eduKateSG Learning Node Series · 0105

How Teacher Workload Works | Protect Teaching Quality by Designing the Work, Not Just Asking Teachers to Cope

A teacher can have the same number of lessons in two different weeks and experience two completely different workloads.

In one week, the lessons are prepared, meetings are purposeful, marking is sampled intelligently, student issues are routed to the right people, and there is enough uninterrupted time to think. In the other, the timetable looks almost identical, but the teacher is interrupted every twenty minutes, must duplicate the same information across several systems, attends meetings without decisions, marks work that nobody uses, responds to messages late at night, covers an absent colleague and rewrites tomorrow’s lesson because today’s evidence arrived too late.

The difference is not commitment. It is work design.

Teacher workload is not simply the number of hours a teacher works. It is the amount, intensity, fragmentation, timing and usefulness of the work that must be completed with finite human attention.

The 50-Second Read

• Workload has at least five dimensions: total time, work intensity, task complexity, fragmentation and emotional demand.
• Two teachers can work the same hours but experience very different cognitive loads because one week contains more interruptions, switching and unresolved decisions.
• Teaching quality depends on protecting high-value work: noticing students, planning, explaining, questioning, feedback, diagnosis and adjustment.
• Administrative work is not automatically waste. The question is whether information collected changes a useful decision.
• Marking should be designed around learning information, not visible volume.
• Meetings should produce decisions, coordination or learning; otherwise they consume scarce professional attention.
• New initiatives create implementation cost even when each initiative is individually sensible.
• Recovery is part of productive capacity, not a reward after productivity.
• Automation should remove clerical friction before it removes professional judgement.
• A strong workload system asks what can be stopped, simplified, standardised, delegated, batched, automated or protected.

Canonical Owner Boundary

This Learning Node owns the design of teacher work as a finite-capacity system. How Organisational Slack Works owns spare organisational capacity for recovery and adaptation. How Initiative Overload Works owns the accumulation of change programmes. How Instructional Time Works owns the conversion of scheduled lesson minutes into usable learning opportunity. This page owns the adult work system around teaching: what work exists, how it arrives, how it competes, and whether it protects or erodes instructional quality.

1. Start With the Work, Not the Personality

When workload becomes unsustainable, organisations often reach first for individual coping strategies: prioritise better, become more resilient, use a planner, answer email more efficiently. Those can help, but they are downstream interventions.

If the work system continuously admits more tasks than available capacity can absorb, personal organisation only changes which task becomes late.

The first question is therefore architectural: what work are we asking teachers to do, why does each task exist, what decision does it support, and what happens if it is removed?

2. Hours Are Only the Outer Shell

The OECD’s TALIS 2024 analysis of the demands of teaching shows why workload cannot be reduced to classroom hours. Across OECD education systems, teaching is the largest single activity but accounts for about 43 per cent of full-time teachers’ self-reported working time. Planning, marking, administration, collaboration, student support and other responsibilities occupy the rest.

That matters because reducing one visible teaching hour does not automatically reduce total work if the surrounding tasks expand to fill the space.

3. Singapore Shows the Shape of the Problem

The OECD’s TALIS 2024 Singapore country note reports that full-time lower-secondary teachers in Singapore self-reported 47.3 total working hours per week, compared with an OECD average of 41. They reported 17.7 hours of teaching, 8.2 hours preparing lessons, 6.4 hours marking and correcting work, and 4 hours of administrative work.

The interesting point is not that one category is inherently too high. It is that the workday is a portfolio. Every hour assigned to one activity has an opportunity cost somewhere else.

4. Workload Has a Composition

Imagine two 47-hour weeks. Week A contains long blocks of lesson preparation and teaching. Week B contains the same total hours but dozens of short administrative requests, messages, meetings, changes, follow-ups and interruptions.

The totals match. The cognitive experience does not.

Workload should therefore be audited by composition: teaching, planning, assessment, feedback, student care, communication, collaboration, administration, professional learning, co-curricular work, change implementation and recovery.

5. Fragmentation Creates Hidden Cost

A teacher who receives a message while planning must stop, read, classify, decide, respond or defer, then rebuild the mental state needed for the lesson. The reply may take ninety seconds. The interruption costs more.

Fragmentation is therefore a multiplier on workload. Ten small tasks can consume more professional attention than one coherent task of the same clock duration.

6. Task Switching Is Not Free

Teaching already requires fast switching: explain, watch, question, diagnose, regulate behaviour, adjust pace, remember who has not answered, and decide whether the misconception is individual or general. Adding avoidable switching outside the classroom compounds the demand.

A workload redesign can therefore improve quality without removing a single major responsibility simply by batching similar work and protecting uninterrupted planning windows.

7. Work Intensity Matters

Forty hours at a steady pace is different from forty hours where nearly every task is urgent, observed, emotionally charged or deadline constrained. TALIS explicitly distinguishes total working time from work intensity because the experience of demand changes when people repeatedly work at speed or under pressure.

A school that measures only hours can miss a workload problem that lives inside those hours.

