Series ID: EDKS-ADV-ZH-0009 · How to Learn Advanced English (Chinese Edition) · Lesson No.009 · C1 → C2

Write Evidence-Based Reports and Recommendations | 用证据写好英语报告与建议

A strong report does not merely sound professional. It reduces uncertainty for the next person who has to think, decide or act.

一份强报告，不只是“看起来专业”。它要降低下一个需要思考、决定或行动的人面对的不确定性。

You may already be able to write essays, explain opinions and discuss evidence. A report asks for a different kind of control. You must decide what the reader needs, separate observation from interpretation, show where the evidence comes from, state what remains uncertain, compare options using stable criteria and make recommendations that are no stronger than the evidence allows.

你可能已经会写 essay、解释观点、讨论证据。但是 report 需要另一种控制能力：先判断读者需要什么；把观察和解释分开；说明证据来自哪里；明确哪些地方仍然不确定；用稳定标准比较选项；最后提出不超过证据强度的建议。

This lesson is for advanced Mandarin-speaking learners who want English that works in universities, research, professional teams, management, consulting, education, technical work and serious project communication.

本课面向希望把高级英语用于大学、研究、专业团队、管理、咨询、教育、技术工作和真实项目沟通的普通话母语学习者。

---

Before we begin: this lesson sits above the school-stage owner | 开始之前：本课不抢学校阶段主页面

eduKate Sengkang already owns stage-specific informational report writing through the Secondary English Learning Hub. That page remains the school-stage owner. Lesson 009 operates at a different layer: C1–C2 academic/professional reporting for readers who may make decisions, allocate resources, approve a trial, review evidence or hand work to another team.

eduKate Sengkang 已经有学校阶段 informational report writing 主页面。本课不复制它。本课负责 C1–C2 学术与专业报告层：读者可能需要作决定、分配资源、批准试点、审核证据，或把工作交给下一支团队。

• Secondary 3 English Learning Guide | Informational Report Writing — school-stage canonical owner.
• Lesson 004 | Read Deeply, Think Clearly, Write Precisely — evidence and reasoning owner inside this series.
• Lesson 008 | Lead Complex Academic and Professional Discussions — discussion-to-decision owner.

1. Report writing begins with a reader decision | 报告从“读者要决定什么”开始

Before opening a document, write one sentence:

The reader needs this report in order to ______.

Examples:

• decide whether to run a pilot;
• understand why performance changed;
• compare three options;
• identify the largest risk;
• choose what to investigate next;
• hand a project safely to another team.

写报告前，先写一句：“读者需要这份报告，是为了 ______。” 如果这句话写不出来，报告往往会变成资料仓库。

2. A report is not an essay with headings | 报告不是“加了小标题的议论文”

Essay	Report
develops a position through continuous reasoning	organises evidence so the reader can inspect and act
often read from beginning to end	often scanned by section
thesis controls argument	purpose/decision question controls evidence
counterargument tests judgement	limitations, risks and alternatives constrain action
conclusion closes the argument	recommendations may be staged, conditional or deferred

报告不是“更正式的 essay”。它的结构要降低查找成本、判断成本和行动成本。

3. A report, proposal, decision brief and minutes are different | Report、proposal、decision brief、minutes 不同

Report: organises findings and supports interpretation or action.

Proposal: argues for a future plan and its feasibility.

Decision brief: compresses options, evidence and a decision request for a busy reader.

Minutes: record what happened, what was decided, who owns actions and what remains open.

四种文件可能互相连接，但不能混为一谈。最常见错误之一，是会议记录里偷偷加入作者自己的建议，或者在 report 里把尚未批准的 recommendation 写成 decision。

4. Cambridge C1 reminds us that purpose and reader matter | C1 写作首先看目的与读者

Cambridge C1 Advanced includes reports and proposals among its Part 2 writing task types. The official format gives the writer a context, purpose and target reader. That matters beyond examinations: professional writing changes when the audience and decision change.

剑桥 C1 Advanced 官方格式把 report 和 proposal 列入写作任务，并明确给出 context、purpose 和 target reader。真实世界同样如此：受众和任务一变，报告结构就应该变化。

Cambridge English C1 Advanced exam format

5. The master chain: Evidence → Finding → Interpretation → Limitation → Recommendation | 总链条

EVIDENCE → FINDING → INTERPRETATION → LIMITATION → RECOMMENDATION.

每一步都不能偷偷跳级。

• Evidence: what information do we actually have?
• Finding: what pattern is directly supported?
• Interpretation: what might the pattern mean?
• Limitation: what can the evidence not establish?
• Recommendation: what action is proportionate to what we know?

6. Keep observation and interpretation separate | 把观察与解释分开

Observation:

Average completion time fell from 18 minutes to 14 minutes during the trial.

