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

An opinion / review paper arguing against Athlete musician scale framing origin timeline, Inquiry-based learning, Situated learning, and other methods which emphasize students solving authentic problems in authentic settings, often using Cognitive scaffolding-laden versions of the authentic disciplinary methods.

Our understanding of expert skill acquisition (e.g. Sweller, J., & Cooper, G. A. (1985). The Use of Worked Examples as a Substitute for Problem Solving in Learning Algebra. Cognition and Instruction, 2(1), 59–89) suggests that expertise can be largely or entirely explained in terms of differences in long-term memory. So, instructional recommendations should be keenly interested in how they’ll change LTM.
We have limited working memory, but those limits don’t really apply to long-term memory (see Ericsson, K. A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102(2), 211–245); when relying on working memory to solve a problem, we can’t actually learn; see author John Sweller’s articles e.g. Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285.
Citing a 2004 review from Richard E. Mayer, it seems that empirical evidence repeatedly demonstrates that unguided approaches are ineffective. By contrast, students—particularly domain-novices—benefit from being “explicitly shown what to do and how to do it.“ (Worked example effect). Problem solving is likely to be effective only when learners are sufficiently experienced.
- In one experiment (Clark, 1982), less-skilled learners tend to prefer less-guided approaches, but they learn less from them.
- In a medical setting, it seems that problem-based learning produces mixed results… but the reply from Schmidt et al suggests that this is a selective reading.

I’m quite sympathetic to constructivist framings (How can I reconcile my views on inquiry learning and memory systems?), but I think the authors are correct in their central criticism that advocates systematically underrate the role of memory (Many people view memory as unimportant to deep creative work) and of the mechanisms of cognition. On the other hand, I accuse the authors of systematically ignoring the role of motivation and meaning (Enabling environments’ activities directly serve an intrinsically meaningful purpose).

Q. Three main theoretical criticisms of constructivism and its peers?
A. 1) expertise relies on LTM, while constructivism mostly ignores it; 2) low-guidance tasks overload WM (per CLT); 3) epistemology != pedagogy (teaching of an inquiry-driven discipline isn’t necessarily best done via inquiry)

Q. My primary criticism of KSC?
A. They’re totally disinterested in motivation, attitude, feeling, meaning. (see also Kuhn’s response)

Commentaries

In 2007’s volume 42 issue 2, there are three commentaries in reply, and then a follow-up from the original trio responding to the commentaries (Schmidt, H. G., Loyens, S. M. M., Van Gog, T., & Paas, F. (2007). Problem-Based Learning is Compatible with Human Cognitive Architecture: Commentary on Kirschner, Sweller, and Clark (2006). Educational Psychologist, 42(2), 91–97).

Schmidt, H. G., Loyens, S. M. M., Van Gog, T., & Paas, F. (2007). Problem-Based Learning is Compatible with Human Cognitive Architecture: Commentary on Kirschner, Sweller, and Clark (2006). Educational Psychologist, 42(2), 91–97

These authors are sympathetic to KSC’s emphasis on cognitive architecture, but they insist that problem-based learning isn’t minimally guided. It’s often quite guided: the selection of the problem, the discussion structure, the feedback from a tutor, etc.
- KSC reply that these structures are fine, but still inadequate for complex material. If you’re going to accede that scaffolding is important, why deny the value of particularly strong scaffolding like worked examples? Besides, “requiring that a learner discover a problem solution always reduces guidance compared to presenting the solution.” What exactly is the point of imposing this heavy extra load?
They cite a few empirical experiments in support of this notion.
- KSC reply that these experiments don’t directly compare higher and lower levels of guidance.
They argue that these discussion-based methods lower cognitive load.
- KSC reply that “it will at best increase germane load with a concomitant (though not necessarily equal) decrease in extraneous load.” The implication is that the total load is still too high.
They point out that PBL emphasizes transfer and flexibility, rather than direct application of knowledge.
- It’s true that in e.g. Sweller, J., & Cooper, G. A. (1985). The Use of Worked Examples as a Substitute for Problem Solving in Learning Algebra. Cognition and Instruction, 2(1), 59–89, the Worked example effect applies only to identical test problems.
- KSC reply (in their response to Hmelo-Silver et al) that it’s not clear what flexibility means, mechanically, and that its advocates have failed to advance a theory of how their methods induce this kind of cognitive flexibility.

