Hands-Off Learning? The Evidence Against Minimally Guided Instruction

Native speaker / non-native speaker. Implicit / explicit. Accuracy / fluency. Dichotomies abound. “False” is a common collocation with “dichotomy,” and this is often because nothing is either black or white but rather falls on a gray-scale continuum between the two ends. Yet, dichotomies are useful for discussion as they represent the crux of a concept. Stepping a bit back from ELT and looking more broadly at pedagogical theory, this research bite will look at Kirschner and colleague’s (2006) exploration of the evidence against minimally guided instruction and in favor of guided instruction, especially in terms of human cognition.

What is minimally guided instruction?

You might know it as discovery learning, problem-based learning, inquiry learning, or experiential learning. It is based on constructivist theory of learning and the assumption that learning occurs when students make their own solutions to “authentic” problems in an information-rich environment. In these situations, guidance on how to complete the task is minimal, if it exists at all.

What is guided instruction?

Guided instruction is direct instruction, modelling, practice, working with examples, and scaffolding that presents relevant information rather than making learners discover it. It is more traditional in the sense that it is less hands-off, though it should not be confused with what we typically believe to be “traditional” teaching.

So, what’s the problem?

According to the authors (p. 77):

The aim of all instruction is to alter long-term memory. If nothing has changed in long-term memory, nothing has been learned.

Much of what we know about our cognitive structure, in particular about how working memory and long-term memory work, shows that long-term memory is the foundation of learning. Moving knowledge from working to long-term memory requires efficient cognitive processing and being within the limits of working memory. Interestingly, research shows that while new information is limited to 4-7 items in working memory, the limits disappear when working with familiar information.

The authors again (p. 77)

Any instructional theory that ignores the limits of working memory when dealing with novel information or ignores the disappearance of those limits when dealing with familiar information is unlikely to be effective.

Empirical studies spanning decades show that minimally guided instruction (when the learners are novices) requires a large cognitive load and, therefore, is not supported by the research on how we learn effectively and efficiently. Solving a problem, specifically “problem-based searching” places a large burden on our working memory, especially by splitting learners’ attention, and therefore takes up valuable resources that are needed for actually learning. It’s possible to search and work on a problem for quite some time without learning a thing. It seems that only those who have extensive experience, schema, and prior knowledge benefit from this type of activity.

Where’s the evidence?

The authors cite a number of studies, qualitative and quantitative, that show the minimal or negative effects of minimally guided instruction across a range of disciplines.

  • In science classes, students have been observed to become lost and frustrated (Brown and Campione, 1994)
  • There are more false starts in this type of instruction and therefore it is less efficient (Carlson et al., 1992)
  • Students in a science class learning under guided instruction could better transfer their learning and the quality of learning was superior (Klahr and Nigam, 2004)
  • Guided learning produced greater recall and longer term transfer (Sweller, 1999)
  • In an algebra class, students learned more studying already solved examples rather than simply solving the problems (Cooper and Sweller, 1987)
  • A meta-analysis from Cronbach and Snow (1977 – yes, quite dated) showed that direct guidance was more beneficial for lower-level learners (something still supported today, Kylllonen and Lajoie, 2003)
  • Another meta-analysis (Clark, 1989) found that “lower-aptitude” students had lower scores after minimally guided instruction, even if they prefer the experience more.
  • A meta-analysis of problem-based learning in medical schools (Albanese and Mitchell, 1993) found many negative effects, including lower scores, no differences, and increased study hours.

What’s the basic takeaway?

Instruction must be rooted in what we know about how we actually learn. We must consider cognitive load (e.g. working memory limits) and how knowledge moves from working to long-term memory: “cognitive load theory suggests that the free exploration of a highly complex environment may generate a heavy working memory load that is detrimental to
learning. This suggestion is particularly important in the case of novice learners” (p.80)

In terms of problem solving, “knowledge organization and schema acquisition are more important for the development of expertise than the use of particular methods” (p. 83). Clearly, the evidence is in support of direct instruction.

How is this related to language teaching?

Minimally guided instruction often appears in language learning as implicit or inductive language instruction, and findings similar to Kirschner’s have been reported: explicit grammar instruction is often superior to implicit instruction (please check out the interesting discussion in the comments). Problem-based learning and similar approaches also appear in ELT. It is important to be aware of the evidence regarding this category of instruction, as it involves all learning and is not necessarily skill specific.


Albanese, M., & Mitchell, S. (1993). Problem-based learning: A review of the literature on its outcomes and implementation issues. Academic Medicine, 68, 52–81.

Brown, A., & Campione, J. (1994). Guided discovery in a community of learners. In K. McGilly (Ed.), Classroom lessons: Integrating cognitive theory and classroom practice (pp. 229–270). Cambridge, MA: MIT Press.

Carlson, R. A., Lundy,D.H.,&Schneider,W. (1992). Strategy guidance and memory aiding in learning a problem-solving skill. Human Factors, 34, 129–145.

Clark, R. E. (1989). When teaching kills learning: Research on mathemathantics. In H. N. Mandl, N. Bennett, E. de Corte, & H. F.
Freidrich (Eds.), Learning and instruction: European research in an international context (Vol. 2, pp. 1–22). London: Pergamon.

Cooper, G., & Sweller, J. (1987). The effects of schema acquisition and rule automation on mathematical problem-solving transfer. Journal of Educational Psychology, 79, 347–362.

