What works? The sage on the stage

Unless they’re experts, students learn more when teachers fully explain the material, write Richard E. Clark, Paul A. Kirschner and John Sweller in the new American Educator.

Discovery learning, problem-based learning, inquiry learning, constructivist learning — whatever the label, teaching that only partially guides students, and expects them to discover information on their own, is not effective or efficient. Decades of research clearly demonstrates that when teaching new information or skills, step-by-step instruction with full explanations works best.

Minimally guided instruction (“the guide on the side”) takes a great deal more time than explicit instruction (“the sage on the stage”). The  brightest and best-prepared students may “discover” what they’re supposed to, but the less-skilled students will fall even farther behind, the authors write.  “Minimally guided instruction can increase the achievement gap.”

In a second story, Principles of Instruction, Barak Rosenshine discusses “highly effective instructional practices, such as teaching new material in small amounts, modeling, asking lots of questions, providing feedback, and making time for practice and review.”

About Joanne


  1. I wish our superiors would actually take a note of that… but it will be just ignored, as usually, in favor of the latest nonsense. In our school that seem to be the internet-connected project-based learning……. Eh…

  2. I can only make the intellectually lightweight statement that “I’m shocked”.

  3. Tom Linehan says:

    In a different context, Project Followthrough came to much the same conclusion.

  4. Although I’m always skeptical of “research” about education by professors, who have never been K-12 classroom teachers, much of this is more about semantics than methods. If “Sage on the stage” means person who stands and lectures, it’s always bad for learning. Similarly, if “Guide on the side” means sit at a desk and read a newspaper, while students work in groups, this too is a poor strategy.

    Neither are meant for this sort of execution. Teaching a student a new equation will obviously require some direct instruction and modeling. Once kids know how to plug in the Xs and Ys, though, the strategy can certainly shift to more collaborative work, providing students with a choice of problems to solve. This method is far more effective than showing them a dozen problems on a white board, followed by a worksheet filled with 20 more.

    Clark, Kirschner and Sweller say that teachers who believe in “minimalist” approaches believe they are based on cognitive science. This couldn’t be more inaccurate. As a 19-year classroom teacher, one who has taught every style imaginable and worked with all kinds of students, I know without a doubt what is effective. Unless you work in a private school, with students who are browbeaten into sitting quietly and listening to one of these PHDs babble endlessly, you have to mix your methods.

    Try lecturing in my classes, and students will fall asleep faster than you can say “Sage on the stage.”

    • The article is not arguing for 40 minute lectures in elementary or high schools. It’s arguing (and documenting research that supports the argument) for direct instruction and guided practice.

      • As I said, it’s all semantics. I use discovery learning all of the time. It doesn’t mean I never teach methods or supply guided practice.

      • Ponderosa says:

        Mark, you continue to caricature lecturing as soporific and stultifying. It doesn’t have to be. How would you like me to caricature group work as “coloring and meaningless chit-chat”?

    • >> Although I’m always skeptical of
      >> “research” about education by
      >> professors, who have never been K-12
      >> classroom teachers, much of this is more
      >> about semantics than methods.

      You mean as opposed to the education professors who have taught K12? Usually, they have no research experience or skill at all, and their education knowledge is limited to the subjects and grade level at which they taught — even though they like to pretend otherwise.

      >> As a 19-year classroom teacher, one who
      >> has taught every style imaginable and
      >> worked with all kinds of students,

      It’s revealing that you don’t mention anything about subject matter, just “styles” and students.

      Let’s try a little counterexample.

      Every human society has language, and every member of every human society is an expert in their native language. And, all this happens without the benefit of any formal instruction. And, this has been going on for tens of thousands of years, if not longer. Now look at science and math, especially the physical sciences. Not every society on earth developed mathematics, and those that did had select groups of people acting as mathematicians. In the case of physical sciences it’s even worse, since it seems that no one prior to Galileo developed any accurate and comprehensive models of motion EVEN THOUGH PEOPLE HAVE WATCHED THINGS MOVE, FALL, AND BE THROWN FOR THOUSANDS OF YEARS. So what does this imply for the ability of physics students to grok kinematics by throwing a tennis ball in the hall? Yet this is the kind of nonsense that the “guide on the side” folks — schooled and experienced generally in “language arts” — demand of physics teachers.

      >> I know without a doubt what is
      >> effective.

      As does every other American self-proclaimed education expert. It funny, though, that the US isn’t exactly well known for stellar K12 education.

      >> Try lecturing in my classes, and
      >> students will fall asleep faster than
      >> you can say “Sage on the stage.”

      My sharpest students asked for MORE LECTURING.

      • I prefer lectures as a student. A teacher guided lecture/discussion open to questions and comments from the students is the great. Everyone is engaged but knowledge transfer is taking place and quick corrections can be made if there is misunderstanding.

        At some point, students have to internalize the knowledge and be able to use it to do something, and at that point, independent or collaborative work work is great. But you have to understand enough about the basics to make meaningful understanding even possible.

        The ed school trends have pushed too much discovery learning and constructivist methodology at the expense of clear exposition of the basics, IMO.

  5. Has anyone ever considered the possibility that different teachers/professors might, I don’t know, have different SKILLS and that some are good at doing engaging lecture (but not with student-directed instruction type stuff) and that others might be good at student-directed-instruction type stuff but can’t speak publicly to save their lives?

    I mean, we put a lot of effort into accommodating different LEARNING styles, why not allow for different TEACHING styles?

