Skip to content

Teacher professional development should include ongoing activities like study teams and peer coaching rather than only occasional large-group sessions

Principle · All principles
Evidence · 50 claims (30 for, 17 mixed, 3 against) · 32 studies (9 design, 8 review, 5 associational, 4 qualitative, 3 causal, 2 quant-synthesis, 1 theoretical), q1–q3 · 2 of 32 report an effect size · 48 claims rest on one study

Conditional relationship

This page owns one structural choice in teacher professional development: whether teachers' learning is carried by ongoing, school-based structures (a study team of colleagues that meets regularly to examine its own practice and students' work, peer coaching in which colleagues observe and debrief each other's lessons, or a school-based coach such as a literacy coach) in addition to occasional large-group sessions, or by the large-group sessions alone. For in-service teachers (adult, working in pre-k-12 schools) whose practice does not change in their own classroom after a workshop or an in-service day, adding such structures over a term-plus span is expected to produce more change in observed classroom practice (skill-performance, observed in lessons) and, less reliably and by less, in students' general-achievement, than occasional sessions alone. The change is expected only where the ongoing time is protected from other duties, the structure focuses on instruction, and participation is not tied to evaluation.

No claim in this wiki compares study teams or peer coaching with occasional large-group sessions alone. The nearest evidence is a meta-analysis of causal coaching studies, reported here on several claim pages: coaching (mostly by a dedicated coach, not by peers) against comparison conditions raised observed instructional practice by a pooled 0.49 SD and student test scores by a pooled 0.18 SD. The same meta-analysis found no dosage relationship and found that coaching paired with group training went with larger effects, which turns the old title's "rather than" into "as well as": the large-group session is not the problem, using it alone is. Peer coaching as such rests on one assertion with no studies behind it.

How this sits beside the converted pages: the Professional Development pattern holds the full cycle for one practice (diagnose, model, rehearse, enact with feedback, fade, check maintenance) and its doses; this page does not repeat it, and says which ongoing structure carries that cycle and how it is kept going. The Communities of Practice principle owns the model of sustained participation in a standing professional group, which a study team is one instance of. The Reflection principle owns the relationship between structured reflection and learning; a study team's examination of assumptions is reflection done with colleagues.

Default design, while the relationship is untested

The earlier page (and the literacy-coach page folded into it this week) gave usable guidance; it is kept here as a concrete default a designer can act on and revise. Each step is a design proposal with its evidence status beside it. Doses are untested proposals unless a claim is named.