Interpretation:

The new interface may have made the task easier to complete.

Alternative interpretation:

Users may simply have become more familiar with the task.

三个句子不是同一层级。报告必须让读者看见这种差别。

7. A finding should survive a source check | Finding 必须经得住来源核对

A strong finding can answer:

• Which source supports this?
• What population does it cover?
• What period does it cover?
• What denominator is involved?
• What was actually measured?

如果一个 finding 回不到来源，它就可能只是作者印象。

8. Build an evidence ledger before drafting | 写正文前先建 evidence ledger

ID	Source	What it supports	Limit	Use
E1	usage log	frequency	four weeks only	finding
E2	survey	reported preference	self-selected respondents	finding + limit
E3	staff interviews	implementation constraints	small group	risk/option analysis
E4	cost estimate	projected cost	assumptions may change	option comparison

Evidence ledger 的价值是 traceability：每个重要句子可以回到证据。

9. Source quality is part of report language | 来源质量必须进入报告语言

Compare:

The system reduces errors.

with:

In the four-week internal pilot, recorded booking errors were lower after the new process was introduced; however, the before/after comparison does not isolate the process from other changes.

第二句不只是“更长”，而是把来源、时间、观察和限制一起写入。

10. Denominators stop percentages from becoming theatre | 分母阻止百分比变成表演

“Satisfaction reached 90%.”

Ask: 90% of whom?

Better:

Eighteen of the twenty survey respondents (90%) selected “satisfied” or “very satisfied”; 42 eligible users did not respond.

百分比没有分母，很容易让读者误判规模。

11. Restore the baseline | 恢复基线

“Complaints fell by 50%.”

From 2 to 1?

From 200 to 100?

Better:

Recorded complaints fell from 24 in June to 12 in July, a 50% reduction.

相对变化和绝对变化最好都让读者看得见。

12. Restore the timeframe | 恢复时间范围

A result over three days and a result over three years should not sound the same.

Use:

• during the four-week trial;
• between January and March;
• across the 2025–2026 academic year;
• at the six-month follow-up.

时间范围是证据边界的一部分。

13. Do not promote correlation into causation | 不要把相关关系升级成因果

Weak:

The new timetable caused attendance to fall.

Safer if evidence is observational:

Attendance fell after the timetable change. The timing is consistent with an effect of the new schedule, but other changes during the same period cannot be ruled out.

报告的可信度，常常来自作者愿意保留“尚未知道”的部分。

14. Qualitative evidence can explain mechanisms | 定性证据帮助解释“为什么”

Numbers might show that completion rates fell. Interviews might reveal that users could not find a required field. Observation might show that staff created workarounds. These evidence types perform different jobs.

定量数据告诉你“有多少”，定性资料可能帮助你理解“为什么”和“怎样发生”。不要让一种证据承担所有任务。

15. Anecdotes illustrate; they rarely establish prevalence | 个案说明，不自动证明普遍性

A user story can show how a failure happens:

One participant abandoned the form after the confirmation screen failed to appear.

It cannot alone justify:

The confirmation screen fails for most users.

高级报告会保护“例子”和“总体频率”之间的距离。

16. Missing data is itself information | 缺失数据本身也是信息

Do not silently remove:

• non-respondents;
• missing records;
• unavailable weeks;
• participants lost to follow-up;
• unknown costs.

可以写：“Cost data were unavailable for two of the four options, so no full cost comparison is possible.” 这比假装表格完整更专业。

17. Negative results belong in the report | 不利结果也要进入报告

A report becomes advocacy when inconvenient evidence disappears.

If response time improved but error rate did not, write both. If users liked a tool but staff workload increased, write both. The reader should see the trade-off, not only the success story.

18. Limitation is not an apology | Limitation 不是道歉

A limitation tells the reader how far the evidence can travel.

Examples:

• small sample;
• short observation period;
• self-selection;
• missing comparison;
• one site only;
• self-reported behaviour;
• unverified cost estimate;
• measurement changed halfway through.

清楚的 limitation 会增强报告，因为它减少错误外推。

19. Put the limitation near the finding it limits | 限制条件要靠近对应 finding

Do not make ten confident claims and hide every caveat in the final paragraph.

Better:

Most respondents preferred the revised process (18 of 20), although the survey response rate was 48%, so the result should not be treated as a complete measure of all users’ views.

20. Cautious language should identify the uncertainty | 谨慎语言要指出“哪里不确定”

The University of Manchester Academic Phrasebank offers useful language for cautious interpretation. The principle is more important than memorising the phrases: connect the hedge to a reason.

Academic Phrasebank | Being cautious

Weak:

It may perhaps possibly be effective.

Better:

The pilot may have reduced processing time, but the lack of a comparison group prevents a causal conclusion.

