這大概是準備 JRE 的人最常問的問題，而答案往往令人意外：公務員事務局從來沒有公開發放過 JRE 歷屆試題。明白這一點，你的備考方向會清晰很多，也不會白花錢買來歷不明的資料。
This is probably the question people preparing for the JRE ask most often, and the answer tends to surprise them: the Civil Service Bureau has never publicly released past JRE papers. Once you accept that, your preparation gets a great deal clearer — and you stop spending money on material of unknown origin.
沒有。公務員事務局並無公開發放 JRE 歷屆試題，坊間亦不存在官方認可的「歷屆試題集」。
官方唯一公開的，是各一條中文和英文的樣本試題， 放在公務員事務局及政務主任招聘的網頁上，用意是讓考生了解題目形式，而非提供操練材料。
你可以省下這個時間了。與其在網上四處搜尋來歷不明的檔案， 不如把時間花在理解題目結構和累積政策知識上——後者才是真正影響分數的東西。
這不是偶然，而是制度設計的結果。
公務員事務局在考試資料頁面明確聲明：所有考生的答卷均屬香港特別行政區政府的專有財產， 答卷的知識產權（包括版權）歸屬政府所有，政府亦保留對不法使用答卷行為作出追究的權利。
這項聲明針對的是答卷，但反映了整個考試材料受嚴格保護的取態。 考生不能把試題帶離試場，也不會獲發試題副本。
JRE 考的是分析和撰寫能力，題目取材自當年的社會與政策議題。 今年考人工智能監管，明年可能考安老服務或土地供應。 即使真的有人記得去年的題目，對今年的幫助也相當有限。
CRE 是標準化的選擇題測驗，題型穩定，官方會提供練習題讓考生熟習形式。 JRE 是開放式寫作，沒有題庫概念，每年由命題人員按當時議題重新擬定。 兩者的公開程度不同，源於考試性質不同。
雖然沒有歷屆試題，官方公開的資料其實足夠讓你掌握考試形式。以下是可以在公務員事務局網頁找到的內容：
| 官方提供 | 用途 |
|---|---|
| 考試時間及結構說明 | 知道三小時、兩題、中英各一、必須全部完成 |
| 題目一樣本（英文個案） | 了解虛構城市設定與提問方式 |
| 題目二樣本（中文論述） | 了解政策議題的提問結構 |
| 評核方向說明 | 知道考的是分析及撰寫能力 |
| 失分警告 | 明言重複背景資料內容及文字不會獲得分數 |
最後一項尤其值得重視。官方直接說明：考生應充分理解和獨立分析議題，提出務實兼具創意的方案； 若只是不假思索地重複背景資料的內容和文字，將不會獲得任何分數。
這句話等於公開了評分邏輯的核心——資料是原材料，不是答案。 閱卷員想看的是你對資料做了什麼加工。
官方的英文樣本題把考生設定為某虛構城市福利部門的人員，要求就精神健康服務撰寫內部討論文件， 並明確列出兩項要求：指出設立綜合中心面對的主要問題並說明其影響，然後建議可行的應對措施。
從這個結構可以歸納出英文個案題的固定骨架：識別問題 → 說明影響 → 提出建議。 無論議題換成什麼，這個骨架幾乎不變。
中文樣本題則以「大牌檔文化」為例，要求分析建議的利弊、表明支持或反對、並按立場提出相應建議。 同樣可以歸納出骨架：分析利弊 → 表明立場 → 提供理據 → 給出方案。
你需要的不是一百條舊題，而是兩個骨架。 掌握了骨架，任何陌生議題都能套用；沒有骨架，做一百條題也只是重複同樣的錯誤。
既然官方沒有發放，那些聲稱有「JRE 歷屆試題」的資料是什麼？大致有三類。
由補習社、教育機構或個人按官方樣本格式編寫的原創題目。 這類材料本身有價值——只要編寫者理解考試方向，模擬題完全可以達到訓練效果。 問題只在於部分賣家把它包裝成「真題」出售，這是誠信問題，不是質素問題。
考完之後在論壇分享大概考什麼議題。這類資訊可以讓你了解出題方向的趨勢， 但細節通常已經失真，而且背景資料、數據、持份者意見這些關鍵部分不可能靠記憶還原。 沒有完整背景資料的題目，無法用來練習個案分析。
有些資料其實是 CRE 練習題、其他政府職位的考試材料，甚至是其他地區公務員考試的題目， 被錯誤標籤為 JRE 真題。購買前如果無法確認來源，風險自負。
如果有人聲稱手上有 JRE 真題，可以問一句：既然考生不能把試題帶出試場， 這份材料是怎樣得來的？答案通常會很含糊。
很多人執著於找真題，是因為習慣了公開試的準備模式——操練歷屆試題、歸納題型、背誦答題套路。 但這套方法在 JRE 上效果有限，原因有三。
就算你拿到去年的題目，也拿不到「標準答案」，因為根本不存在。 兩個立場相反的考生，都可能取得高分——關鍵在於論證質素，不在於選對邊。
JRE 取材自當前的社會與政策議題。與其研究前年考什麼， 不如把時間花在追蹤今年的施政報告、財政預算案和立法會討論—— 那才是今年題目的來源。
多數考生的問題不是「沒見過題目」，而是：
這幾個問題，多做十條舊題都解決不了，只能靠限時練習加上有人批改。
以下是在沒有歷屆試題的前提下，實際有效的準備方式。
不要只是讀一遍。把兩條樣本題拆解成骨架，理解每個部分在問什麼。 這是唯一百分百確定符合官方格式的材料，值得反覆研究。
按範疇整理筆記：房屋土地、人口老化、公共衛生、經濟轉型、教育、 扶貧、公共財政、環境、公眾參與、科技監管。 每個範疇掌握核心矛盾、主要政策工具、正反論點。
資料來源建議：最近一份施政報告、財政預算案演辭、立法會相關文件。 這些都是公開資料，比任何「真題」都更貼近今年的出題方向。
找符合官方格式的模擬題來練習。重點不是題目本身，而是訓練你的骨架反射—— 看到題目能在五分鐘內列出大綱。
這是最多人略過、卻最重要的一步。三小時寫兩篇長文的體力和節奏， 不練是撐不住的。建議考前至少完成三次完整計時模擬。
自己看自己的答卷，很難發現「這段只是複述資料」或「這個建議根本無法執行」。 一次有質素的批改，往往比自己做五條題收穫更大。
JRE 沒有歷屆試題，這對所有考生都一樣——沒有人靠真題取得優勢。 真正拉開分數的，是政策知識的厚度、論證結構的清晰度，以及限時寫作的熟練程度。
把找真題的時間拿去讀施政報告、練限時寫作、找人批改，回報會高得多。
還未清楚整個考試流程？看 JRE 考試懶人包。
搞不清 CRE 和 JRE 的關係？看 CRE 與 JRE 有什麼分別。
本文關於官方樣本試題、評核說明及知識產權聲明的內容， 整理自公務員事務局考試資料頁面。詳情請參閱 www.csb.gov.hk。
