本文由 UniWriterPro 學術規例研究組 verify · Last updated: 2026-08-26 · 本文係按 PolyU ELC FYP Literature Review guide、HKUST CSE Literature Survey guide、HKUST MASS LibGuide、University of Westminster LibGuide 逐條解讀。冇公開文件嘅位一律標「搵唔到官方文件」，唔搬其他學校套用。
目錄
- Supervisor 打「too descriptive」返嚟嘅具體意思
- Technique 1：由 author-driven 改成 theme-driven paragraph 結構
- Technique 2：Reporting verb 升級——由 stated 升到 argued／challenged
- Technique 3：擺明 sources 之間嘅 disagreement，唔再 list agreement
- Technique 4：每個 theme 結尾直接對應返 gap statement
- Technique 5：Signpost 你自己嘅 position——用 however／nevertheless／this study extends
- Rewrite 完最後 self-check
- FAQ
Supervisor 打「too descriptive」返嚟嘅具體意思
Supervisor 喺你 literature review 邊位打「too descriptive」／「reads like a summary」／「where is your critical engagement」，多數同學嘅第一反應係「我要多讀啲 paper」。呢個判斷錯咗。問題唔係篇數，係組織方式同語氣。你可能已經讀咗 40 篇 paper，但呢 40 篇喺 draft 裡面每篇獨立一段（「Chan (2020) found…」「Wong (2021) reported…」），就會 read like a summary。
PolyU English Language Centre 對 literature review 嘅任務定義得好清楚——唔係 report，係 critique：
It provides a critique of the existing literature and shows how it contributes to the development of your particular study.
PolyU ELC — FYP Literature Review Guidelines
「Critique」呢個字係 key。HKUST MASS LibGuide 對呢個對比講得更直接：
You effectively evaluate (appraise critically, not just summarize) the documents you selected in relation to the research you propose.
HKUST MASS — Library Literacy: Literature Review
「Appraise critically, not just summarize」——呢句直接標明分別。University of Westminster LibGuide 用返一句同「too descriptive」對頭嘅描述：
It should not be purely descriptive: a literature review needs to include summary and critical evaluation.
University of Westminster — Literature Reviews LibGuide
留意 Westminster 嘅講法：唔係要拎走 summary，係要 summary + critical evaluation 兩層都有。你 draft 只有 summary 冇 critical evaluation = descriptive。加返 critical evaluation 落每個 section = analytical。
下面 5 個 rewrite technique，係將現有 descriptive draft 改成 analytical draft 嘅具體 pattern。每個 technique 都示範 before／after 對照，直接可以套返你 draft 嘅段落。
都可以直接交俾我哋幫你做 literature review：FYP / Final Year Project 全程支援
好擔心你份 FYP? 即時睇吓每一 part 點做：FYP 題目點揀、FYP Proposal 點寫、FYP 研究方法點揀、FYP Discussion 點先展示到 interpretation?、SPSS 結果論文點寫、論文剩返兩個禮拜？即睇點應對
Technique 1：由 author-driven 改成 theme-driven paragraph 結構
Descriptive literature review 最典型嘅 structural signature：一段一個 author。段落 topic sentence 開頭係「Chan (2020) examined…」，第二段「Wong (2021) explored…」，第三段「Lee (2022) reported…」。呢個結構本身就係 descriptive——你喺為每個 author 做 mini-summary，而唔係為每個 theme 做 argument。
Before（author-driven，descriptive）：
「Chan (2020) examined the impact of mobile trading apps on retail investor behaviour among Hong Kong undergraduates. Using a survey of 300 participants, Chan found a positive correlation between app usage frequency and trading volume. Wong (2021) explored similar phenomena but focused on U.S. college students. Wong reported that daily app usage predicted higher risk-taking behaviour. Lee (2022) conducted interviews with 15 young investors and described their emotional response to real-time price notifications.」
After（theme-driven，analytical）：
「Existing research on mobile trading apps has converged on two main themes: usage-behaviour correlations at the population level, and the individual emotional mechanisms driving these behaviours. On the correlational front, surveys in Hong Kong (Chan, 2020) and U.S. contexts (Wong, 2021) both report positive associations between app usage frequency and trading volume, though Chan’s sample was limited to undergraduates while Wong extended this to a broader college-age population. On the mechanism front, Lee’s (2022) qualitative work with 15 young investors suggests that real-time price notifications may be a mediating factor—an angle that quantitative surveys have not directly examined.」