21. Use reporting verbs to calibrate evidence | 用 reporting verbs 调节证据强度

Verb	Typical force
records / reports	source says or records
shows	directly establishes the stated observation
suggests	supports an interpretation
indicates	points towards
is consistent with	fits an explanation but does not prove it
demonstrates	strong; use only when justified

“demonstrates” 不是“更高级”的 shows。它是更强的证据承诺。

22. Recommendation strength should match evidence strength | 建议强度要匹配证据强度

Evidence state	Possible action
strong repeated evidence, low risk	direct action may be justified
promising evidence, important uncertainty	pilot / staged action
conflicting evidence	clarify conditions / gather data
weak anecdotal evidence	investigate, do not generalise
critical missing information	defer decision until resolved

23. Recommendations can be direct | 建议可以直接

Replace the outdated form because it no longer records the mandatory field required by the current process.

Direct action is appropriate when the requirement and evidence are clear.

24. Recommendations can be conditional | 建议可以带条件

Run the pilot in October, provided that support coverage and rollback procedures are confirmed before launch.

条件不是软弱，而是把已知依赖写进建议。

25. Recommendations can be staged | 建议可以分阶段

1. Run a small trial.
2. Measure defined outcomes.
3. Review risks and user groups.
4. Expand only if thresholds are met.

证据有希望但不完整时，staging 可以把学习和行动连接起来。

26. Recommendations can be deferred | 有时最好的建议是“暂不决定”

Do not select a supplier until the missing maintenance-cost estimates are available.

这不是拖延，而是识别“决定缺少必要信息”。

27. A recommendation needs an owner, trigger or next step | 建议要能被执行

Weak:

Communication should be improved.

Better:

The project lead should publish one weekly status note containing decision changes, unresolved risks and next actions, beginning with the next reporting cycle.

越接近执行，责任、动作和时间越需要清楚。

28. But do not assign authority you do not have | 不要在报告里发明权限

If your role is advisory:

The team recommends…

not:

The organisation will…

unless that decision has actually been authorised.

29. Separate recommendation from approval | Recommendation 不等于 approval

A professional report should distinguish:

• analyst recommendation;
• management decision;
• approved action;
• planned implementation;
• completed implementation.

这些状态如果混在一起，会制造严重误解。

30. Option comparison requires stable criteria | 选项比较必须使用同一套标准

Do not praise Option A for speed, Option B for quality and Option C for cost without showing the common decision frame.

Define criteria first:

• cost;
• time;
• reliability;
• risk;
• user impact;
• reversibility;
• evidence strength;
• implementation burden.

31. Use an option matrix | 用 option matrix 比较

Option	Benefit	Constraint	Evidence	Reversibility
A: keep current	low disruption	known inefficiency remains	strong current baseline	high
B: limited trial	learn before full change	requires support	moderate	high
C: full change	fast scale	high transition risk	weak before pilot	lower

表格的价值不是“看起来专业”，而是让比较维度一致。

32. Reversibility changes recommendation strength | 可逆性会改变建议强度

If an action is cheap and reversible, a smaller amount of evidence may justify a controlled trial. If an action is expensive, irreversible or high-risk, the evidence threshold should usually be higher.

高级报告需要把“证据强度”与“错误决定的代价”一起看。

33. Feasibility is part of evidence-based writing | 可行性也是证据的一部分

A recommendation can be desirable but infeasible.

Ask:

• Do we have staff?
• Do we have authority?
• Do we have time?
• Do we have budget?
• Do dependencies exist?
• Can the change be reversed?

34. Prioritise recommendations | 建议要有优先级

A list of twelve equal actions transfers the work back to the reader.

Rank:

• Immediate: fix a known defect.
• Next: run a trial.
• Later: consider expansion after evidence.

Priority can reflect impact, urgency, feasibility, risk and dependency.

35. Executive summaries are not mini introductions | Executive summary 不是“缩短版开头”

An executive summary should often answer:

1. Why was the report written?
2. What are the most important findings?
3. What limitations materially affect interpretation?
4. What is recommended?
5. What decision or next action is required?

忙碌读者可能只读 executive summary，所以它不能只写背景。

36. Write the executive summary last | 最后写 executive summary

You cannot accurately compress a report whose reasoning is still changing.

Draft body → verify evidence → finalise recommendation → write executive summary.

37. A four-sentence executive-summary model | 四句执行摘要

1. Purpose: This report examines…
2. Finding: The evidence indicates…
3. Boundary: However…
4. Action: It therefore recommends…

不是每份报告都必须四句，但这个练习能检查中心是否清楚。

38. Headings are navigation tools | 小标题是导航系统

Weak:

• Background
• Other Information
• Discussion

Stronger:

• Current Demand and Usage
• Implementation Constraints
• Comparison of Options
• Recommended Pilot and Conditions

让标题告诉读者“这一节负责什么”。

39. Build sections around reader questions | 按读者问题分 section

Instead of organising by collection order:

survey → interview → log → cost sheet

organise by decision questions:

demand → performance → constraint → cost → risk → option.