Cronbach, L. J., & Snow, R. E. (1977). Aptitudes and instructional methods: A handbook for research on interactions. New York: Irvington.

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.

Klahr, D., & Nigam, M. (2004). The equivalence of learning paths in early science instruction: Effects of direct instruction and discovery learning.
Psychological Science, 15, 661–667.

Kyllonen, P. C.,&Lajoie, S. P. (2003). Reassessing aptitude: Introduction to a special issue in honor of Richard E. Snow. Educational Psychologist,
38, 79–83

Sweller, J. (1999). Instructional design in technical areas. Camberwell, Australia: ACER Press.

Featured image by HelloImNik

Anthony Schmidt on TwitterAnthony Schmidt on Wordpress
Anthony Schmidt
English language Instructor at University of Tennessee, Knoxville
Anthony Schmidt is editor of ELT Research Bites. He also has his own blog at anthonyteacher.com. Offline, he is a full-time English language instructor in a university IEP program. He is interested in all aspects of applied linguistics, in particular English for Academic Purposes.

7 thoughts on “Hands-Off Learning? The Evidence Against Minimally Guided Instruction”

  1. 1. “Empirical studies spanning decades show that minimally guided instruction (MGI) (when the learners are novices) requires a large cognitive load and, therefore, is not supported by the research on how we learn effectively and efficiently.” How about the evidence of how 99% of the world’s population have learned their L1 over the past 3,000 years?

    2. The evidence of the negative effects of MGI found in the works cited doesn’t use a commonly-agreed definition of MGI.

    3. MGI in instructed SLA does not, of course, = implicit learning, and doesn’t seem to me to be a helpful way of looking at teaching practice, since even Krashen agrees that scaffolding and explicit instruction have their places.

    1. Thanks for commenting Geoff!

      1. Minimally-guided instruction is clearly about classroom instruction – specifically guided discovery, inquiry-based learning, etc., not natural learning situations such as learning an L1, or even learning to walk. Clearly, we can learn our L1s without much effort, but where classroom instruction is concerned, it does require some effort and explicitness, especially later in life and at lower levels of proficiency. The current study focused heavily on studies that look at math and science instruction. This bite is only *suggesting* that it may be *generalizable* to all types of instruction, including language learning, as working memory is an underlying facet in all areas of learning. In addition, such learning practices as discovery and problem-based learning also exist with ELT.

      2. As the article states, MGI has been called many things, so of course, there may not be an agreed-upon definition. As the authors define it, MGI is “rather than being presented with essential in-formation, must discover or construct essential information for themselves” and included research that fell into these parameters. Keep in mind, this was a literature review, not a meta-analysis, for which an operational definition would have been established.

      3. It’s helpful if there is an overemphasis on these MGI approaches in one’s ELT program or teaching. The takeway and practical application is that, for novice learners, it may be better to do more explicit work than the inductive teaching activities.

      1. hi
        if i recall +guided+ instruction is based on the assumption that the subject matter (i.e. subjects such as maths, physics etc.) can be broken down and then taught in a direct way?

        if so the notion of breaking down language (as a system) and teaching it, is, from what most informed SLA voices agree, not what we want to do if we want to use language ultimately in a spontaneous way; which is point 3 that Geoff makes above

        although certainly we need to approach novice and less novice language learners differently i don’t think the terms used by Sweller and chums field (cog sci?) can straightforwardly be used with terms from SLA such as explict/implicit teaching, deductive/inductive activities

        i assume seeing learning as a change in LTM would put some of their assumptions in the skill acquisition camp in language learning? hence be open to the various criticisims that skill aquisition in lang learning faces?

        1. Mura,

          I think you are right that this article, since it is based on cognitive science, would fall into the skill acquisition camp – there is a reliance on WMC, LTM, and for language learning, a focus on the process of declarative to procedural knowledge to automaticity, though this article doesn’t address that aspect.

          I don’t think MGI is saying to teach a subject piecemeal, and I agree that SLA research would certainly be against this. I think what can be suggested from the article is that, when it is time to teach a grammar point, a writing technique, or some vocabulary, for example, it would be better to provide direct instruction for lower-proficiency students rather than put too much emphasis on a implicit or inductive methodology for the learning point.

          Finally, in my own view, I do not necessarily separate language learning and other subjects, at least not always. Learning math, martial arts, critical thinking, or a language requires similar underlying mechanisms and the ultimate goal is automaticity – spontaneous production. So, I see an article such as this, as informative to all learning (under certain conditions) not just language learning.

  2. A convincing retort to constructivism and problem-based learning! Thanks for sharing.

    There’s surely a middle way that includes elements from both directed and problem-based learning? Particularly if the student can freely choose a problem domain that personally interests and motivates them? I take the point that less-experienced students will benefit less in this situation.

    Also, taking a global overview of the purpose of education, long-term, maybe it’s a good idea for students to experience the uncertainty, inefficiency, and discomfort of flailing in an unknown problem space (even if it is less efficient on a material_learned/hour metric?) Might stand them in good stead in the future?

    (No curriculum, however well conceived, can be entirely comprehensive…)

    1. Hi! Thanks for reading and commenting.
      I’d like to note that the findings of the study hold true for “novices,” so the “experience the uncertainty, inefficiency, and discomfort” may be better saved for those who have learned the basics!

Leave a Reply