C&T 5004 Raghu Betina

Blog #3

What is your theory of learning? What learning goals/purposes do you care about, and how can they be achieved?

This reflection includes the basic elements of a theory of action:

  1. 2–3 specific goals
  2. A small number of activities/strategies that will help participants achieve those goals
  3. A small number of specific short-term outcomes that can serve as the basis for assessments that make progress visible
  4. A small number of related long-term outcomes ("stretch goals") worth striving for but unlikely to be achieved without additional time and support

Approached as a sketch or draft rather than a final representation.

In Blog #1 I described problems I keep running into teaching programming to beginners, and sketched a redesign with a main campaign and optional side quests. In Blog #2 I looked at what would make that hard to put in place. This post is my attempt to state the theory of action behind both, and to use the frameworks from class to find its weak points.

My theory of learning

Programming is sequential. Nearly every unit assumes the ones before it, so a student who half-learns arrays finds everything that uses arrays harder. I think the same is true of most technical subjects. It was true for me in math, where I had to keep to a schedule set for students who came in far more prepared than I was, and each topic I only half understood made the next one harder.

Students arrive with very different preparedness, whether from prior study, help at home, or luck. In my course, students with some background finish the early assignments in one or two hours, and true beginners can need six to ten. The beginners who end up with an A tend to go back over the early material on their own time and ask for help. That looks to me like the strategy that works, and until this term my course left it for each student to discover.

Weeks are uneven, too. My students have many other demands on their time, so a bad week is normal, and in a course paced to a calendar a bad week can leave a gap the student never closes. Gaps may also compound, each one making the next unit harder to learn. The closest evidence I found is about children learning to read, and it’s mixed. A review of twenty-five years of that research found no single pattern overall, but in studies with precise measures, the gap between strong and weak readers widened or held steady far more often than it narrowed. Most of the apparent narrowing came from imprecise tests (Pfost et al., 2014).

Students also spend their limited hours in a predictable order: on what’s required, to a lesser extent on what earns credit, and almost never on anything else. That holds even for students who love the material. When advanced students asked me for more to work on and I offered extra topics as optional assignments, very few did them, and they told me they had a thousand and one demands on their time, so anything optional fell to the bottom of the list. When I offered beginners extra practice the same way, very few did that either, including the ones who needed it most.

The standard course as a theory of action

Most courses, including earlier versions of mine, run on something like this:

If every student covers the same material in the same order on the same schedule,

and the class moves on when the schedule does,

then the students who are good at this will master it, the rest won’t, and the grades will show which is which.

I don’t think many teachers would say some share of their students can’t learn the material, but the design acts as if that were true, by treating whatever spread comes out as a fact about the students. In the terms we used in class, it’s a design theory nobody espouses.

The amount of material in a course like this is calibrated to some student, the median or the top. Everyone below that point gets moved on before they’ve mastered what came before, and the gaps stay with them into later units. Everyone above it finishes early, and nothing else they do counts.

Reversing each assumption gives roughly the changes I’ve made:

The standard course assumes What I changed
Everyone should cover the same material A shorter required path, sized to what every student can master in the hours they have, with the rest elective
Everyone should move at the same pace Deadlines slower than usual, and unlimited resubmission, with late penalties only on points earned after the deadline
Extra practice is the student’s own business Supplemental practice on earlier material that earns credit
Anything beyond the required work is optional Electives that earn credit, including building your own app idea

Every student still follows the same required path, and what varies is what else they choose to do. That’s narrower than what I described in Blog #1, where students would “choose their own path through the content.”

A least-prepared student under both regimes Under the traditional regime the student reaches all four units but masters only the first, with the rest falling progressively below the mastery bar. Under the proposed regime the same hours buy the three required units at full mastery, plus supplemental practice, and the elective fourth unit goes unreached. Traditional mastery bar A B C D Four units reached, three of them short of the bar This design elective from here mastery bar not reached A B C D The three required units, all mastered, and the practice it took
The same student, the same hours. Small blocks are supplemental practice on the unit above them.

None of this is new. Individualized pacing is a century old, and attempts at it have rarely lasted. The Dalton Plan is the closest precedent to what I’m doing, and it met resistance partly because teachers “objected to the massive amount of paperwork and time for individualization that the plan required” (Tyack & Cuban, 1995, p. 96). That labor is cheap now: automatic grading handles varied submission schedules, and Study Buddy, our course-integrated AI tutor, helps students get unstuck.

The design also stays inside what Tyack and Cuban call the grammar of schooling. They describe how the Carnegie unit became part of it by “influencing three fundamental resources in secondary education: instructional time, specialized subjects, and academic credits” (1995, p. 88). Individualized pacing usually asks a school to let instructional time vary from student to student. Mine keeps the same term, the same credit hours, and one grade at the end, and changes only what’s required inside them. It would work under a mandatory curve, too.

Gains and costs

I expect the least prepared students to gain the most. They won’t get as far as what used to earn an A, but they should understand what they cover, and for the same hours they should end with a grade at least as good as before.