40. A methods/source section should be proportionate | 方法说明要和任务比例匹配

For a short professional report:

This review draws on four weeks of usage logs, a voluntary user survey, six staff interviews and current supplier cost estimates.

If important limitations exist, state them.

不要把简短内部报告写成研究论文 methodology，也不要完全隐藏来源。

41. Findings sections should lead with the finding | Finding section 先给 finding

Weak:

We collected data from several users. Many comments were made…

Better:

Most reported delays occurred at the approval stage, not during form completion.

Then show evidence.

42. Use finding → evidence → interpretation → limit | 一个稳定 finding 段落

Finding: Approval was the main delay point.

Evidence: 14 of 18 delayed cases remained pending at approval for more than 24 hours.

Interpretation: This suggests that improving form speed alone is unlikely to remove the largest delay.

Limit: The review covers one month and does not test seasonal workload.

43. Recommendation paragraphs need a different architecture | 建议段落结构不同

ACTION → REASON → CONDITION → MEASUREMENT.

Example:

Run a four-week approval-workflow pilot because the current evidence identifies approval as the dominant delay point. The pilot should proceed only if escalation ownership is defined, and it should be evaluated using approval time, error rate and staff workload.

44. Risk language should identify consequence and likelihood separately | 风险要分 consequence 与 likelihood

“High risk” can hide two dimensions.

Risk	Likelihood	Consequence	Response
support unavailable	medium	high	confirm backup before launch
minor form confusion	high	low	clarify instructions

这样读者能看出为什么某个风险优先级更高。

45. State assumptions | 把假设写出来

A cost comparison might assume:

• current staffing levels;
• no licensing increase;
• fixed user volume;
• no migration cost.

If the recommendation depends on assumptions, the report should say so.

46. A condition is not an assumption | Condition 与 assumption 不同

Assumption: something treated as true for analysis.

Condition: something that must be true before action proceeds.

Mixing them can make a recommendation unsafe.

47. A dependency is not merely a risk | Dependency 不只是 risk

If approval cannot begin until legal review is complete, legal review is a dependency.

If legal review may reveal a problem, that possibility is a risk.

专业英语要把关系类型写准。

48. Distinguish current state, target state and transition | 分清 current、target、transition

Reports often describe a desirable target but ignore transition cost.

• Current: manual process.
• Target: shared automated process.
• Transition: migration, training, parallel running, support.

很多失败发生在 transition，而不是 target design。

49. Handover writing must preserve state | 交接必须保留状态

A good handover tells the next team:

• what is complete;
• what is approved;
• what is only recommended;
• what is blocked;
• what evidence exists;
• what assumptions remain;
• what happens next.

这和 Lesson 008 的 discussion close 直接连接。

50. Minutes should not invent commitments | Minutes 不要发明承诺

If someone said:

“I can probably review it on Thursday.”

Do not write:

“Alex will review it Thursday.”

Better:

“Alex will confirm Thursday availability.”

书面摘要不能把 tentative commitment 升级成 confirmed commitment。

51. Professional neutrality is not emotional emptiness | 专业中性不等于没有判断

Loaded:

The terrible current process wastes enormous amounts of time.

Professional:

The current process requires duplicate entry at three stages and adds a median six minutes per case.

第二句更有力量，因为它可检查。

52. Replace vague evaluation with criteria | 把 vague judgement 换成标准

Weak:

Option B is much better.

Better:

Option B has the lowest implementation cost and remains reversible, although it provides less information than the longer trial.

53. Avoid “we can see that” when the evidence can speak directly | 少用空泛 we can see that

Translation-shaped writing often says:

From the data we can see that users prefer Option B.

Better:

Option B was preferred by 18 of 27 respondents.

然后再解释范围。

54. Avoid “from the aspect of…” as a universal bridge | 不要把“从……方面”都翻成 from the aspect of

Instead:

• In terms of cost… when the dimension itself matters;
• Option B is cheaper… when direct comparison is clearer;
• From an operational perspective… when viewpoint is genuinely relevant.

55. Avoid bureaucratic noun piles | 避免官样名词堆叠

Heavy:

The implementation of the improvement of the management of the process…

Clearer:

To improve process management, the team introduced…

高级报告通常让动作和责任更可见。

56. Use nominalisation when it genuinely helps comparison | 名词化有用，但要有任务

implementation cost, approval rate, completion time and risk exposure are useful analytical categories.

