GAC Academic English · ⁨GAC Tiếng Anh Học thuật⁩

GAC002 develops reading and writing skills through communicating, culture, leisure, travel, living things and technology. The skills include understanding general and academic texts, using vocabulary and dictionaries, constructing sentences and paragraphs, avoiding plagiarism, adapting text types, and planning a short essay.

Your teaching centre supplies the actual assessment brief and referencing rules. reading tasks 阅读任务, writing 写作 and coursework 平时作业 are useful practice formats; this guide does not establish official task weights, an official final essay 论文 format, or a rule that organisation always outweighs language accuracy. The six accompanying sheets are original local practice. Their scenarios, sources, word ranges and analytic marks are teaching conventions, not official GAC papers.

A practical message can contain several dates with different jobs. In the fictional Writing Desk message, Monday noon is the booking deadline; Wednesday is the session. A number is not an answer until you know what it refers to.

• Skimming 略读 gives an initial overview using headings and selected sentences.
• Scanning 查读 locates a target detail, such as a date, place or number.
• Close reading 精读 checks exact meaning, qualifications and relationships.

Choose a technique for the question. After scanning for a date, read the surrounding sentence closely. Skimming may suggest a text’s structure without revealing every qualification. A topic sentence 主题句 states a paragraph’s main point; it can occur at the beginning, later in the paragraph, or be implicit in the supporting material.

Worked example. The message says that unbooked visitors may request a walk-in slot, although a slot is not guaranteed. A faithful linked sentence is: “If you have not booked, you may request a walk-in slot, although availability is not guaranteed.” Keep both the condition and the limitation. Neither a request nor a possible slot is a confirmed appointment.

Context clues 上下文线索 include explanations, examples and contrasts. In the fictional craft-exchange note, “shared, not exclusive to one group” suggests that exclusive concerns restricted access. The note does not establish a national rule about how everybody makes a craft.

Word parts 词根词缀 can also help: bio- commonly relates to life and -logy to study. Treat them as clues, then check the particular word and sentence. An unfamiliar word may not affect the task; an essential unclear word may need a prompt lookup.

An English-English dictionary 英英词典 can supply numbered senses, usage labels, examples and a part of speech 词性. Choose the sense matching the context, rather than the first or last one automatically. Grammar belongs to the whole phrase: in “the reservoir”, reservoir is a noun; in “the large reservoir”, large follows the but is an adjective. Position after the alone is not a noun test.

Worked example. The sheet’s original classroom entry gives demonstrate two senses: show how something works by doing it, and give evidence for a claim. The craft task asks a participant to show a process, so the first sense fits. A display is one example, rather than proof of a universal cultural practice.

Build a paragraph around that supported point. Describe what a speaker shows, what visitors can ask and what the example can establish. An ending claiming that all people from a country use the same method goes beyond the supplied evidence.

A simple sentence 简单句 has one independent clause. A compound sentence 并列句 combines independent clauses. A complex sentence 复合句 contains a main clause and one or more subordinate clauses. Subordination 从属 places a clause within another sentence structure; it does not make its information automatically unimportant.

Connectives 连接词 express relationships: because gives a reason, although a concession, therefore a supported result and for instance an example. Choose a link that fits the actual evidence. Grammar can express a causal claim without proving it.

Worked example. In the fictional chess-club note, floor repairs explain the hall closure. Room 12 has eight seats, while ten students may attend. Their possible attendance does not cause the room’s seat count. A faithful sentence is: “Although ten students may attend, room 12 has only eight seats.” The coordinator needs confirmed numbers before choosing an arrangement.

A run-on sentence 流水句 incorrectly joins independent clauses. A comma splice joins them with a comma alone; use a suitable full stop, semicolon or conjunction. Check the purpose and reader too. A friendly class notice should make actions clear; an academic paragraph should support its claims precisely. Neither style permits invented attendance, approval or outcomes.

In the fictional visit brief, the meeting is at 08:10 and departure is scheduled for 08:20. Separate travel requires teacher permission. If rain closes the boardwalk, an indoor workshop is offered instead; the whole visit is not automatically cancelled. Adapt these facts into a practical passenger message without removing their conditions.

