught, it immediately begins to fade from memory. There are a number of strategies that a student can use in order to delay this process of fading, for example by rereading notes, highlighting sections or summarising key passages. Retrieval practice involves any technique where the student actively tries to summon the information up from their long-term memory. One way to do this is to simply try and answer a question: as soon as you see the question, you try to actively summon up the answer from long term memory. This process strengthens that memory and slows its decay.
We now have lots of evidence that retrieval practice is a highly effective memory tool, especially when compared to other techniques. In one now-classic review, psychologist John Dunlosky summarised the evidence surrounding various techniques as:
How does Carousel use retrieval practice?
Retrieval practice can come in many forms, and Carousel primarily uses “self testing”, in which students first use flashcards and then do a test. The flashcards are “zero stakes”: nobody ever sees how well the student does and there are no direct ramifications for their performance. The test is slightly “higher stakes,” as teachers will be checking and evaluating student performance.
This retrieval is Carousel’s bread and butter, and the evidence behind it is very strong, with a recent review by Professor Paul Kirschner and others exploring more of the evidence. As such, we feel like Carousel’s most fundamental features are well supported by the evidence.
Improving retrieval practice: spacing
There are a number of ways to improve retrieval practice. The simplest way is by using “spaced” retrieval practice, which is where the retrieval episodes are spread out over time. For example, if you teach something on Monday, doing ten minutes of retrieval practice on Wednesday, Friday, and then Tuesday will probably lead to better long term learning than if you do thirty minutes on any one of those days.
How does Carousel use spacing?
Knowing that “spacing” is better than “not spacing” doesn’t provide teachers with a huge amount of useful information in terms of how large those spaces should be. When we set up Carousel, one of the first things we wanted to do was find a way to automatically generate spaces on a student-by-student basis to optimise their learning. However, we quickly realised that this would be an extremely complex endeavour. Any optimum spacing would be highly idiosyncratic, and depend on the:
- Content being learnt (some material is inherently harder than others);
- Amount of content being learnt (a list of 5 items might need different spacing to a list of 3 items);
- Student (their motivation, prior knowledge, ability to self-regulate and execute retrieval correctly).
Getting a “perfect” solution would have required a huge amount of work and research, and we weren’t convinced that it was actually worth it. Instead, we built in methods that might not be “perfect”, but should be “good enough.”
The main mechanisms we use are the question randomiser and the number of quizzes and questions set. If a teacher sets quizzes regularly and uses the randomiser, the odds are very good that students will be presented with spaced retrieval. The resultant spaces might not be perfect for learning, but they would certainly be good enough. Dr Pooja Agarwal’s recent large scale academic review somewhat vindicated our “good enough” approach:
We conclude that educators should implement retrieval practice, with less concern about the precise format or timing of retrieval interventions. Almost all effect sizes…indicated a positive benefit from retrieval practice under wide-ranging conditions, and retrieval practice improved student learning to a greater extent than time spent on other classroom activities (e.g., reviewing material, lectures without quizzes).
Improving retrieval practice: feedback
Though retrieval practice is effective without feedback, it is better when corrective feedback is provided. At the very least, students should receive “correct answer feedback” which shows them what the correct response was.
How does Carousel deliver feedback?
There are three main ways in which feedback is given to students via Carousel:
- Directly from Carousel’s flashcards: whilst a student is using flashcards to self-test, the flashcards can be flipped to show the correct answer.
- Directly from Carousel’s quizzes: once a student has completed a quiz, they begin marking it. Some responses are automatically marked (e.g. word for word correct answers are marked right, “I don’t know” or blank responses are marked wrong), and students will self-assess the remainder. They are shown their response and the “correct” response, and based on this are prompted to mark their answer as right or wrong.
- From their teacher: teachers have access to the students’ responses as well as analytics on their overall performance. Teachers can select individual student’s answers for discussion via the “feedback” interface, which will allow them to explore students’ ideas and deliver corrective and constructive feedback.
Challenges to Carousel’s use of retrieval practice:
Training students to self-quiz
In order for students to effectively self quiz, they need to:
- Give themselves time to try and generate an answer when using flashcards;
- Go back to flashcards they got wrong;
- Also go back to flashcards they got right, just at a later point;
- Be honest when marking.
Students do not naturally meet these criteria, and must be properly trained to do so. In time, we are going to build in tracking algorithms to present students with content they have previously faltered on, but for the minute teachers will need to make sure that their students understand how to effectively self-quiz, and why it is important that they do so.
Retrieval on longer items
A lot of the research on retrieval practice deals with short items: single words, phrases or sentences. Some of the question banks have much longer responses than others, and it isn’t yet clear if this is the best way for students to do retrieval.
Using multiple choice is an effective way to do retrieval practice in some circumstances, but Carousel does not (yet) support multiple choice. We explained why that is here.
Part 1 conclusion
With all the above taken into account, the evidence base on retrieval practice strongly indicates that the way Carousel is currently set up and used should be effective in terms of promoting long term memory retention. In part 2, we will look at the evidence base on homework and whether it supports Carousel’s primary use as a homework tool.
In Part 2, we will look at the evidence around homework and how it relates to Carousel.
☞ TO ENJOY A FREE TRIAL OF CAROUSEL GOLD OR TO PURCHASE A CAROUSEL GOLD SUBSCRIPTION, LOG IN TO YOUR ACCOUNT HERE AND GO TO THE ‘MY ACCOUNT’ SECTION.
☞ TO FIND OUT ABOUT PRICES FOR CAROUSEL PLATINUM, SEND US AN EMAIL AT HELLO@CAROUSEL-LEARNING.COM.
☞ TO LEARN MORE ABOUT CAROUSEL’S GOLD AND PLATINUM TIERS, CLICK HERE.