不要为了“正式”把每个动词都变成名词。

57. Avoid “suggest somebody to do” | 注意 suggest 的英语补语结构

Weak:

We suggest the manager to extend the trial.

Possible:

We suggest extending the trial.

We recommend that the manager extend the trial.

We recommend extending the trial.

需要完整 grammar mechanism 时，请回到语法主页面；这里强调专业报告中的稳定表达。

58. “According to the data” is not required in every paragraph | according to the data 不要每段都写

Use direct evidence statements when source context is already clear.

The usage log records 317 bookings during the trial.

比重复写 “According to the data…” 更具体。

59. Tables should compress comparison, not hide reasoning | 表格压缩比较，不隐藏逻辑

Use a table when readers need to compare the same dimensions across items.

Do not put every sentence into a table because it looks official.

60. Charts need a sentence that states the finding | 图表后要有 finding 句

Never assume the reader will infer exactly the pattern you mean.

Figure 2 shows that most of the reduction occurred in the first two weeks, after which performance remained broadly stable.

图表展示数据，正文负责解释它与问题的关系。

61. Appendices should remove detail from the main decision path | Appendix 用来移走细节，不是藏关键证据

Move:

• full survey instrument;
• detailed raw table;
• technical specification;
• long interview schedule.

Keep decision-critical findings in the main report.

62. References need traceability | 引用要能追踪

Do not write:

Research shows…

without enough source information for the reader to identify the research.

In internal reports, this may mean a linked source register. In academic reports, follow the required citation style.

63. Paraphrase sources without strengthening them | 改写来源时不能偷偷升级

Source:

The findings suggest a possible benefit.

Bad paraphrase:

The study proves the intervention is beneficial.

意义保留包括 certainty level。

64. Multi-source synthesis should be issue-led | 多来源综合要按问题组织

Weak:

Source A says… Source B says… Source C says…

Stronger:

Both usage logs and interviews identify the approval stage as a bottleneck, while survey responses suggest that users perceive the form itself as simpler than the previous version.

来源开始围绕一个问题对话。

65. Preserve disagreement between sources | 来源不一致时不要强行统一

Operational logs show fewer errors, whereas staff interviews describe greater workload during peak periods.

两个结果可以同时成立。

66. Ask whether the sources measure the same thing | 先检查来源是不是测同一件事

A satisfaction survey and an error log do not conflict merely because one is positive and one is negative. They measure different outcomes.

高级 synthesis 先检查 constructs，再比较结论。

67. Decision thresholds make recommendations testable | 决策阈值让建议可检验

Instead of:

Expand if the trial goes well.

write:

Consider expansion if task completion remains above 90%, critical errors stay below 2% and support demand can be covered within current staffing.

阈值不是必须数字化一切，而是把“什么算成功”提前说明。

68. Do not invent thresholds after seeing the result | 不要看完结果才临时改成功标准

If possible, define key success criteria before the trial.

否则，团队很容易在结果出来后挑最漂亮的指标。

69. Recommendations should state what would change your mind | 建议最好说明“什么会推翻它”

The recommendation should be reconsidered if support demand exceeds available staffing or if critical booking errors remain unresolved after the first week.

这让报告从“主张”变成可监控决策。

70. Build a recommendation ladder | 建议强度阶梯

1. Investigate.
2. Run a limited pilot.
3. Expand under conditions.
4. Adopt as standard.
5. Retire the old process.

不是所有 evidence 都应该直接跳到第 4 或第 5 级。

71. Worked case: a fictional feedback system review | 完整案例：虚构反馈系统评估

Fictional case. A learning centre tests a new digital feedback form for six weeks. Forty students are invited to use it. Thirty-two use it at least once. Twenty-six complete the final survey. Median teacher response time falls from 48 hours under the old process to 30 hours during the pilot. Revisions are submitted by 19 of the 32 users, compared with 13 of 35 comparable users in the previous six-week period. Teachers report that short comments are easier to send, but three teachers say longer explanations take more time because they need to copy examples into a separate field. There is no random assignment, the two periods include different students and no long-term learning measure exists.

虚构案例：某学习中心测试新的数字反馈表六周。40 名学生获邀，32 人至少使用一次，26 人完成最终问卷。教师反馈的中位响应时间从旧流程 48 小时降至试点期间 30 小时。32 名使用者中有 19 人提交修改稿；前一个六周周期中，可比较的 35 人有 13 人提交修改稿。教师认为短评更容易发送，但三名教师表示，长解释更耗时，因为要把例子复制进另一个栏位。没有随机分组，两段时间的学生不同，也没有长期学习测量。

72. Step 1: write only what is directly observed | 第一步：只写直接观察

• Median response time was lower during the pilot.
• A higher proportion of users submitted a revision than in the previous comparison period.
• Some teachers reported increased effort for longer explanations.
• Survey coverage was incomplete.