  1. Open with a short large-group session that sets a shared language, then keep group sessions as one part of the year. Use one day (about 3–6 hours, a proposal) at the start of the year to introduce the practice or framework all teams will work on, and one shorter group session each term to add the next piece. For a literacy initiative, this is where the shared language of evidence-based reading instruction is established for teachers and administrators before coaches or teams build on it (from the folded literacy-coach page). Evidence status: pairing coaching with group trainings is associated with larger effects [+M] (an exploratory regression across the meta-analysis's studies; the authors say it is descriptive, not a causal effect of the feature).
  2. Form study teams of 3–6 teachers who share a grade or subject, meeting every one to two weeks for 45–60 minutes. Each meeting has a facilitator and a written agenda in which instruction is a formally led item, not left to spare time: one teacher brings a piece of student work or a short lesson video, the team names what the students did, asks what assumption about teaching it reflects, and agrees one change each member will try before the next meeting. Evidence status: team size, frequency and agenda are untested proposals. Daily teaching demands competed with and distracted from a sustained developmental focus [~M] (interviews in seven small Chicago high schools: teachers valued instructional discussion but rarely initiated it) is the reason to lead the item formally rather than hope it happens. Collaboration with colleagues around instruction has consistently positive relationships with math and science instructional practices [+M] (a survey association in one district, not a comparison). The schedule is feasible where teams already exist: in one large district, around half of teachers reported meeting in grade-level teams at least weekly and observing colleagues' classrooms at least quarterly [+W] (a description of how common the structures were, not of what they did). Treat each agreed change as a small test: the teacher plans it, tries it, brings back what students did, and the team decides to keep, adapt or drop it. In one five-year network of high-school mathematics teachers, teachers credited the focus on instructional routines and plan–do–study–act testing cycles with improving their instruction [+W] (interviews with 24 teachers, all crediting routines and 11 the cycles; self-attribution, no comparison group).
  3. Run peer coaching in pairs or trios, one cycle every two to three weeks. A cycle: a 10–15 minute planning conversation in which the observed teacher names one thing they want watched; a 20–30 minute observation (live or on video) recording only that; a 20–30 minute debrief within two school days that starts from what students did, names one thing to keep and one to try, and is kept between the two teachers. Evidence status: cycle shape and timings are untested proposals. Peer coaching is asserted, citing no studies, to be a highly effective form of teacher professional development [+W] is an assertion only. Peer reflective groups encouraged student teachers to challenge existing theories [+W] (one study of student teachers' writing and interviews, reported in a digest).
  4. Keep participation voluntary and separate from appraisal. Say in writing that what is seen in a peer observation or study-team meeting is not used in evaluation; let teachers choose their own focus within the shared practice. Evidence status: coaching framed as a partnership, with voluntary non-evaluative participation [+M] (correlations from one program's teacher survey); a district culture of coaching makes teachers more comfortable participating [+W] (the authors' own qualitative finding).
  5. Where the school has a coach, use the coach to run training, demonstrate, co-teach and observe with immediate feedback, and protect that time. The folded literacy-coach page gave the coach's role: facilitate teacher training on evidence-based reading instruction and data-based decision making; demonstrate lessons; co-teach or observe and give immediate feedback. Spend most coaching time on instructional delivery rather than on behaviour management alone, and keep the coach's caseload and other duties small enough that weekly contact happens. Evidence status: coaching time focused on instructional delivery improved practice, while time focused on behavioural management was associated with a decrease [~M] (one study, reported second-hand); time spent between teachers and coaches is one of the most important factors [+W] (qualitative, from the authors of one three-year study); the complexity of coaching responsibilities may fragment coaching and reduce its impact [-W] (an author's rationale from the literacy-coaching literature, with no data) is the reason to keep the coach's role narrow rather than add programme administration to it, as the district-programme row of the Professional Development pattern also says. A literacy-coaching randomised trial with early childhood educators, reported in a review [+M], found larger classroom-quality and child language and literacy gains than controls, with 16 hours of workshops and seven coaching sessions over a semester.
  6. Pair the ongoing work with the instructional materials teachers will use. The study team and the coach work on lessons from the curriculum the school has adopted, not on generic examples. Evidence status: pairing coaching with instructional resources and materials is associated with larger gains [+M] (same exploratory regression; a video library alone was associated with smaller effects). The high-quality instructional materials principle holds the materials side.
  7. Plan the year as an arc in which the source of support moves from coach or facilitator to the teachers. Proposed default, untested:
  8. Weeks 1–10: launch session; the coach or a trained facilitator leads every study-team meeting with a fixed protocol; the coach models or co-teaches once in each teacher's class; peer coaching starts in week 4 with the facilitator present at the first debrief.
  9. Weeks 11–25: teachers take turns leading the study-team protocol; peer coaching carries most observation; the coach observes each teacher once a month and joins a study-team meeting once a month.
  10. Weeks 26 onward: the team chooses its own next focus from what its student work shows; peer observation every three to four weeks; the coach is on call. A teacher or team moves to the next practice, or to less coach support, when the practice is seen at the agreed level in two consecutive observed lessons and the teacher can say why they made the choices they did (a proposal). Total coach contact under this arc is roughly 15–20 hours per teacher across the year. Evidence status: dose untested; no specific coaching dosage is established as necessary, and quality appears to matter more than dosage, with sustained duration favouring comprehensive coaching [~M] (a review: studied dosages ranged from 3 to 32 sessions; early-learning coaching ran 5 to 7 months on average); coaching dosage shows precisely estimated null effects [~M]. In one program, more than 81% of coached teachers received at least 30 minutes a week and averaged more than 16 hours over the year [+W]; that is a description of one program's dose, not a tested minimum.
  11. Check practice again after support ends, and give later adopters the same start. Observe each teacher with the same look-for list at least six weeks after coaching stops and fold the practice into the study team's standing routine. When the practice spreads beyond the first teachers, give each later group the same launch, facilitated team meetings and coached first weeks rather than a short orientation. Evidence status: untested here; most coaching studies did not measure maintenance, and where it was measured it was not always kept [~M]. In one university, student gains from a pilot taught by faculty who had intensive professional development were not consistently sustained when later adopters did not get it [+W] (a campus report, no statistics for the later years, higher education).

