eduKateSG Learning Node Series · 0088

How Knowledge Brokering Works | Move Research Across the Boundary Between Evidence and Educational Practice

Education does not suffer from a shortage of information.

There are journals, evidence reviews, policy reports, data dashboards, teacher resources, conferences, professional networks, university centres, think tanks and thousands of new studies every year. Yet a school leader with a real decision on Tuesday morning can still struggle to answer a basic question: what evidence is relevant here, how trustworthy is it, and what should we do differently because of it?

That gap is where knowledge brokering lives.

Knowledge brokering is the work of helping research travel between communities that produce evidence and communities that must use it. The broker may be a person, team, organisation, teacher leader, ministry unit, research centre, professional association or intermediary network. Good brokerage does more than forward a paper. It finds, appraises, translates, contextualises, connects, questions, updates and helps people turn evidence into better decisions without pretending research can make the decision for them.

Research becomes useful when someone can connect the evidence, the decision, the context and the people who must act.

The 50-Second Read

• Knowledge brokering connects research evidence with educators, leaders and policy-makers who need to make decisions.
• The broker is not merely a messenger. Brokerage can include searching, synthesis, appraisal, translation, contextualisation, relationship-building and capacity development.
• A strong broker begins with the user’s decision problem, not with a favourite study.
• Evidence quality and evidence relevance are different questions. A rigorous study may still answer the wrong question for a local context.
• Translation should simplify access without deleting uncertainty, limitations or conditions.
• Good brokers help practitioners ask better questions and help researchers understand the constraints of practice.
• Trust matters because evidence use depends partly on whether the intermediary is seen as credible, useful and sufficiently independent.
• Research should be combined with professional knowledge, local data, values, feasibility and student needs rather than treated as an automatic command.
• Knowledge mobilisation is broader than dissemination. Sending information is only one possible activity.
• Versioning matters because evidence changes. A recommendation should have a date, source trail and update route.
• The strongest systems build brokerage capacity into schools and organisations rather than depending forever on one heroic expert.
• The goal is not “evidence-based obedience.” It is better reasoning with evidence.

Canonical Owner Boundary

This Learning Node owns knowledge brokering as the mediation layer between research and educational decision-making: finding, appraising, translating, contextualising and connecting evidence with users. How Research–Practice Partnerships Work remains the owner for sustained collaborative relationships in which researchers and practitioners jointly work on problems. How Teacher Inquiry Works owns practitioner inquiry from classroom question to evidence and action. How Implementation Support Systems Work owns the infrastructure that helps an adopted educational change survive real conditions. This page owns the boundary work that helps evidence reach, inform and be interpreted by those systems.

1. Evidence Does Not Move by Itself

A study can be publicly available and practically invisible. It may sit behind technical language, unfamiliar statistics, disciplinary jargon, paywalls, weak indexing or a publication format designed for researchers rather than teachers.

Access is the first boundary. Understanding, relevance and use come after. Knowledge brokering exists because making research available is not the same as making it usable.

2. Start With the Decision Problem

Weak brokerage begins with a paper and asks where to send it. Strong brokerage begins with a decision: Should we adopt this programme? Why are reading outcomes falling? What kind of professional development is most defensible? Which students are not benefiting?

The decision defines what evidence is relevant. Otherwise research search becomes information accumulation rather than problem solving.

3. Turn the Problem Into an Answerable Question

“How do we improve teaching?” is too broad. “What approaches help novice Secondary mathematics teachers improve questioning during their first year, and under what support conditions?” is more searchable and more actionable.

A broker helps users narrow without shrinking the problem so far that the answer becomes irrelevant to real practice.

4. Search Beyond the Famous Study

Decision-makers often encounter research through one memorable article, conference speaker or viral summary. Knowledge brokering asks what the wider evidence base says.

That may require systematic reviews, evidence syntheses, meta-analyses, practice guides, implementation studies and high-quality primary research rather than relying on a single result.

5. Evidence Quality and Relevance Are Different

A randomised trial can provide strong causal evidence for a specific intervention under studied conditions. It does not automatically tell a school whether its staffing, curriculum, timetable, age group or implementation capacity match those conditions.

Brokerage therefore asks two questions: how credible is the evidence, and how relevant is it to this decision?

6. Translate Without Flattening

Translation means making research intelligible to a new audience. It should reduce unnecessary complexity, not erase important complexity.

“This intervention works” is often too flat. A better translation might be: studies in these contexts found an average improvement under these implementation conditions; effects varied; the evidence is stronger for this age group than that one; the mechanism appears to depend on regular practice and feedback.

7. Preserve Uncertainty

Decision-makers often want a clear answer. Research often provides a bounded answer. Effect estimates have uncertainty. Contexts differ. Measures are imperfect. Some findings fail to replicate.

