FYP 研究方法點揀：由 research question 型態反推 qualitative／quantitative／mixed methods

本文由 UniWriterPro 學術規例研究組 verify · Last updated: 2026-08-26 · 本文係按 CityU Run Run Shaw Library Research Methods LibGuide、CUHK 教育學院 EDM6402 Mixed Methods 課程材料、PolyU ELC FYP Methodology guide、HKU Libraries Research Support 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

• 錯嘅起點：由「你識做邊個」揀
• 先 anchor 返三個 method 嘅官方 definition
• Axis 1：由 research question 動詞反推
• Axis 2：由 data type 反推
• Axis 3：由 sample access reality 反推
• Axis 4：由你識做嘅 analysis skill 反推
• Mixed methods 幾時值得做
• FAQ

錯嘅起點：由「你識做邊個」揀

好多同學諗 fyp 研究方法點揀嘅時候，第一句自問係「我 SPSS 識用啲，唔識 NVivo，咁我做 quantitative」。呢個起點錯咗。由 method skill 反推 method design，等於用工具倒推問題，會出現「拎住鎚仔睇乜都似釘」嘅情況——你條 research question 明明係「understand how HK Year 3 students experience burnout」（呢條 question 天然係 qualitative），但你為咗用 SPSS 就砌一份 Burnout Inventory survey，最後 collect 到 quantitative 數據都答唔到條 question 想問嘅 lived experience。

CUHK 教育學院 EDM6402 Mixed Methods 課程材料對呢層講得好直接：

A primary consideration is that the design should match the research problem

CUHK Faculty of Education — EDM6402 Mixed Methods

正確起點：由 research question 型態反推 method design。Method skill 唔夠？學。Sample access 唔夠？調整 scope。但唔可以由 skill 或 access 倒轉揀 method，然後夾硬將 research question 拉去對應個 method——marker 一睇就見到 mismatch。

HKU Libraries Research Support 亦強調 methodology 係 research cycle 一部分：

Designing an appropriate research methodology to collect and analyze data is an important component of your research cycle.

HKU Libraries — Research Support: Research Methods

Component 呢個字係 key——methodology 唔係獨立揀嘅選擇，係整個 research cycle 裡面對應返 research question 嘅一環。如果你未 lock research question，返返上一步。可以睇 FYP Proposal 點寫：4 段 workflow 由 background 推到 research question，focus 喺點 draft 一條 clear／focused／researchable 嘅 question。

都可以直接交俾我哋幫你做 research method：FYP / Final Year Project 全程支援

好擔心你份 FYP? 即時睇吓其他 part 點做：FYP 題目點揀、Literature Review 點先展示到你嘅 critical engagement? 、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

先 anchor 返三個 method 嘅官方 definition

反推之前，先擺清楚三個 method 官方點定義。CityU Run Run Shaw Library 嘅 Research Methods LibGuide 對三個都有標準定義：

Qualitative Methods – Involve a researcher describing kinds of characteristics of people and events. Qualitative researchers study things in their natural settings, attempting to make sense of, or interpret phenomena in terms of the meanings people bring to them.

CityU Run Run Shaw Library — Research Methods: Key Resources

Quantitative Methods – Focus attention on measurements and amounts (more and less, larger and smaller, often and seldom, similar and different) of the characteristics displayed by the people and events that the researcher studies.

CityU Run Run Shaw Library — Research Methods: Key Resources

Mixed Methods – Involve collecting, analyzing and mixing both qualitative and quantitative data in a single study or a series of studies.

CityU Run Run Shaw Library — Research Methods: Key Resources

三個 keyword：qualitative 係「meanings people bring」——側重意義同詮釋；quantitative 係「measurements and amounts」——側重量度同比較；mixed methods 係「both qualitative and quantitative data in a single study」——同時做兩邊。

三個 definition 擺喺度做 anchor，下面 4 個 axis 逐個對返你自己嘅 research question。

Axis 1：由 research question 動詞反推

Research question 用邊個動詞，直接暗示 method design。

• 動詞係「measure」／「correlate」／「compare」／「predict」／「test whether」／「to what extent」→ 應該考慮 quantitative。呢類動詞天然要求數字上嘅比較同 statistical inference。例：「To what extent does weekly social media usage predict self-reported anxiety scores?」——「To what extent」＋「predict」＋「scores」= quantitative signal 齊全。
• 動詞係「explore」／「understand」／「describe how」／「interpret」／「give voice to」／「what does X mean to Y」→ 應該考慮 qualitative。呢類動詞要求 lived experience、meaning、subjective interpretation，統計數字答唔到。例：「How do Hong Kong Year 3 undergraduates experience burnout during FYP submission?」——「How do」＋「experience」= qualitative signal。
• 動詞係「examine both the extent of X and how X is experienced」→ 應該考慮 mixed methods。有 quant 部分（extent）＋ qual 部分（experienced），一個 method 答唔晒。

