Step 3: IMRaD 與學位論文章節 — Academic Writing Tutorial

概覽

為什麼結構這麼重要？

IMRaD（Introduction–Methods–Results–and–Discussion）是現代生醫期刊論文的「通用合約」：作者照這個順序寫，讀者就知道在哪裡找什麼。讀者掃 Introduction 找「為什麼做這個研究」、跳到 Methods 確認「實驗設計可不可信」、看 Results 抓「具體數字」、再讀 Discussion 確認「這對我有什麼意義」。

這個合約讓讀者 30 秒內就能定位到他關心的段落——這對被引用、被同儕審查、被期刊接受都至關重要。學位論文則衍生出更多變體，因為國家、學科、學位等級不同，組織方式也不一樣。

IMRaD (Introduction–Methods–Results–and–Discussion) is the universal contract of modern biomedical journal writing: if you write in this order, the reader knows exactly where to look for what. Readers scan the Introduction for "why this study," jump to Methods to verify "is the design credible," read Results for "the actual numbers," and finally Discussion for "what this means for me."

This contract lets any reader locate the paragraph they care about in 30 seconds — crucial for citation, peer review, and journal acceptance. Theses have more variants because country, discipline, and degree level all influence how chapters are organized.

💡

核心原則：結構不是裝飾，是讓讀者「讀得快、信得過、引得到」的工具。寫之前先問：我的讀者會用什麼順序提問？答案就是你的章節順序。 Core principle: Structure is not decoration — it's the tool that makes readers read faster, trust more, and cite easier. Before writing, ask: in what order would my reader ask their questions? That order is your chapter order.

IMRaD 逐節

一、IMRaD 四節各自回答什麼問題？

IMRaD 不是隨意排序——它對應讀者腦中的四個問題。記住「沙漏 (hourglass)」形狀：Introduction 從寬到窄（領域→ gap →本研究問題）；Methods 與 Results 保持窄（只談你做了什麼、發現什麼）；Discussion 從窄到寬（你的結果→與文獻對比→領域意義）。

IMRaD isn't arbitrary — it mirrors four questions in the reader's head. Remember the hourglass: Introduction goes wide → narrow (field → gap → your question); Methods and Results stay narrow (only what you did and found); Discussion goes narrow → wide (your result → vs literature → field implications).

Introduction （緒論）

回答：「為什麼要做這個研究？」
定位領域 → 已知什麼 → 還不知道什麼 (gap) → 本研究問題 / 假設。形狀：寬 → 窄。

Answers: "Why this study?"
Position the field → what's known → what's missing (gap) → your question / hypothesis. Shape: wide → narrow.

Methods （方法）

回答：「你怎麼做的？我能不能複製？」
樣本、材料、實驗設計、統計、軟體版本——每一步都要讓另一個實驗室能照著做。形狀：窄而精確。

Answers: "How did you do it? Could I reproduce it?"
Samples, materials, design, statistics, software versions — each step must let another lab repeat it. Shape: narrow and precise.

Results （結果）

回答：「你發現了什麼？」
只陳述事實與數字（含圖表），不解釋、不推論。順序與 Methods 對應，每段先 topic sentence 再給數據。形狀：窄。

Answers: "What did you find?"
State facts and numbers (with figures/tables) — no interpretation, no speculation. Order should mirror Methods; each paragraph leads with a topic sentence, then data. Shape: narrow.

(and) Discussion （討論）

回答：「這代表什麼？跟別人比起來如何？接下來呢？」
Key finding → 與文獻對比 → 機制解釋 → limitation → 未來方向 / 意義。形狀：窄 → 寬。

Answers: "What does it mean? How does it compare? What's next?"
Key finding → vs literature → mechanism → limitations → future directions / implications. Shape: narrow → wide.

⚠️

常見違規：(1) 在 Results 寫「This suggests…」（推論屬於 Discussion）；(2) 在 Discussion 第一次出現新數字（應該都已在 Results）；(3) Methods 寫成日記體（「我們先試了 A 不行，又試了 B…」）。三者都會讓審稿人皺眉。 Common violations: (1) writing "This suggests…" in Results (interpretation belongs in Discussion); (2) introducing new numbers in Discussion (they should all already be in Results); (3) writing Methods like a lab diary ("We first tried A but it failed, then tried B…"). All three irritate reviewers.