The broker’s job is not to make uncertainty disappear. It is to make uncertainty understandable enough that people can act responsibly despite it.

8. Contextualise the Evidence

Contextualisation asks what changes when evidence crosses settings. Does the school have the staff expertise assumed by the study? Is the curriculum comparable? Are class sizes different? Does the intervention depend on technology or protected time? Is the outcome measure meaningful locally?

Context does not invalidate external research. It determines how directly the research can travel.

9. Professional Knowledge Is Evidence Too

Teachers possess knowledge about students, task demands, curriculum timing and classroom feasibility that may not exist in published studies. Leaders know staffing, budgets and organisational constraints. Students know aspects of the learning experience adults may miss.

Strong evidence use combines research with professional expertise and local evidence rather than treating published research as the only legitimate knowledge source.

10. Local Data Needs Interpretation Too

A school’s own test scores feel immediately relevant, but local data can also mislead. Cohorts differ. Assessments change. Small samples fluctuate. Scores can move for reasons unrelated to the programme under review.

Knowledge brokerage can help teams combine the external evidence base with a disciplined reading of their own data.

11. Brokering Is a Two-Way Function

Researchers need information from practice too. They need to know which questions matter, which implementation barriers are common, which outcomes practitioners care about and why a theoretically strong intervention may be unusable in ordinary conditions.

A broker can translate practitioner constraints back toward research so evidence production becomes more relevant over time.

12. Dissemination Is Only One Activity

Sending newsletters, reports and summaries is dissemination. It can be useful, but knowledge mobilisation is broader. OECD work distinguishes activities that include evidence generation, dissemination, collaboration, capacity building, organisational support and system-level support.

The difference matters because a school drowning in summaries may need facilitated interpretation, not another PDF.

13. Synthesis Reduces Search Cost

Practitioners rarely have time to read fifty primary studies. High-quality synthesis compresses the evidence base while preserving the important distinctions: population, intervention, outcome, uncertainty and conditions.

A good broker knows when a trustworthy synthesis already exists and when a local question requires more targeted searching.

14. Appraisal Protects Against Attractive Weak Evidence

An intervention can have persuasive testimonials, polished branding and impressive before-and-after graphs while lacking a credible comparison. A study can be technically sophisticated but too small or too indirect for the decision.

Brokerage includes checking design, sample, measurement, comparison, attrition, conflicts, replication and the match between claim and evidence.

15. Evidence Hierarchies Are Helpful but Not Sufficient

Different questions need different methods. Randomised trials are powerful for some causal questions. Qualitative studies can illuminate implementation, meaning and experience. Longitudinal data can reveal patterns over time. Design studies can expose mechanism.

The broker should match method to question rather than rank all research with one universal ladder.

16. Build an Evidence Chain

A mature recommendation can show its chain: decision question → evidence search → synthesis or study → quality judgement → contextual assumptions → proposed action → local evidence plan.

This makes the reasoning inspectable. If the recommendation later fails, the organisation can ask which link was weak rather than blaming “research” in general.

17. Provenance Matters

Where did the claim come from? Is the source primary research, synthesis, advocacy material, commercial marketing or commentary? Has the claim changed as it passed through summaries?

Good brokerage preserves provenance so users can trace important recommendations back to the strongest available source.

18. Date the Recommendation

Evidence evolves. A strong conclusion in 2022 may need revision after later trials or systematic reviews. Technology domains can change even faster.

Every important evidence brief should show when the search was last updated and what would trigger a review. Knowledge that cannot be versioned eventually becomes folklore.

19. Separate Evidence From Recommendation

Research might show that an approach tends to improve an outcome. The recommendation to adopt it also depends on cost, feasibility, opportunity cost, existing provision and strategic priorities.

Keep those layers visible. “The evidence says” should not be used to hide a value judgement or resource decision.

20. Ask Who Benefits and Who Bears the Cost

An intervention can raise average attainment while adding substantial teacher workload, widening gaps or depending on resources unevenly available across schools.

Brokerage should surface distributional questions: for whom, under what conditions, at what cost, and compared with what alternative?

21. Relationships Reduce Translation Friction

A broker who understands both research and school operations can translate more accurately because they know the languages, constraints and norms of both communities.

Trust is built through repeated usefulness: answering the real question, acknowledging uncertainty, not overselling evidence and returning with updated information when the evidence changes.

22. Neutrality Is Valuable but Never Perfect

Brokers make choices about which evidence to search, how to frame questions and which outcomes matter. Complete neutrality is unrealistic.

What matters is transparency about roles, interests, funding, assumptions and uncertainty. Perceived independence can strengthen trust, but transparent reasoning is more durable than a claim of objectivity.

