When Learning Requires Uncertainty
Systems built for standardization alone often struggle to support understanding
Researchers from the University of Virginia recently released a study: “This 100-year-old teaching method is beating modern preschools.” The main takeaway is that public Montessori preschool students are entering more prepared for kindergarten than other children and at costs significantly less per child than traditional preschools.
When I shared this with my colleague Steve at Learning Tapestry, his reaction was to doubt that this would lead to any kind of large scale increase in Montessori methods being used in preschools. Montessori methods require teachers to design a prepared environment that supports self-directed work and child agency, embracing individual timelines, uneven pacing, and emergent outcomes rather than pre-canned performance and standardized milestones. While someone could attempt to use Montessori methods within a traditional school setting, if the conditions that support those methods are not present, the benefits would likely remain unrealized.
The doubt isn’t whether Montessori “works,” but whether our traditional school systems are capable of supporting forms of learning that require ambiguity, judgment, and responsiveness.
Here are snippets of our conversation:
Steve: I think there are two ways to understand how to operate at scale: remove all ambiguity, or train on how to deal with ambiguity.
Me: I wonder if there is actually a dichotomy here or there is a third option that acknowledges when to remove ambiguity and when to manage it? I guess that’s the ideal situation. This is how education is. There are some things that need no ambiguity, but issues arise when the approach doesn’t fit the kind of thinking needed.”
Steve: Totally -- we all have to have ground rules for reducing ambiguity. We don’t need everyone to have “corner case expertise.” Some ambiguity can be considered “exceptional” and not part of initial training. So I agree. This all ties into the “testing outcomes” incentives too I think.
Me: Maybe that’s part of the inherent flaw with our system, though. We need teachers to be able to handle exceptional cases daily. I think it may also be due to our misunderstandings of learning. This post popped up today, which you may enjoy:
Me: I think that’s why you mention testing outcomes: If we define learning in a particular way, then we end up creating a bunch of outliers that teachers have to get in line. If we allow for a broader definition of learning that is more suited to personal intellectual curiosity, handling ambiguity might be more manageable.
Standardization Has a Purpose
It’s easy to suggest that we should just do away with standardization in K-12 education. But as I suggest above, I think reality is more nuanced. There is a time for standardization and a time for ambiguity. Systems that work well understand when to use both.
For example, we need attendance systems in schools. We need children to move through a cafeteria line efficiently. Some subjects have clear learning targets that are easily measurable and have very little uncertainty (e.g., letter and number recognition or math facts). We need bus schedules. Imagine letting a child choose how she might get home, choose the bus she wants to ride, share her address with a bus driver she has never met, and let the bus driver figure out how to get there even if it’s 10 miles out of the way? That’s rife with inefficiencies and potential risks of failure and even harm.
In these situations, having shared expectations for success, a standard of care, and a well thought out process that can be easily monitored and fixed should something not go according to plan are essential. Some school systems do this better than others. While the “window dressing” to actual learning, these things matter. They are part of the whole educational experience, and they make a difference in whether staff, parents, and students believe they are getting a quality education. And they do factor into learning as well. If taking attendance is disorganized or kids aren’t on time for school, learning is affected. It also speaks to confidence in management. If teachers believe things are handled well by their administrators, they will have more trust in their decisions overall.
Standardization Can Also Impede Learning
When considering complex learning like reading comprehension, I can’t and should not want to standardize meaning making. If I read a text, define a theme, and then teach my students how to regurgitate that exact meaning, that’s not true reading and it also turns out that it is really boring for kids. Developing understanding is a creative act in our brains. It requires a confluence of many complex factors:
Our proficiency with reading foundational skills;
Our knowledge of words, the world, and ourselves;
Our cognitive processing (e.g., attention spans and focus, sensory and working memory processing, retrieval and storage from long-term memory and schema development);
Our mindsets and dispositions toward reading generally, the text topic, and the style of writing;
The context in which we’re reading (e.g., what is happening in our lives at the time, who is or is not around us when we’re reading)
Our purpose for reading, or what we plan to do (if anything) with the knowledge or experience we gain from the text.
(I’ve written about this extensively. If you want more information, check out: “Reading Misunderstood,” “Comprehension Is…,” or “Stop Teaching Standards. Start Teaching Comprehension.”)
When teaching comprehension, atomizing meaning into measurable skills and trying to standardize, control, and predict outcomes actually gets in the way of comprehension. We need to allow for ambiguity and uncertainty. We want students to recognize confusion and seek understanding. We want them to persevere through frustration, knowing that with time, support, and continued attention and thought, the fog will lift and they will develop clarity and understanding. That payoff is the reward of reading.
(Related side note: We’ve come to misunderstand that payoff and substitute quick dopamine hits for slow, deliberate meaning making. I’ve also written about the drawbacks of gamification of learning. If you want more information, check out: “The Learning We Design For,” “How Learning Feels,” and “Let the Kids Play.”)
