I had a high school Social Studies teacher who would write on the whiteboard at the beginning of the semester “AQuAInt”. Acknowledge but Question Authority Intelligently. “Starting with yourself,” he would add. Clearly something I remember, but not always something that I listen to.
I am finally reaching the point where I feel an educational philosophy forming within me. The one I wrote in teachers’ college doesn’t count; it was written out of obligation, not inspiration. So, now that I am feeling it, this is the right moment to challenge it. To seek debate and discussion.
I want to be true to the tagline on my blog. I want to recognize the ideas and assumptions I hold dear, and challenge them. I want to read arguments from people with whom I disagree, to determine why I disagree or if I even disagree at all. I want to be able to say that I follow educational research, and not have that be a euphemism for just reading that with which I already agree.
So, for that purpose, I read Constructivist Instruction: Success or Failure? (follow the link for the Table of Contents). What follows here is not a review of the book in a traditional sense. I don’t think anyone will deny that it is a good book, regardless of their educational philosophical leanings. Instead, I pick out the major themes and common trends revisited over and over. I weigh their merits. And I try to see where I end up at the end of all of it, so I can better know myself as a teacher and fledgling researcher.
So, without further ado, here I go.
Define Your Terms
The book started with an article, critiquing constructivist pedagogy, which prompted several response articles, which prompted a response from the authors of the original article, which prompted a debate at some conference, which then lead to the book. All of this focused on the effectiveness of constructivism as an approach to teaching and learning.
The problem being, there is little agreement on what constructivism even means.
When I teach debating skills, one of the things I emphasize for my students is to define the terms in the resolution. This isn’t an exercise in mere dictionary definitions; by defining the terms, you define the parameters of the debate. And it is clear that much of the debate about constructivism is born from a lack of clarity on what the term even means.
Kirschner et al’s (2006) argument, which spurred the debate, framed constructivist approaches as “minimally guided”, which was challenged by several constructivists. Amount of guidance does not seem to define the distinction between constructivism and direct instruction, because it just places them on a spectrum, but there is more to distinguish the two approaches than just degree of support. So what is it?
Reading through all the different ways constructivism is defined, a few common themes emerge.
First, all learning is constructivist, because “we surely must construct mental representations of the external world that we can use to function in that world” (127). This cognitive constructivism seems to be accepted by both sides of the debate.
Second, constructivists define learning as participation. Learning occurs if there is “change in ability to interact with resources in the environment” (265) and as “individuals participate in the practices of the community” (268). It is not individual, but interactive. This participatory element also provides motivation, as “learning is stimulated by the desire to make sense of the world or to be able to participate/”do”” (358).
Third, constructivists measure learning through transfer, which holds implications for how we assess learning. We want to assess whether students can apply what they’ve learned in new situations (268) and their ability to create new knowledge (35).
Finally, constructivism is broad, and this broadness leads to a lack of precision in definition, which itself leads to constructivists complaining they have been misrepresented. Which some constructivists fully admit, that constructivism remains “too large and general a philosophy to be useful for the precise handling of the many specific ways and reasons that people learn…[and] too macroscopic a level for deriving specific instructional decisions” (36).
These definitions provide some degree of clarity. If constructivism is defined by its goals, then we can debate whether these goals are valuable or debate the best ways to reach these goals. This helps focus the debate, but at the same time, constructivism remains ill-defined enough that it continues to be a slippery target for criticism.
The other side of the debate also suffers from not being clearly defined. I say “the other side”, because the book cannot even settle on a name. Direct instruction advocates. Instructivism. Information processing perspective. For simplicity’s sake, I’ll just go with instructivism. Beyond nomenclature, instructivists also suffer from having their approach misrepresented, as did constructivism in being represented as minimally guided. Gresalfi and Lester describe an example of direct instruction in their chapter, but I don’t think it is an example of particularly good direct instruction (276). Klahr had problems with the responses to his research, with people challenging whether his direct instruction group really received direct instruction and arguing it is actually an example of guided inquiry (297-298).
The best definition provided for instructivism in the book is probably the entirety of Rosenshine’s chapter. He emphasizes that direct instruction entails systematic teaching, with an emphasis on mastery, and includes features such as review, carefully designed presentation of material, and supporting student practice.
Overall, the lack of clear definitions is holding back the quality of the debate. It also makes the next cornerstone of debating much more difficult…
One area of agreement between constructivists and instructivists is the importance of evidence. “Empirical evidence on effectiveness matters” (256); our kids deserve the best education, and research can help us decide what is best for them.
