Switch course ▾
Session 2. Designing a Pilot in One or Two Classes
Objective. Design a one-page pilot with one change, a baseline, outcome, process and balancing measures, realistic dates and safeguards for pupils.
A pilot is the safest way to find out whether an idea helps your pupils, because it is small enough to stop and short enough to learn from. But a pilot that changes too much, has no starting point or is planned on an imaginary calendar tells you nothing. This session shows how to turn one problem into a one-page plan that a colleague can read in two minutes and that protects your pupils and your time.
Key ideas
A pilot is small and bounded
- A pilot tests one change in one or two classes for four to six weeks, or one unit. It is a trial that you can stop, not a new rule for the school.
- Change one thing and keep the rest the same. If you change seating, homework and marking together, you will never know which one helped.
- Describe the change as actions: what the teacher does, what pupils do, how many minutes, how often, with which materials. 'Use group work' is not a change idea; 'pairs retell for 5 minutes after each text' is.
The one-page pilot plan
- AIM: who will improve, by how much, by when. CHANGE: the one thing. BASELINE: the starting measure. MEASURES: outcome, process, balancing. DATES. SAFEGUARDS. DECISION RULE.
- If the plan does not fit on one page, the pilot is too big. Plan the real calendar: count lessons, and subtract Navruz, exam weeks and school events (course 8.8).
- Write the aim with numbers you can check: 'raise the pupils who retell the main idea correctly from 9 of 28 to at least 16 of 28 in five weeks'.
Three kinds of measure
- OUTCOME: did pupils learn more? Use a short, repeatable tool such as the same type of exit ticket or a rubric on a work sample (course 5.5). PROCESS: did we actually do the change, and how often? Keep a tally.
- BALANCING: what does the change cost or disturb? Extra preparation minutes, lesson time taken from other parts, noise, marking load. A change that helps pupils but doubles your evenings will not last.
- Take the baseline with the SAME tool and conditions as the end measure. Prefer measures you already collect: a new test just for the pilot is an extra cost.
Comparison, safeguards and short cycles
- A comparison class (a parallel class of a similar level) helps you see what would have happened anyway, but it is imperfect. Offer the comparison class the change afterwards if it helps. Do not give any class weaker teaching.
- Safeguards: the pilot ticket is not graded, struggling pupils are supported, the standard is not lowered, parents are informed according to school rules, and pupils know why you are trying something new. Avoid exam-year classes.
- Use a Plan-Do-Study-Act loop: plan, try for one or two weeks, study what happened, then adjust. Small cycles inside the pilot let you fix problems early instead of finding them at the end.
Plan, Do, Study, Act
Why one change and one baseline
When several things change at once, a rise in results has many possible causes and you cannot tell which one to keep. When nothing was measured before, you can only say that results are 'better' by memory. A baseline taken with the same tool you will use at the end, in the same conditions, turns an impression into a comparison.
Choose measures you already collect where possible: the same kind of exit ticket, a rubric on a piece of work (course 5.5), or a tally of how many pupils asked a question. A special test built only for the pilot costs time and may measure something else.
Small cycles inside the pilot
The Plan-Do-Study-Act cycle fits well inside a pilot. Plan what you will try and write down what you expect to happen. Do it for one or two weeks. Study what actually happened, looking at your process tally and a few pupil comments. Act: keep, change or stop, and begin the next short cycle.
Writing a prediction first is useful, because it makes you notice surprises. If you expected a quiet lesson and the room was noisy, that is information about the design, not a failure. Short cycles let you repair a weak plan in week two instead of discovering the problem in week six.
Examples from the classroom
Copied homework in Fergana
Scene. Aziz finds that 12 of 30 Grade 8 algebra notebooks contain the same wrong step. He first checks the cause: pupils copy because they do not know where to start. He chooses one change only: a 5-minute worked-example start to each homework. The baseline is the number of identical wrong steps in the next homework check. He keeps marking, seating and homework length the same and tallies the check each week.
Insight. One change plus a baseline from an existing routine makes the result readable.
Counting real lessons in Bukhara
Scene. Malika plans a six-week pilot in Grade 9 physics in Bukhara. She counts her calendar: Navruz, a school exam week and an excursion remove nine lessons from what looked like thirty. She reduces the pilot to the 21 real lessons, chooses the unit on measurement, and sets the process target at 14 lessons of use. She tells her principal the exact dates.
Insight. A plan built on the real calendar avoids a pilot that is cut short by events.
Protecting the comparison class in Samarkand
Scene. Dilnoza pilots paired retelling in 5A and keeps 5B as comparison. She tells her colleague who teaches 5B that the routine will be offered to her class in week seven if it helps. Pilot tickets are not graded in either class, and weaker pupils in 5A are paired with supportive partners. The head knows the dates and what is being tested.
Insight. Fairness to the comparison class and to weaker pupils is part of good design.
Common pitfalls
- Changing seating, homework and marking together and then crediting one of them.
- Taking the baseline with an easier or different tool from the end measure.
- Planning from the textbook's lesson count instead of the real calendar.
- Skipping a balancing measure, so the extra workload is only noticed in month two.
Write a One-Page Pilot Plan
- Open the Activity Kit. Pairs (10 min): read the SAMPLE pilot in Table A. Find at least six flaws using the plan parts on the cards.
- Pairs (10 min): rewrite the aim, the change idea and the three measures of the sample pilot so that it passes. Count the real lessons before choosing the dates.
- Alone (20 min): take your problem statement from Session 1 and write your own one-page plan in Table B. Use one change only, and measures you can collect in a few minutes.
- Second activity, trios (10 min): a pre-mortem in Table C. Imagine it is week 5 and the pilot has failed. Each person names three reasons for your plan; you add one safeguard for each.
- Trios (13 min): check each plan with the five-point check in Table D. Do not argue the idea; test the design. Note one change for each plan.
- Alone (7 min): write your school-based task from the Kit.
Your facilitator will guide you through this activity in the session.
Questions to discuss
- Which flaw in the sample would have been easiest for you to make in your own plan?
- What did the pre-mortem show that your plan did not protect against?
- Is your measure something you will really collect in the middle of a busy week?
Reflect
- Have you ever started something new without a starting measure? What could you not say afterwards?
- Which of your pilot's costs would a colleague notice first?
- Which class could serve as a fair comparison, and how will you treat it?
Try it tomorrow
- Write your aim with numbers and read it to a colleague: can they say what you will do differently, for how many minutes?
- Run your baseline ticket in your pilot class this week and count the results without names.
- Draw a calendar of the pilot weeks and cross out holidays, exams and events before you fix the dates.
Further reading and links
One change, a baseline taken with the same tool, three measures, a calendar that is real, and pupils protected.