Rewrite 動作：拎張紙將 draft 每段嘅 core argument 用一句寫低，如果一句寫落係「Chan 研究 X」，改成「Existing research on X shows Y」呢類 theme 起首句。Authors 由主語變 evidence，theme 由 background 變 argument backbone。
Marker 睇到嘅 signal：author-driven draft 讀完之後，讀者只記得「有一堆 author 做過類似嘢」；theme-driven draft 讀完之後，讀者可以講出「呢個 field 現時有邊 2–3 個 theme／main finding」。呢個係最快嘅 self-check。
Technique 2：Reporting verb 升級——由 stated 升到 argued／challenged
Descriptive literature review 嘅第二個 signature：reporting verb 全部係 neutral——「stated」、「reported」、「found」、「examined」、「noted」。呢類 verb 只表達「有人講過」，冇 evaluation。
Analytical literature review 嘅 reporting verb 有 3 個層次：
- Neutral（描述）：state、report、find、examine、note、describe、observe。用喺純粹交代 empirical finding。
- Evaluative（帶立場）：argue、contend、demonstrate、propose、suggest、posit。表明作者係 propose 一個 claim，唔係單純描述現象。
- Contestive（表 disagreement）：challenge、refute、question、dispute、call into question、problematise。用喺 authors 之間 disagreement 或者你 push back 一個 established view。
Before：「Chan (2020) found that app usage correlates with trading volume. Wong (2021) also reported similar findings.」
After：「Chan (2020) argued that app usage causally drives trading volume through a habit-formation mechanism. Wong (2021) demonstrated a similar correlation but challenged Chan’s causal interpretation, proposing that pre-existing risk preferences may confound the observed relationship.」
兩個對比嘅分別：after version 讓 marker 見到你認得每個作者係喺 argue 咩、彼此立場點對頭。「Found」全篇一律 = 你冇分辨作者係做 empirical description 定 theoretical argument；升級 verb 就係 show 你有 read into 每篇 paper 嘅立場。
Rewrite 動作：Ctrl+F 你 draft 裡面所有「stated」、「reported」、「found」，逐個問「作者實際上係做緊咩」——係 empirical measurement（keep neutral verb）定 theoretical claim（升 evaluative verb）定 challenge 前人（升 contestive verb）。三分一 verb 換晒都唔算多。
Marker 睇到嘅 signal：全篇 verb 得一種 flavour = descriptive；verb 有 texture／有升降 = 你識分辨 authors 之間立場。
Technique 3：擺明 sources 之間嘅 disagreement，唔再 list agreement
Descriptive literature review 第三個 signature：只 list agreement 唔標 disagreement。「Chan (2020), Wong (2021), and Lee (2022) all found positive correlations between X and Y」——讀落好齊整，但呢個 sentence 講嘅完全係 consensus，冇 tension。Marker 會問：「Then what is there to research further?」
HKUST CSE 對 literature survey 呢層要求好直接：
It shows that you are intelligent enough to evaluate the quality of the other work done on the subject, i.e., it shows that you are capable of thinking critically and identifying strengths and weakness.
HKUST CSE — 1.3 Literature Survey
「Strengths and weakness」——每篇 paper 都有 weakness，你 draft 冇提任何 weakness = 冇 evaluate。真正 field literature 幾乎冇 100% consensus。你搵唔到 disagreement，代表未讀夠深，或者只讀咗同一個 school of thought。
Before：「Studies have consistently shown that mobile trading apps increase retail investor risk-taking (Chan, 2020; Wong, 2021; Lee, 2022).」
After：「While quantitative studies (Chan, 2020; Wong, 2021) suggest that mobile trading apps are associated with increased risk-taking, qualitative work by Lee (2022) complicates this picture: her interviewees described the apps as reducing rather than amplifying impulsive decisions, because real-time price data gave them a sense of control. This tension between aggregate-level correlation and individual-level experience remains largely unresolved in the current literature.」
Rewrite 動作：每個 theme 裡面主動搵一個 counter-finding／limitation／methodological caveat。冇搵到就 explicit 標明「few studies have questioned this framing」——呢句本身就開闢咗你嘅 gap。
如果 draft 已 submit 過去被 push back「where is the tension between sources」，可以搵 UniWriterPro 由 human writer 起草嘅 FYP 代寫——我哋 review 過幾百份 lit review，可以直接指出邊個 theme 段落 tension 冇 surface，而唔係逐句改文法。
Technique 4：每個 theme 結尾直接對應返 gap statement
Descriptive literature review 通常會將所有 gap 塞喺一個獨立 section（「Research Gap」或者「Research Problem」）擺喺 chapter 結尾，前面每個 theme 段落自己讀完好似冇立場、冇方向。呢個結構令 marker 讀到中段完全唔知你係想 land 邊。
PolyU ELC 對 literature review 嘅其中一個功能講得好清楚：
It helps to refine the direction of your own research.
