Discovery, experiential or constructivist learning has long been promoted in teacher training as the best form of instruction for students. The following is a review of Clark, Kirschner and Sweller’s (2012) article in which they argued students learn best when they are given guided direct instruction by an expert, NOT when they attempt to ‘discover’ things by themselves.
Direct instruction fully and explicitly explains the concepts and skills that students are required to learn. This instruction can be provided via lectures, modelling, videos, computer presentations, demonstrations and class discussions, just to name a few types of lessons. For example, to teach a new type of maths problem a teacher may begin by explicitly showing and explaining to students step-by-step how to solve this particular type of problem. Further examples would continue to be provided, but gradually reducing the amount of support provided, yet providing the necessary feedback to help students move from step-to-step until they can solve that type of problem independently.
Discovery, problem-based, enquiry, experiential or constructivist learning (different names for the same idea) provide minimal guidance and students are expected to discover the information to be learned on their own. Using this type of teaching, students would receive partial instructional guidance regarding the new type of maths problem and then asked to brainstorm possible solutions in small groups, with or without hints or prompts. Through the process of trying to solve the problem and discussing various solutions students are expected to independently discover the steps necessary to solve the problem.
Clark et al. point to decades of research (from the 1980s onwards) to demonstrate that for novices (the state of most students), direct explicit instruction is more effective and efficient than expecting students to discover the information for themselves.
The research shows that when discovery learning is used in classrooms:
- Only the brightest and best prepared students make the ‘discoveries’.
- Many student become frustrated, confused and disengaged.
- Students often develop misconceptions because they incorrectly believe they have discovered the correct solution and even after being shown the correct answer are more likely to remember their own incorrect discovery.
In one study, researchers found unambiguously that compared to discovery learning, direct instruction resulted in more learning occurring, and the relatively few students who had managed to learn the concept via the discovery learning approach did not have a superior understanding of the concept compared to the direct instruction students. In addition, a review of 70 studies found that discovery learning actually increased the achievement gap between weak and strong students.
Using direct instruction, a concept can be taught in a 25 minute demonstration with discussion, followed by 15 minutes of independent practice with good teacher feedback. Learning the same concept via a discovery methodology may take several classes.
To understand why direct instruction is more effective for novice learners, you need to understand two essential components of how we learn – long-term memory and working-memory.
Long-term memory is a limitless space in which information can be stored. This information is stored in the form of concepts and procedures (similar to how you would store information in a filing cabinet).
Everything we see, hear and think about is dependent on and influenced by long-term memory. If we are skilful in a particular area, we have a lot of information we can draw on quickly and accurately. As a result, in any given situation that occurs in that area of knowledge, we are able to quickly and unconsciously recognise the situation and know what to do. Without this vast store of knowledge we would be unable, or find it difficult, to do simple activities such as avoiding traffic while crossing the street through to complex activities such as interpreting a timetable or playing games like chess.
In comparison, working-memory is quite limited and we are conscious of the information being processed. When processing uniquely new information, working-memory is very limited in duration and capacity. Almost all information stored in working-memory is lost within 30 seconds if it is not rehearsed and is limited to around 7 items.
The limitations of working-memory only applies to new information. When dealing with previously learned information, the working-memory limitations disappear because information can quickly be brought back from long-term memory as required. Therefore, the ultimate aim of instruction should be to add to our long-term memory data base so we can process information more effectively.
This means that novices (i.e., a student learning a new concept) have difficulties solving novel problems because they must do so within the constraints of their limited working-memory. In contrast, experts are more capable of solving similar problems because they can draw down on all the relevant knowledge and skills that they have stored in their long-term memory.
For example, if novices are trying to solve a problem which requires discovering the solution, the only thing they can do is blindly search for possible solution steps that might help them solve the problem. This places a large burden on their working-memory because they have to continually hold and process the current step in their potential solution while remembering previous steps and thinking about future steps and analysing the outcomes of the current step and comparing it other alternative possible steps. Consequently, the limited working-memory is over-burdened and resources are directed away from storing information in long-term memory, which is the ultimate goal.
In comparison, if novices are provided with a ‘worked example’ (for example the steps needed to solve a particular maths problem), the burden on working-memory is reduced because the solution only has to be understood and not discovered. This enables attention to be directed towards storing the key problem-solving steps in long-term memory. As more and more of the steps are stored in long-term memory, the learner can solve similar problems which have less and less information provided (partially worked examples).
Interestingly, the reverse is true for experts in that experts with a lot of knowledge, solving a problem in their area of expertise can be more effective strategy for gaining further knowledge compared to studying a worked example.
Clark et al. argue that the reason for a strong leaning towards discovery or constructivist learning in education has resulted from a confusion between constructivism as a theory of how we learn to see the world versus constructivism as a teaching methodology. Constructivism is a widely accepted theory of learning in which it is claimed that learners must construct mental representations of the world by engaging in active cognitive processing. The confusion arises because this cognitive processing can occur with or without behavioural activity and behavioural activity does not guarantee cognitive activity. In fact, the kind of active cognitive processing that students need to engage in to ‘construct’ knowledge can occur through reading a book, listening to a lecture or watching a demonstration.
Evidence from controlled, experimental studies almost uniformly support direct explicit instruction whereas there is no body of sound research demonstrating that ‘discovery learning’ is effective for anyone other than experts.
Clark, R., Kirschner, P. & Sweller, J. (2012) Putting students on the path to learning: The case for fully guided instruction. American Educator. 36 (1), 6-11: https://files.eric.ed.gov/fulltext/EJ971752.pdf