arrow_back Back to Research

Student Self-Explanations

Previously, I thought the main benefit for getting students to explain their thinking was for me, so I could get a better sense of their levels of understanding of a given topic or concept. However, it seems it is so much more than this. According to Chi (see first paper), the term self-explaining refers to “the activity of generating explanations to oneself, usually in the context of learning from an expository text. It is somewhat analogous to elaborating, except that the goal is to make sense of what one is reading or learning, and not merely to memorize the materials (as is often the case when subjects in laboratory experiments are asked to elaborate). In this sense, self-explaining is a knowledge-building activity that is generated by and directed to oneself”. Hence, we have something similar to the Testing Effect, whereby self-explaining can actually cause learning, and is not just an indication of it. Likewise, in the language of Cognitive Load Theory, the act of self-explaining is likely to increase cognitive load (as it is cognitively demanding), but in a way that contributes to schema acquisition, and hence may be described as Germane Load.  this section I survey the research into student self-explanations and look at practical ways we can make use of it in the classroom.

Research Paper Title: Self-explaining: The dual processes of generating inferences and repairing mental models
Author(s): Michelene T.H. Chi
My Takeaway:
Whilst not maths specific like some of the other papers in this section, this paper is an excellent starting point into the world of student self-explanations.  According to Chi, the self-explanation effect is a dual process - involving two key elements:
1) Generating inferences - this involves the learner inferring information that is missing from a text passage or an example’s solution
2) Repairing the learner’s own mental model - here it is assumed that the learner engages in the self-explanation process if he or she perceives a divergence between his or her own mental representation and the model conveyed by the text passage or example’s solution.
According to Chi  “each student may hold a naive model that may be unique in some ways, so that each student is really customizing his or her self-explanations to his or her own mental model”. Chi is careful to point out that self-explaining is different from talking to or explaining something to others, which is something I had not considered before. The focus in self-explaining is simply to understand or make sense of something, while the purpose of talking or explaining to others is to convey information to them. Talking or explaining to others adds the requirement to the learner of monitoring the listener's comprehension, which might prevent the learner from acquiring the knowledge if cognitive load becomes a problem. It is reasonable to assume that the cognitive capacity that is taxed through talking may hinder the learner from engaging in critical self-explaining behaviours. This has made me much more selective in my use of "convince the person next to you", suggesting this should only be used one students have engaged in a period of self-explanation to convince themselves. As well as providing plenty of evidence in support of the power of student self-explanations (findings which are matched by other papers in this section), this paper also confronts a questions I have long had: what if students’ self explanations are wrong? Obviously, assessing a class-full of answers alone is quicker and easier than assessing explanations, so if it is not practical to assess explanations, and if incorrect explanations are damaging to learning, then we should not bother with them. Interestingly, Chi argues that if anything incorrect student self-explanations may actually be beneficial to learning! The reason: because they are likely to provide an opportunity for cognitive conflict later in the learning process.  I am a little dubious about this, and I always strive to elicit and discuss student explanations via the process of Formative Assessment. However, it is somewhat reassuring that incorrect explanations may not be as damaging as one might think, and the benefits appear to significantly outweigh the costs.
My favourite quote:
In conclusion, self-explaining seems to be an effective domain-general learning activity. If psychologists and educators had heeded Ben Franklin’s wise remarks, we would not have had to waste our time studying learning (as measured by remembering and forgetting) in the context of telling and teaching. We should have known that we needed to focus on involvement, a form of which is self-explaining, in order to achieve learning. However, perhaps Franklin could have gone one step further, and added, “Challenge me and I understand.”

Research Paper Title: Microgenetic studies of self-explanation
Author(s): Robert Siegler
My Takeaway:
This paper gives a lovely introduction to Siegler’s Overlapping Waves Theory of learning, relating it to the importance of student self-explanation. In short, the overlapping waves theory states that individuals know and use a variety of strategies which compete with each other for use in any given situation. With improved or increased knowledge, good strategies gradually replace ineffective ones. However, for more efficient change to occur, learners must reject their ineffective strategies, which can only happen if they understand both that the procedure is wrong and why it is wrong (i.e. which problem features make the strategy inappropriate). Indeed, the author suggests one contributing factor to the high levels of maths performance of Japaneses students compared to their US and English counterparts is the emphasis on on generating explanations for why mathematical algorithms work. In Japanese classrooms, both teachers and students spend  considerable time trying to explain why solution procedures that differ
superficially generate the same answer, and why seemingly plausible approaches yield incorrect answers. Encouraging children to explain why  the procedures work appears to promote deeper understanding of them  than simply describing the procedures, providing examples of how they  work, and encouraging students to practice them — the typical approach to mathematics instruction in the US.