8. Emotional Labour Is Real Work

A teacher may spend twenty minutes with a distressed student, call a parent about a difficult situation, then walk directly into a lesson and produce calm, attentive instruction for thirty other students.

The timetable records forty minutes. The nervous system records something else.

Workload design should recognise that student care, conflict, safeguarding and emotionally difficult conversations create recovery requirements even when they do not generate visible paperwork.

9. Protect High-Value Teacher Work

Not every teacher task has equal leverage on learning. Some work sits close to the student: analysing misconceptions, planning explanations, designing examples, asking questions, giving feedback, adapting instruction and building relationships.

Other work may be necessary but farther away from the instructional core. Workload design should protect the work with the greatest educational value from being crowded out by lower-value coordination and reporting.

10. Administrative Work Needs a Decision Test

Administration is not automatically bureaucracy. Attendance records matter. Safeguarding records matter. Assessment records can matter. The better question is: what decision will this information change?

If a data field is collected repeatedly, stored carefully and never used, the organisation has created work without creating information value.

Apply a decision test: who reads this, what will they decide differently, how quickly must they know, and could the system obtain the same signal more cheaply?

11. Duplicate Entry Is Pure Friction

When the same student information must be entered in a spreadsheet, learning platform, messaging system and reporting portal, no new educational value is created by the second, third or fourth entry.

Duplicate entry is a useful workload target because eliminating it can reduce effort without reducing service.

12. Marking Should Be Designed Backwards From Learning

A teacher can spend three hours writing comments that students glance at for twenty seconds. That is not evidence that feedback is unimportant. It is evidence that the feedback system may be poorly designed.

Start with the instructional decision: what does the teacher need to know, what does the student need to know, and what action should follow? Sometimes that requires detailed written feedback. Sometimes whole-class feedback, coded errors, verbal conferencing, self-checking or targeted re-teaching produces more learning per minute.

13. Visible Labour Is Not the Same as Effective Labour

Schools can accidentally reward workload theatre: every page marked, every box filled, every meeting attended, every initiative documented. These activities are visible and therefore easy to audit.

Thinking is harder to see. A teacher who spends forty focused minutes redesigning one explanation may create more learning than another who spends two hours producing documentation.

Do not confuse auditability with educational value.

14. Meetings Need an Output

A meeting should do at least one of four jobs: make a decision, coordinate interdependent work, solve a problem that requires multiple perspectives, or create shared professional learning.

If the purpose is merely to transmit information, another channel may be cheaper. If no decision can be made, no work coordinated and no learning produced, the meeting is occupying capacity without changing the system.

15. Every New Initiative Carries Shadow Work

A new programme rarely costs only the training session. It creates planning, coordination, documentation, troubleshooting, data collection, parent communication, resource preparation and follow-up.

This is why a school can approve five individually reasonable initiatives and create one collectively unreasonable workload.

Before adding an initiative, estimate its shadow work and identify what will stop, shrink or wait.

16. Workload Has a Peak Problem

Average workload can look acceptable while peak workload is destructive. Examination marking, report writing, parent conferences, curriculum planning and events can cluster into the same fortnight.

Capacity planning therefore needs a calendar view. Move predictable work before the peak instead of asking every adult to absorb the collision.

17. Workload Has a Variability Problem

Unexpected absence, student incidents, technology failure and urgent family matters cannot be scheduled neatly. A system running at one hundred per cent planned utilisation has nowhere to put variation.

This is where workload meets organisational slack. Spare capacity looks inefficient only until the first disruption arrives.

18. Recovery Is Part of Capacity

Human systems cannot operate like storage shelves. Attention degrades. Patience thins. judgement becomes noisier. Creative planning becomes harder when cognitive and emotional resources are repeatedly depleted.

Recovery is not a luxury added after all useful work is complete. It is one of the conditions that makes tomorrow’s useful work possible.

19. UNESCO Places Working Conditions Inside the Teacher Supply Problem

UNESCO’s Global Report on Teachers frames teacher shortages not simply as a recruitment issue but also as a professional-conditions and retention issue. The report calls for better working conditions, professional development and greater teacher participation in decision-making.

That is an important systems insight: if the work itself is badly designed, recruiting more people into the same design does not remove the underlying leak.

20. Teacher Voice Is a Workload Sensor

Leaders cannot see every friction point from the timetable. Teachers know where systems duplicate work, where deadlines collide, where data is requested but unused, and which processes create twenty minutes of effort for thirty seconds of value.

Teacher consultation is therefore not only a morale practice. It is operational sensing.

21. Standardise the Repetitive, Preserve Judgement

Templates, shared resources, common assessment structures and routine workflows can reduce repeated invention. But standardisation becomes harmful when it reaches into decisions that genuinely require professional judgement.

The design rule is simple: standardise low-value variation; preserve high-value adaptation.

22. Automation Should Remove Clerical Friction First

Technology is most useful when it removes copying, formatting, routing, scheduling, searching and repetitive classification. These are areas where reducing effort usually does not reduce professional quality.