Hmelo-Silver, C. E., Duncan, R. G., & Chinn, C. A. (2007). Scaffolding and Achievement in Problem-Based and Inquiry Learning: A Response to Kirschner, Sweller, and Clark (2006). Educational Psychologist, 42(2), 99–107

Like Schmidt et al, they argue that PBL / IL aren’t minimally guided. Scaffolding is provided, but sometimes only when students need to know.
- KSC’s reply is similar to above: scaffolding is good, but why not consider “full” scaffolding?
They point out that PBL and IL are designed to emphasize “disciplinary epistemologies and investigative strategies” (in addition to content knowledge)
- KSC also accuse Hmelo-Silver et al for failing to ground their preferred methods in any explicit cognitive architecture, and for emphasizing general problem-solving skills which don’t appear to exist.
- “The techniques favored by Hmelo-Silver et al. (2007) were developed in an era when it was thought that the central component of human cognition was not knowledge in long-term memory, but rather the ability to devise novel, general problem-solving and thinking strategies. IL was intended to foster this skill. The failure over many decades to isolate a single, novel, teachable, general problem-solving or thinking strategy has inhibited that pursuit.”
- KSC argue that it’s not clear we can teach such strategies directly; or at least, H-S offers no evidence that we can.
The offer many examples of empirical studies in which PBL and IL demonstrate learning benefits.
- KSC retort that these studies aren’t well-controlled enough; they’re comparing totally different curricula with different students in different places. “We might expect that meta-analyses of such studies would yield mixed, ambiguous results, precisely the findings of the meta-analyses cited by Hmelo-Silver et al.”

Kuhn, D. (2007). Is Direct Instruction an Answer to the Right Question? Educational Psychologist, 42(2), 109–113.

The author claims that KSC are asking the wrong question. The real challenge is figuring out what to learn, not how to learn it.
- KSC reject this premise; irrespective of the answer to “what to learn?” we’ll still need good answers to “how?”
The author says that we should emphasize “the skills of knowledge acquisition”
- KSC say that it’s not clear what those skills are, how to teach them, or how they might relate to cognitive architecture.
The author suggests that content knowledge in science is less important than attitudes and values. “Are scientific topics worth learning about, knowing about, or inquiring about any more deeply? And, as the motivation theorists highlight, another critical question: Am I someone who is competent to engage in such learning?”
- KSC reject this out of hand as, basically, a difference of opinion.
- I think this objection is actually quite important, though I view it as complementary to KSC’s concerns. Reminds me a lot of Dan Meyer’s arguments.
Most compellingly, the author emphasizes motivation.
- “Activities centered on inquiry and argument enable students to appreciate the power and utility of these skills as they practice them.”
- “…key to predicting students’ performance is examining what it is that they are undertaking to learn in school, what they think it means, how they construe the meaning of this material in relation to themselves, and whether they can see it as worth learning”
- “…students need to learn what it is scientists do and why they bother to do it. Students can develop that understanding only by engaging…in the practice of science.”
- KSC see this as irrelevant. I see it as central and quite appropriate.
Kuhn cites a bunch of empirical papers and KSC ignore them; I didn’t dig into them.

For another very interesting later response, see Schwartz, D. L., Lindgren, R., & Lewis, S. (2009). Constructivism in an age of non-constructivist assessments. In Constructivist instruction: Success or failure? (pp. 34–61). Routledge\/Taylor & Francis Group.

For another, more balanced review, see Lee, H. S., & Anderson, J. R. (2013). Student Learning: What Has Instruction Got to Do With It? Annual Review of Psychology, 64(1), 445–469

Last updated 2024-04-22.