For an academic comparison, discuss what the brief actually supports: travelling together and permission-based individual arrangements. No comparative fare or emissions data are supplied, so “always cheapest and greenest” is unsupported.

Supporting sentences 支撑句 develop a point with relevant explanations, evidence or examples. A concluding sentence 结尾句 can connect the support to the point or lead to another paragraph. These functions need not always follow one fixed order. Cohesion 衔接 uses clear references, key terms and links so ideas are easy to follow. Unity 统一性 means the sentences serve the paragraph’s overall focus.

Check the reasoning. Supporting details may reveal a paragraph’s focus even if its topic sentence is deleted. That alone does not prove the topic sentence is useless or the paragraph is poorly built. An academic paragraph can appropriately start with a source or statistic when it serves the purpose; explain relevance and provide real source credit rather than enforcing an absolute rule against such openings.

Plagiarism 抄袭 presents another person’s words or ideas as your own. Academic source use must distinguish borrowed work from your contribution. Follow your centre’s instructions on referencing, permitted collaboration and assistance. Consequences depend on the institution and case; this guide does not claim one universal penalty.

• Quoting 引用原文 uses exact words, quotation marks and source credit.
• Paraphrasing 改写 uses your own wording and sentence structure, retaining the source’s meaning and limitations. Borrowed ideas still need credit.
• Summarising 概括 condenses a longer text to its main points, with source credit.

A citation 引用 identifies the source in the text; the reference list 参考文献表 provides the required full details. Record these while taking notes. Label exact words immediately so they cannot later be mistaken for your own wording. A two-column source/comment method can help; it does not make missing credit impossible.

Worked example using an original fictional source. Lin, M. (2026), Seedling observation notebook, classroom practice note, page 2, describes a seven-day comparison: shaded seedlings grew more slowly than those beside the window, but conditions were not controlled. This note is not published scientific research.

A faithful paraphrase is “One shaded tray grew more slowly than the window tray, but uncontrolled conditions limit causal conclusions (Lin, 2026, p. 2).” The citation gives credit; the qualification preserves the meaning. “Lin proves that shade always prevents growth” is inaccurate even with a citation. Translating a borrowed passage also does not remove the need for credit. The author-year-page format here is a local practice convention; use the style required for your actual assignment.

Analyse the question 审题 before choosing evidence: identify its command, topic and scope. Compare asks for relevant similarities and differences; evaluate asks for a judgement against suitable criteria. Discuss need not mean two equal sides. Read the specific brief rather than assuming a universal format.

Brainstorm then select 筛选: collect possible relevant ideas, then retain those supported by the available evidence. An Outline 提纲 names paragraph points, source support and connections to the answer. There is no fixed requirement for three body points or five paragraphs in every essay.

The introduction 引言 gives context and a thesis 论点 answering the question. A thesis should guide the argument at an appropriate level of certainty. It need not be controversial or exaggerate what the sources establish. Synthesis 综合 combines source ideas around your argument, retaining their credit and limits. The conclusion 结论 answers the question from the reasoning already developed, without adding unexplained new evidence or claiming unobserved outcomes.

Worked outline for the fictional device-free-room task.

1. Introduction: propose a limited optional trial while retaining a device-permitted room.
2. Distraction paragraph: 18 of 30 self-selected respondents reported interruptions (Rao, 2026); this is not representative proof or measured learning improvement.
3. Access paragraph: some students use devices for dictionaries, accessible text or necessary messages (Chen, 2026); connect this to voluntary participation.
4. Conclusion: propose recording use and inviting access/interruptions feedback before deciding whether to continue. Do not invent marks, attendance or trial results.

The two supplied notes are original fictional classroom sources. Combining their concern and access condition creates a reasoned limited proposal; copying each note in turn does not supply that synthesis. The accompanying sheet asks for an outline and a short credited paragraph, not an official full GAC assessment essay.

GAC007 develops listening and speaking through comparing spoken text types, applying listening strategies, taking notes and summarising, describing visuals and processes, giving presentations with relevant digital media, and encouraging interaction. The themes provide material for these skills rather than establish official task weights.