暂时不要写“新系统提高学习”“学生更有动力”。

73. Step 2: quantify carefully | 第二步：数字要完整

Pilot revision rate:

19/32 = 59.4%

Previous comparison:

13/35 = 37.1%

Difference:

about 22.3 percentage points

但这仍然不是因果证明，因为两组不是随机、也不是同一批学生。

74. Step 3: identify plausible interpretations | 第三步：列出解释，不急着选一个

• faster feedback may have made revision easier;
• students in the pilot may have been more willing to revise;
• teachers may have changed their behaviour;
• the different cohort may explain part of the change.

报告可以提出多种 plausible explanation，而不是为了“有结论”过早关闭。

75. Step 4: identify decision-relevant limitations | 第四步：只保留影响决定的 limitation

• non-random comparison;
• different cohorts;
• incomplete survey response;
• teacher workload trade-off;
• no long-term learning outcome.

不是列越多 limitation 越好。重点是哪些会改变 recommendation。

76. Step 5: choose recommendation strength | 第五步：选择建议强度

Too strong:

Adopt the system permanently because it improves learning.

More defensible:

Continue the system for one further term as a controlled extension while measuring revision behaviour, teacher workload and an independent learning outcome. Permanent adoption should be reconsidered after those data are available.

77. Model executive summary | 示例 executive summary

This report reviews a six-week pilot of a digital feedback form. During the pilot, median teacher response time was lower and a larger proportion of participating students submitted revisions than in the previous comparison period; however, the groups were not randomly assigned, survey response was incomplete and some teachers reported greater workload for longer explanations. The evidence therefore supports a controlled extension rather than a permanent adoption decision. A further term should measure revision behaviour, teacher workload and an independent learning outcome before the system is considered for standard use.

注意：这段没有声称“系统提高学习”，因为案例没有相应证据。

78. Model finding paragraph | 示例 finding 段

Feedback was delivered faster during the pilot. Median response time fell from 48 hours under the previous process to 30 hours during the six-week trial. This indicates a shorter turnaround for participating users, although the comparison does not show which element of the new process produced the change.

79. Model trade-off paragraph | 示例 trade-off 段

Teacher workload effects were mixed. Short comments were reported as easier to send, while three teachers described additional effort when giving longer explanations because examples had to be copied into a separate field. Any wider rollout should therefore evaluate response speed together with the time required for complex feedback rather than treating faster median response as a complete measure of staff efficiency.

80. Model recommendation paragraph | 示例建议段

The centre should extend the pilot for one term rather than adopt the system permanently at this stage. The extension should retain the current user-support process, add a simpler method for inserting examples and collect three outcome groups: response time, revision behaviour and an independent measure of subsequent performance. Permanent adoption should be considered only after the workload trade-off and learning outcome are clearer.

81. Audit the model: every sentence has a job | 审计示例：每句话都有任务

• purpose;
• finding;
• boundary;
• trade-off;
• recommendation;
• measurement;
• decision threshold.

高级报告不是堆正式句型，而是让每一句执行一个可解释任务。

82. Practice A: finding or interpretation? | 练习 A：finding 还是 interpretation？

Classify:

1. “Average waiting time fell by seven minutes.”
2. “The new schedule made staff more efficient.”
3. “Users reported fewer navigation problems.”
4. “The interface is easier to use.”

83. Practice B: repair the percentage | 练习 B：修百分比

“Eighty per cent of users preferred Option B.”

Given: 20 of 25 respondents preferred B; 60 people were invited.

Rewrite accurately.

84. Practice C: repair causation | 练习 C：修因果

“The training caused error rates to fall because errors were lower after the workshop.”

Rewrite for an observational before/after comparison.

85. Practice D: recommendation strength | 练习 D：建议强度

Evidence: one small pilot, promising result, no long-term follow-up.

Choose: investigate / pilot / full adoption / permanent retirement of old system.

Explain why.

86. Practice E: authority boundary | 练习 E：权限边界

The group can recommend but not approve. Rewrite:

“The organisation will implement Option C next month.”

87. Practice F: missing data | 练习 F：缺失数据

Two suppliers have full maintenance cost estimates; a third does not.

Write one sentence that prevents a false complete comparison.

88. Practice G: executive summary | 练习 G：执行摘要

Compress a fictional report into four sentences: purpose, finding, limitation, action.

89. Practice H: stakeholder conflict | 练习 H：相关方证据冲突

Users prefer a new process. Staff report higher workload.

Write a synthesis that preserves both.

90. Practice I: conditional recommendation | 练习 I：条件建议

Write a recommendation that proceeds only if support coverage is confirmed.

91. Practice J: staged recommendation | 练习 J：分阶段建议

Create three stages for a promising but uncertain new system.