When a condition is not met: if no time can be protected for teams or coaching, the ongoing structures will be the first thing daily demands push out, so shrink the design (one focus, one cycle a month) rather than keep a schedule that will not run; if participation is tied to evaluation, separate the peer structures from appraisal before starting.

Fitting the design to a situation

The three facts that most change the decision:

  • How much protected time exists for ongoing work, and whose time it is. A study team or peer-coaching cycle that competes with meetings about attendance and test preparation rarely survives. If unstated, ask: how many minutes per fortnight are set aside for instructional work, inside the timetable, and is there a coach whose time is protected from other duties?
  • Whether the ongoing work is tied to evaluation. Voluntary, non-evaluative participation goes with more reported change; observation that feeds appraisal changes what teachers show. If unstated, ask: will anything seen in a peer observation or team meeting reach a teacher's appraisal?
  • What the target is: observed practice, or student achievement, on what horizon. The pooled effect on practice is much larger than on achievement, and the link between them is weak across studies. If unstated, ask: what will an observer see in a lesson when the practice is in place, and which student measure, if any, is expected to move, and when?
If the brief says… Then change… Basis
Learners are student teachers or teachers in their first years Put them in a study team with experienced colleagues and a peer reflective group that reads their own lesson notes; pair each with a mentor or coach who models in their class before peer observation starts; observe weekly for the first term Peer reflective groups encouraged student teachers to challenge existing theories [+W] (student teachers; outcome is reported views, not practice); weekly observation untested
Learners are experienced teachers who chose to take part Let each teacher set their own focus within the shared practice; start with peer coaching rather than a coach; keep group sessions to the one-day launch Voluntary, non-evaluative participation correlated with improvement [+M] (one program's survey, correlations)
Goal is early reading (K–3, foundational skills or the science of reading) Establish the shared language first in training for teachers and administrators; then a school-based literacy coach trains on evidence-based reading and data use, demonstrates, co-teaches and observes with immediate feedback; study teams look at students' reading data each fortnight A literacy-coaching randomised trial found larger classroom-quality and child literacy gains [+M] (early childhood educators, reported in a review; carried to K–3 untested); sequence from the folded literacy-coach page
Goal is standards-aligned math or science teaching Make the study team's work collaborative planning and review of lessons in the subject; prefer a coach with expertise in the subject Colleague collaboration associated with standards-aligned practice [+M] (survey association); schools with math-expertise coaches reported greater gains [+W] (one administrator's report); 31 of 35 reviewed effective-PD studies had a specific content focus, most often job-embedded [~W] (prevalence among studies a review selected, not a comparison)
Some teachers join remotely, or coaching must be at a distance (phone or video only) Run peer coaching on short filmed segments shared privately with the partner; hold study-team meetings online with the same protocol; keep one in-person session per term if possible Coaching can be effective face-to-face, at a distance and combined [~W] (a review: most studies were in person); no significant in-person versus virtual difference [~M] (too imprecise to rule out moderate differences); in a web-mediated coaching programme for secondary teachers (about 20 hours of in-service over 13 months), students of participating teachers gained more than controls, the equivalent of moving from the 50th to the 59th percentile [+W] (one trial, reported in a review)
Teachers are the only one of their subject or grade in the school, or the plan is a network across schools (a networked improvement community, a district subject network) Join teachers across schools into a standing network around one shared instructional aim; each teacher tests one change in their own class between meetings and brings back student work; share tasks and materials across schools; make sure network meetings are for teachers, not only for school leaders 20 of 24 interviewed teachers said networking with teachers in other schools did most to change their instruction [+W] and about 8 in 10 attributed changes in their instruction to the network [+W] (one network of about 100 hours a year; teachers' own reports); teacher-level collaboration stayed low where collaborative activities targeted school leaders [-W] (focus groups in one grant programme)
One coach for many teachers, or facilitator time is scarce (a coach shared across schools, a busy principal) Let peer coaching carry most observation cycles; the coach trains the peer coaches, leads study-team meetings by rotation, and checks in with each teacher once a month for 15–20 minutes; cap the caseload rather than spread visits thin Teacher–coach time is a key success factor [+W]; effects shrink as programs grow [~M] (a descriptive pattern across studies); the peer substitution is untested