PolyU ELC — FYP Literature Review Guidelines
Refine your direction 呢個 outcome 唔係讀完全 chapter 之後突然發生嘅——係 chapter 每個 theme 段落結尾都要 signal 一次「呢個 theme 未 close 嘅位」，讀者才會累積出方向感。
Before（gap 集中喺 chapter 尾）：Chapter 有 5 個 theme section，每個 section 結尾都係一句「Studies in this area have made significant contributions to our understanding.」，然後 chapter 最尾一個 Research Gap section 一次過 dump 所有 gap。
After（gap 散喺每個 theme 尾）：每個 theme section 結尾寫類似「However, existing research on [this theme] has largely focused on [specific angle], leaving [specific dimension] under-examined—an issue this study directly addresses through [your approach].」
Rewrite 動作：每個 theme 段落結尾加一句「gap bridge」——由 theme 直接連去你嘅 research question。呢句唔洗長，30–60 字夠。5 個 theme 段就有 5 條 gap bridge，累積起嚟 chapter 讀完 marker 已經 orient 到你研究方向。
Gap statement 點寫，可以 cross-reference 返 FYP Proposal 4 段 workflow 裡面 gap 段嘅 template（Study A → Study B → However, X remains unexamined）。
Marker 睇到嘅 signal：Draft 每個 section 結尾都 land 到具體 gap = 你 chapter 有方向感；每個 section 結尾都係 generic「contributes to our understanding」= chapter 冇 argument arc。
Technique 5：Signpost 你自己嘅 position——用 however／nevertheless／this study extends
Descriptive literature review 最後一個 signature：作者聲音消失。全篇淨係 authors X、Y、Z 講咗咩，讀完 marker 唔知你自己 stance 係咩。HKUST MASS LibGuide 對呢層要求好 explicit：
You write it from specific standpoints (perspective) – express views on the nature of the topic & how you will investigate the topic or question(s)
HKUST MASS — Library Literacy: Literature Review
「Specific standpoints」意思係讀者要能夠指出你嘅 position——你認為現有 literature 哪個 stream 更 convincing？你 build on 邊個 tradition？你 push back 邊個 assumption？
Signpost 你自己 position 嘅 4 個 phrase pattern：
- 「However, [specific limitation of prior work]」——標明你認為現有工作嘅 weakness。
- 「Nevertheless, [nuance that current literature underplays]」——標明你認為被 underplay 嘅 dimension。
- 「This study extends [Author, Year] by [specific extension]」——標明你 build on 邊個 tradition。
- 「Building on [theoretical framework], this review adopts the position that…」——一次過 declare 你嘅 theoretical stance。
HKUST CSE 用另一個角度講同一件事：
In each document, identify the approach(es)/method(s) for solving problem(s), and compare this/these with what you already know.
HKUST CSE — 1.3 Literature Survey
「Compare with what you already know」——呢句嘅 subject 係「you」，唔係「the literature」。Literature review 唔係 detached 匯報，係你同 existing work 對話。
Rewrite 動作：Chapter 每個 major section（唔係每段）至少一次用上面 4 個 phrase pattern 之一。整份 chapter 至少 3–5 次 signpost 你 position——太少會被讀成 detached summary，太多會讀成 opinion piece，3–5 次係一般 undergraduate 到 taught master’s dissertation lit review 嘅合理密度（呢個係 rule of thumb，冇一個全港統一嘅官方 threshold，要對返你 supervisor field 嘅 convention）。
Marker 睇到嘅 signal：Chapter 讀完 marker 可以講出「作者嘅 position 係 X」= 過關；讀完只記得「有一堆 author」= descriptive。
Rewrite 完最後 self-check
5 個 technique 走完之後，做一個 5 分鐘 final check：
- Random pick 3 個段落，讀第一句 topic sentence——係 theme（「Existing research on X shows…」）定 author（「Chan (2020) found…」）？三段有兩段以上係 author 開頭 = Technique 1 未做透。
- 整章 Ctrl+F「stated」、「reported」、「found」——出現超過 20 次？Technique 2 未做透。
- 每個 major section 有冇至少一次講 sources 之間 disagreement／limitation？冇 = Technique 3 未做透。
- 每個 theme section 結尾有冇 gap bridge 句？冇 = Technique 4 未做透。
- 整章讀完，你可唔可以用一句講出你自己嘅 position？講唔到 = Technique 5 未做透。
5 條 self-check 有兩條以上 fail = draft 未 ready re-submit。做多一輪 rewrite 再交，比 supervisor 打第二次「too descriptive」返嚟慳好多 cycle。