Your teaching centre provides the actual assessment brief. A listening test 听力测试 needs spoken input; a presentation 演讲 needs actual delivery. The six original local sheets provide written scripts, evidence records and rehearsal plans. They support preparation but do not replace reviewed recordings or assess pronunciation, one-pass comprehension, vocal fluency or measured delivery time. Their word ranges and analytic points are local teaching conventions, not official GAC specifications.

Compare purpose, audience, style and response expectations. A coordinator’s announcement can inform visitors about a time and request suggestions. A seminar contribution on the same topic can evaluate an observation, acknowledge its limit and recommend a check. The shared topic does not make their functions identical.

Speech functions 言语功能 describe what a statement does: inform, request, question, recommend, concede or qualify. A hedge 模糊限制语 such as might or tends to preserves uncertainty. Register 语域 concerns language suited to the situation; speakers can vary their language within a talk. Do not assume an informal aside will contain an exam answer or that academic speakers hedge every statement.

Worked example. In the original discussion-table script, students exchanged notes on three observed afternoons, but learning was not measured. “The table might support discussion” is a possibility, not proof of improved learning. A useful response acknowledges the exchange, asks what needs checking and invites a relevant contribution from a quieter participant. Silence alone does not show agreement.

Selective listening 选择性听力 focuses on information needed for the task while following the message. Before a rehearsal, identify whether the question needs a main point, a particular time, a condition or a comparison. Prediction can prepare likely vocabulary; it cannot supply a confirmed answer.

Discourse markers 语篇标记 such as first, however and finally signal relationships. Retain the actual content that follows them. An Inference 推断 is a supported conclusion from the text and context, rather than a plausible story added by the listener.

Worked example. A fictional garden speaker says that first-year funding was unusually generous, then says later budgets and harvest records are not supplied. The later-year results are unknown. Neither “later years failed” nor “funding ended and caused failure” is established. However introduces the limit; it does not automatically make the earlier statement false. Likewise, “popular in the nineties” alone does not establish current unpopularity.

Keep dates with their actions: Tuesday noon is the access-request deadline; Thursday 18:30 is the meeting in room 6. A replacement room is conditional, not already booked. After a permitted rehearsal, compare notes with the original script and clarify essential gaps. Do not present a reading-check score as a listening-proficiency score.

Notes should retain the hierarchy 层级 of a message: main proposal, supporting figures, decision limits and next actions. Abbreviations and arrows help when their meaning remains clear. Preserve important negatives, units, owners and permission conditions. Summarising 概括 selects those relevant points rather than repeating every sentence.

Worked example. The fictional workshop poll invited forty staff; thirty replied; eighteen respondents preferred Tuesday. That is not “all forty prefer Tuesday” and it does not establish availability. Room A seats twenty at sixty credits per session; room B seats thirty at ninety. Proposed two room-A sessions cost 120 credits before unknown extra fees. The manager must approve the plan before booking.

A simple digital comparison visual can show capacity and cost per session. Explain how those quantities bear on attendance, then identify the availability, fee and approval checks. Do not label a proposal as a booking or infer productivity gains the script never measured. Rehearse the actual spoken update with a timer; written length alone does not establish whether it fits a three-minute slot.

For a spoken review, prepare notes under finding, method, limitation and next action. A preference, an association and a causal effect are different claims. A qualification can determine what the speaker’s evidence supports, so keep it attached to the finding.

Worked example using original fictional data. Twenty volunteers chose a discussion-room arrangement: fourteen preferred quiet and six preferred music. This is 70% of those twenty respondents, not 70% of all people. The review measured no stress or health outcomes and did not recruit a random sample. It provides no treatment claim or clinical advice.

A faithful summary reports the preference counts, their scope and the next actions: feedback forms are due Friday noon, the meeting is Monday 15:00, and participants should disclose if they did not try both arrangements. The coordinator checks forms before a recommendation. Do not convert that future step into an approved permanent policy.

The separate lesson’s fictional diary example also distinguishes an observed association from a cause. These are language-practice records, not real clinical evidence. A summary should neither exaggerate a result nor dismiss the source merely because it has limitations.

Name the chart type 图表类型, the variables and units before interpreting a comparison. A categorical bar chart compares categories; it does not automatically show a continuous time trend. Keep totals distinct from percentages and measured records distinct from future predictions.