A fixed short time budget, such as 45 minutes a fortnight in an existing meeting Use it whole for one instructional item: 5 minutes to recall last time's commitments, 25 minutes on one piece of student work or a video clip, 10 minutes to agree one change each, 5 minutes to set the next peer observation untested proposal; daily demands crowd out developmental work [~M] is the reason to keep the item fixed
Only one session is available Do not present it as ongoing PD. Use it to form the teams, practise one peer-observation cycle in the room on a video clip, and set the dates of the first two meetings; record that no change in practice can be claimed untested proposal
The brief names its own routine (a PLC protocol, a district observation tool, a coaching cycle) Keep it; change only what the evidence here speaks to: a formally led instructional item, non-evaluative use, protected time, and focus on instruction untested proposal; the four changes cite the claims in the default design
The brief gives its own criteria, framework or rubric for teaching Use it as the look-for list in peer observation and study-team review; do not substitute another untested proposal
Peer coaching or coaching is linked to appraisal, or stakes are high (a school under review) Separate peer observation from appraisal in writing; if that cannot be done, use study teams on student work rather than observation of individual lessons, and measure both teacher and student change Partnership framing and non-evaluative support correlated with improvement [+M]; a district coaching culture eased participation [+W]; supervisory coaching shows mixed effects [~M] (16 K–12 studies, 6 meeting all quality indicators)
Teachers lack confidence, or what is discussed is sensitive (a struggling class, a video of their own lesson) The teacher decides who sees a video and which segment; peer debriefs stay between the two teachers; the study team discusses student work, not named teachers untested proposal
Administrators are among the learners (principals, assistant principals) Include them in the launch training; make their part protecting team and coaching time and framing coaching as improvement, not inspection; they do not attend peer debriefs; leaders' own meetings do not replace teachers' team time A district culture of coaching [+W]; teacher-level collaboration stayed low where activities targeted leaders [-W]; target learners from the folded literacy-coach page
The goal claimed is student achievement Keep observed practice as the first outcome, add an aligned student measure, and expect a smaller student effect than the practice effect Pooled effect on achievement 0.18 SD [+M] against 0.49 SD on practice [+M]; large practice gains are needed for modest achievement gains [~M]
Teaching is to be kept as a whole performance, not split into fragments Observe and discuss whole lesson segments (a launch, a discussion) rather than isolated moves, and judge change in the lesson as taught untested proposal
Teachers are earning micro-credentials or working in a cohort Give the cohort collaborative planning time as its study-team slot Collaborative planning time was the top support cited by micro-credential earners [~W] (a survey of preferences, not outcomes)
The goal is integrating a classroom technology tool or platform into teaching Make the ongoing work the tool in use in teachers' own lessons: monthly coaching on how it fits their instruction, and a study-team item reviewing one lesson that used it; do not stop at a one-off training on its features. The coaching cycle itself is in the materials-and-tools row of the Professional Development pattern Monthly coaching helped teachers use and integrate a writing tool with their instruction [+W] (one implementation study, grades 7–8, no comparison); coached teachers reported greater impact of their technology use on students [+W] (self-report survey against non-coached peers)
Learners are higher-education faculty (a department, a teaching-and-learning centre programme) Use a department-based learning community in place of a grade-level team: about six meetings across a term, every two weeks, 60–75 minutes, on one shared teaching question; add peer observation of a class, seminar or lab by a colleague from the same department; make the first training extended rather than a brief session A six-session biweekly faculty learning community on AI ran in one physics department [+W] (a description of format and attendance, no outcome); ePortfolio outcomes were stronger at several universities where faculty had extended intensive PD than brief training or none [+W] (a report's summary of other campuses, no statistics); the coaching claims above are from K–12 classrooms
A practice piloted with volunteers is being spread to later adopters Give later adopters the launch, the facilitated team meetings and the coached first weeks the pilot group had, rather than a short orientation from the pilot teachers; expect smaller effects as the programme grows and keep observing practice Pilot gains were not consistently sustained when later adopters did not get intensive PD [+W] (one campus, higher education); coaching effects shrink at scale [~M]

