Turn a supervisor meeting into clearer research decisions and accountable next steps
A graduate research supervisor meeting connects progress, evidence, interpretation, obstacles, decisions, ethics, milestones and the next piece of work. English matters because a student must distinguish data from inference, discussion from approval and advice from an agreed action. This guide helps students prepare a concise agenda, communicate uncertainty and keep an accurate follow-up record. It does not replace programme rules, research ethics approval, specialist methods advice or the supervisor’s academic judgement.
Choose the route closest to your task, or read straight through for the complete system.
People search for research supervisor meeting, PhD progress meeting, thesis meeting agenda and questions to ask a supervisor because research grows through recurring decisions rather than one perfect conversation.
A useful meeting pack states the research aim, work completed since the last meeting, evidence produced, interpretation, unresolved question, decision needed, ethical or practical risk, next milestone and requested feedback.
Clear English protects the research relationship. “We discussed it” is not the same as “the supervisor approved it”, and “the results look promising” is not the same as evidence that supports a conclusion.
This guide was checked on 14 October 2026. Programme milestones, reporting systems, committee roles, meeting frequency and ethics procedures vary. Follow the current graduate-school, faculty, supervisor and ethics-board requirements for the actual project.
Did You Know? The NUS Graduate School research-progress guidance, checked on 14 October 2026, requires graduate research students to submit a progress report every semester and asks for a clear, structured account of progress. Related NUS thesis-advisory guidance uses meetings and recorded progress to support candidature. The practical discovery is that a good supervision conversation is not only memorable; its claims, decisions and actions should be reconstructable later.
Find the section you need
Define the meeting purpose · 2 chapters
Build a progress packet · 2 chapters
Run the discussion · 2 chapters
Record agreements · 2 chapters
Transfer the result · 3 chapters
1. Connect the meeting to candidature milestones
Back to contentsA useful agenda begins with the programme’s real calendar, not only the student’s immediate frustration.
List upcoming progress report, proposal, ethics, confirmation, committee or submission milestones and identify what this meeting must unlock.
The student needs agreement on a pilot design before an ethics amendment and schedules that decision first.
What to check
- Check milestone
- Check deadline
- Name dependency
- Name decision
Apply this chapter to a stalled recruitment plan. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: A routine supervision meeting does not replace a formal committee or programme approval.
Learning transfer: Milestone mapping supports long projects.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
Supervisors need enough context to think, but a diary of every action hides the pattern.
Summarise the aim, period covered, work completed, evidence produced, change from plan, current obstacle and requested decision. Link detailed files separately.
The first page shows three completed interviews, a coding issue and one question about revising the sampling plan.
What to check
- State aim
- Quantify work
- Name deviation
- Ask decision
Apply this chapter to a disputed method choice. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: Activity is not automatically progress; connect work to the research question and plan.
Learning transfer: Executive summaries support workplaces.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
3. Separate evidence, interpretation and confidence
Back to contentsResearch discussion becomes unreliable when observation, explanation and certainty merge into one sentence.
Use a three-part note: what the data or source shows, what you currently infer and what would test that inference. State limitations.
“Four participants raised access barriers; this may indicate a recruitment bias; I propose checking the non-response pattern.”
What to check
- Show evidence
- Label inference
- State confidence
- Propose test
Apply this chapter to a draft needing focused feedback. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: A confident presentation cannot repair weak evidence or an unsuitable method.
Learning transfer: Evidence language supports science.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
4. Show what changed since the previous meeting
Back to contentsA research plan evolves, but unexplained drift makes supervision and reporting difficult.
Compare last agreed action with actual outcome. For each deviation, state reason, consequence and proposed repair.
A dataset arrived two weeks late, so the student proposes moving a literature-synthesis task forward while analysis waits.
What to check
- Quote action
- State outcome
- Explain change
- Propose repair
Apply this chapter to an ethics-sensitive protocol change. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: Do not rewrite history to make the original plan look successful.
Learning transfer: Change reporting supports project management.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
“I am stuck” invites sympathy but not necessarily a useful academic choice.