Looking at the evidence in the debate, “there is stimulating rhetoric for the constructivists position, but relatively little research supporting it” (346). The best example of this paucity of evidence is the chapter by Spiro and DeSchryver, where the “argument is not made on an empirical basis, that such approaches work better than direct instructional guidance approaches. Rather the argument is made in principle” (italics in original) (113). Yes, lots of the ideas sound nice, such as the importance of context in learning, and the goals of transfer and learning how to learn, but “without supporting research these remain merely a set of interesting hypotheses” (346). Tobias puts it quite well, in his summarizing chapter:
In comparison to constructivists, advocates for explicit instruction seem to justify their recommendations more by references to research than rhetoric. Constructivist approaches have been advocated vigorously for almost two decades now, and it is surprising to find how little research they have stimulated during that time. If constructivist instruction were evaluated by the same criterion that Hilgard (1964) applied to Gestalt psychology, the paucity of research stimulated by that paradigm should be a cause for concern for supporters of constructivist views. (346)
However, the research supporting instructivism is not watertight either. It is rather narrow in scope, focused on particular phenomenon like the worked example effect. They don’t assess transfer as much, as “what instructionists take as evidence of learning (at least in the experiments cited in the recent debate) is usually performance on a school-like task performed very shortly after initial training” (84). The research cited in the Rosenshine chapter includes studies where there is just one experimental group and a control group, rather than two experimental groups receiving two different treatments, a method criticized by some instructivists.
There are promising new developments for both constructivism and instructivism in terms of evidence. The evidence cited in the Klahr study on teaching students how to control variables is very interesting, as it focuses on assessing transfer (296). Schwartz et al. discuss ways to measure constructivist outcomes, specifically students’ ability to construct new knowledge in a domain (38).
On balance, I would say the evidence is in favour of the instructivists. It is far from perfect, but if constructivists want to support their theory, it would be better done so by finding evidence for their approach, rather than just focusing on poking holes in the evidence for instructivism.
More importantly, however, we need collaboration between researchers from different schools of thought in codesigning experiments. Most research pertaining to this debate is conducted inside of theoretical bubbles, where “most of us find evidence for our own theory when we design the treatments that represent not only our approach but also the views of people who disagree with us” (59). Several authors in the book called for collaboration on research design and implementation between researchers, to “give up the comfortable isolation of like-minded groups” (175); Duffy in particular encourages constructivists and instructivists to make more use of each other’s research methods (365).
The main barrier to this collaboration is rooted in one of the fundamental differences between constructivism and instructivism: desired outcomes. The two sides measure learning differently, and if “the goals for instruction are not the same, then the comparison is meaningless” (271). Agreement may be difficult to obtain. If we were to “find the world’s greatest constructivist instructional designers and the world’s greatest direct instructional designer and put them head-to-head in a big clinical trial,” it is likely that the “negotiation would stall on the outcomes to be measured” (58). Ultimately, choosing a measure of learning is not an evidence-based discussion, but a values-based one: “Scientific data cannot prove the principles that define the outcomes we hold most dear. Data can only help us determine how to achieve those outcomes” (51).
Theories of teaching and learning must be founded, not only on a base of evidence, but also on theories of cognition and an understanding of how the mind works. Ultimately, however you are defining learning and desired outcomes for education, they are all held in the minds of students: knowledge, skills, habits of mind, and attitudes towards learning.
The Kirschner et al (2006) paper is orientated around a theory of human cognitive architecture, particularly the importance of long-term memory, the limited capacity of working memory, and the implications of the latter in developing skills and building knowledge. Their argument for fully guided instruction is based on this model.
Jonassen, in his chapter, seeks to challenge this representation of cognitive architecture. He argues that Kirschner et al “make no attempt to articulate what is stored in long-term memory, how it gets there, or what learners do with it, except to retrieve it,” and that “[l]ong-term memory is not the only component or mechanism of cognition” (13).
First of all, this is a chapter that really needed a diagram, like the one in Mayer’s chapter (187). It was hard for me to see how the different types of knowledge described related to each other; really, it seemed they could all be categorized under explicit and implicit memory, with some spanning both. Despite Jonassen’s assertion that we should be concerned with more than long-term memory, I do not see how these types of knowledge exist outside long-term memory. I agree with Mayer than Jonassen is arguing more that Kirschner et al (2006) are incomplete in their theory of how learning works, not incorrect (28). He discusses more about what is stored in long-term memory, but does not challenge the structure of memory discussed by Kirschner et al (2006).