Fourteen of the twenty rows cite a claim; none of those claims compares ongoing structures with occasional sessions, and most are single surveys, teachers' own reports or qualitative reports carried from coaching programs to peer structures, or from higher education to schools and back. The rows for a single session, a fixed time budget, the brief's own routine or criteria, sensitive material and whole-lesson observation are untested proposals.

Observation, state and explanation

What can be observed: attendance at team meetings and peer cycles; what a teacher brings and what the team agrees; the practice as seen in an observed lesson against the look-for list; students' work. "Has changed practice", "is reflecting on assumptions" and "has bought in" are inferred states. Talk in a study team is not change in a lesson.

Response Plausible explanations Observation that could change the interpretation
Study-team meetings happen, but talk is about logistics, attendance or test dates Daily demands crowd out instruction; no one leads the instructional item; the team does not trust each other enough to show work Read the last three agendas and notes; ask who led the instructional item; run one meeting with a facilitator and a protocol and compare what is discussed.
Teachers describe the practice well in meetings, but it does not appear in observed lessons They do not yet know how to do it; materials or pacing block it; the observation came too soon Ask the teacher to plan the next use of it aloud; if they can, find the constraint; if not, the coach models it in their class.
Peer debriefs are uniformly positive and nothing changes Peers avoid criticism of colleagues; the look-for list is vague; the observation is felt as evaluation Ask each pair what one change was agreed last time and whether it happened; check whether anything reaches appraisal.
Practice changes, but student work does not The practice is enacted in form but not purpose; the student measure is not aligned; the horizon is too short Look at what students did in a filmed segment; check whether the student measure tests what the practice targets.
Practice in place at the end of the year, gone the next term The team structure stopped; staff turnover; the practice was tied to one coach Ask when it stopped and what replaced it; check whether the team still meets.

These rows are proposals, untested as diagnostics. Record each probe, because asking a team about its agenda changes the next agenda.

Evidence and qualifications

In this wiki no claim compares ongoing study teams or peer coaching with occasional large-group sessions alone, and none tests peer coaching. Seven of the claims below come from one meta-analysis (Kraft, Blazar and Hogan 2018), six directly and one through a practitioner review, so they are one study, not seven.

Learners, settings and outcomes differ across these claims (pre-kindergarten to high school teachers, technology coaching, early literacy, charter schools, a commercial coaching provider), so do not rank them. None measures observed practice a year after support ends.

The evidence by question

Further evidence, not yet read against this model

Read against the model on 2026-10-19. Of the 25 claims that had gathered here in 24 bullets, 22 now sit in the model above: three were already cited in it and were read again (the informal-PD survey in the learner-valued goal section, the coaching-hours survey in step 7, the subject-coach report in the standards-aligned row), and nineteen were moved into the default design (steps 2, 5 and 8), the situation table (four new rows: technology tools, higher-education faculty, spreading a pilot to later adopters, and networks across schools; and additions to the remote, subject, appraisal and administrators rows), ## Evidence and qualifications, the evidence-by-question list and the learner-valued goal section. The three below were read and left here, because they do not change a design decision on this page.