如果 rewrite 完自己讀依然覺得未 land、supervisor 已經退咗第二輪 draft、或者 deadline 已迫在眉睫，可以搵 UniWriterPro 提供嘅 academic proofreading 服務，FYP 代寫 同 Thesis 代寫由第三隻眼 tighten paragraph argument structure，唔係逐字改文法。發現有 draft 但只想有 human writer review 一輪，可以直接發 chapter draft 過嚟 quote。
另外，如果你打算用 GenAI 輔助 literature review rewrite（例如叫 AI paraphrase paragraph、summarise paper），有兩層要留意：（1）AI 產生嘅 citation 常有 hallucination 風險，author + year + journal 睇落好真但實質查 Google Scholar 搜唔到；（2）每間大學嘅 GenAI declaration 要求唔同，用之前對返你 department 政策，詳細睇 香港八大 GenAI 使用政策同 declaration 指引。
FAQ
Literature review 究竟要引幾多篇 paper 先算夠？
冇一個 magic number，篇數睇 topic maturity 同 review scope。Undergraduate FYP proposal 階段常見 15–30 篇，dissertation chapter 常見 50–100 篇。搵唔到官方文件講明全港統一數字，要對返你 supervisor 同 department 慣例。真正 signal 唔係篇數，係你有冇 map 到主要 stream 同 disagreement。
如果 supervisor 話 draft 太 descriptive，係唔係代表我要重寫成篇？
唔一定。5 個 rewrite technique 裡面，Technique 1（theme-driven 結構）通常要動段落順序，可能重寫；Technique 2（verb 升級）、3（tension）、5（signpost position）多數係逐段插入，唔洗重寫；Technique 4（gap bridge）係每個 section 結尾加句。實際 rewrite scope 應該由問返 supervisor「which section specifically」開始，唔好 assume 要重寫成篇。
Chapter 已經有一個 Research Gap section 喺結尾，仲需要每個 theme 尾加 gap bridge 嗎？
需要。結尾嘅 Research Gap section 係 synthesis（將所有 gap 收埋一齊），theme 尾嘅 gap bridge 係 local signposting（讀者讀到中段已經 orient 到方向）。兩者作用唔同，唔互相 replace。
用「I」呢個 pronoun 講自己 position 得唔得？
睇 department 同 field convention。Humanities 同 qualitative social science 越嚟越 accept 第一人稱，quantitative science 傳統 avoid「I」用 passive 或者「this study」。搵唔到官方文件講明全港統一標準，要對返你 supervisor 同 style guide（APA 7 accept 第一人稱，其他 style 可能保守）。
Reporting verb 升級之後，會唔會 introduce bias？
如果你 verb 選錯（例如 empirical finding 用 argued），的確會 misrepresent 作者。原則：作者實際做咗 empirical measurement用 neutral verb（found、reported）；作者提出 theoretical claim 或 interpretation用 evaluative verb（argued、proposed）；作者明確 push back 前人用 contestive verb（challenged、questioned）。改前先返返 source 睇下佢實際講咩，唔好單憑 title 判斷。
可以喺 lit review 直接批評一篇 paper 嗎？
可以，但要基於具體 methodological／theoretical grounds，唔可以係人身或者情緒。有效嘅 critique 例：「While Chan’s (2020) survey design captured cross-sectional patterns, the absence of longitudinal follow-up limits causal inference」；無效嘅 critique：「Chan’s study is not convincing」。前者 marker 認，後者會 flag「unsupported claim」。
Turnitin similarity 對 lit review 有咩特殊考慮？
Lit review paraphrase 多、reporting verb 多，similarity score 一般會偏高。單一長 quote 或者連續 paraphrase 未加引號都會抬高 percentage。實際 similarity threshold 同 declaration 處理，可以睇 Turnitin Similarity Score 香港指引。
Systematic literature review 同 narrative literature review 有咩分別？
Systematic review 有明確 search protocol、inclusion／exclusion criteria、通常配 PRISMA flow diagram，係 methodology 嚴格嘅 secondary research。Narrative review（一般 FYP／dissertation chapter）冇 systematic protocol 要求，但都需要有 theme 組織同 critical evaluation。搵唔到官方文件講明 undergraduate FYP 一定要 systematic review。
Supervisor 話「too descriptive」但 rewrite 完仲係 land 唔到？
好多時作者自己讀完再讀都 spot 唔到邊段仲係 summary tone。第三隻眼 review 一次 chapter 5 個 technique 有冇 land，通常 30 分鐘可以指出邊個 theme 段仲係 author-driven／verb 未升級／position 冇 signpost。
UniWriterPro 提供由 human writer 起草嘅 FYP 代寫 同 Thesis 代寫。
先發 chapter draft 或 supervisor feedback 我 quote —— 唔啱 timeline 我唔會硬啃你單。
相關文章：FYP Proposal 點寫：4 段 workflow 由 background 推到 research question、FYP 研究方法點揀、FYP 題目點揀？交 proposal 前 10 分鐘可行性檢查、Turnitin Similarity Score 香港指引。