Effective teacher professional development: what the evidence shows

Effective teacher professional development is not a format you can buy; it is a property a programme either has or does not. What the evidence supports is narrow enough to state in a sentence: teacher professional development changes practice when it contains a cycle of attempt, observation and second attempt, and when the school permits the new behaviour afterwards.

What is being claimed, and by whom

Teacher professional development is one of the few areas of education where the funder, the provider and the recipient want different things and rarely say so. Systems buy it to improve pupil outcomes. Providers sell content. Teachers, reasonably, want something usable on Monday. A programme can satisfy any two of those and fail the third, and most evaluations only ask the recipient whether the day was worthwhile.

This brief inherits its frame from the adult-education brief, because teachers are adult learners in exactly the sense that matters: what they bring — a working model of their own classroom, built from experience and partly wrong — is the variable that decides what any programme can do.

What teacher professional development changes, in what order

The causal chain from a programme to a pupil is long, and each link is a place to lose the effect:

1. The programme changes what a teacher knows or believes. This is the easiest link and the one most programmes stop at.
2. The teacher attempts the practice, in their own room, with their own class.
3. The attempt is observed and adjusted, by a colleague, a coach or a recording.
4. The practice stabilises as something done by default rather than by effort.
5. The stabilised practice is the kind that affects learning.

Most disappointing programmes are complete at step one and absent from step three onward. That is also why format comparisons are so noisy: two programmes with the same label can differ entirely in whether they include steps two to four, and the ones that do outperform the ones that do not regardless of what they are called.

Decay, and why it is expected

A new practice is effortful before it is automatic, and effortful practices lose out to established ones under pressure: the pressure of a difficult class, a short week, an inspection. Decay is therefore the default and does not indicate that the programme was wrong; it indicates that nothing maintained the change past the effortful phase.

Two design consequences follow. Maintenance should be planned as part of the programme rather than hoped for, and it can be small: a scheduled review, a colleague who asks, a recurring point on a meeting agenda. And evaluation timing should be chosen to match the claim: a measure taken immediately after the programme reports enthusiasm, not durability.

The formats, and what makes one effective

Read the third column first: it is the difference between effective teacher professional development and an enjoyable day. It is a better predictor of whether practice changes than the format label in the first, and it is the column that costs money, which is precisely why programmes are so often bought from the rows that do not have it.

Why the one-off workshop survives

The single-day workshop is the format with the weakest claim on changing practice and the strongest institutional position, and it is worth being clear about why rather than treating it as irrationality. It is procurable: it has a price, a date and an invoice. It is coverable: every member of staff can be recorded as having received it. It is schedulable inside the one slot the timetable already reserves. And it produces immediate positive feedback, because a good speaker is enjoyable.

Coaching has none of those properties. It is expensive, it is hard to schedule, its effects appear late, and it makes visible that different teachers need different things, which is administratively inconvenient. The workshop wins on procurement, not on evidence, and naming that is the first step to arguing for anything else.

The defensible use of a training day is as step one of the chain: shared vocabulary and a stated expectation, explicitly followed by something that carries steps two and three. Sold as the whole intervention, it is the most reliably wasted line in a school budget.

The conditions that decide everything

The training-transfer literature and school-improvement research converge on the same short list of conditions, and they sit outside the programme:

• Permission to try and fail. A new practice performed badly the first time has to be safe, or teachers will run the old one when observed.
• Time that already exists in the timetable. Development that depends on goodwill hours reaches the teachers with the fewest competing demands.
• Coherence. A programme that contradicts last term's initiative teaches staff that initiatives are weather.
• A leader who does the thing. Modelling by senior staff is the cheapest available signal that the change is real.

Where these are absent, the honest recommendation is to fix one of them rather than to buy a better programme, and any proposal that cannot say which of the four it depends on is a proposal to spend money on step one.

Induction and early-career retention

The period with the clearest return is the first two years, for a structural reason: early-career teachers are still forming default practices, so change is cheap, and they are the group most likely to leave, so retention effects compound. Structured induction with a trained mentor, protected time and a reduced timetable is one of the few interventions in this field that plausibly pays for itself in avoided recruitment.

The failure mode is mentoring in name only: a named colleague with no time allocation and no training in giving feedback on practice. That arrangement produces the appearance of induction and the workload of it, without the observation-and-second-attempt cycle that makes it work.

Evaluating a programme against the evidence

Four questions separate an evaluation that means something from one that does not:

1. Which link in the chain is being measured? Satisfaction, knowledge, practice, or pupil outcome. Name it, and do not let a measure of one stand in for another.
2. When was it measured? Immediately, or after the effortful phase where decay happens.
3. Who did not participate? A voluntary programme evaluated on its volunteers reports the effect of being the kind of teacher who volunteers.
4. What else changed that term? Schools rarely run one initiative at a time, which is a real limit on attribution rather than a reason to skip the question.

Those four questions are what separates evidence about teacher professional development from testimony about it. An evaluation that answers all four can be small and still useful. One that answers none can be large and tell you only that the speaker was good.

Works referred to

1. Military Pedagogies: And Why They Matter — Sense Publishers, 2013, on training cultures outside schooling.
2. Leading Educational Change Wisely — Sense Publishers, 2009.
3. Lesson Study: Professional Learning for Our Time — a standard account of the Japanese lesson-study cycle.
4. Understanding and Facilitating Adult Learning — Stephen D. Brookfield, 1986, for the adult-learning frame this brief inherits.

Questions readers send

Does professional development improve pupil outcomes?

It can, and the chain is long: the programme has to change what a teacher knows or believes, that has to change classroom practice, and the changed practice has to be the kind that affects learning. Each link loses signal, which is why effects on pupil outcomes are typically small even when effects on practice are clear.

How long should a programme run?

Long enough to include cycles of trying something and reviewing it, which in practice means terms rather than days. Duration on its own is not the active ingredient (a long programme with no application cycle is simply a long workshop), but no short programme can contain a cycle.

Is instructional coaching the best-supported format?

It has the strongest evidence base of the common formats, and the reason matters: coaching is the format that most reliably includes observation, feedback and a second attempt. Programmes that include those elements under another name tend to perform similarly.

Why do teachers report that development days are useless?

Usually because the day was aimed at awareness while the teacher needed a change in practice, and nothing in the following weeks supported the attempt. The judgement is about the design, not about the content, and it is often correct.

Can development be voluntary and still work?

Voluntary participation raises the average commitment and lowers the reach, which is the same trade adult education faces everywhere. Mandatory participation reaches everyone and imports resistance. Neither is free, and pretending otherwise is how programmes get designed for the people who would have come anyway.

Where to go next

1. Learning theories The four families of learning theory compared on the same three questions: what each one can explain, what it cannot, and which teaching decisions it licenses.
2. Adult education What adult education research supports and what it only asserts: the six andragogical assumptions checked one at a time, plus the participation evidence.
3. Curriculum design Backward design, spiral sequencing and constructive alignment explained as a coverage-versus-depth trade, with the cost of each choice stated in teaching hours.