Name the obstacle, evidence, options considered, trade-off and decision or feedback needed. Ask one question at a time.
The student compares two coding frameworks and asks which better fits the stated research question and available reliability checks.
What to check
- Name obstacle
- List options
- Show trade-off
- Ask one decision
Apply this chapter to a stalled recruitment plan. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: A supervisor may need time, data or another expert before advising; not every question can be settled immediately.
Learning transfer: Question framing supports consultations.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
A meeting can be wasted when the supervisor receives an unlabelled draft too late to review it.
Send the agreed material with version, purpose, requested feedback and realistic lead time. Highlight the sections that need attention.
The student sends “Methods v0.4” and asks for comments on sampling logic, not a vague review of forty pages.
What to check
- Name version
- State purpose
- Mark sections
- Respect lead time
Apply this chapter to a disputed method choice. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: Sending a file does not mean it was reviewed; confirm the meeting expectation.
Learning transfer: Review requests support collaboration.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
Consent, privacy, access, storage and protocol changes can be hidden inside what looks like a method detail.
Flag any change that may affect ethics approval or data governance and use the authorised review route before implementing it.
A proposed new interview platform is paused until security and consent implications are checked.
What to check
- Check approval
- Check consent
- Check storage
- Check change
Apply this chapter to a draft needing focused feedback. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: Supervisor discussion alone may not authorise a protocol amendment or data transfer.
Learning transfer: Governance language supports responsible innovation.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
A supervision meeting should allow evidence-based challenge without turning every difference into conflict.
Restate the recommendation, name the evidence or constraint creating concern, ask about the underlying criterion and compare options.
The student says, “I understand the preference for a larger sample; may we compare that with the recruitment and ethics limits?”
What to check
- Restate view
- Name evidence
- Ask criterion
- Compare options
Apply this chapter to an ethics-sensitive protocol change. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: Disagreement does not remove role responsibilities or programme procedures.
Learning transfer: Constructive challenge supports teams.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
A supervisor may brainstorm, recommend, agree, defer or formally approve, and these verbs have different consequences.
During the meeting, summarise what you heard and ask which status applies. Record any further authority or document required.
The supervisor supports an analysis idea but says the thesis advisory committee must review the candidature change.
What to check
- Name status
- Confirm wording
- Find authority
- Record condition
Apply this chapter to a stalled recruitment plan. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: Friendly discussion or silence should not be represented as approval.
Learning transfer: Decision-status literacy supports governance.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
A productive discussion can still dissolve when the next action has no owner or observable completion.
Read back each action with owner, deliverable, date, dependency and evidence of completion. Schedule the next checkpoint when appropriate.
The student will send a revised sampling memo by Friday; the supervisor will comment on its two decision points by the following meeting.
What to check
- Name owner
- Name output
- Set date
- Name dependency
Apply this chapter to a disputed method choice. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: A date should be realistic and updated transparently when circumstances change.
Learning transfer: Action tracking supports operations.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
11. Write the follow-up while meaning is fresh
Back to contentsA concise written record protects both learning and continuity across long projects.
Within the agreed practice, send decisions, open questions, actions and next meeting purpose. Invite correction and update the formal progress record separately.
The follow-up labels one decision, two actions and an ethics question still awaiting the authorised route.
What to check
- Record decisions
- Record actions
- Mark open items
- Invite correction
Apply this chapter to a draft needing focused feedback. Write the purpose, reader, verified facts, uncertainty, boundary and next action before polishing the language.
Create an ordinary case, a difficult case and a missing-information case. Preserve the governing meaning while changing detail, tone and support for the situation.
Ask an independent reader or listener to explain what they understood, what they would do next and where they hesitated. Revise, keep the evidence and compare both versions.
Boundary: The note is not a substitute for formal progress reports, ethics records or committee minutes.
Learning transfer: Meeting records support institutional memory.
Write one sentence that states the evidence, one that states its limit, and one that states the next responsible action. Then test each noun, number, condition and source against the current task and verified source. Replace vague confidence with a traceable reason.