Objective and learner-valued goal

The designer's objective is usually a named practice observed in teachers' lessons, sometimes with an aligned student outcome. Teachers value other things: time with colleagues, help with a hard class, materials that work, not being judged, or meeting a mandate. Ask each teacher, and record where it differs from the objective. Teachers report more satisfaction with informal learning than with in-service days (informal PD more satisfying than in-service days [+W], a survey); in a nationally representative survey of 856 teachers, almost all took part in formal PD, averaging 30 required hours a year, but far fewer were satisfied with it [+M] (84 percent attended in-service days, 20 percent were satisfied with them). Satisfaction is an observation about what they value, not evidence that practice changed. A study team that teachers enjoy and that never discusses instruction serves the valued goal and not the objective; one that reviews instruction under appraisal pressure may serve neither.

For example, a K–3 teacher may want a calmer reading block, while the school's objective is consistent phonics instruction. A literacy coach who co-teaches the phonics segment and helps with transitions in the same visit serves both; record both outcomes.

What would revise this model?

The expectation should weaken if studies comparing ongoing school-based structures plus group sessions with group sessions alone, with observed practice as the outcome, show no difference, or if peer coaching without a trained coach shows no effect on observed practice. If peer coaching works only with a coach training the peers, the default should make the coach the carrier and peers the supplement. If the meta-analytic association with group training reverses in direct comparisons, drop step 1's emphasis. If maintenance studies show practice holding without a standing team, the sustain step is a preference, not a requirement. Do not protect the model by explaining every null result as "not enough time" or "not enough buy-in" after the fact.

Observed practice, student achievement, teacher satisfaction and maintenance after support are separate claims. The present evidence in this wiki establishes no dose, no team size and no peer-coaching cycle.

Claims

Examples

Key Sources

  • Ferraro, Joan M. (2000). Reflective Practice and Professional Development. ERIC Digest. https://eric.ed.gov/?id=ED449120
  • Comprehensive Early Literacy Policy: Fundamental Principles. (2024). ExcelinEd Policy Toolkit. https://www.excelined.org
  • Merueña, J. A. A. (2023). Dynamic Language Learning Framework (DL2F) Instructional Model. EPRA International Journal of Research and Development (IJRD). https://eric.ed.gov/?id=ED637673
  • Boise Cascade Center for Community Development. (1972). Indian Economic Development: An Evaluation of EDA's Selected Indian Reservation Program. Volume I: Narrative. https://eric.ed.gov/?id=ED134349
  • CEO Forum on Education and Technology. (1999). School Technology and Readiness Report. Professional Development: A Link to Better Learning. The CEO Forum on Education and Technology, Year Two. https://eric.ed.gov/?id=ED428747
  • Smith, T. M., Walters, K., Griffin, M., Jones, W., Lennon, V., & Sanders, Z. (2021). The Better Math Teaching Network: Lessons Learned From a 5-Year Instructionally Focused NIC. American Institutes for Research & WestEd. https://www.bettermathteachingnetwork.org/
Page metadata
Field Value
type principle
id ongoing-study-teams-peer-coaching-in-service
aliases literacy-coaches-job-embedded-support, dl2f-teacher-development-sustainability-principle, on-the-job-specialist-training-for-tribal-development-personnel, sustained-follow-up-technology-pd, nic-instructionally-focused-pd-model
title Teacher professional development should include ongoing activities like study teams and peer coaching rather than only occasional large-group sessions
description For in-service teachers whose practice does not change after occasional large-group sessions, adding ongoing school-based structures (study teams, peer coaching, a school-based coach) with protected time, an instructional focus and non-evaluative participation is expected to change observed classroom practice more than the sessions alone; no claim in this wiki tests that comparison, and the nearest evidence (a coaching meta-analysis) favours pairing coaching with group training rather than replacing it.
status review
canonical True
generated by claude/unspecified on 2026-10-19
trust unverified — no one has confirmed this page's content
sources 6 — ferraro-2000, comprehensive-early-literacy-policy-fundamental-principles-2024, merueña-2023, boise-cascade-center-for-community-development-1972, ceo-forum-on-education-and-technology-1999, smith-2021

Frontmatter is the page's machine-readable record — see the Schema & Guide.