Parts of his assertions seem untestable as well, and I have little interest in untestable theories. For example, his claim that “different learning outcomes, especially more complex and ill-structured kinds of problem solving, call on different knowledge types which are accessed by working memory in different ways” (18). There is no explanation on how we could know that information is accessed by working memory in different ways, and it seems like something that is unprovable, considering we cannot study the workings of the mind directly, but only infer them through behaviour. He asserts at other points in the chapter that “as experimental researchers, we will never know how learning is affected” (26), and that “[s]ome kinds of learning (tacit knowledge, for instance) cannot be known in any external way” (29). Then how do we know that this kind of knowledge exists? How do we know when learning has occured? I am open to a range of research methods, but I am not open to untestable hypotheses.
Finally, Jonassen’s human cognitive architecture lacks coherence. It seems to be more of a collection of different ideas of cognition, from a range of theorists, with no sense of how all these ideas fit together. He does not describe a single architecture, and brings confusion, not clarity, to the discussion of human cognition. While Kirschner et al’s (2006) architecture may be narrower in scope, I can see how the different parts fit together.
This lack of coherence leads to an absence of clear instructional implications. Sweller, in response to Jonassen, argues that “we need distinctions and categories that have instructional implications and instructional implications are completely missing from the analysis,” and asks for evidence that “some categories of knowledge identified in the chapter should be taught using explicit, direct instruction but others should be taught using discovery learning/constructivist teaching techniques” (25). The question of instructional implications and evidence is not answered.
The instructivist perspective on human cognitive architecture is not the be-all and end-all for theories for teaching and learning. Information processing theory is just that: a theory about how the mind processes information. Other psychology theories are important, such as those of motivation (66-67). However, information processing theory does offers a clear and coherent theory of cognition, one which is supported in the field of cognitive psychology (and not just by educationalists). The alternative offered by Jonassen is just not viable.
Complexity and Structure of Domains
One question raised in the book is how the structure of a discipline or domain should inform how that subject is taught. Which I think is a brilliant question. More and more I am convinced that we need to shift away from generic teaching methods, and let the structure of the discipline inform how we approach instructional design.
Specifically, the book discussed how the extent to which a domain has a clear structure; constructivists such as Spiro argued that well-structured domains were well-suited for direct instruction, while ill-structured domains were better suited for constructivist methods. In these ill-structured domains, a focus on teacher explanations and knowledge acquisition is insufficient; students “cannot have pre-packaged prescriptions in long-term memory for how knowledge in those domains is to be applied across any reasonably large range of situations” because “circumstances for knowledge application in [ill-structured domains] are characterized by considerable variability from one instance to another, and thus pre-specifiability of the conditions for knowledge use is not possible” (107).
While the distinction between well-structured domains and ill-structured domains is interesting, a particular teaching method does not follow. Recall that this is the chapter argued, not on evidence, but “in principle” (106). Kintsch tackles the topic of reading comprehension in his chapter, which is a fairly ill-structured domain: how we come to understand a text. However, what he suggests, such as teaching sentence connectives and text structure, seem fairly explicit and teacher-led to me.
We can still explicitly teach students procedures, expose them to and discuss a range of exemplars from the discipline, and provide lots of practice, to help them manage some of the more ill-structured aspects of a domain. Clark suggests we explicitly teach students the “expert-based procedures for inventing solutions to problems in the domain” (165); these procedures can be derived from cognitive task analysis (177).
The writing process is a good example of this. For skilled writers, it is very non-linear and messy. This post is a good example of that. My outline consisted of possible headings and relevant page numbers. I wrote the sections out of order, and actually drafted this one last. I revised both as I went, and more fully once the first draft was done.
But my students are not skilled writers. I have read submissions that were entire walls of text, repeated ideas, or were so riddled with errors that it was obvious the student did not proofread. Teaching them some kind of process helps them internalize habits of planning, drafting, and revising. Just because the work of experts is messy, does not mean we can’t make it tidy for students, for now. This Twitter thread here shows some interesting discussion about how, when we have internalized these writing habits, we may not be aware that students still need to be taught them.
Teaching is another good example, and one used by Spiro and DeSchryver; it is a domain where there are not always clear prescriptions, and very often “it depends.” Yet, what I sought most after my initial training was explicit instruction and procedures, such as from Teach Like a Champion, to help me get by until I gained the “accumulation of considerable experience, the exposure to many examples, and an appreciation for multiple interacting contextual features” (108-109).