How to build this kind of English before the job title arrives
Back to contentsStudents do not need professional authority to practise the underlying language. They can learn to define scope, annotate conditions, compare versions, rebuild a timeline, match a number to its unit and explain uncertainty without embarrassment. Parents can ask calm questions: What decision is this document supporting? Which sentence carries the strongest claim? What would make that claim false? Which source would you open next?
Use a three-pass routine. First, map the document: title, owner, audience, date, sections and decision. Second, trace one claim from source to conclusion. Third, explain the limit and next action. This routine strengthens comprehension, science, mathematics, humanities and workplace readiness because it turns English into an evidence tool.
A student should never imitate professional authority. The useful goal is to recognise when a text needs a qualified reviewer, current rule or official decision. Knowing when to pause is part of literacy.
Purpose
Choose a defensible sampling adjustment before an ethics amendment.
Progress
Completed work and deviation from the last plan summarised.
Evidence
Observed recruitment pattern separated from interpretation.
Decision
Options and trade-offs discussed; required authority identified.
Actions
Owner, output, date and next checkpoint written in the follow-up.
Aim
Current research purpose.
Milestone
Programme event the work supports.
Progress
Completed work since last meeting.
Evidence
New data, sources or analysis.
Interpretation
Current meaning and confidence.
Deviation
Change from agreed plan.
Obstacle
Specific unresolved problem.
Options
Plausible routes and trade-offs.
Decision
Feedback or choice needed.
Governance
Ethics, data or authority issue.
Actions
Owner, output and date.
Record
Follow-up and formal reporting route.
Use this roadmap to see how the chapters connect. Each decision names a purpose, action, observable example and limit rather than repeating a generic language rule.
1. Connect the meeting to candidature milestones
A useful agenda begins with the programme’s real calendar, not only the student’s immediate frustration. The practical move is to list upcoming progress report, proposal, ethics, confirmation, committee or submission milestones and identify what this meeting must unlock.
Test the chapter through this example: The student needs agreement on a pilot design before an ethics amendment and schedules that decision first. Keep the boundary visible: A routine supervision meeting does not replace a formal committee or programme approval.
2. Write a one-page progress snapshot
Supervisors need enough context to think, but a diary of every action hides the pattern. The practical move is to summarise the aim, period covered, work completed, evidence produced, change from plan, current obstacle and requested decision. Link detailed files separately.
Test the chapter through this example: The first page shows three completed interviews, a coding issue and one question about revising the sampling plan. Keep the boundary visible: Activity is not automatically progress; connect work to the research question and plan.
3. Separate evidence, interpretation and confidence
Research discussion becomes unreliable when observation, explanation and certainty merge into one sentence. The practical move is to use a three-part note: what the data or source shows, what you currently infer and what would test that inference. State limitations.
Test the chapter through this example: “Four participants raised access barriers; this may indicate a recruitment bias; I propose checking the non-response pattern.” Keep the boundary visible: A confident presentation cannot repair weak evidence or an unsuitable method.
4. Show what changed since the previous meeting
A research plan evolves, but unexplained drift makes supervision and reporting difficult. The practical move is to compare last agreed action with actual outcome. For each deviation, state reason, consequence and proposed repair.
Test the chapter through this example: A dataset arrived two weeks late, so the student proposes moving a literature-synthesis task forward while analysis waits. Keep the boundary visible: Do not rewrite history to make the original plan look successful.
5. Turn obstacles into decision-ready questions
“I am stuck” invites sympathy but not necessarily a useful academic choice. The practical move is to name the obstacle, evidence, options considered, trade-off and decision or feedback needed. Ask one question at a time.
Test the chapter through this example: The student compares two coding frameworks and asks which better fits the stated research question and available reliability checks. Keep the boundary visible: A supervisor may need time, data or another expert before advising; not every question can be settled immediately.
6. Prepare documents for close reading
A meeting can be wasted when the supervisor receives an unlabelled draft too late to review it. The practical move is to send the agreed material with version, purpose, requested feedback and realistic lead time. Highlight the sections that need attention.