Worked chart example. Three fictional batches each contain 100 inspected packs. The counts set aside for rework are 8,5 and 11. Batch B has the lowest recorded count, and C the highest. The counts do not fall steadily, and set-aside packs are not automatically permanent waste. No causes or rework outcomes are supplied.

A process 过程 description retains actions, sequence and decision points. Inspect the seal; passing packs enter the ready tray; failing packs are labelled for rework; reworked packs require another inspection before release. Record first and later results separately. A passive voice 被动语态 sentence such as “The pack is labelled for rework” can focus on the action. Active wording may be clearer when responsibility matters. Changing voice does not change the order or conditions.

A comparison visual should label the quantity, unit, scope and period. Include relevant limits rather than hiding them in a confident conclusion. Distinguish what was recorded from what the group proposes next.

Worked example. One garden plot used 400 litres across four observation days before a trial and 250 litres across four different days after it began. The recorded difference is 150 litres, or 37.5% of the first total. Rainfall differed and planted area was not recorded, so the comparison does not isolate the trial’s effect or supply litres per square metre. It does not establish the same result everywhere.

For a formal presentation, show the two labelled totals, explain their limits and invite concerns about plant condition and missing records. A digital slide supports the spoken reasoning; it cannot prove pronunciation, audience understanding or rehearsal duration.

Encourage discussion with relevant questions, responses and an honest summary. An open question can invite explanation; a closed question can usefully confirm a detail or check agreement. Neither is always superior. Invite a quiet participant’s view instead of treating silence as approval. Summarise the proposed recommendation and unresolved checks, ask whether the group accepts it, then seek the supervisor’s required approval before continuing. A discussion that ends with an honest unresolved issue need not be a failure for everybody.

GAC014 develops critical listening, academic interaction, evaluation of arguments, oral presentations using relevant media and formal class seminars. Its education, medicine, legal-studies, political-studies, communications and future themes provide contexts for those skills. They do not prescribe a universal assessment weighting, ten-minute presentation or compulsory consensus in every task.

Your teaching centre supplies the actual brief, criteria and timing. A seminar presentation 研讨课展示 needs real delivery and interaction. The six accompanying sheets provide original written scripts, evidence records and rehearsal tasks. Their word ranges, analytic points, roles and timings are local practice conventions, not official GAC assessment rules. Printed-script checks do not replace reviewed audio or measure one-pass listening, pronunciation, vocal fluency or actual presentation duration.

Critical listening 批判性倾听 tracks a claim, its support, its limits and unanswered questions. A qualification 限定 such as in this sample narrows the scope. A strong summary retains the qualification rather than changing after into because of.

Before a rehearsal, choose note headings suited to the question. Claim, evidence, limitation and next check work for an argument. On a permitted second reading, verify numbers, negatives and conditions. A confident voice does not strengthen the evidence, and a sentence without a hedge is not automatically an exaggerated claim.

Worked example. Sixteen fictional volunteers used a teacher checklist and exchanged peer comments; twelve made a relevant revision. The sheet defines that revision as clarifying a claim or adding checklist-required support, not just correcting spelling. There is no checklist-only comparison or earlier revision rate. The record supports revisions after a combined activity, not an isolated causal effect of comments.

The remaining four did not meet the criterion; that does not establish they changed no words. A useful next comparison could keep the checklist and revision criteria consistent across groups. Limited evidence is not the same as no useful evidence, and a proposed comparison is not proof that the effect has been established.

An abstract 抽象的 discussion can concern access, fairness or evidence rather than one concrete object. Ask what the term means in the task before agreeing or disagreeing. Clarification should identify the uncertainty, then allow the speaker to explain it. The invented medical-communication examples here are language tasks, not clinical studies, health advice or authority to change care.

Worked example. Five fictional volunteers each score ten on a notice-comprehension quiz before reading. Their after scores are nine, nine, nine, nine and twenty-four. The group total rises from fifty to sixty, so its mean rises from ten to twelve. Only one individual score rises and four fall. An average increase therefore does not establish that most people improved. The quiz also measures wording comprehension, not health, attendance, consent or treatment.

