All resources · 1,649 words · 7 minute read

Lesson observation that produces evidence, not paperwork

Most lesson observation systems produce paperwork, not evidence. What to record, how to cover every department, and how to close the loop to CPD and back.

Most observation systems generate a great deal of paper and very little evidence. The reason is that the form was designed for the observer sitting at the back of the room, not for the person who will need, a year later, to show that teaching in a department changed and that the school caused it.

A lesson observation becomes evidence when three things are true: what gets recorded is consistent enough to compare, coverage reaches the departments nobody visits, and there is a traceable line from an observation to training to a later observation of the same teacher. Without those, you have a filing cabinet.

Why most lesson observation forms fail a year later

The form usually serves the conversation that follows the lesson. It has space for narrative, a judgement, and two or three targets. That is exactly right for the feedback meeting on Thursday, and close to useless in June when someone asks what changed across the year.

Three failure modes recur.

• Free text that cannot be aggregated. Forty pages of prose contain the answer, but no one can extract a pattern from them without reading all forty and forming a private impression, which is an assertion again.
• Every observer records something different. One writes about behaviour, another about questioning, a third about differentiation. Two rounds of this cannot be compared, because they are not measuring the same thing.
• Coverage follows relationships. Observations cluster where the observer is comfortable and where timetables are convenient. The department that most needed a visit is the one with none.

None of this is a failure of effort. It is a design problem, and it is fixed at the point where you decide what the form asks for.

What "usable a year later" actually means

Consistency of what is recorded

Agree a short fixed spine that appears on every record without exception: date, teacher, class and subject, observer, the agreed focus for this round, what was seen against that focus, what was agreed next, and the date of the follow-up. Everything else can stay free text.

The spine is what makes aggregation possible. If every record for the year carries the same eight fields, a middle leader can answer "what did we see about questioning in science this year" in ten minutes instead of a weekend. The free text still matters for the individual conversation; it just cannot carry the whole load.

Agreeing the focus each round is the part schools skip. A round without a shared focus produces twenty observations of twenty different things.

Coverage across departments

Coverage is a question about blanks, not totals. Build a simple grid of every teacher against the rounds in the year, and look at the empty cells rather than the full ones.

The cells that are usually empty are predictable: part-time staff, teachers who joined after the start of the year, the specialist subjects that sit outside the main timetable blocks, and any department whose head is also the person doing most of the observing. A grid with visible blanks is more honest than a tally, and a reviewer reads it faster.

A line from observation to CPD and back

This is the part that turns records into evidence of improvement. One observation is a description of a lesson. An observation, followed by a piece of training or coaching chosen because of what that observation found, followed by a second observation with the same focus, is an argument that the school acts on what it sees.

The middle term is what most schools lose. The training happened, but nothing connects it to the observation that prompted it, so the two sit in different systems and the link exists only in someone's memory. Recording the connection takes one line at the time: which observation this training answers, and when the follow-up will be. The same principle applies to CPD records that show effect rather than attendance, which is the other half of this loop.

A cycle that actually closes

The following is a full round. Most schools can run two or three of these in a year; running one properly beats announcing three and finishing one.

1. Set the focus for the round. One thing, agreed by the leadership team and written down in a sentence a teacher would recognise. Publish it before the round starts, so nobody is observed against a criterion they had not seen.
2. Publish the schedule. Every teacher, with a window rather than a fixed slot, so the round survives absence and timetable changes.
3. Calibrate the observers. Two observers watch one lesson together, write independently, then compare. Record where they disagreed and how they resolved it, before the round begins.
4. Observe, and record on the spine. Same eight fields, every time, completed before the observer leaves the room or within the same day.
5. Hold the conversation within a few days. Agree one change, not five. Write the agreed change in the teacher's own words on the record.
6. Aggregate immediately after the round. Read across all the records for one thing only: what recurs. Write half a page for each department saying what was seen and what is missing.
7. Choose the response from that half page. Coaching for two teachers, a department meeting for one issue, a training session for a pattern across the school. Note against each observation which response it triggered.
8. Deliver the response, and record it with dates and names. This is the middle term. Without it the next observation proves nothing.
9. Observe again with the same focus. Same teachers, same criterion, several weeks or months later, long enough for a change to be real.
10. Write what changed and what did not. Including the teachers where nothing changed and what you decided to do about that. A cycle that only reports successes reads as selective, and reviewers notice.
11. Feed the aggregate into self-evaluation and the improvement plan. The half-page departmental summaries are the raw material for the teaching section of your self-evaluation form, and the response decisions belong in an ADRI improvement plan as deployment and results.

Steps 6 to 8 are where cycles die. They are also the only steps that make the exercise evidence rather than paperwork.

The arithmetic before you publish a schedule

Work out whether the round is deliverable before you announce it. If your school has sixty teachers and you want two observations each, that is a hundred and twenty observations. Against roughly thirty teaching weeks, and if you allow an hour and a half for the observation, the note and the conversation together, that is around six hours of leadership time every week, spread across whoever observes.

Run that calculation with your own numbers. If it does not fit, reduce the number of observations rather than the quality of the record. Sixty observations that close the loop are worth more than a hundred and twenty that do not, and a schedule that visibly stalls in February is itself a piece of evidence, of the wrong kind.

What to do with the records you already have

If you are reading this mid-year with a drawer of inconsistent forms, do not rewrite them. Retro-fitting records is both a waste of time and easy for a reader to spot.

Instead, write a dated summary of what the existing records show, note honestly what they cannot show, and start the fixed spine from the next round. A school that says "our records before January did not let us compare across departments, so we changed the form, and here is the first comparable round" is describing improvement. That is a stronger position than a uniform folder nobody believes.

The general test for any of these records (dated, attributable, showing change, findable by a stranger) is covered in more detail in what counts as evidence in a school inspection, and the term-by-term sequence sits in our guide to Gulf school inspection readiness. If the aggregation step is what keeps failing, that is usually a systems problem rather than an effort problem, and the guide to inspection preparation software sets out what to look for.

Frequently asked questions

How many lesson observations should a school do in a year? Enough to show a pattern rather than a sample. Two observations of each teacher, several months apart and sharing a focus, demonstrate more than six scattered visits. Count backwards from the number of teaching weeks you have and check the total is actually deliverable before publishing a schedule.

Should lesson observations be graded? That is a school decision and both approaches are defensible. Whichever you choose, the aggregate record needs the same fixed fields: date, teacher, class, observer, agreed focus, what was seen against it, and the follow-up date. A grade without those fields aggregates into nothing useful.

What is the difference between a learning walk and a lesson observation? A learning walk samples several classrooms briefly, usually to check one thing across a department. An observation looks at a whole lesson for one teacher. Both can be evidence, but only if what is recorded is consistent enough to be compared with the next round.

How do you moderate lesson observations between observers? Two observers watch the same lesson, write their notes independently, then compare before agreeing anything with the teacher. Record where they disagreed and how it was resolved. That written record of disagreement is itself strong evidence that your judgements are moderated.

How Asabya Academy helps

Asabya Academy records lesson observations against a consistent set of fields, links each one to the CPD that followed and to the next observation of the same teacher, and rolls the results up by department into the self-evaluation and ADRI improvement plans held in the same place, in Arabic and English.