Test the chapter through this example: The student sends “Methods v0.4” and asks for comments on sampling logic, not a vague review of forty pages. Keep the boundary visible: Sending a file does not mean it was reviewed; confirm the meeting expectation.
7. Discuss ethics and data governance explicitly
Consent, privacy, access, storage and protocol changes can be hidden inside what looks like a method detail. The practical move is to flag any change that may affect ethics approval or data governance and use the authorised review route before implementing it.
Test the chapter through this example: A proposed new interview platform is paused until security and consent implications are checked. Keep the boundary visible: Supervisor discussion alone may not authorise a protocol amendment or data transfer.
8. Use respectful disagreement as research work
A supervision meeting should allow evidence-based challenge without turning every difference into conflict. The practical move is to restate the recommendation, name the evidence or constraint creating concern, ask about the underlying criterion and compare options.
Test the chapter through this example: The student says, “I understand the preference for a larger sample; may we compare that with the recruitment and ethics limits?” Keep the boundary visible: Disagreement does not remove role responsibilities or programme procedures.
9. Distinguish advice, decision and approval
A supervisor may brainstorm, recommend, agree, defer or formally approve, and these verbs have different consequences. The practical move is to during the meeting, summarise what you heard and ask which status applies. Record any further authority or document required.
Test the chapter through this example: The supervisor supports an analysis idea but says the thesis advisory committee must review the candidature change. Keep the boundary visible: Friendly discussion or silence should not be represented as approval.
10. End with owners, outputs and dates
A productive discussion can still dissolve when the next action has no owner or observable completion. The practical move is to read back each action with owner, deliverable, date, dependency and evidence of completion. Schedule the next checkpoint when appropriate.
Test the chapter through this example: The student will send a revised sampling memo by Friday; the supervisor will comment on its two decision points by the following meeting. Keep the boundary visible: A date should be realistic and updated transparently when circumstances change.
11. Write the follow-up while meaning is fresh
A concise written record protects both learning and continuity across long projects. The practical move is to within the agreed practice, send decisions, open questions, actions and next meeting purpose. Invite correction and update the formal progress record separately.
Test the chapter through this example: The follow-up labels one decision, two actions and an ethics question still awaiting the authorised route. Keep the boundary visible: The note is not a substitute for formal progress reports, ethics records or committee minutes.
The completed route should let another person reconstruct what the communication is for, what evidence supports it, what remains uncertain and what happens next.
Each case asks the learner to compare a quick response with a version another person can actually use.
Case 1: Connect the meeting to candidature milestones
Start with the worked situation: The student needs agreement on a pilot design before an ethics amendment and schedules that decision first. Write the most tempting shortcut, then mark the missing reader or listener need, condition, source, authority or next action.
Repair the communication by applying this method: List upcoming progress report, proposal, ethics, confirmation, committee or submission milestones and identify what this meeting must unlock. Finish by stating why this boundary still matters: A routine supervision meeting does not replace a formal committee or programme approval.
Case 2: Show what changed since the previous meeting
Start with the worked situation: A dataset arrived two weeks late, so the student proposes moving a literature-synthesis task forward while analysis waits. Write the most tempting shortcut, then mark the missing reader or listener need, condition, source, authority or next action.
Repair the communication by applying this method: Compare last agreed action with actual outcome. For each deviation, state reason, consequence and proposed repair. Finish by stating why this boundary still matters: Do not rewrite history to make the original plan look successful.
Case 3: Discuss ethics and data governance explicitly
Start with the worked situation: A proposed new interview platform is paused until security and consent implications are checked. Write the most tempting shortcut, then mark the missing reader or listener need, condition, source, authority or next action.
Repair the communication by applying this method: Flag any change that may affect ethics approval or data governance and use the authorised review route before implementing it. Finish by stating why this boundary still matters: Supervisor discussion alone may not authorise a protocol amendment or data transfer.
