Throw them in the deep end! (but keep a life preserver handy)

What happens if you let learners try to figure it out themselves first, and only then teach them? They could learn much more deeply, according to several studies. Learn more.

Throw them in the deep end! (but keep a life preserver handy)

By Cathy Moore

As an Amazon Associate, I earn from qualifying purchases.

Child swimming in deep end“You’re setting them up to fail!” You’ve probably heard this if you’ve proposed starting with an activity instead of first providing instruction.

“Everyone knows” that people should be carefully shown how to do something and only then allowed to practice doing it. If you just throw them in the deep end, frustration and cognitive overload and squashed self-esteem will supposedly inhibit their learning.

However, several studies suggest that when we first challenge learners and then give them instruction, we can improve their ability to apply and extend their new knowledge. They could more effectively apply what they’ve learned to their jobs and to new situations.

In Scenario-Based Elearning, Ruth Clark and Richard Mayer point out this study on “productive failure,” which led me to several others.

In these and similar studies, students with some knowledge of a discipline were given a problem without first being told how to solve it. They floundered, usually in groups, and then their solutions were examined and they were taught the correct process.

These “productive failure” groups were slightly weaker at applying the new process than were the “direct instruction” groups who were first taught what to do. But the former flounderers were clearly better at applying what they learned to other situations and at developing additional models that they hadn’t been taught.

Chart comparing direct instruction with productive failure
From this summary of this study

It’s not clear how much support is best during the initial challenge. Collaboration with other learners seems to help, so in lonely, asynchronous elearning you’ll want to provide at least some scaffolding, such as hints or questions that guide learners to the correct steps to take. If I were queen, this scaffolding would be optional, it wouldn’t teach the content, and it would be provided in the activity, not as pre-activity instruction.

Like most research in instruction, these studies were done on elementary and university students, not adults in the working world. But in contrast to many studies, the researchers went beyond assessing the correct regurgitation of facts and looked at how well learners applied and extended their knowledge, which is our goal in business training.

Some papers are available as full text:

When you think about the lessons you’ve learned, which are the most memorable — the ones in which someone first taught you everything you needed to know, or the ones in which you at first floundered and even failed?

Photo by anuarsalleh


Scenario design toolkit now available

Design challenging scenarios your learners love

  • Get the insight you need from the subject matter expert
  • Create mini-scenarios and branching scenarios for any format (live or elearning)

It's not just another course!

  • Self-paced toolkit, no scheduling hassles
  • Interactive decision tools you'll use on your job
  • Far more in depth than a live course -- let's really geek out on scenarios!
  • Use it to make decisions for any project, with lifetime access



12 comments on “Throw them in the deep end! (but keep a life preserver handy)

Comments are closed.

  1. Hi Cathy –
    This is a very interesting post and it makes me think of something somewhat similar — the process of geocaching! If you’re not familiar with it, geocaching is a real life treasure hunt: people hide caches and then log the coordinates of the cache on the geocaching website, then “geocachers” use their GPS to locate the cache, sign a “visitor’s log,” and log their find on the website. The thing about geocaching is that the coordinates can only get you so close to the cache. Once you’re in the general location, you have to use your wits, common sense, intuition, keen eye sight (if you’re lucky enough to have it ;)), and problem solving skills to figure out where the cache is hidden. If you’re stuck, there is usually a single hint available to help you along the way. The process of finding a cache would be completely unrewarding if you were given precise instructions on the exact location of the cache. Equally unrewarding would be being given a larger search area in which the chances of finding the cache were discouragingly low. The gratification of the find is in having *just enough* guidance to get you close, and letting your mind do the rest. I just couldn’t help but find this paralleled with the information on training/instruction in this post. It also makes me wonder if the reward centers of our brain are stimulated more when we attempt — but don’t necessarily succeed at — solving a problem, receive further guidance, and finally “find our way.” It would be interesting to see some neuroscience research on the subject (but that’s just my area of interest :)).

    1. Emily, I very nearly became a geocacher when I lived in an area that had it! I agree that it seems like the combination of just enough information + figuring out the rest on your own would be most rewarding and addictive. If anyone can point to more research on this, please share it in the comments!

  2. Thanks Kathy,

    Another really valuable blog for IDs everywhere.

    I’ve long been a big advocate of the deep end approach to learning, because all my experience tells me it works. If you want to see it in action, give any 2 year old an iPad or look over the shoulder of any teenage or would-be teenage video gamer.
    Just last week I ran a 2 day workshop for new graduates in their first week at a leading London market insurance company. We set them loose on a business simulation, running an insurance company in a competitive market place. This is a very tough challenge as it involves coping with large amounts of data, much of it noise, unfamiliar concepts and jargon, working as a team, getting to grips with a planning model, and making coherent business decisions.
    They get very little in the way of upfront teaching, but instead learn from making mistakes and using their intelligence to interpret the results and adjust their decision making accordingly. The teams are not in direct competition – just trying to get better results than the other teams on the same challenge, but that is a sufficient element of competition to add considerably to the motivation and sense of challenge.
    Inevitably, teams run into trouble – to take one example from last week, the team who adopted an aggressive market share growth strategy quickly ran into solvency problems that in turn put serious constraints on their decision making in the next period. A powerful lesson that would have bored them to death as a PowerPoint presentation from an expert.