A separate fictional date-label check changes from 1% to 2%. The difference is one percentage point 百分点; relative to the initial rate, it is a doubling or 100% increase. Keep the measured outcome and denominator attached to either description. Different collections do not isolate a cause. Neither these formatting rates nor the quiz scores can be treated as clinical risks.

For discussion, acknowledge the contribution and correct the specific distinction: “Do you mean the mean score or the number of people improving?” Follow with a useful question about access or the design of a suitable comparison. Do not invent missing data to make the answer sound decisive.

When evaluating an argument, ask whether the reasons support its conclusion. For a measurement, Validity 效度 concerns whether the evidence measures the intended thing. Reliability 信度 concerns consistency under suitable repeated conditions. It cannot be inferred merely because a result comes from an official-looking page.

Bias 偏差 can arise through recruitment, wording, measurement or selective reporting. Describe the possible mechanism and what could check it. A speaker’s role or personal stake may affect emphasis without making every statement false. This module’s invented procedure examples are reasoning practice, not actual law or statements of legal rights.

Worked example. An invented library wants all five hundred students’ views about its review procedure. Twenty voluntary website respondents answer “How helpful is our fair new review procedure?” Sixteen choose helpful. This is 80% of those respondents, not observed support from 80% of all students. Helpful is also not a measurement of every disputed case’s fairness.

Website-only self-selection may omit other views; fair and helpful frame the question positively. There is no repeat or comparable measurement, so reliability is unestablished. Do not say it would probably repeat or that every response is false. A stronger review could recruit through accessible routes, use neutral questions matching the intended judgement and examine consistency under suitable comparable conditions. Each change addresses a concern without guaranteeing perfect representation or validity.

Represent a viewpoint fairly before judging it. In the fictional community-hall case, quiet access matters to residents without a calm home space, while group activities also need a room. Neither viewpoint means banning the other activity everywhere. State the comparison criteria explicitly: access, noise, demand or setup effort.

A measured preference and a priority need different support. Twenty-four of forty voluntary respondents choosing quiet study is a bounded empirical result. The value placed on quiet access is a judgement about what should matter. The count alone does not settle how much weight to give either group’s needs.

Worked comparison. The hall serves two hundred residents. Scaling 24/40 to that population gives a calculation, not an observed census or proof that 120 residents want quiet-only use. A quiet-only option prioritises quiet access; a group-only option prioritises shared activities. A scheduled-use trial could explore both needs, but actual time-slot demand, noise and switching effort remain unmeasured.

Ask what evidence could address those questions and give a conditional recommendation. The original script approves no schedule. Record suggestions and priorities without attributing motives or claiming that an untried schedule has already solved the problem.

Digital media 数字媒体 should clarify the point you explain aloud. Use readable headings, contrast, clear labels and a text description where available. Colour alone should not carry the distinction. Include the quantity, denominator, scope and source. Use real source credit for real evidence and label these supplied classroom records as fictional; do not make them look like published evaluations.

Worked presentation plan. In the original notice example, eight of twelve first-day volunteers answer the event-time question correctly, compared with ten of twelve different second-day volunteers viewing a revised notice. That is two additional correct responses, not a doubling. Different people and days do not isolate the revision’s effect, and one event-time question does not measure overall understanding of the event.

1. First slide: the question and the notice change being considered.
2. Second slide: a clear table retaining 8/12 and 10/12, with group labels and the fictional source.
3. Final slide: visible comparison limits and a proposed next check.

Explain both the result and uncertainty aloud. Do not hide the qualification in unreadable print. Prepare brief speaker notes and a usable static backup. Time the actual rehearsal and ask a listener to restate the result and its limit; written length alone does not establish duration or comprehension.

A formal seminar connects preparation, evidence-based contributions, questions and an accurate synthesis around a focused task. Roles can include chair, prepared speakers and recorder; their exact allocation and timing depend on the actual brief. Protect discussion time instead of treating several consecutive speeches as the whole seminar.

Worked seminar brief. The fictional equipment log records forty requests and six double bookings across two weeks. Waiting time, abandoned attempts and unique-user counts are missing. No interface trial has occurred. The question is what a limited trial should measure before wider adoption, not whether an untested interface has already succeeded.