Case 4: End with owners, outputs and dates
Start with the worked situation: The student will send a revised sampling memo by Friday; the supervisor will comment on its two decision points by the following meeting. Write the most tempting shortcut, then mark the missing reader or listener need, condition, source, authority or next action.
Repair the communication by applying this method: Read back each action with owner, deliverable, date, dependency and evidence of completion. Schedule the next checkpoint when appropriate. Finish by stating why this boundary still matters: A date should be realistic and updated transparently when circumstances change.
A parent, tutor, teacher or colleague can ask questions, but the learner should make the final revision in their own words and keep evidence of what changed.
Clear English is reliable when the reader can see not only what the sentence says, but also where it stops.
Boundary 1: Separate evidence, interpretation and confidence
Research discussion becomes unreliable when observation, explanation and certainty merge into one sentence. Before treating the wording as complete, identify an exception, missing fact, reader who may need another route and change that would force revision.
Apply the observable signal—“Four participants raised access barriers; this may indicate a recruitment bias; I propose checking the non-response pattern.”—and underline the exact words carrying scope, time, permission, comparison or uncertainty. A confident presentation cannot repair weak evidence or an unsuitable method.
Boundary 2: Prepare documents for close reading
A meeting can be wasted when the supervisor receives an unlabelled draft too late to review it. Before treating the wording as complete, identify an exception, missing fact, reader who may need another route and change that would force revision.
Apply the observable signal—The student sends “Methods v0.4” and asks for comments on sampling logic, not a vague review of forty pages.—and underline the exact words carrying scope, time, permission, comparison or uncertainty. Sending a file does not mean it was reviewed; confirm the meeting expectation.
Boundary 3: Distinguish advice, decision and approval
A supervisor may brainstorm, recommend, agree, defer or formally approve, and these verbs have different consequences. Before treating the wording as complete, identify an exception, missing fact, reader who may need another route and change that would force revision.
Apply the observable signal—The supervisor supports an analysis idea but says the thesis advisory committee must review the candidature change.—and underline the exact words carrying scope, time, permission, comparison or uncertainty. Friendly discussion or silence should not be represented as approval.
Boundary 4: Write the follow-up while meaning is fresh
A concise written record protects both learning and continuity across long projects. Before treating the wording as complete, identify an exception, missing fact, reader who may need another route and change that would force revision.
Apply the observable signal—The follow-up labels one decision, two actions and an ethics question still awaiting the authorised route.—and underline the exact words carrying scope, time, permission, comparison or uncertainty. The note is not a substitute for formal progress reports, ethics records or committee minutes.
A bounded statement is useful because a reader can test it, ask a focused question and act without treating confidence as proof.
Audience adaptation changes emphasis, vocabulary and next steps; it does not authorise a different truth.
Learner
A useful agenda begins with the programme’s real calendar, not only the student’s immediate frustration. This person may need a different level of detail or next action.
Preserve check milestone, check deadline, name dependency, name decision while adapting examples and support. Do not change the underlying fact, decision or limit simply to make the short version easier.
Teacher or adviser
Research discussion becomes unreliable when observation, explanation and certainty merge into one sentence. This person may need a different level of detail or next action.
Preserve show evidence, label inference, state confidence, propose test while adapting examples and support. Do not change the underlying fact, decision or limit simply to make the short version easier.
Parent or colleague
“I am stuck” invites sympathy but not necessarily a useful academic choice. This person may need a different level of detail or next action.
Preserve name obstacle, list options, show trade-off, ask one decision while adapting examples and support. Do not change the underlying fact, decision or limit simply to make the short version easier.
Authorised reviewer
Consent, privacy, access, storage and protocol changes can be hidden inside what looks like a method detail. This person may need a different level of detail or next action.
Preserve check approval, check consent, check storage, check change while adapting examples and support. Do not change the underlying fact, decision or limit simply to make the short version easier.
Future reader
A supervisor may brainstorm, recommend, agree, defer or formally approve, and these verbs have different consequences. This person may need a different level of detail or next action.