Automating judgement is different. A machine can help surface patterns, but decisions about a learner’s need, context, readiness or welfare often require human interpretation. The first automation target should be friction around expertise, not expertise itself.

23. Delegation Needs Authority as Well as Tasks

Poor delegation moves the task but keeps every decision with the teacher. The teacher still answers all questions, approves every step and repairs every exception.

Good delegation transfers a bounded decision right along with the work so the coordination cost does not exceed the labour saved.

24. Novice Teachers Need a Different Workload Shape

Experienced teachers possess routines, examples, lesson structures and pattern recognition that compress work. A novice may need much longer to prepare the same lesson and may spend more mental effort managing the classroom while teaching it.

A fair workload is therefore not always an identical workload. Early-career support, mentoring, shared planning and protected preparation can compensate for the fact that the same visible task may have very different internal cost.

25. Cross-Domain Comparison: Aviation

Aviation does not assume that skilled professionals can safely absorb unlimited interruptions and procedural complexity. Checklists, standard operating procedures, role clarity and sterile-cockpit rules exist because human attention is finite.

Education is not aviation, but the transferable principle is strong: protect attention around consequential work and remove unnecessary variation around routine work.

26. Cross-Domain Comparison: Hospitals

Clinical workload is shaped not only by patient count but by case complexity, interruptions, documentation, handovers and coordination. A ward can become unsafe before anyone reaches an official maximum number of hours.

Teacher workload behaves similarly. Headcount and lesson count are useful signals, but they are not complete workload measures.

27. Cross-Domain Comparison: Software Engineering

Software teams distinguish feature work from technical debt, interruptions, maintenance and incident response. A team can appear fully staffed while most capacity is consumed keeping existing systems running.

Schools need the same visibility. Teaching capacity is not whatever remains after everything else is scheduled.

28. A Practical Teacher Workload Audit

• List the work: teaching, planning, feedback, meetings, administration, communication, student support, activities and change projects.
• Measure peaks: map when major tasks collide across the term.
• Find duplication: identify information entered or communicated more than once.
• Find fragmentation: count avoidable interruptions and handoffs.
• Apply the decision test: ask what each data collection or report changes.
• Protect instructional work: ring-fence planning, diagnosis and high-value feedback.
• Stop before adding: require every major new initiative to identify displaced work.
• Standardise wisely: share templates and routines where variation adds little value.
• Automate clerical friction: reduce copying, formatting, routing and scheduling first.
• Plan recovery: avoid sustained operation at peak intensity.
• Recheck outcomes: verify that workload reduction did not simply transfer the work to somebody else.

29. Failure Mode: The Resilience Answer

Teachers report overload. The response is a wellbeing workshop.

The workshop may be useful, but if the work system remains unchanged, the intervention asks individuals to become better at absorbing the same structural demand.

Repair: redesign the demand before training people to tolerate it.

30. Failure Mode: Efficiency Creates More Work

A new system saves thirty minutes, so another task is added immediately. The teacher never experiences the benefit because every efficiency gain becomes new utilisation.

Repair: decide in advance whether saved capacity will reduce overload, increase teaching quality or support a genuinely higher-priority task.

31. Failure Mode: Workload Is Reduced on Paper

A form is removed, but the information is still requested by email. A meeting is cancelled, but every participant now writes an individual update. A policy becomes optional, but staff still believe it is expected.

Repair: track whether the underlying work disappeared, not whether the named process disappeared.

32. Questions for a Workload Review

• Which five tasks consume the most teacher time?
• Which five create the most interruption?
• Which five create the most emotional demand?
• Which recurring data collections actually change decisions?
• Where is the same information entered twice?
• Which meetings could become asynchronous?
• Which work peaks collide during the year?
• Which tasks require a teacher’s expertise and which do not?
• What high-value instructional work is regularly postponed?
• What should stop before the next initiative starts?
• Where could a shared resource remove repeated invention?
• What capacity is reserved for unpredictable events?
• How does workload differ for novice and experienced teachers?
• Do efficiency gains reduce overload or simply invite more demand?

33. The Missing-Node Scan

If a school has capable teachers, strong curriculum and reasonable staffing but instructional quality still feels fragile, the missing node may be workload architecture.

Look for clues: teachers planning late at night, repeated duplicate entry, meetings that end without decisions, marking policies measured by volume, new initiatives without subtraction, constant messaging, no protected preparation, peaks where several major deadlines collide, and experienced staff spending increasing time on coordination rather than teaching.

These are not isolated productivity problems. They are signals that professional attention is being allocated badly.

34. The Return Path

Return to the two teachers with identical timetables.

The first spends the afternoon analysing student work, redesigning tomorrow’s explanation and calling one family whose child needs help. The second spends the same afternoon entering duplicate data, attending an information-only meeting, replying to fragmented requests and completing documentation that nobody uses before starting lesson preparation at home.

Both are busy. Only one work system protects the activity that students actually need.

Teacher workload works when schools treat attention as a scarce educational resource: remove friction, protect judgement, control peaks, and spend teacher time where teacher expertise changes learning.