23. Commercial Evidence Needs Extra Care

Education vendors may produce useful research. They also have an interest in adoption. Brokerage should therefore distinguish independently replicated evidence, vendor-funded evaluation, case studies and marketing claims.

The existence of a commercial interest does not automatically invalidate evidence. It raises the importance of transparent appraisal.

24. Build Capacity, Not Dependence

If every evidence question must go to one specialist, that specialist becomes a queue and the organisation remains dependent. Strong brokers teach others how to search, appraise and use evidence.

The long-term goal is distributed evidence literacy supported by expert help for harder questions.

25. Teacher Leaders Can Broker Knowledge

A teacher who reads research, understands colleagues’ practice and can translate evidence into usable instructional discussion can function as a local knowledge broker.

OECD work on teaching as a knowledge profession has explicitly discussed the possibility of knowledge-broker roles that facilitate knowledge exchange and keep colleagues connected to developing evidence.

26. School Leaders Broker Too

Leaders constantly translate external information into local priorities. They decide what reaches staff, which evidence gets airtime, which expert is invited and which programme is funded.

That makes evidence brokerage part of leadership even when nobody uses the term. The question is whether it is done deliberately and transparently.

27. Research Organisations Broker Too

Universities and research institutes increasingly produce summaries, tools, webinars, partnerships and practice guides. Their proximity to evidence is an advantage; their distance from everyday school conditions can be a limitation.

Effective brokering often requires interaction rather than one-way simplification.

28. Ministries and System Agencies Can Broker at Scale

OECD work has documented policy and analysis units inside education ministries that gather, translate and disseminate research for decision-makers. Formal brokerage agencies also operate across some education systems.

System-level brokerage can reduce duplication and create shared evidence infrastructure, but it still needs routes for local questions to travel upward.

29. Networks Can Broker Across Organisational Boundaries

Professional associations, school networks and research–practice communities can connect people who would otherwise search separately. They can circulate tested resources, compare implementation experiences and identify common questions worth researching.

Network value grows when information is curated and challenged rather than merely amplified.

30. Artificial Intelligence Changes the Search Cost, Not the Epistemic Duty

AI systems can accelerate search, summarisation and comparison. They can also hallucinate citations, flatten evidence quality and present weak claims with confident language.

The core brokerage duties remain: provenance, appraisal, relevance, uncertainty, context and traceability. Faster retrieval increases the value of stronger verification.

31. Evidence Briefs Should Be Decision-Shaped

A useful brief is not a miniature literature review. It should answer the decision-maker’s question: what is known, how confident should we be, what conditions matter, what are the likely trade-offs, and what evidence should we collect locally if we proceed?

Good format reduces cognitive load without hiding the chain of reasoning.

32. Use Evidence Menus When There Is No Single Best Answer

Some problems have several defensible approaches. A broker can present options with evidence strength, resource demands, likely mechanisms and risks rather than manufacturing false certainty.

This turns evidence use into structured choice rather than expert decree.

33. Create a Decision Record

Record the question, evidence considered, uncertainties, contextual assumptions and chosen action. Months later, the organisation can compare what it expected with what happened.

Decision records create institutional learning. Without them, schools can repeat the same evidence search because the reasoning behind earlier choices disappeared with staff turnover.

34. Link Brokering to Implementation

Evidence selection is not the end. Once an approach is chosen, the organisation needs implementation support: training, time, materials, coaching, data and feedback.

A broker should therefore identify implementation conditions in the evidence and hand the decision forward with those conditions attached. Otherwise a promising intervention may be stripped of the support that made it promising.

35. Link Brokering to Evaluation

If a school adopts a research-informed practice, define what local evidence will test whether it is working here. That may include implementation data, student work, attainment, attendance, workload or other outcomes relevant to the decision.

Research informs the starting hypothesis. Local evaluation updates it.

36. Cross-Domain Comparison: Clinical Evidence Translation

Healthcare systems use guideline developers, evidence syntheses and knowledge translation structures to connect research with practice. Clinicians still exercise judgement because evidence must be applied to individual patients and local conditions.

The educational lesson is not to medicalise teaching. It is to recognise that a mature evidence system needs structures between research production and frontline decisions.

37. Cross-Domain Comparison: Intelligence Analysis

An intelligence analyst combines sources of different reliability, distinguishes observation from inference, preserves uncertainty and briefs a decision-maker who must act before perfect information exists.

Knowledge brokers face a gentler version of the same epistemic problem: evidence is incomplete, sources differ, and the user still needs a decision. Structured uncertainty is better than confident oversimplification.

38. Cross-Domain Comparison: Engineering Standards

Standards bodies transform large technical evidence bases into specifications that practitioners can use, while maintaining revision processes as knowledge changes.

Education rarely needs standards at the same level of prescription, but the idea of traceable synthesis, versioning and scheduled review is highly transferable.