The paper provides comprehensive research to reach six fascinating conclusions related to the importance of self-explanation:
1) Encouragement to explain other people's statements is causally related  to learning;
2) Five-year-olds as well as older children can benefit from encouragement to explain;
3) Explaining other people's answers is more useful than explaining your own, at least when the other people's answers are consistently correct and your own answers include incorrect ones;
4) Variability of initial reasoning is positively related to learning;
5) Explaining why correct answers are correct and why incorrect answers are incorrect yields greater learning than only explaining why correct answers are correct;
6) The mechanisms through which explaining other people's reasoning exercises its effects include increasing the probability of trying to explain observed phenomena; searching more deeply for explanations when such efforts are made; increasing the accessibility of effective strategies relative to ineffective ones; and increasing the degree of engagement with the task.
My favourite quote:
One way in which encouragement to explain exercises its effects is  to increase the probability of the learner seeking an explanation at all.  When people are told that an answer is wrong, they often simply accept  the fact without thinking about why it is wrong or how they might generate correct answers in the future. The number conservation data provide evidence regarding this source of effectiveness. Children who were told that their answer was wrong and which answer was right, but who were not asked to explain why the correct answer was correct, did not increase the accuracy of their answers over the course of the four sessions. In contrast, children who received the same feedback but who also were asked to explain how the experimenter had generated the correct answer, did increase their accuracy. Further, those children who showed the largest increases in successfully explaining the experimenter's reasoning also showed the largest increases in generating correct answers on their own. Thus, encouragement to generate self-explanation seems to work partially through encouraging children to try to explain observed outcomes.
Research Paper Title: Promoting Transfer: Effects of Self-Explanation and Direct Instruction
Author(s): Bethany Rittle-Johnson
My Takeaway:
This study looked at third- through fifth-grade students (ages 8 – 11) learning about mathematical equivalence under one of four conditions varying in (a) instruction on versus invention of a procedure and (b) self-explanation versus no explanation. The results found that both self-explanation and instruction helped children learn and remember a correct procedure, and self-explanation promoted transfer regardless of instructional condition. Neither manipulation promoted greater improvements on an independent measure of conceptual knowledge. Analysing the results of their study, the authors suggest the following benefits of self-explanation:
1) Self-explanation aided invention of new problem-solving approaches. Children who self-explained in the intervention condition were more likely to invent at least one correct procedure, and those in the instruction condition were more likely to invent a second correct procedure, compared with children who did not self-explain.
2) Self-explanation broadened the range of problems to which children accurately applied correct procedures. This is crucial. When people learn new ideas, they often use them on an overly narrow range of problems. Recognising
the range of problems to which an approach applies, regardless of changes in surface features, is a critical component of learning and development
3) Self-explanation supported the adaptation of procedures to solve novel problems that did not allow rote application of the procedure. Solving some of the transfer problems required considerable insight into the rationale behind the procedure, and self-explanation supported such flexible adaptations
4) Self-explanation supported retention of correct procedures over a 2-week delay. There were no effects of test time in any analysis, indicating that the benefits of self-explanation observed immediately after the intervention were maintained on the delayed posttest
Hence, once more we have support for getting students to explain their answers, and an extra notch on rhe bed-post for direct instruction versus inquiry.
My favourite quote:
Prompts to self-explain seem to facilitate transfer equally well under conditions of invention or instruction,and these benefits persist over a delay. A growing body of research indicates that there is indeed a time for telling;invention is not necessary for children to be productive and adaptive. What may be necessary is for people to engage in effective cognitive processes, such as generating self-explanations.
Research Paper Title: An effective metacognitive strategy: learning by doing and explaining with a computer-based Cognitive Tutor
Author(s): Vincent A.W.M.M. Aleven and Kenneth R. Koedinger
My Takeaway:
This is another study which supports the view that students’ self-explanations are a key component to improving learning, specifically looking at its application for mathematics and the use of technology to support this. The authors investigated whether self-explanation can be scaffolded effectively in a classroom environment using a Cognitive Tutor, which is intelligent instructional software that supports guided learning by doing. Basically, students answer a question and then explain their problem-solving steps by selecting from a menu the name of the problem-solving principle that justifies the step. In two classroom experiments, the authors found that students who explained their steps during problem-solving practice with a Cognitive Tutor learned with greater understanding compared to students who did not explain steps. Students who were promot to give self-explanations better explained their solutions steps and were more successful on transfer problems. The authors interpret these results as follows: “by engaging in explanation, students acquired better-integrated visual and verbal declarative knowledge and acquired less shallow procedural knowledge”. Of course, this study relied on computer-based software (and indeed, reading this paper was one of the key reasons I brought the student explanation aspect into my Diagnostic Questions website), but the principle of getting students to explain each of their solution steps can be done on paper as well as it can on a screen. For those students who struggle to come up with explanations, or provide shallow ones, a key principle from this study could be borrowed - a list of relevant explanations could be placed on the bard for students to choose from. These could be topic specific, such as angle facts (“corresponding angles are equal”), or steps for solving equations (“divide both sides by…“). As well as supporting students, this would have the added advantage of making the explanations far easier to mark and assess.
