Is a blueprint required?

It varies by school. Many ask for one in the exam file and some do not. Where it is not required it is still worth building, because the point is not the paperwork but seeing the imbalance before the exam is written.

How many columns should it have?

As many levels as the exam actually uses. Two columns can be enough for a four-question quiz. An empty column does not fill the table, it just makes it harder to read.

Can one question go into two cells?

Better not. The totals break and the cognitive profile of the exam looks wider than it is. Place the question at the single level of the mental work it asks for.

Is a blueprint the same as a question distribution chart?

Close but not the same. A distribution chart usually shows only the question count per topic. A blueprint adds the cognitive level and becomes two-dimensional.

2026-09-02|6 min read

How to build a test blueprint

In short

A test blueprint is a two-way table with topics down the rows and cognitive levels across the columns, and each cell holds how many questions you ask from that topic at that level. Topic weights are set in proportion to teaching hours. Building the grid before writing the exam, rather than filling it in afterwards, turns it from a record into a plan.

A test blueprint is a two-way table showing how many questions an exam asks from each topic and at which cognitive level. Topics run down the rows, levels across the columns, and each cell holds a question count.

The table is usually filled in after the exam is over. In that order it stops being a planning tool and becomes paperwork. What follows is how to build it before the questions are written.

What the table proves

For an exam to count as valid it has to sample the area it means to measure in a balanced way. Measurement literature calls this content validity.

If sixteen of twenty questions come from one topic, the exam measures that topic rather than the term. A student who studied it scores high, and everything else they learned never reaches the grade. A blueprint makes that imbalance visible before the paper is printed.

The same table helps with appeals. When a parent says the exam was too hard, a chart showing how the questions spread across levels moves the discussion off opinion.

Rows: topics and their weight

The rows hold the topics covered during the term. How many questions a topic gets is set in proportion to the time spent on it.

A worked example: if the term ran 40 teaching hours and one topic took 10 of them, that topic should take a quarter of the exam. On a twenty-question paper that is five questions. A topic covered in two hours needs one.

Teaching hours spent on the topic is the most reliable measure of weight
The number of learning objectives comes second; a topic with many objectives needs more questions
Topics that later units build on are nudged upward

Columns: cognitive levels

The columns come from the levels of Bloom taxonomy: remember, understand, apply, analyse, evaluate, create. Not every exam needs all six.

Do not open a column for a level the exam never uses. Six columns on a four-question quiz leaves five of them empty and the table says nothing. Write only the levels actually in use.

A practical balance: about half the questions at remember and understand, a third at apply, the rest at analyse and above. The weight shifts upward as the grade level rises.

Cells: question counts

Each cell holds how many questions you ask from that topic at that level. Row totals give the weight of the topic, column totals give the cognitive profile of the exam, and both add up to the number of questions on the paper.

If the totals do not add up, the problem is in the exam rather than the table. Instead of writing one question into two cells, decide what the question really measures. A question sits at the level of the mental work it asks for.

Three common mistakes

Filling the table in after the exam. In that order it corrects nothing and only records what happened.
Piling every question into the remember column. Those are the easiest questions to write, and the whole exam drifts toward recall without anyone noticing.
Writing rows as the sub-headings you happened to ask about rather than as units. Rows that fine make the table unreadable.

Building it before the exam

The table earns its place in this order: build the empty grid, write target counts into the cells, then pick or write questions against those targets. The blueprint becomes the plan for the exam.

That order also speeds up question selection. Needing two analyse-level questions from one unit is a far easier task than staring at a blank page.

For teachers building exams from this pool the table is not filled in by hand at all. Every question carries its topic and cognitive level as tags, so when the exam comes down as a PDF the blueprint comes with it, together with the objectives covered.

Frequently asked

Is a blueprint required?: It varies by school. Many ask for one in the exam file and some do not. Where it is not required it is still worth building, because the point is not the paperwork but seeing the imbalance before the exam is written.
How many columns should it have?: As many levels as the exam actually uses. Two columns can be enough for a four-question quiz. An empty column does not fill the table, it just makes it harder to read.
Can one question go into two cells?: Better not. The totals break and the cognitive profile of the exam looks wider than it is. Place the question at the single level of the mental work it asks for.
Is a blueprint the same as a question distribution chart?: Close but not the same. A distribution chart usually shows only the question count per topic. A blueprint adds the cognitive level and becomes two-dimensional.