No. The Civil Service Bureau does not release past JRE papers, and no officially sanctioned collection of them exists anywhere.
The only thing published officially is a sample question in each language, on the Civil Service Bureau and Administrative Officer recruitment pages. Their purpose is to show you what the questions look like, not to supply practice material.
You can stop. Rather than combing the internet for files of unknown provenance, put that time into understanding the structure of the questions and building up policy knowledge — those are what actually move your score.
This is not an accident. It follows from how the examination is set up.
The Civil Service Bureau states on its examination information pages that candidates' scripts are the exclusive property of the Government of the Hong Kong Special Administrative Region, that intellectual property in the scripts — copyright included — belongs to the Government, and that the Government reserves the right to pursue unlawful use of them.
That statement is about scripts, but it reflects the posture towards the examination material as a whole. Candidates cannot take question papers out of the hall and are not given copies.
The JRE tests analysis and drafting, and the subject matter is drawn from the social and policy issues of the year. If this year is artificial intelligence regulation, next year might be elderly services or land supply. Even if someone did remember last year's question, it would be of limited use.
The CRE is a standardised multiple-choice test with stable question types, and the Bureau publishes practice questions so candidates can get used to the format. The JRE is open-ended writing with no concept of a question bank; it is written afresh each year around the issues of the moment. The difference in what is published follows from the difference in what is being tested.
There may be no past papers, but what is published is enough to understand the format. The following can all be found on the Civil Service Bureau website:
| Published officially | What it tells you |
|---|---|
| Timing and structure | Three hours, two questions, one in each language, both must be completed |
| Sample question one (English case) | How the fictional-city setting and the task are framed |
| Sample question two (Chinese essay) | How a policy issue is put to you |
| Assessment focus | That analysis and drafting ability are what is marked |
| Warning about lost marks | That repeating the content and wording of the background earns none |
The last of these deserves particular attention. The official wording is that candidates should fully understand and independently analyse the issue and put forward practical and creative proposals, and that unthinking repetition of the content and wording of the background material will receive no marks.