My favourite quote:
A surprising finding is the fact that self-explanation is effective even when students are asked only to name the problem-solving principles that were involved in each step, but not to state the problem-solving principle or to elaborate how it applies to a problem. Perhaps equally surprising is the fact that self-explanation was scaffolded effectively by a computer tutor. Thus, other aspects of self-explanation that have been hypothesized to be crucial, such as the presence of a human instructor or the fact that students explain in natural language (as opposed to a structured computer interface) are not a necessary condition for effective support. It is an interesting question whether these aspects are even relevant. For example, will students learn more effectively when they are asked to provide more complete explanations or state explanations in their own words? This is a focus of our on-going research.
Research Paper Title: Is self-explanation worth the time? A comparison to additional practice
Author(s): Katherine L. McEldoon, Kelley L. Durkin, and Bethany RittleJohnson
My Takeaway:
We have seen the benefits of students self-explanation throughout the papers in this section, but there is no doubt that it takes longer to explain and answer a question than just to answer it. When you combine this observation with the clear benefits of regular practice, is begs the obvious question: Is self-explanation worth the time, or should we just get our students to practice more instead? This paper seeks to provide an answer. The authors compared the effectiveness of self-explanation prompts to the effectiveness of solving additional practice problems (to equate for time on task) and to solving the same number of problems (to equate for problem-solving experience). The authors found that compared to the control condition, self-explanation prompts promoted conceptual and procedural knowledge, hence once again confirming the benefits of student self-explanation. However, compared to the additional-practice condition, the benefits of self-explanation were more modest and only apparent on some subscales. The findings suggest that self-explanation prompts have some small unique learning benefits, but that greater attention needs to be paid to how much self-explanation offers advantages over alternative uses of time. So, we are left without a clear answer, so I must conclude that a balance is needed. We cannot afford to miss out on the benefits of self-explanation, but we also need to ensure students gain sufficient practice. Perhaps the answer is to focus on self-explanation in class time, where the teacher has a more important role in evaluating the students’ explanations, teasing out the meaning and correcting where needed. Whereas, depending on the nature of the problems, practice could be done at home with the students being supplied with the answers so misconceptions do not get reinforced.
My favourite quote:
The findings suggest some small, unique benefits of self-explanation relative to an alternative use of time. At the same time, it is important to consider potential benefits of this alternative - supporting additional practice, particularly on unfamiliar problems. Both activities are constructive learning activities, as they each require responses that go beyond what is provided in the original material. Both the self-explanation prompts and additional practice provided more opportunities for thinking about correct procedures (describing or implementing them) than the control condition. In turn, this should strengthen a procedure’s memory trace and related relevant knowledge, increasing the likelihood that the procedure will be selected in the future. Consequently, self-explanation prompts and additional practice can both provide opportunities for students to improve their knowledge, although there may be some benefits specific to self-explanation prompts.
Research Paper Title: Learning from Worked-Out Examples: A Study on Individual Differences
Author(s): Alexander Renkl
My Takeaway:
This is a key paper both to end our look at self-explanations and to tie this in with how we can really make the most of worked examples that will be the focus of the next section. Renkl’s research found two types of successful self-explaining strategies:
1) Anticipative Reasoners. These are learners who tended to self-explain by anticipating the next step in an example solution, then checking to determine whether the predicted step corresponded to the actual step. Crucially, these learners tend to have high levels of relevant prior knowledge.
2) Principle-based explainers. These learners tended to identify the essential meaning of a problem by attempting to articulate its goal structure—including the application of operators—while also elaborating on the principle that the example was intended to convey. Learners who adopted this strategy tended to have low prior knowledge.
However, perhaps the most important finding was that the majority of learners do not spontaneously engage in successful self-explanation strategies. This is crucial, as there is a danger in assuming that any spare working memory capacity will be used up for things that contribute towards learning (i.e. germane load, in the language of Cognitive Load Theory). If this is not the case, and students do not naturally engage in self-explanations, then their learning will not be as effective as it could be. This has huge implications for the classroom, and suggests that we should prompt students to provide self-explanations during instruction. Taking this together with Renkl’s other findings implies that these prompts should differ depending on the prior knowledge level of the students we are working with. Hence we should attempt to elicit principle-based explanations to learners with low prior knowledge while encouraging anticipative reasoning to learners with high prior knowledge. This important finding will be discussed further in the section on Making the most out of Worked Examples.
My favourite quote:
The finding that more than half of the subjects had to be assigned to the group of unsuccessful learners, reaffirms research findings that learners, left to their own devices, typically fail to show effective learning behaviors when no external support (e.g., teacher guidance or scaffolding) is present