    • Great point Ricki,
      I love to lecture. There I said it. I bounce all around the room, stand on desks, yell, whisper, use the kids, etc. I have a 13×10 projection screen to illustrate what I say, with a slide clicker that allows me to stay on the move and have good timing. I KNOW its effective, but I rarely do it. I am a progressive educator through and through, so I repress my “sage” tendencies and run an inquiry-based or discovery-based classroom. Now, I’m experimenting with a flipped classroom. I’m always looking for new methods, but lecture is always hanging in the back of my mind. I really like the idea about focusing on teaching styles–good comment.

  6. As with all things in life, I am not sure if there is a recipe that guarantees 100% success. However, as a student and a teacher, I must admit that I am for the ‘more traditional’ style of instruction. In my opinion, students do learn more when teachers fully explain the material. Step-by-step instruction does not necessarily have to be boring – good communication and presentation skills are required. It also does not mean that elements of ‘problem-solving’ or ‘discovery’ should be excluded from the classroom either. However, I am very skeptical of lessons that are void of or lacking instructional guidance. In fact, I can think of very specific examples of ‘discovery learning’ or ‘problem-based learning’ from my own school days which were, quite sadly (and at the expense of the students), complete failures.

  7. My sharpest students asked for MORE LECTURING.

    I prefer lectures as a student.

    You guys apparently didn’t read the article closely.

    Better qualified students would actually benefit from more open-ended lessons, but they were generally too lazy to take the harder route.

    So if you want to be congratulated for preferring lecturing, you need to be a low ability student. Otherwise, you’re just someone fairly bright who likes taking the easy way out.

    Similarly, Barnes’ students prefer discovery because they are lower ability, something Barnes continually reminds everyone of, the better to vaunt his fabulouso teaching skills. But the reason low ability students prefer discovery is because it’s not too hard and it doesn’t remind them of their deficiencies.

    • Ponderosa says:

      The article said lecture was better for NOVICES –not the less-capable per se. Harvard and Yale rightly ply their bright novices with lectures.

      • Wrong. It said lectures were better for the LESS-SKILLED. And no, it’s not necessarily “rightly” to “ply” their novices with lectures. Just efficient.

        • “Efficient” is the key word. I do not want to spend the lifetime “discovering” the basics. The time for scholling is limited (for objective and subjective reasons), so why don’t we spend it as efficiently as possible?

        • Ponderosa says:

          What is a novice but someone who is less-skilled than an expert?

          • A less-skilled person is someone who doesn’t have the cognitive ability to learn the material easily. A more-skilled person is someone who just doesn’t know the material yet, but has the cognitive ability to learn it.

  8. How much of the difference of opinion here depends on the subject that we suppose is being taught?

    I present perhaps 2 hours of material (at most) in a year-long Alg. I class, five minutes’ worth of new material every two weeks. The rest of the time we spend on practice (worksheets). Hard to see how a US History class could operate that way.

    As ever: “What works?” is an empirical question to which an experiment (in social policy, federalism or a competitive market in goods and services) will provide a more accurate answer than will a State-monopoly enterprise.

    • >> How much of the difference of opinion
      >> here depends on the subject that we
      >> suppose is being taught?


      “Educators” love to make the most outrageous generalizations based on narrow experience. Take for example someone like Alfie Kohn who apparently taught existentialism at a hoity-toity prep school and then, based on that, decides that no one should assign homework. AND HE’S TAKEN SERIOUSLY.

      Portfolios make perfect sense for art subjects but are a useless distraction for things like physical science, yet “educators” push them for all subjects and all age groups.

      During my last year teaching I attended an inservice specifically for science teachers. We were supposed to learn to identify students’ “Bloom’s level” by the words they use. So, apparently, a student who’s applying Newton’s second law will be using words like ….. “apply”. This inservice for SCIENCE teachers seems to have been conceived by a an English teacher.

      People like Mark Barnes continue that tradition.

      In other professions people don’t speak to areas outside of their experience. A drummer doesn’t say “melody doesn’t matter” to a violinist, and drummers don’t train violinists. “Educators” should take note.

  9. This is the same argument Kirshner et al. been making for years now, and other researchers have already shown the flaws in their arguments:

    See also research on the value of ‘productive failure’ by Manu Kamur and others.

    ‘Unguided’ and ‘minimally guided’ instruction is a strawman. Most inquiry-guided approaches to instruction such as problem-based learning, projects, simulations, labs, and so forth require MORE guidance and effort on the part of the teacher to help students, more even than during a lecture: http://docs.lib.purdue.edu/ijpbl/vol1/iss1/4/

    Kirshner et al. hold up controlled, experimental lab studies as the ‘gold standard’, but these are students we are talking about, not mice. When you are doing a study with some freshman psych student, that has little external validity to the classroom. Some of the findings from these ‘gold standard’ lab experiments, when applied in the classroom turn out to be not strongly supported and sometimes the effects are even reversed, such as the modality effect: http://edtechdev.wordpress.com/2009/11/16/cognitive-load-theory-failure/
    This is also an old debate, and researchers in the learning sciences have already developed better and more ecologically valid research techniques, such as design-based research: http://edutechwiki.unige.ch/en/Design-based_research

    Basically Kirshner et al. seem to want to turn back the clock and ignore the last 30 years of research in the learning sciences, physics education, math education, and even psychology (e.g. research in situated and embodied cognition and learning).

    • Ponderosa says:


      A student once told me she learned more in the first two weeks of my direct-instruction history class than she had all year in a projects-based history class.