學位論文變體

二、學位論文三種主要架構

期刊論文幾乎都用 IMRaD，但學位論文（碩/博）有三種主流架構：傳統 5 章、延伸 6 章、以及 paper-based「三明治」論文。選錯了會讓口委困惑、也讓未來投稿很痛苦。

Journal articles almost universally use IMRaD, but theses (master's / PhD) come in three main flavors: traditional 5-chapter, extended 6-chapter, and paper-based "sandwich" thesis. Picking the wrong one confuses your committee and makes future submission painful.

	5 章傳統式	6 章延伸式	Paper-based / 三明治
典型章節	Ch1 緒論 Ch2 文獻回顧 Ch3 方法 Ch4 結果 Ch5 討論（含結論）Ch1 Introduction Ch2 Literature Review Ch3 Methods Ch4 Results Ch5 Discussion (incl. Conclusion)	Ch1 緒論 Ch2 文獻回顧 Ch3 方法 Ch4 結果 Ch5 討論 Ch6 結論與未來方向Ch1 Introduction Ch2 Literature Review Ch3 Methods Ch4 Results Ch5 Discussion Ch6 Conclusion & Future Work	Ch1 General Introduction Ch2 Paper 1 Ch3 Paper 2 Ch4 Paper 3 (±4) Ch5 General DiscussionCh1 General Introduction Ch2 Paper 1 Ch3 Paper 2 Ch4 Paper 3 (±4) Ch5 General Discussion
何時用	單一大型專題；多數生醫碩士；第一次寫長文獻Single large project; most bioscience master's; first long-form writing	PhD 或較深的碩士；想把「貢獻」明確獨立成章PhD or in-depth master's; want a stand-alone "contributions" chapter	已發表/投稿 1+ 篇論文；多個獨立子計畫；歐盟、澳洲、北歐 PhD 常見1+ paper already published/submitted; multiple independent subprojects; common in EU, Australia, Nordic PhDs
優點	敘事連貫；單一故事線；最多人寫過、模板多Cohesive narrative; single storyline; most templates available	結論獨立易凸顯貢獻；方便寫 implications & future workStand-alone conclusion highlights contributions; easy to expand implications & future work	寫論文 = 投期刊（雙重產出）；畢業後立刻有發表紀錄；每章可獨立評審Writing = submitting (dual output); publication record at graduation; each chapter assessed standalone
缺點	討論章可能過長；轉成期刊論文要重新切割Discussion can balloon; converting to journal papers needs re-slicing	討論與結論章易內容重複；要小心切分Discussion and Conclusion easily overlap; careful partitioning needed	章節間風格、格式不一致；General Intro / Discussion 容易變空話；需學校允許Inconsistent style/format across chapters; General Intro/Discussion can feel empty; school must permit

💡

實務建議：動筆前去你系上找近 3 年至少 5 本通過的論文，看你們系實際採用哪一種，並影印封面、目次當模板。學校 handbook 寫的「規定」與系上實際慣例經常不同。 Practical tip: Before writing, find at least 5 successful theses from your department in the last 3 years to see which structure they actually use, and copy their cover/TOC as templates. Departmental practice often differs from the official handbook.

篇幅與深度

三、PhD vs 碩士 vs 期刊論文

🎓

碩士論文 Master's

字數：約 30k–50k 字（中文常見 60–120 頁）
深度：展現「會做研究」
讀者：口委 3–5 人
壽命：少數人下載；除非後續轉為期刊論文，影響力有限

Words: ~30k–50k (60–120 pages)
Depth: show you "can do research"
Audience: 3–5 committee members
Lifetime: few downloads — limited impact unless converted to a paper

👨‍🎓

博士論文 PhD

字數：約 80k–200k 字（生醫常見 150–300 頁）
深度：對領域有「實質、原創貢獻」
讀者：口委 + 同領域博士生
壽命：機構庫永久收錄；常被同領域引為背景

Words: ~80k–200k (150–300 pages)
Depth: "substantial original contribution" to the field
Audience: committee + future PhD students
Lifetime: permanently archived; often cited as background

📄

期刊論文 Journal

字數：3k–6k 字（不含圖表/補充材料）
深度：單一明確發現 + 充分證據
讀者：全球同領域研究者
壽命：數十年可被引；引用次數定義學術影響

Words: 3k–6k (excluding figures & supplements)
Depth: one clear finding with full evidence
Audience: the entire global subfield
Lifetime: citable for decades; citation count defines impact

⚠️

關鍵差異：學位論文要「展示思考過程」（包含 negative results、嘗試過的死胡同），期刊論文要「壓縮成最有說服力的故事」。同一個研究改寫成這兩種文體時，不能直接複製貼上——通常字數要砍一半以上、補實驗、補圖。 Key difference: A thesis must "show your thinking" (negative results, dead ends, etc.); a paper must be "compressed into the most persuasive story." Converting one to the other is not copy-paste — usually you cut more than half the words, add experiments, and rebuild figures.