92. Practice K: correct a meeting record | 练习 K：修会议记录

Original spoken statement:

“I may be able to review this Friday, but I need to check.”

Bad minutes:

“Li will review on Friday.”

Repair it.

93. Practice L: Chinese-to-English report style | 练习 L：中文思路转专业英语

Chinese concept:

从效率方面来看，新方案有一定优势，但仍需要进一步观察其长期效果。

Write two natural professional-English versions with different levels of formality.

94. Model answers | 练习解析

A: 1 and 3 are direct findings if the measurements are accurately reported; 2 and 4 are interpretations unless the evidence directly establishes efficiency/usability through defined measures.

B: Twenty of the twenty-five respondents (80%) preferred Option B; 35 invited users did not respond.

C: Error rates were lower after the workshop. The timing is consistent with a training effect, but the before/after comparison alone does not establish causation.

D: a controlled pilot or extension is normally more proportionate than full adoption when evidence is promising but long-term evidence is absent.

E: The group recommends implementing Option C next month, subject to approval.

F: A complete maintenance-cost comparison is not yet possible because Supplier C has not provided an estimate.

G: answer varies; preserve purpose, finding, material limitation and next action.

H: User preference favours the new process, while staff interviews identify a workload cost; the adoption decision should therefore consider both usability and operational burden.

I: Proceed with the two-week pilot only after support coverage is confirmed for the full period.

J: small pilot → measured extension → wider adoption if thresholds hold.

K: Li will confirm whether Friday review is possible.

L: The new approach appears more efficient, although its longer-term effect remains uncertain and should be monitored. / From an operational perspective, the new approach offers an efficiency advantage; however, further observation is required before its long-term performance can be assessed.

95. The 20-minute report drill | 20 分钟报告训练

Time	Task
4 min	read source pack
4 min	build evidence ledger
6 min	write two findings + limits
4 min	write one recommendation
2 min	trace recommendation back to evidence

96. The 45-minute growth session | 45 分钟增长模式

Time	Task
8 min	define decision question
10 min	source/evidence mapping
12 min	finding paragraphs
8 min	option/recommendation writing
7 min	executive summary + audit

97. The 90-minute deep session | 90 分钟深度训练

Time	Task
15 min	read multi-source pack
15 min	evidence ledger + source quality
20 min	draft findings/options
15 min	recommendations + risk/conditions
10 min	executive summary
15 min	traceability, scope and language audit

98. Seven-day report cycle | 七天报告循环

1. Day 1: decision question + audience.
2. Day 2: source quality + evidence ledger.
3. Day 3: quantitative findings.
4. Day 4: qualitative findings + limitations.
5. Day 5: options + recommendation strength.
6. Day 6: executive summary + handover.
7. Day 7: unfamiliar complete report + audit.

99. Twelve-week C1–C2 progression | 12 周 C1–C2 报告路线

Weeks	Focus	Evidence
1–2	finding vs interpretation vs recommendation	short reports
3–4	source quality, numbers, limitations	evidence ledgers
5–6	options, trade-offs, recommendation strength	decision matrices
7–8	executive summaries, professional register	compressed briefs
9–10	risk, assumptions, conditions, handover	project reports
11–12	multi-source independent transfer	full report portfolio

100. Monthly benchmark | 每月基准任务

Use one unseen source pack containing:

• one quantitative table;
• one survey/interview extract;
• one conflicting source;
• one implementation constraint.

Produce:

1. 200-word executive summary;
2. three findings;
3. one option matrix;
4. two recommendations;
5. one evidence gap;
6. one condition that would change the recommendation.

101. Measure more than grammar | 不要只看语法

Track:

• source traceability;
• claim scope;
• numerical accuracy;
• limitation quality;
• recommendation proportionality;
• reader navigation;
• authority boundaries;
• professional clarity;
• Mandarin-to-English naturalness;
• handover accuracy.

102. First weak-link diagnosis | 第一个薄弱环节诊断

Symptom	Likely weak link	Repair
report sounds like essay	reader decision unclear	write decision question first
recommendations appear from nowhere	traceability	evidence → finding → action map
numbers sound impressive but vague	denominator/baseline/timeframe	restore context
everything sounds certain	evidence calibration	separate observation and interpretation
too many formal nouns	translation/bureaucratic register	restore actors and verbs
good findings, weak action	recommendation design	condition, stage, owner, measure
meeting report invents agreement	state preservation	separate proposed/agreed/approved

103. What not to do | 不要这样写高级报告

• Do not write the recommendation first and search for evidence later.
• Do not hide contrary evidence.
• Do not present percentages without denominators.
• Do not turn before/after into automatic causation.
• Do not treat limitations as embarrassment.
• Do not use formal vocabulary to hide weak reasoning.
• Do not call a recommendation an approval.
• Do not make every action equally urgent.
• Do not let tables replace interpretation.
• Do not let translation-shaped English obscure responsibility.

