Good work does not speak for itself

If they have to find the point, you haven’t made one.

Two openings, same evidence, same author

Predict before you look · an invented case

“Background.”

In March the team began a review of teacher use of the assessment module.
We interviewed 24 teachers in 8 schools and analysed six months of usage logs.
Several themes emerged and are discussed below.
Section 5 sets out recommendations for consideration.

“Decision sought: approve a redesign of the marking flow before the Term 3 release.”

Teachers abandon the rubric step more than any other, and most don’t come back that day.
A redesign tested with 24 teachers cut abandonment in testing.
We’re fairly confident it will hold across schools, less sure about very large classes.
Detail and risks follow.

Most people can tell which one a busy decider would endorse. So here is the harder question.

Which one took longer to write?

Most work is told the way it was done

We report in the order we found things. Readers need things in the order they will use them.

Linda Flower called the default writer-based prose: organised as a story of the writer’s discovery, full of the writer’s own shorthand, surveying the topic instead of making a point. Expert writers do something different. They decide the goal first, then select and rebuild the content to serve it.

Writer-based first drafts are fine. Drafting is how you find out what you think. Shipping the draft is the problem.

If you can’t write the claim first, the analysis isn’t finished

Analysis climbs from evidence up to a claim. Reading runs the other way. That inversion is why communication costs effort: the order you thought in is the wrong order to be read in, so the work has to be rebuilt, not reformatted.

How you built it

Interviews and logsPatterns in drop-offRubric step stands outRedesign before Term 3

How they read it

Redesign before Term 3Rubric step stands outPatterns in drop-offInterviews and logs

When the claim won’t come, go back to the analysis, not the wording. A draft that opens with background is often a sign that no conclusion has been reached yet. The fix is more thinking, not better prose.

Writing the claim is a test of the analysis, not a summary of it.

Name the reader, the act and the date

Before a word, write who the work is for, what they can do with it, and by when. Write the act as a verb: approve, fund, defer, direct, decline.

“For the deputy director: approve or defer the extension of the reading pilot, before the budget meeting on 30 June.”

One line, and every later paragraph has a job: help that person make that call by that date. A brief that can’t name the act is a paper for the file, and should say so.

Some work asks for a decision. Some work changes how people think about a problem. Both are legitimate. The failure is writing one as if it were the other. Think pieces read better when they say they are think pieces.

Headings should carry the argument even if nobody reads the body

Most readers read the headings and first lines, then decide whether to go on. If the headings name topics, those readers learn the table of contents. If the headings state claims, they learn the argument. In one study, students shown slides with sentence headlines understood more and held fewer misconceptions than students shown topic headings.

× Topic headings label

Background
Findings
Next steps

✓ Claim headings argue

Teachers skip the rubric step
Shorter rubrics keep them in
Redesign before Term 3

Try the ghost deck. Write every heading first, read the column top to bottom, and only then fill in the evidence. If the column doesn’t argue, no amount of body text will.

You may have noticed every heading in Recess is a claim. Scroll up and read only those. The argument survives.

Take this with you

Name the reader, the act and the date before you write.
Put the claim first. If you can’t, the thinking isn’t done.
Write headings as claims, and check the argument survives on headings alone.

The work is done when someone else can use it, not when you have finished doing it.

Try this week

Ten minutes on a draft you are working on now.

Write the reader, the act as a verb, and the date. Two minutes.
Write the claim in 25 words or fewer. Ask a colleague what you want them to do. If they can’t say, it isn’t a claim yet.
Rewrite three headings as claims, then read only the headings.

To mull over

Which recent piece of your work was finished in your head and unfinished on the page?
When a draft won’t open cleanly, how often is the real problem two findings you haven’t ranked?
Who reads only your headings, and what did they take away last time?

Sources

Flower, L. (1979). Writer-based prose: A cognitive basis for problems in writing. College English, 41(1).
Bereiter, C., & Scardamalia, M. (1987). The Psychology of Written Composition.
Minto, B. (2009). The Pyramid Principle. Rozenblit, L., & Keil, F. (2002). The misunderstood limits of folk science. Cognitive Science, 26(5).
Garner, J. K., & Alley, M. P. (2013). How the design of presentation slides affects audience comprehension. International Journal of Engineering Education, 29(6).
Weiss, C. H. (1979). The many meanings of research utilization. Public Administration Review, 39(5).