The operations speaker proposes recording conflicts; the access speaker proposes waiting times and abandoned attempts. A synthesis connects them: fewer recorded conflicts may coexist with access difficulties. The group explicitly agrees to recommend measuring conflicts and access difficulties, while duration and a numerical success threshold remain undecided. Keep specific proposed indicators distinct from the broader agreed categories.

Fielding a question 应对提问 means responding to the question actually asked. Clarify an unclear term, answer from available evidence and identify what remains unknown. A short restatement can help; do not repeat the whole talk to avoid the question. When the record covers a limited period, it cannot establish an unmeasured long-run outcome.

Close with what was agreed, what remains unresolved and what planning comes next. Invite correction from the recorder and participants. Silence is not consensus, a recommendation is not rollout approval, and criteria should be defined before observing trial results. An honest unresolved decision need not mean every participant failed the task. For a real seminar, conduct the session, time it and preserve an accurate record of the interaction.

GAC015 develops critical reading and writing, selection and evaluation of sources, logical arguments, annotated bibliography work, information literacy and an in-depth academic essay. Use the actual teaching-centre brief for assessment requirements, referencing style, length and weighting. This guide does not claim that one piece is the programme’s largest assessment or prescribe a fixed paragraph count.

An annotated bibliography 注释性参考书目 evaluates sources for a question; an essay 论文 connects their evidence within an argument. The six accompanying sheets use original fictional reports and cards as classroom practice. Their local analytic points and word targets are not official GAC specifications. Actual research coursework requires genuine in-depth sources that you have retrieved, read, evaluated and referenced accurately. Do not submit these fictional records as real educational, medical, legal or policy evidence.

Critical reading 批判性阅读 identifies what an author wants the reader to accept and tests the support. Analytical 分析性 reading separates the issue, claim, evidence, reasoning and limitations. An assumption 假设 is a premise the argument depends on; it may be explicit or unstated. Do not invent an assumption simply to criticise it.

Worked example. Fictional Report E compares thirty optional workshop attendees with thirty non-attendees. Their later means are seven and six out of ten. Earlier scores were not compared, and attendees reported more revision time. The recommendation that compulsory attendance will raise every score depends on unestablished causal and individual-benefit assumptions.

The later mean difference does not show that every attendee improved by one point. Individual before-and-after scores are missing. Voluntary participation and different revision time provide alternative explanations to investigate, not explanations proven by this text. A fair response can acknowledge possible workshop usefulness while rejecting a guaranteed universal effect. Critique the demonstrated inference rather than accuse the author of fabricating scores without evidence.

The accompanying analytical-response task asks for 150–180 words with accurate source labels, evidence limits, a fair counterargument and a proportionate recommendation. These are local practice conditions. Further evaluation can be proposed without being misreported as a completed study.

Select a source for the claim it can support, not for an impressive headline. Check its question, measured outcome, participants, method, comparison, limitations and source trail. A peer-reviewed 同行评审 publication may provide useful scrutiny, but that label alone does not guarantee relevance, validity or an appropriate conclusion. Use the required kind of source for the particular research question and verify its actual content.

Worked source selection. Fictional Observation O records ten missed appointments out of fifty before reminders and six out of fifty afterwards. Different months and unrecorded group differences prevent isolating a reminder effect. It measures missed attendance, not recovery, safety or treatment quality. Advertisement A asserts improved care without a method. Methods R proposes consistent measurement and a specified comparison; it is guidance, not a completed experiment.

O and R can support a limited attendance-investigation paragraph with their respective roles made clear. A can be discussed as a promotional assertion, not independent proof of care improvement. Four fewer missed records are not four cures. This is academic language practice in a health-service setting, not a clinical review or recommendation.

Keep source details while reading. In actual research, verify author or organisation, title, publication details, date, locator and link against the original source. Follow the required style when information is absent; do not invent it. For the supplied cards, use the honest fictional label and section locator rather than a fabricated journal entry.

An argument connects a conclusion to reasons the reader can check. A warrant 论据与结论之间的理由 explains why a fact supports a proposed action. Evidence describing a problem does not automatically show that the writer’s preferred solution will work. These invented procedure examples are reasoning practice, not actual law or legal rights.

Worked example. A four-week makerspace log records forty bookings, six unused slots and four equipment conflicts (L, §1). It does not establish ten different affected users, because overlap and identity are unknown. A proposal to confirm bookings may help test whether unneeded time can be released (P, §1), but no trial has occurred.