Preserve name status, confirm wording, find authority, record condition while adapting examples and support. Do not change the underlying fact, decision or limit simply to make the short version easier.
Successful transfer preserves identity, source, status, uncertainty, boundary and responsible action while making the route easier for this reader to follow.
A polished paragraph or fluent conversation can still conceal an unclear purpose, missing condition, invisible assumption or stale source.
Purpose test
State in one sentence what the other person should understand, decide or do. Delete a detail only if its removal does not change that outcome or a necessary boundary.
Write the expected interpretation, the reader’s or listener’s actual interpretation, the difference and one precise revision. Keep the earlier version so the learning remains visible.
Condition test
Circle words such as if, unless, before, after, may, must and only. Add the missing role, date, version or exception beside each one.
Write the expected interpretation, the reader’s or listener’s actual interpretation, the difference and one precise revision. Keep the earlier version so the learning remains visible.
Independent-person test
Give the communication to someone who was not present. Ask what they think is true, what is still unknown and what action they would take next.
Write the expected interpretation, the reader’s or listener’s actual interpretation, the difference and one precise revision. Keep the earlier version so the learning remains visible.
Change test
Imagine that a deadline, interface, source, role or policy changes tomorrow. Mark every sentence needing review and name who will update it.
Write the expected interpretation, the reader’s or listener’s actual interpretation, the difference and one precise revision. Keep the earlier version so the learning remains visible.
Close reading is practical care: it keeps language from outrunning the audience, evidence or authority behind it.
Answer each prompt in a complete sentence tied to the current task, source, reader or authorised clarification. Write “unknown” when the available information does not support an answer.
- Check milestone? State the evidence, limit and narrow next action it supports.
- Check deadline? State the evidence, limit and narrow next action it supports.
- Name dependency? State the evidence, limit and narrow next action it supports.
- Name decision? State the evidence, limit and narrow next action it supports.
- State aim? State the evidence, limit and narrow next action it supports.
- Quantify work? State the evidence, limit and narrow next action it supports.
- Name deviation? State the evidence, limit and narrow next action it supports.
- Ask decision? State the evidence, limit and narrow next action it supports.
- Show evidence? State the evidence, limit and narrow next action it supports.
- Label inference? State the evidence, limit and narrow next action it supports.
- State confidence? State the evidence, limit and narrow next action it supports.
- Propose test? State the evidence, limit and narrow next action it supports.
Read or rehearse it aloud
Listen for sentences that postpone the answer, hide the actor, overload one step or imply certainty the source does not provide. Revise for meaning before tone.
Ask a real reader or listener
Invite a learner, parent, colleague or mentor to identify the purpose, strongest claim, important condition and next action. Treat hesitation as useful evidence.
Keep a small revision note
Record the source, date, change and reason. A short history helps the next writer avoid restoring an earlier ambiguity.
How much should I send before the meeting?
Follow the agreed supervision practice. A one-page snapshot plus clearly labelled documents and focused feedback requests is often easier to use than an unstructured file dump.
What if I have little progress?
Report the situation honestly: work attempted, evidence, obstacle, consequence, help needed and a realistic recovery plan.
Can I disagree with my supervisor?
Use evidence and respectful questions, understand role responsibilities, and follow programme processes for unresolved concerns.
Does supervisor advice count as ethics approval?
Not necessarily. Research ethics and data changes may require a separate authorised review or amendment.
Should I send meeting notes?
Use the agreed practice. A concise record of decisions, open questions and actions can prevent later ambiguity, but it does not replace formal reports.
Can AI summarise my research meeting?
Only with appropriate confidentiality and human verification. Do not upload sensitive or restricted research data to an unapproved system.
- NUS Graduate School Research Progress Report Information
- NUS Research Progress Report Guide
- NUS ISEP Thesis Advisory Committee
- NUS ISEP Thesis Advisor FAQs
- Why English? Group Projects and Useful Feedback
Cycle-sensitive facts were checked against current official pages on 14 October 2026. Always reopen the authority’s live page before a real decision.