Fixing Lessons With Data: One Unit's PDCA Cycle From Start to Finish

Looking at data does not make a lesson better on its own. Many teachers stop at the point of "I did the analysis, but I don't know what I'm supposed to change." For data to lead to better teaching, you need a cycle that connects observation to a hypothesis, the hypothesis to an experiment, and the experiment back to observation. Instead of talking about this in the abstract, let us follow one full turn through a middle school math unit on linear equations.

Following PDCA through one unit

It does not have to be grand. One unit and one change are enough.

1. Plan: In last semester's data you confirm that this unit's formative assessment pass rate, at 58%, was the lowest of the semester. Analyzing the most-missed items, the wrong answers cluster on the concept of transposing terms. You set a concrete goal: "a 75% pass rate."
2. Do: You make a three-minute mini video covering transposition and nothing else, and add one check-quiz period in the middle of the unit. Rather than changing several things at once, you put in only these two.
3. Check: After the unit ends, the pass rate has risen to 71%. It falls short of the goal, but it is a 13 percentage point improvement. More interesting still is that the pass rate among students who watched the mini video all the way through was far higher, at 82%.
4. Act: The video works, so keep it, but add five minutes of face-to-face support for the students who did not watch it into next semester's plan. This is how one cycle's conclusion becomes the next cycle's starting point.

Data-informed instructional improvement is not one enormous innovation but the craft of accumulation, stacking half a point at a time, one unit at a time.

Principles for turning the cycle

To keep the cycle from spinning in place, hold to the following. Turn it blindly and the wheel goes around without carrying you forward.

• Change only one thing at a time: Change three things at once and even if the pass rate rises, you will never know which one did the work. You have to control the variables to be able to reproduce the result next time.
• Always record the baseline you will compare against, first. Without the number from before the change (58%), you can neither prove the improvement nor take credit for it.
• Record the failed cycles too. Even a note saying "this method had no effect" is precious data that saves the next teacher time.
• Keep the cycle short, on the scale of a unit. Look back once after the semester ends and it is already too late for those students.

Another value in this approach is that the accumulated record becomes the school's asset rather than one teacher's. Stack up a single line like "in the linear equations unit, the transposition mini video was worth 13 percentage points," and the newer teacher who takes the same unit next year starts from a tested starting point instead of from nothing. Instructional improvement that used to rest on a teacher's intuition alone turns, one cycle at a time, into data that can be handed on. One small record saves the next group of students time.

Key takeaways

The essence of data-informed instructional improvement lies not in analytical skill but in the cycle of testing a small hypothesis quickly and checking the result back against the data. Start with one unit, one change, one baseline. Only the teacher who has turned one full round of PDCA gets to feel, in their own classroom rather than in a report, that data really does change teaching.