決策引導

四、學位論文架構決策樹

🌳 3 分鐘決定你該用哪一種架構

Q1:

你是否已有 1 篇以上發表/投稿/接受中的論文（你是第一作者）？→ 是 → 強烈考慮 paper-based，論文等於章節、效率最高。

Q2:

學校或系上 handbook 是否硬性規定某種格式？→ 是 → 直接照規定（多數台灣/亞洲生醫所偏 5 章傳統式；部分歐洲博班強制 paper-based）。

Q3:

你的研究是「圍繞一個大問題的單一專題」嗎？→ 是 → 用 5 章（碩士）或 6 章（博士，凸顯結論）傳統式。

Q4:

你的研究是「3+ 個彼此弱關聯、各自獨立完成的子計畫」嗎？→ 是 → 用 paper-based，General Intro/Discussion 負責串接邏輯。

Q5:

還是不確定？→ 找主導指導教授 + 同實驗室近 3 屆畢業生各問一次，他們的偏好權重大於書本。

Q1:

Do you already have 1+ first-author paper (published / submitted / in review)? → Yes → Strongly consider paper-based — each paper becomes a chapter, maximally efficient.

Q2:

Does the institution or department mandate a format? → Yes → Follow the rule (most TW/Asian biomed prefer 5-chapter; some EU PhDs mandate paper-based).

Q3:

Is your research a single project around one big question? → Yes → Use 5-chapter (master's) or 6-chapter (PhD, with stand-alone conclusion).

Q4:

Is it 3+ loosely-connected, independently completed subprojects? → Yes → Use paper-based; let the General Intro/Discussion stitch them together.

Q5:

Still unsure? → Ask your main advisor + last 3 graduates from your lab — their preference outweighs any textbook.

互動工具

五、各章節字數分配計算器

調整總字數與論文類型，看每節建議字數。常見比例：Intro 15%、Methods 25%、Results 30%、Discussion 25%、其他 5%（包含 Abstract、Conclusion、References metadata 等）。

Adjust total word count and paper type to see suggested length per section. Rule of thumb: Intro 15%, Methods 25%, Results 30%, Discussion 25%, Other 5% (Abstract, Conclusion, References metadata, etc.).

總字數： 40,000

論文類型：

💡

注意：這只是粗略基準。Methods-heavy 的 omics / wet-lab 論文，Methods 可能佔 30–35%；純 review / theory 論文則 Methods 比例會大幅縮小。實際以指導教授與口委的習慣為主。 Note: This is just a rough baseline. Methods-heavy omics / wet-lab papers may push Methods to 30–35%; pure review / theory papers shrink Methods drastically. Defer to your advisor and committee's preferences.

📝 自我檢測

1. 在以下哪種情況下，paper-based（三明治）論文最適合？

1. In which scenario is a paper-based (sandwich) thesis most appropriate?

A. 你只有一個圍繞單一假設的大型專題A. You have one large project around a single hypothesis

B. 你還沒做任何實驗、文獻也沒看完B. You haven't run any experiments and haven't finished the literature

C. 你已有 2–3 篇第一作者論文、且學校允許 paper-based 格式C. You have 2–3 first-author papers and your school permits paper-based format

D. 你想最快畢業，所以章節越少越好D. You want to graduate fastest, so fewer chapters is better

2. 「我們使用 Seurat v5 對 12,438 個細胞進行分群，並以 resolution=0.6 執行 Louvain 演算法。」這句話應該寫在 IMRaD 的哪一節？

2. "We used Seurat v5 to cluster 12,438 cells and ran Louvain at resolution=0.6." Which IMRaD section does this belong to?

A. Introduction（緒論）A. Introduction

B. Methods（方法）B. Methods

C. Results（結果）C. Results

D. Discussion（討論）D. Discussion

3. IMRaD 的「沙漏 (hourglass)」比喻指的是什麼？

3. What does the "hourglass" metaphor for IMRaD describe?

A. 寫論文時間越久越好A. Spending as much time as possible writing

B. 中段（Methods/Results）字數要最多B. The middle (Methods/Results) should have the most words

C. Introduction 從寬到窄、Methods/Results 保持窄、Discussion 從窄到寬C. Introduction goes wide→narrow, Methods/Results stay narrow, Discussion goes narrow→wide

D. 字數要對稱（前後章節字數一樣）D. Word counts must be symmetric across chapters