39. Example: Choosing a Reading Intervention

A school sees weak reading outcomes and considers three commercial programmes. Instead of comparing marketing brochures, the broker clarifies the learner group, reading component, staffing capacity and timetable. They search independent syntheses, inspect the population and implementation conditions, compare cost and dosage, and identify what local data should be collected during rollout.

The final recommendation may still involve judgement. The difference is that the judgement is connected to an explicit evidence chain.

40. Example: A Research Finding That Does Not Travel Cleanly

A study reports strong outcomes from intensive coaching. A school wants to replicate it but has one coach for eighty teachers and no protected observation time.

The broker should not simply say the intervention is evidence-based. The relevant evidence includes the support conditions. The school may need a different model, staged rollout or a decision not to adopt until capacity exists.

41. Failure Mode: PDF Dumping

Staff receive a folder of reports and are told to become evidence-informed. Information volume rises. Decision quality does not.

Access without curation can shift the search burden to the people with the least time to carry it.

42. Failure Mode: One-Study Certainty

A single exciting study becomes the basis for a system-wide claim. Replication, synthesis, implementation conditions and alternative evidence are ignored.

Brokerage should widen the evidence base before narrowing the decision.

43. Failure Mode: Translation That Becomes Advocacy

The broker becomes attached to a programme and begins selecting only supportive evidence. Uncertainty disappears from summaries. Questions are treated as resistance.

At that point brokerage has turned into promotion. Preserve the ability to change the recommendation when the evidence changes.

44. Failure Mode: Research as a Trump Card

“Research says” is used to end professional discussion. Local constraints, student needs and competing evidence are dismissed.

Evidence should improve reasoning, not terminate it.

45. A Practical Knowledge Brokerage Stack

• Decision: define what must be decided and by when.
• Question: convert the problem into an answerable evidence question.
• Search: locate relevant syntheses and primary evidence.
• Appraise: judge credibility, design and claim strength.
• Synthesise: identify patterns across the evidence base.
• Translate: make findings usable without deleting uncertainty.
• Contextualise: compare evidence conditions with local conditions.
• Combine: integrate research with professional knowledge and local data.
• Options: make trade-offs and alternatives explicit.
• Provenance: preserve source trails and conflicts of interest.
• Version: date the recommendation and define an update route.
• Implement: carry evidence-informed conditions into rollout.
• Evaluate: collect local evidence about what happened.
• Learn: update the recommendation and organisational memory.

46. What OECD and IES Work Adds

The OECD’s current Evidence Web for Education defines knowledge mobilisation broadly as intentional efforts to increase the use of research in education policy or practice, from synthesis and translation to dialogue and organisational conditions. It describes knowledge or evidence intermediaries as actors that help practitioners and policy-makers engage with research and apply it to decisions.

The OECD’s 2025 report Everybody Cares About Using Education Research Sometimes emphasises that knowledge mobilisation includes multiple activities and that intermediary work takes different forms across organisations and systems. A companion chapter maps activities across evidence generation, dissemination, partnerships, capacity building, organisational support and system support.

Earlier OECD analysis of who facilitates research use in education systems documents roles played by ministries, universities, teacher-education institutions and formal brokerage agencies. The U.S. Institute of Education Sciences has likewise studied research use in education, including organisational routines, knowledge exchange and the infrastructure needed for evidence to travel.

47. Missing-Node Scan: Questions a Mature Evidence System Should Ask

• What decision are we actually trying to make?
• Are we using the best available synthesis or the easiest study to find?
• Can users trace important claims to their sources?
• Have we separated evidence strength from local relevance?
• What uncertainty has been lost during translation?
• Which implementation conditions travelled with the evidence?
• What professional or local knowledge must be combined with research?
• Who funds or benefits from the evidence source?
• When was the recommendation last updated?
• What local evidence will tell us whether the decision is working here?
• Are we building evidence-use capability or dependence on one expert?
• Can practitioners send important questions back toward researchers?

48. The Return Path

A school leader begins with a familiar request: “Find me the best programme.”

The broker changes the question. Which students? Which outcome? Compared with current practice? What staffing and time are available? What does the wider evidence base say? Which studies resemble our context? What conditions made the strongest results possible? What are the costs and opportunity costs? What evidence should we collect if we adopt?

The answer is no longer a product name. It is a decision model. One option has stronger evidence but requires coaching capacity the school does not have. Another has smaller average effects but fits the current timetable and has lower implementation burden. A third lacks enough independent evidence to justify adoption.

The leader still has to decide. That is the point. Knowledge brokering does not remove judgement. It makes judgement more informed, more traceable and easier to revise when new evidence arrives.

Knowledge brokering works when evidence stops being something schools merely receive—and becomes something they can interrogate, contextualise, act on and update.