    The results are remarkable in terms of the quality of learning, its relevance to the real world (in this case, of insurance underwriting), and the enjoyment that the participants get from a memorable learning experience.

    Not all of those elements can be integrated into every e-learning course, but the underlying principles should be embedded in ID good practice.

  3. I have used this productive failure when teaching new doctors on how to perform minor surgical procedures. They are asked to first focus on the task of planning and cutting a cyst out in a simulation exercise. On successful completion of removing a planted stone within the simulation, they are then asked how they will close the wound. Usually there is a large wound, tissue damage and skin that will not pull together well, so they then ask for a second chance to perform the procedure. This time it is done with care, planning and increased awareness of all the necessary steps in performing this task. There is more to this task then just removing the cyst. I also ask them if they were this patient how would they feel about the surgery they had just performed and how do their patients measure their success. This short discovery shows the lights on moments and allows the doctors a greater sense of awareness than I could ever teach didactically or even giving them all the information prior to the procedure for a step by step process on how to. As all circumstances and patients are different they need to be able to plan individually for each person – and that is what I want them to learn.

  4. The example that sticks in my mind, has stuck better than the source. I thought it came from Earl Stevick ‘Teaching Languages: A Way and Ways’.

    The story’s about a mother hippopotamus teaching her new born to swim. The mother stands by and watches her calf as it struggles to the surface. If it looks like it’s not going to make it, she’ll give it a helping nudge.

    While I was looking for the story I skimmed through the chapter I thought it had come from. Stevick was writing about ‘control’ and ‘initiative’ in a ‘world of meaningful action’. Although perhaps harder to think of these in elearning terms, I’m sure they’re there.

  5. After reading this article yesterday, I was inspired to make some changes in my training this morning!! Woohoo! Isn’t that what it’s all about?!

    I did a combo of instruction and participant exploration. It worked out well because the training participants had the opportunity to explore our website without having someone directly show them what to do. I think this is essential because there are many who are still fearful with technology and this gave them an opportunity to look around and realize that they won’t break anything in our system!

    Thanks for the inspiration, and I’ve signed up for your course! I’m excited about it.

  6. Hello Cathy this was very interesting article allowing learners to fail before providing them with instruction is something new to me. I would think that people would become frustrated and confused because of the lack of training. According to this article people who “flounder” in the beginning are then shown the proper way have better results and can implement new ideas and strategies to the problems. I feel this article will be helpful in providing learning styles and techniques once I become a instructional designer. Thank you for this article.

  7. It is great to have statistics on this subject. In the NASAGA Training Activity Book (2012), I wrote up an activity that I’ve run called “The End in Mind”. In this, I’m teaching a technical training (in my case it was a sales order class for SAP), and after introductions I have my phone ring and I go on a Bob Newhart style conversation with the class listening in to inform them that we have a large rush order to create for MegaCorp in the next 30 minutes, but everyone outside the class is busy and we need to do it. The class then has to perform one of the final exercises (create a rush order for a one-time company). We debrief and see if they succeeded or failed (the classes that worked together seemed to succeed and the ones where everyone worked individually mostly failed). But then we continue with the class from the beginning, but everyone has now seen the final exercise and we debrief their experience.

  8. As a school teacher, I believe it is beneficial to most of the students to start with an activity as it sets the stage for the lesson that is going to be taught. It gives students the chance to try something new and problem solve even if they have not directly been taught the specific skill needed to complete the task. It also allows students the chance to process what they are going to be learning and relate to activities. Students can then ask questions about the material and I have found as a teacher that the students ask more probing question when they get hooked in through a do now or starting activity.
    Scenario-Based Elearning seems to allow students to first flounder but in the long term help to retain information and understand better how to apply concepts to solve a problem. They are able to connect how to approach the situation based on using skills they just learned and skills they are aware do not work to complete the project. Based on how one’s brain works, Scenario-Based learning can be beneficial because it helps with processing and with acquiring knowledge based on having stored information in their working or long term memory. It also helps with encoding since students had been exposed to the problems or concepts previously (Ormrod, 2009). The brain is complex and I believe further studies need to be done to explain how the brain processes information with scenario-based learning, looking at student’s memory as well could be helpful. These studies will also help teachers with modifications for instruction of students.

    Here is the reference to the above information about the brain and information proccessing.

    Ormrod, J. (2009). “Information Processing and Problem Solving.Laureate Education [Video webcast}. Retrieved from

  9. Hi Cathy,
    I really enjoyed this article. As a teacher, I find that it is so often that the higher ups want us to first model, give instructions and then let the students attempt. However, as a teacher and as an adult, I can think of so many times that I have learned through failure first. This can be with anything, not just in the classroom. We remember what it was like to fail and what the outcomes were of the choices we made. After the first attempt (which is not always failure) we then reevaluate and analyze the process we took to solve the problem the first time. Not always, but sometimes it is important to let exploration happen before explaining how to do something. This can create metacognitive learning, as the learner is understanding more how to solve problem through their own solutions (figuring out what works best for them). As I said, this method does not always need to be used, there are some things I do not want my students to learn through failure.

    Cori Romanowski