For ill-structured domains, explicit instruction can build a foundation of skills and knowledge, so students don’t flounder when left to grapple with ill-structured problems later. Experience with those ill-structured problems is still valuable, but the timing must be right. Which is to say, instructional design should consider the…
Of the many new ideas I’ve learned in my reading on educational theory, one of the most powerful I’ve found is the novice-expert continuum. Simply put, novices and experts learn differently. Novices need more structure, to have complex ideas broken down, content sequenced and skills practiced. Experts have extensive amounts of domain-specific knowledge in long-term memory and have practiced key skills to the point of automaticity; what they need is freedom to explore and grapple with problems and questions. (See Adam Boxer’s post for more explanation.)
The novice-expert continuum is revisited at various points in the book, mostly by instructivists. First, there is the attempt to define expertise, and what distinguishes an expert from a novice. For example, there is the extent of the expert’s knowledge: having “an enormous amount of information in long-term memory” allows experts to “recognize a large number of problem states and the problem-solving moves appropriate to those states” (131). As well, this knowledge is retrieved effortlessly and organized into meaningful schema (147-148). Kintsch discusses particularly the role of expertise for reading within a given domain. Experts “have available retrieval structures that link the information in working memory…automatically with relevant information in their long-term memory,” and so “a situation model [of the text] is formed, largely without conscious effort” (226).
Of course, therein lies the question of how to best develop expertise. The point I think Jonassen was trying to make in his chapter was that some of the types of knowledge used by experts cannot be taught and can only be developed through exploration of problems, but it is an assertion made in the absence of evidence.
There is nothing to suggest that novices can develop expertise by participating in unstructured exploration of a problem that is the hallmark of working within a domain. In other words “how to learn or be taught in a domain is quite different from how to perform or “do” in a domain ” (151).
The second question raised is at what point we become experts. The answer is, not during our K-12 schooling. Students simply do not accumulate the practice they need in a given subject to reach the depth of knowledge required to claim expertise. However, they are given a breadth of knowledge and skills to pursue expertise in one of those domains with further training (231).
My discussion here of the novice-expert continuum certainly draws more on arguments from instructivists, which is unsurprising; it supports the assertion that novices need the kind of guidance and support promoted by instructivism. This view is challenged by constructivists, arguing that it does not matter the form of the instruction, experts will always learn better in their domain than novices. If the style of instruction is a lecture, an expert audience would have the prior knowledge needed to make sense of the material and to ask clarifying questions afterwards. Experts manage constructivist-style activities better as well. Thus, the main effect is “experts are better than novices when learning in their domain…[which] is likely to swamp any of the subtle differences in the relative benefits of one type of instruction over another for novices versus experts” (57).
This assertion, that it doesn’t matter which form of instruction is used for novices versus experts, those who have a strong knowledge base and those who don’t, is not supported by the evidence. Tobias sums this up nicely, in his overview of the debate:
Thus, converging evidence from different fields such as text processing, ATI research, work using programmed instructional materials, and cognitive-flexibility theory (Spiro, this volume) suggest a general hypothesis that constructivist materials may be differentially effective for knowledgeable students, whereas explicit instruction may be more beneficial for their less knowledgeable counterparts. That hypothesis is also supported by the unchallenged assumption that constructivist instruction requires more working memory from those with little domain knowledge than it does for more knowledgeable students. (342).
Ultimately, the continuum holds implications for when we use instructivist and constructivist approaches, depending the the prior knowledge and skills of our students. Movement along the continuum can be seen in a single lesson, as we shift from guided to independent practice, as suggested by Rosenshine. It can be seen over the course of unit, where we build up to a culminating, open-ended task or project. It can be seen over the course of a degree in university, where we start with lecture-based courses and memoization, and end with leading our own research project.
The novice-expert continuum shows us that, when we ask whether we should use constructivist or instructivist approaches, the answer is…
Not Either-Or, But Both-And
No one in the book is arguing entirely one approach or the other. It is an indefensible position. The question being asked is not which is better, but “under which circumstances, and for whom, is one kind of instruction superior to the other” (xi).
Both sides bring good ideas to teaching. For the instructivists, they help take away the stigma from teacher-led instruction. Even the most dedicated constructivists use explicit instruction at times, but when teachers are made to feel ashamed for talking too much, we do it poorly. By talking about how teacher-led instruction can be beneficial, we can also talk about how to do it well.
For constructivism, it helps us set up a purpose for learning, beyond acing the exam. Interesting problems or projects give students a reason for the knowledge and skills they are learning, and “learners will construct a different understanding if they are given an explanation in isolation, versus first having an experience that gives them a “need to know”” (91). It sets up a context for learning.