104. CEFR mediation connects directly to reporting | CEFR mediation 与报告高度相关

The Council of Europe describes mediation as helping create meaning and communication, including conveying ideas and information and working across linguistic or cultural barriers. Advanced report writing is often a mediation task: complex evidence must become usable for another person.

欧洲委员会把 mediation 描述为帮助建立意义与沟通，包括传递信息、协作和跨语言/文化障碍。高级报告经常就是 mediation：把复杂材料转换成另一个人可以使用的形式。

Council of Europe | Mediation

105. Reference floor | 参考基础

• Cambridge English C1 Advanced format — report/proposal writing, purpose and target reader.
• University of Manchester Academic Phrasebank | Being cautious — cautious claims and limitations.
• Academic Phrasebank | Being critical — evaluating sources, limitations and arguments.
• Council of Europe | Mediation — conveying and co-constructing meaning.
• CEFR qualitative aspects — range, accuracy, fluency, interaction and coherence, useful when reports originate in discussion.

Official descriptors and academic resources provide a quality floor. They do not turn every teaching scaffold in this lesson into an official formula.

106. Canonical eduKate routes | eduKate 主页面路由

• Secondary 3 Informational Report Writing
• Source Credibility, Reliability and Evidence Evaluation
• Synthesis Across Sources and Viewpoints
• Statistics, Percentages, Averages and Quantitative Claims
• Editing for Concision, Redundancy and Clarity
• Lesson 004 | Read Deeply, Think Clearly, Write Precisely
• Lesson 008 | Lead Complex Academic and Professional Discussions

107. SEO language map | 本课自然覆盖的搜索意图

This lesson naturally serves learners searching for advanced English report writing, C1 report writing, C2 academic writing, professional report English, evidence-based report, executive summary, report recommendations, findings and limitations, report writing for Chinese speakers, English reports for Mandarin speakers, academic report writing, business report English, decision brief English, professional recommendations, 高级英语报告, 英语报告写作, C1 英语写作, C2 英语写作, 学术英语报告, 英文执行摘要, 英文建议, 英文 findings, limitation 怎么写, 中文母语英语写作, 英文专业报告 and 英文报告怎么写.

108. The one-page evidence-to-action operating system | 一页 evidence-to-action 系统

1. Name the reader decision. 先写读者需要决定什么。
2. Build an evidence ledger. 证据、来源、范围、限制。
3. Write findings before opinions. 先写观察。
4. Separate interpretation from finding. 不把解释装成事实。
5. Restore denominator, baseline and timeframe. 让数字完整。
6. State material limitations near the claim. 限制靠近 finding。
7. Compare options using stable criteria. 同一标准比较。
8. Match recommendation strength to evidence and risk. 建议强度匹配证据与代价。
9. Preserve authority and state. 分清建议、同意、批准、执行。
10. Write the executive summary last. 最后压缩。
11. Trace every recommendation back to evidence. 每条建议能回到 finding。
12. State what would change the recommendation. 让建议可监控。

109. Final assignment | 最终作业

Create a fictional professional source pack or use a non-confidential real learning problem. Include at least:

• one quantitative table;
• one interview/survey extract;
• one contradictory or limiting piece of evidence;
• one cost, staffing or implementation constraint.

Then produce:

1. a one-sentence decision question;
2. an evidence ledger;
3. three findings;
4. one limitation for each finding;
5. an option matrix;
6. two recommendations of different strength;
7. one explicit condition;
8. one risk that would reopen the decision;
9. a 150–200 word executive summary;
10. a final traceability check showing which evidence supports each recommendation.

最后问：

If the reader disagrees with my recommendation, can they still see exactly what evidence I used, what I assumed, what remains uncertain and what would change the decision?

如果读者不同意我的建议，他仍然能不能清楚看见：我用了什么证据、做了什么假设、哪里仍不确定，以及什么信息会改变决定？

If yes, your report is doing professional work.

如果可以，这份报告就不只是“英语写得好”。它已经开始真正工作。

---

Continue | 继续

The next lesson moves into a major C1–C2 mediation skill: how to take one complex idea and adapt it accurately for different audiences without making it childish, vague or misleading.

下一课进入一个重要的 C1–C2 mediation 能力：怎样把同一个复杂思想准确转换给不同受众，同时不把内容说幼稚、不说模糊，也不说错。

Next: EDKS-ADV-ZH-0010 · Lesson No.010 · Adapt Complex English for Different Audiences · 把复杂英语准确讲给不同受众

Back to Advanced English Chinese Edition Hub · 返回高级英语中文版主页