That sentence effectively publishes the core of the marking logic: the material is raw input, not the answer. What the marker wants to see is what you did with it.
The official English sample casts the candidate as an officer in the welfare department of a fictional city, asked to write an internal discussion paper on mental health services, with two explicit requirements: set out the main problems facing the establishment of an integrated centre and explain their implications, then recommend workable measures.
From that structure you can read off the fixed skeleton of the English case question: identify the problems → explain the implications → recommend measures. Whatever the issue, the skeleton barely changes.
The Chinese sample uses dai pai dong culture, asking candidates to analyse the merits and drawbacks of a proposal, state support or opposition, and make recommendations consistent with that position. Again a skeleton emerges: weigh the arguments → state your position → give your reasons → propose measures.
What you need is not a hundred old questions but two skeletons. With the skeletons, any unfamiliar issue can be handled. Without them, a hundred questions only repeats the same mistakes a hundred times.
If nothing is released officially, what is being sold as "past JRE papers"? Broadly, three things.
Original questions written to the official sample format by tutorial centres, education businesses or individuals. This material has real value — provided the writer understands what the examination is testing, mock questions train exactly what needs training. The problem is only that some sellers package them as authentic past papers, which is a question of honesty rather than of quality.
Posts on forums after the examination describing roughly what came up. These give you a sense of the direction questions are taking, but the detail has usually drifted, and the parts that matter most — background material, figures, stakeholder views — cannot be reconstructed from memory. A question without its background material cannot be used to practise case analysis.
Some of what circulates is CRE practice material, examination material for other government posts, or civil service examinations from other jurisdictions, mislabelled as JRE past papers. If you cannot verify the source before buying, the risk is yours.
If someone claims to hold authentic JRE papers, ask one question: given that candidates cannot take the question paper out of the hall, where did this come from? The answer is usually vague.
People fixate on finding past papers because that is how public examinations are prepared for: drill the past papers, classify the question types, memorise the moves. On the JRE that method does only limited work, for three reasons.
Even with last year's question in hand, there would be no model answer to go with it, because none exists. Two candidates taking opposite positions can both score well — what matters is the quality of the argument, not picking the right side.
The JRE draws on current social and policy issues. Rather than researching what came up two years ago, put that time into this year's Policy Address, Budget speech and Legislative Council discussions — that is where this year's questions come from.
For most candidates the problem is not that they have never seen the question. It is that:
Ten more old questions solve none of these. Timed practice and someone marking the result do.
Here is what actually works when there are no past papers to drill.
Do not just read them once. Break both sample questions down into their skeletons and work out what each part is asking for. This is the only material guaranteed to match the official format, and it repays repeated study.
Organise notes by area: housing and land, an ageing population, public health, economic transition, education, poverty alleviation, public finance, the environment, public participation, technology regulation. For each, know the central tension, the main policy instruments, and the arguments on both sides.
Sources worth using: the most recent Policy Address, the Budget speech, and the relevant Legislative Council papers. All public, and all far closer to this year's questions than any "past paper".
Find mock questions written to the official format. The point is not the question itself but training the reflex — seeing a question and having an outline down within five minutes.
The step most often skipped and the one that matters most. The stamina and pacing needed to write two long pieces in three hours do not arrive on the day. Aim for at least three full timed simulations before the examination.
Reading your own script, it is very hard to notice that a paragraph is just restating the material, or that a recommendation could not actually be implemented. One good round of marking is usually worth more than five questions done alone.
There are no past JRE papers, and that is true for everybody — nobody is gaining an advantage from them. What separates scripts is the depth of your policy knowledge, the clarity of your argument structure, and how practised you are at writing against the clock.
Trade the time you would spend hunting for past papers for reading the Policy Address, writing to time, and having your work marked. The return is far higher.
Not clear on the process as a whole? See The JRE Explained.
Unsure how the CRE and the JRE relate? See CRE vs JRE.
The official sample questions, assessment notes and intellectual property statement described here are drawn from the Civil Service Bureau examination information pages. For details see www.csb.gov.hk.
看懂考試格式只是第一步。JRE 考的是三小時內的實際產出，唯一有效的準備方式是限時寫、有人批。 本站提供 10 個政策範疇的模擬題庫、線上限時練習系統，以及導師逐份批改服務。
Understanding the format is only the first step. The JRE tests what you can actually produce in three hours, and the only preparation that works is writing against the clock and having it marked. This site provides mock question banks across 10 policy areas, a timed online practice system, and paper-by-paper marking by a tutor.