Address the real objection: some users may not check email daily (S, §1). Do not replace that access concern with “students hate every rule”. Another response route could be investigated, but is not already proven effective or free. A conditional trial argument should state access, conflicts and manageable effort as evaluation criteria and identify what the present sources cannot settle.

The local 160–190 word task develops a thesis, source support, a warrant, a fair objection and a conclusion matching the evidence. Explain why each source bears on the question. Do not invent successful results merely to make the ending sound complete.

Different sources may inform different parts of one decision. Keep their populations, methods and outcomes visible instead of merging them into an unsupported total. A preference count can inform a convenience question without settling how much weight to give access needs or staffing resources.

Worked example. Thirty of forty voluntary online respondents prefer digital booking (S, §1). Eight of twenty logged desk visitors request help in one week (L, §1). The samples may overlap, so thirty plus eight is not a known count of distinct people. Neither 75% nor 40% describes all service users. The desk record does not establish that every user lacks digital access.

A planning note adds a staffing-time concern for a telephone route (N, §1), not a cost estimate or proven performance. A reasoned paragraph can propose investigating an online route with accessible assistance, acknowledging that an alternative consumes resources. Record assistance demand and staff effort before treating feasibility as settled.

Synthesis 综合 connects these contributions to a conditional recommendation: convenience, access and resources interact. It retains source credit, unknown overlap and a fair resource objection. A voluntary majority does not justify an invented universal approval.

An annotation 注释 adds a faithful summary, question-specific evaluation and intended use to a reference. It may identify a claim for which a source should not be used. Neither praise nor a fixed number of words establishes a good annotation. Read the relevant complete source and follow the assignment’s actual format.

Worked extended-source example. Fictional Report C contains three sections on an event-notice comparison. Different groups of twelve volunteers view two versions on consecutive days. Eight first-group readers and ten second-group readers answer all four details correctly (§2). Both use the same key, but layout and wording change together; convenience recruitment, different participants and days limit the comparison.

A shared key supports a consistent scoring procedure; it does not establish reliability across repeated studies or a causal effect of layout. Missed details, later retention and event attendance are unmeasured. The proposed follow-up in§3 is a proposal, not completed research. An annotation can use C as an evaluation-design example while rejecting it as proof that headings alone improve everyone’s attendance.

The sheet supplies a complete 120–160 word model using honest fictional label/title/section references. For real coursework, retrieve in-depth appropriate texts, check bibliographic details and locate each finding. Compare sources’ contributions and limits for the actual question; do not cite a text you have not read or invent its author or publication.

Narrow the question before collecting sources. A thesis 论点 should answer it with conditions that the evidence can support. An evidence matrix 证据矩阵 can organise themes, source contributions and missing information before paragraphs are written. The plan’s length and order follow the argument and actual brief, not a rule that 2,000 words must have six body paragraphs.

Worked evidence matrix. For the fictional shared-equipment question, U records interest from twenty of thirty voluntary current users; B records six double bookings in forty requests; M identifies cleaning, repair and replacement without cost or staffing estimates. Organise these under demand/access, reliability and resources. Missing information includes non-user needs, use frequency, waiting and abandoned attempts, maintenance effort and costs.

Explain the interactions: interest alone does not justify expansion without workable access and maintenance. Fewer visible conflicts could coexist with abandoned attempts. A serious counterargument is that even a trial consumes scarce resources without a proven benefit. A response should state when a limited evaluation is feasible, and when further planning is preferable to immediate action.

The local 650–800 word essay task includes a complete 699-word model, a conditional thesis, connected source reasoning, a fair counterargument, proposed evaluation and honest local source entries. Use the worksheet lines for planning and additional paper for the full essay. The model does not approve a trial or claim measured results; actual research needs genuine sources instead of these cards.

Revise the argument and structure, then review source accuracy, references and language. If a sentence reveals a reasoning problem, return to the argument rather than following an inflexible editing order. Reading aloud can help identify unclear phrasing, but it neither has a guaranteed duration nor replaces source and logic checks. Finish with a conclusion that answers the focused question without new invented evidence.