Ultimately, our decisions about instructional design should be shaped by the situation we are in, rather than our allegiance towards a particular philosophy. Our approach should depend on “the nature of the material, the background of the learner, as well as the stage of learning” (235). Questions of the structure of the material, necessary skills and knowledge, and academic readiness of the students should be discussed, to decide the best path. If we accept Spiro and DeSchryver’s argument that professions like teaching are ill-structured domains, then we cannot apply philosophies like constructivism or instructivism wholesale, without adaptation or integration of other theories; “generalizable principles have to be combined and tailored in the context of highly variable cases of application in the real world” (110).
Fletcher captures this mixing of approaches, in his chapter on training pilots:
The instruction we now provide to produce combat pilots uses a bottom-up process, providing drill and practice for the discrete knowledge and skills required. This process is complemented with top-down, simulation-based learning that situates the learner in approximations of the ultimate performance environment. (252)
He also highlights how “there is an interaction between knowledge of the subject matter and the value of simulation alone” and “that the less the student knows about the subject matter, the greater the need for tutorial guidance in simulation” (253). So, we can see here an analysis of the:
- desired learning outcomes, found in the structure of the domain: the knowledge, skills, and experiences needed to become a flight pilot
- a selection of instructional strategies that best support those outcomes
- and a consideration of the academic readiness of the learner, or their position on the novice-expert continuum, which informs the amount of guidance and scaffolding needed.
The are other considerations to make in instructional design, particularly the practicality of a given approach. Klahr suggests that constructivist approaches depend “much more heavily on the skill, knowledge, and pedagogical acumen of the teacher than does direct instruction” (303). Tobias raises the concern that if “an instructional method requires unusually able and/or knowledgeable teachers, questions can be raised about how effective it may be in schools where most of the teachers are likely to cluster around the mean in terms of teaching ability” (345). Duffy asks whether we are “simply designing an ideal but unrealistic approach to instruction…[and] whether the prescriptions are great in principle but a failure in general practice” (364).
It may be that teachers need to use an balance of the two approaches, simply because it is unfeasible to be constructivist all the time. I am fond of Tom Sherrington’s 80/20 balance, with simple, teacher-led lessons making up the majority, building knowledge and skills, with a role still available for ‘wow’ lessons to inspire curiousity and create a purpose for learning (but which take more time to plan). With the research evidence (currently) leaning towards instructivists approaches, this seems to be a fair and effective balance.
Debating the Merits of the Debate
I certainly think it has its benefits. My teaching training had a strong constructivist leaning, but this was never made explicit, and I was completely unaware I was receiving guidance on teaching and learning that was filtered through a particular theoretical lense. Being introduced to the debate through blogs and Twitter showed me that theories of teaching and learning were not “settled” but subject to contention and debate. I had some beliefs challenged, such as those on group work, some demolished, like learning styles, and new ones formed, like the importance of a knowledge-rich curriculum.
At the same time, there are limitations to the debate. Many teachers pull from both constructivist and instructivist practices, and resent being asked to sit on one side of the fence or the other; they prefer to select strategies based on the needs of their students, rather than identifying wholly with one side of the argument.
Kintsch’s chapter serves, unintentionally, as a strong challenge to the value of the debate. He skillfully demonstrates how we construct our understanding of the text, and describes strategies that support this construction of meaning. There is nothing strongly constructivist or instructivist about his analysis; rather, he carefully defines what it means to obtain an outcome (reading comprehension) and discusses ways to reach that goal. So, instead of debating broad educational theories, maybe we should just focus on determining which methods best help us reach our goals. As stated by Mayer:
Overall, the search of “schools of learning” has been an unproductive approach for the science of learning. In my opinion, our field would be better served by trying to figure out research-based answers to how learning and instruction work rather than by engaging in high-level philosophical arguments about which “ism” is best. (197)
However, framing the discussion as a debate helps us to see the philosophical roots of our beliefs on education to identify the sources of our disagreement. It helps us remember the ‘why’ behind the ‘what’, the reasoning behind the different approaches we take.
So, is the debate good? Yes. And it is best where the two sides meet, like two rivers. The skills and knowledge developed through direct instruction can serve as the roots and support for constructivist exploratory learning. The goal of the debate, then, is not to “fashion an easy answer but to hold the competing ideologies together in an awkward contradictory balance.”
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why Minimal Guidance During Instruction Does Not Work: An Analysis of the Failure of Constructivist, Discovery, Problem-Based, Experiential, and Inquiry-Based Teaching. Educational Psychologist, 41(2), 75-86. doi:10.1207/s15326985ep4102_1
Tobias, S., & Duffy, T. M. (Eds.). (2010). Constructivist instruction: Success or failure? New York: Routledge Taylor & Francis Group.