The best-prepared students should lose nothing, as long as they use the time the shorter path frees up. They finish the required path early and can take the units I made elective, so they end up where they used to or further along. If they don’t use the time, they could lose more than anyone, because they have the most of it.

A well-prepared student under both regimes Under the traditional regime a prepared student masters all four units and runs out of course. Under the proposed regime they clear the shorter required core early; directing the surplus takes them through the dropped unit and a further elective, while leaving it undirected wastes it. Traditional mastery bar A B C D Masters all of it, then runs out of course This design, surplus directed mastery bar elective from here A B C D E Clears the core early, spends the surplus on what was dropped This design, surplus not directed mastery bar A B C D E Clears the core early, surplus goes nowhere
One prepared student under both regimes. In the bottom row, the time the shorter path frees up goes unused.

Getting them to use it is a pricing problem of the kind game designers deal with. Electives have to be worth enough that prepared students take them, and supplemental practice has to be worth little enough, or capped, that nobody earns a grade by repeating material they’ve already mastered.

The cost should land in the middle, on students who could just keep up with the old pace. They used to be required to cover everything and mostly managed it. Now some of it is elective, and they may not take it.

My theory of action

Goals Activities Short-term outcomes Long-term outcomes
Every student masters the required core A required path sized to the hours students have, with resubmission Does every student finish it at full marks? Students who finish are ready for later courses and technical work
Each student’s time goes where it helps most Credit for supplemental practice and for electives Do the students who need practice elect it, and do later units go faster for them? Less prepared students catch up in later courses
Fluency that students keep using Electives built around their interests Do students keep doing electives once they have the grade they want? Students keep building after the course, with nothing graded

As a chain:

If the required path is sized so that few students are carried past what they understand,

and supplemental practice earns credit,

and students can tell when they need it,

and a failed attempt costs them little,

then the students who need more practice will do it and ask for help, which until now only the ones who end up with an A have worked out on their own.

Weak points

The chain starts with sizing the required path, and so far I’ve done that by judgment, making elective the units that felt closest to optional. The rule should be to require what later units and later courses assume, and nothing else. That matters more in my design than in a traditional one. When a course requires everything, a B tells the next instructor that a student partly mastered all of it. When finishing my required path is supposed to mean mastering it, a mistake in what I require doesn’t show up anywhere, and students would arrive at the next course confident and missing something.

The deadlines work against the same condition. They exist to stop procrastination, as I found when an earlier version of the course had none, and to keep everyone ready for shared class activities. So the shared part of the course is also what moves some students on before they’re ready.

In class we discussed how schooling has mostly focused on the average student, and on individuals more than groups. My design is built against the first but fits the second: every outcome in my table is about one student at a time. The best evidence I know of for this kind of approach comes from untracked settings where students “conduct inquiries together, present and vet their answers to one another, and continue to revise their findings and products until they deeply understand the concepts,” with results that held “regardless of whether participating students entered high school with low or high levels of prior achievement” (Darling-Hammond & Darling-Hammond, 2022, p. 12). Working together is part of what produced those results. I’d like to make class time work for students who are at different points in the material, so the class doesn’t need everyone at the same point, but I don’t know what that looks like yet.

Credit raises the question of whether an extrinsic reward can turn into something more like intrinsic motivation. The advanced students I described above already loved the work and wanted credit because of time, so I read credit as letting motivation students already have compete with everything else that’s graded. I’m assuming beginners skip practice for the same reason. Students who keep doing electives after they have the grade they want would be an early sign that it works.

Whether students can tell when they need practice is the condition I’d bet against. A student who is behind when a deadline arrives has to choose between the new unit and going back. Before this term the practice earned nothing and the shaky students didn’t go back, so I can’t tell yet whether they didn’t know they needed it or had no reason to. If it’s the first, credit won’t help. Something has to tell them at the moment it matters, for example a message at the end of each unit, based on how they did, that points them to the practice and says what it’s worth. Study Buddy could do this better, and we haven’t built that part yet.

Evidence that would change my mind

If the students who need practice don’t elect it once it earns credit, credit wasn’t what was stopping them.

If they do the practice and later units don’t go faster for them than for similar students in earlier years, the practice isn’t transferring. We have data on how long students take on each lesson, so this is checkable.

If the gap between more and less prepared students in those earlier years stayed the same from the early units to the late ones, gaps don’t compound in my course, and the case for the design rests on uneven weeks alone. I’d measure the gap in time per lesson, since scores top out at full marks and the reading review found that ceilings make gaps look like they’re closing.

If students who finish the required path struggle with whatever comes next, I required the wrong things.

References

Darling-Hammond, K., & Darling-Hammond, L. (with Byard, E.). (2022). The civil rights road to deeper learning. Learning Policy Institute. https://doi.org/10.54300/462.143

Pfost, M., Hattie, J., Dörfler, T., & Artelt, C. (2014). Individual differences in reading development: A review of 25 years of empirical research on Matthew effects in reading. Review of Educational Research, 84(2), 203–244. https://doi.org/10.3102/0034654313509492

Tyack, D., & Cuban, L. (1995). Tinkering toward utopia: A century of public school reform. Harvard University Press.