We Need a Balance
Educational systems work best when they use methods that fit their goals. This requires balancing a desire for certainty AND embracing and leaning into ambiguity.
We want schools to have processes and rules and standards for places where the results are predictable and repeatable and require efficiency. We want schools to be safe and for kids to be accounted for and on time. We want schools to be a place where children learn and grow. We want children to have knowledge and skills and we also want them to feel confident and have a high self-efficacy for learning simple, complicated, and complex ideas across all subjects.
Most school systems handle this by creating more controls and standards that can be measured: Can you answer this question with a specific fact that you memorized? Can you perform this skill in the exact way you were taught? Can you meet a predictable outcome? Anything outside of those bounds is not only difficult to handle or wrong, but often seen as a threat and rejected. As new ideas or innovations come along which push on or threaten those boundaries, if it is not rejected outright from the start, it will then be molded to fit in the system. What that means, then, is that it will no longer resemble where it started. (Read more about systems and standards in “Learning is Not a System.”)
As Bernard asks in his post: “If we use the word ‘learning’ everyday without a problem, why are there such acrimonious disagreements about what it means?”
Learning is complex and has unique characteristics. We can’t fully predict what students will understand or not. Struggle can be productive. Understanding builds over time.
If we focus on creating simplicity and control, we’re only addressing a part of learning. We say to teachers, “Your students must pass this test for you to keep your job.” The natural reaction then is to figure out how to get students pass the test. And then, if the assessment only measures a small portion of things and it is overly standardized, we end up teaching that narrow portion in an overly standardized way.
It makes sense as much as it is problematic. Teachers are trying to reduce all risks up front: if I teach this content in this way, then students will learn it, and then they will pass the test. But, when we try to eliminate all risks up front, we end up reducing productive struggle, overscaffolding, and missing opportunities for teaching what students are really interested in and excited to learn. Not only that, but some uncertainty is unavoidable, and it provides useful data.
This means that teachers must allow for more uncertainty, rather than trying to reduce it or get rid of it. This is risky: What would it mean for us to allow students to show up as they are and teach them to be even better versions of themselves? What knowledge might they need to gain? What skills? What kinds of experiences might each student need to have? What kinds of interactions will support their individual growth? And how would we know they have grown if we can’t compare them against someone else? What do we strive for with each child if we’re not striving for some common defined set of knowledge and skills?
Agile Risk Management to Support Uncertainty
To design learning systems that better handle uncertainty and ambiguity, we have to build our capacity for risk tolerance and management. Instead of planning everything in advance, we have to learn our way forward.
One way to do this is to sequence and manage uncertainty over time rather than trying to eliminate it at the start. Teachers can reduce the risks that might block access to initial learning and then stay responsive to the risks that emerge as understanding develops.
For example, if students are reading a complex text that includes challenging vocabulary and syntax and is on a obscure topic, instead of trying to predict and prevent all problems up front, identify the biggest risks early, include small tasks as formative assessments to test assumptions, and adjust based on real feedback.
Instead of pre-teaching a vocabulary list, that might look like, saying, “This text is going to have a lot of challenging words, but we’re going to work through them together. Let’s read this first section together and reflect on what we think it means. Then, when we reread it a second time, I want you to underline any words that you don’t understand.” Presumably, you will likely have a sense of what words will be challenging, and so this is a test of your assumption.
Once you have a better sense of the words that are challenging for students, identify which words to define for them and which words to teach. For the words to teach, come up with a task to teach those words, whether focused on determining meaning using context or looking at relationships. (For more information about teaching vocabulary, refer to this Vocabulary Guide from the Louisiana Department of Education.)
Managing risk over time means that you accept that new problems will emerge, students’ needs will change, and some misunderstandings won’t be visible until later. So, you check for understanding frequently and adjust instruction accordingly. This is what risk management looks like in learning. It is not prediction and control, but responsiveness informed by real evidence.
The benefits of this is that students have experiences with the text first and know you will support them through the challenge, so they are more likely to stay cognitively engaged. Additionally, you’re modeling for them how they might read a text on their own and address challenging vocabulary in the future.
Returning to the impetus for this post and my conversation with Steve: I’m claiming that traditional schools are built to remove uncertainty. When uncertainty enters into learning, it is quickly excised like a demon from an unwitting host. But, allowing for more uncertainty and ambiguity in learning is healthy. It can encourage students to develop more independence and self-efficacy to handle complex learning tasks. It more accurately reflects different facets of learning. It also helps teachers widen the scope of what they might teach, as it acknowledges that learning happens beyond the blueprint of the mandated statewide assessment. As the study showed, adopting different methods can lead to more learning, but this isn’t about Montessori versus traditional schools. It is about ensuring that whatever methods we adopt fit our goals and that our goals reflect the full breadth of learning we want our students to have.