Switch course ▾
Session 2. Planning the Test: The Blueprint
Objective. Build a test blueprint (table of specifications) that weights marks by teaching time, importance and thinking level, chooses item types, and fits the real time of the lesson.
A blueprint feels like extra paperwork until the day a test goes wrong: the last chapter takes 12 of 20 marks, or pupils run out of time on the final task, or two colleagues test the same unit and cannot compare their classes. A blueprint is the cheapest protection against all three. This lesson goes beyond weighting by lesson counts and shows how to handle essential topics, review of earlier work, parallel forms for two classes and the difference between a short quiz and an end-of-unit test.
Key ideas
Why plan before writing items
- Without a plan, a test drifts towards what is fresh in your mind or easy to write. A blueprint, also called a table of specifications, is a grid of topics against thinking levels with the marks for each cell. You fill it in before you write a single item.
- The aim is validity: the test measures what was taught and what the standard asks, in the proportions it was taught. It also helps reliability, because two classes taking tests built from the same blueprint are compared more fairly.
- A blueprint takes 15-20 minutes and saves hours of rewriting. It also gives you something to show a colleague or the head of studies when they ask why the test looks as it does.
How to weight the marks
- Start from teaching time: if topic C took 4 of 12 lessons, it gets about a third of the marks. Adjust up for outcomes that later units depend on, and down for optional material. Never give marks to content that was not taught.
- Then split each topic across thinking levels, guided by the verbs of the standard. A common first target is about 40 percent know, 40 percent apply and 20 percent reason; change it if your standard asks for more reasoning.
- Round to whole marks and check that rows and columns add up to the total. If a cell is empty, ask whether the outcome is really untested or the plan is wrong.
Item types and the clock
- Closed items (multiple choice, matching) cover many topics quickly and are easy to mark. Short answers and extended tasks show reasoning but take longer and need a rubric. Mix them: closed for breadth, open for depth.
- A working rule to check with your own pupils: about 1 minute for a closed item, 2-3 minutes for a short answer, 8-10 minutes for an extended task, plus 5-8 minutes for handing out, reading and checking. Time two or three pupils on sample items before trusting the rule.
- A test that is too long measures speed. If the time does not fit, cut items but keep the proportions of the blueprint.
Balance, sharing and using the plan
- Spread difficulty on purpose: some easy items early so that every pupil can show something, most of moderate difficulty, and a few demanding ones. A test of only hard items tells you little about most pupils.
- After you write the items, return to the blueprint and tick every cell. Many teachers find the plan and the draft do not match; fix the draft, not the plan, unless the plan was wrong.
- Share the topics and levels (not the items) with pupils beforehand. Knowing what will be tested is fair and helps them study the right things.
Building a blueprint
When teaching time is not the whole story
Weighting by lesson count is a good start, but it is not a law. A topic that took two lessons but is the foundation for the next three units may deserve more marks than its time suggests. A topic that took four lessons only for repetition may deserve fewer. Write the reason beside each adjustment so that a colleague sees a decision, not an accident.
Earlier work belongs in the blueprint too. If a September topic is needed in March, give it 2 or 3 marks in the term test; otherwise pupils learn that anything taught earlier can be forgotten. And do not give marks to content that pupils did not have a chance to learn, even if it is in the textbook: the blueprint must match the lessons you actually taught.
Short quizzes, unit tests and parallel forms
Not every assessment needs a full grid. A five-minute quiz after a lesson may check just one outcome at one level; a quick plan on a card is enough. Black and Wiliam argued that short, frequent tests with specific feedback support learning better than rare long ones, so keep the full blueprint for the end of a unit or term and use the card version for the rest.
If two teachers teach the same grade, one blueprint can serve both classes. Each teacher writes a different form (A and B) of the same cells: for example, Form A cell 'C, apply, 3 marks' and Form B the same cell with different numbers. The two forms reduce copying, allow a retest later, and let the teachers compare results by outcome. Agree in advance on one marking guide for both.
Examples from the classroom
Adding a spiral item
Scene. Sardor teaches Grade 6 natural science in Tashkent. His unit had 10 lessons: 6 on plants and 4 on water. A 20-mark test by lesson count would give 12 and 8 marks. He gives 10 marks to plants, 8 to water and 2 to a question that uses measuring from the previous unit, because water experiments need it. He writes the reason beside the 2 marks.
Insight. Weights follow teaching time, then are adjusted for importance and review, with the reason written down.
Cutting without breaking the plan
Scene. Malika's draft test for a 40-minute lesson in Bukhara adds up to 55 minutes. Instead of dropping the last topic, she shortens each topic's closed items by two, merges two short answers into one and keeps the extended task. The grid still shows all four outcomes, and her total falls to 38 minutes. She tells pupils which topics will be assessed.
Insight. When time is short, trim every cell proportionally rather than deleting a whole topic or the reasoning item.
One blueprint, two classes
Scene. Aziz and Dilnoza both teach Grade 7 maths in different Tashkent classes. They agree one blueprint for the percentage unit: 20 marks, four outcomes, 8 / 8 / 4 across know, apply and reason. Each writes their own form, and they swap and check each other's items. After the test they compare results by outcome, not by class average, and find that outcome C was weak in both.
Insight. A shared blueprint makes results comparable by outcome and lets colleagues learn from each other.
Common pitfalls
- Writing the test first and filling in the blueprint afterwards to make it look planned.
- Giving marks to a topic because it is fresh in your mind rather than because it was taught.
- Making the test long enough that pupils who write slowly lose marks for speed.
- Changing the blueprint during marking to make a result look better.
Blueprint a 40-Minute Test
- Pairs (10 min): open the Kit. The unit is Grade 8 maths, Ratio and Percentages, with four outcomes taught for 3, 3, 4 and 2 lessons. Work out each outcome's share of 20 marks and round to whole marks.
- Pairs (15 min): fill the grid. Spread each outcome's marks across KNOW, APPLY and REASON. Target about 8 / 8 / 4 for the whole test. Check that every row and column adds up.
- Pairs (10 min): choose item types for each cell (closed, short answer, extended task). Use the time table to add up the minutes. Does the total fit 40 minutes with 5-8 minutes for handing out and checking?
- Two pairs (10 min): swap blueprints and audit. Is any outcome untested? Does the marks share follow teaching time? Is the time realistic? Give two written suggestions.
- Second short activity, alone then pairs (15 min): take your own next unit and draft a blueprint: outcomes, lessons taught, marks, levels. Show it to a partner who teaches the same grade and ask what is missing.
- Alone (10 min): write your school-based task from the Kit and agree a date with your colleague.
Your facilitator will guide you through this activity in the session.
Questions to discuss
- Where did you change a mark share away from pure teaching time, and why?
- Which cut would you make first if the test turned out to be 50 minutes long?
- What did the other pair see in your blueprint that you had missed?
Reflect
- Which topic of my last test got more marks than the time I spent on it deserved?
- Have I ever had pupils run out of time in a test? What did the test then measure?
- Would my head of studies understand why my test looks the way it does if I showed the blueprint?
Try it tomorrow
- Take the next test you will write, draw the grid first and fill in marks by lesson count before writing any item.
- Add one spiral item from an earlier unit and write the reason beside its marks.
- Time two pupils on a sample closed item and a short answer to check your minutes rule.
Further reading and links
Plan marks by teaching time and thinking level before writing items, check the clock, then write.