Self-test：將你 draft 條 research question 嘅動詞 highlight 出嚟。上面三 category 對應唔到？可能你條 question 未夠 specific。返返 proposal 段 refine question，唔好硬砌 method。

Marker 睇到嘅 signal：如果你條 question 用「explore」但 method 用 survey with Likert scales，marker 會寫「the method does not match the exploratory nature of the research question」。呢個 mismatch 唔係文法問題，係 design 問題，重寫成本高。

Axis 2：由 data type 反推

你要 collect 嘅原始 data 天然係邊個 type，直接決定 method。

• Data 天然係 numerical（例：GPA、股價、hours spent、frequency count、pre-existing scale score）→ quantitative。
• Data 天然係 textual／observational／experiential（例：interview transcript、field observation、diary entry、social media post 內容分析）→ qualitative。
• Data 有 numerical component 亦有 textual component 都需要（例：Likert scale 分數 + open-ended follow-up 訪問）→ mixed methods。

呢層有一個常見誤區要 flag：把 open-ended 問卷答案當成「qualitative data」加去 quantitative study，然後聲稱做咗 mixed methods。呢個 framing marker 唔會認——open-ended 問卷嘅 3 句短答案唔等於 in-depth interview 嘅 45 分鐘 transcript。真正 qualitative data 有 depth 要求，唔係任何 non-numeric text 都算。

Self-test：假設你聽日開始 data collection，你 physically 落 spreadsheet／transcript 嘅第一行係咩樣？如果係一堆 number → quantitative；如果係一段對答 verbatim → qualitative；如果兩樣都要 collect 完整 dataset → mixed methods。

Axis 3：由 sample access reality 反推

Method 設計嘅第三閘：你 realistic 上可以接觸到嘅 sample size 同 depth。

• 大 sample 有機會（n ≥ 100，經 mailing list、online panel、public dataset）→ quantitative 做得起（regression、correlation 等 statistical test 有 power）。
• 大 sample 冇機會但可以 recruit 到 8–15 個願意畀 45–90 分鐘 in-depth interview 嘅 participant→ qualitative 做得起（thematic analysis、grounded theory 等一般 saturation 喺 8–15 人達到）。
• 兩邊都收得到（例：60 個 survey response + 8 個 follow-up interview）→ mixed methods 做得起。
• 兩邊都收唔到→ 唔係揀 method 嘅問題，係 topic 本身 access reality 有問題，返返 feasibility check 第 3 項重新 verify data availability。

Undergraduate FYP 特別要留意：quantitative 嘅 n 少於 30 通常做唔到有 meaningful statistical inference（除非用特定小樣本設計例如 case study within-subject design），呢個係一個常見 rewrite trigger。同樣，qualitative 得 3 個 interview 通常唔夠 saturation。呢兩個 threshold 冇一個全港統一嘅官方 minimum，要問返 supervisor 對你 field 嘅 convention。

Self-test：由今日起 4 星期，你 physically 攞到手嘅 sample size 底線係幾多？如果連底線都答唔到具體數字，method design 未 ready，返 pilot recruitment。

Axis 4：由你識做嘅 analysis skill 反推

Skill 唔係揀 method 嘅起點（呢個係開頭講嘅錯誤），但係揀 method 之後最後嘅 sanity check——你要 realistic 睇下你能唔能夠喺 timeline 內做完 analysis。

PolyU ELC 對 methodology 段有一個好嚴嘅 replication test：

Provides sufficient detail for another researcher to replicate the study precisely.

PolyU ELC — FYP Methodology

意思係你要能夠寫得夠具體，令另一個 researcher 跟得返。如果你自己都唔識做，寫都寫唔具體。

• Quantitative：SPSS／R／Stata 嘅 regression assumption（linearity、multicollinearity、homoscedasticity）你 check 唔 check 得掂？t-test／ANOVA 嘅 post-hoc 你識唔識揀？如果冇實際跑過完整 exercise，你 proposal 寫「will conduct multiple regression」係空口。
• Qualitative：thematic analysis 六階段（Braun & Clarke 或者其他 established framework）你識唔識行？coding scheme 點建、inter-coder reliability 點量？NVivo／MAXQDA 用過未？
• Mixed methods：要兩邊都做得掂。做唔掂唔好賭。

如果 skill 有 gap，兩條路：（1）由而家開始 upskill，YouTube、DataCamp、Coursera 都有 tutorial，預兩到四星期學到 baseline；（2）換去你識做嘅 method（前提係 axis 1–3 都對得返，唔係為咗 skill 而 twist question）。唔應該有嘅第三條路：proposal 寫住個 method 名，賭 supervisor 睇唔穿。

Self-test：由 raw data 落 software、run analysis、output 圖表、interpret result，中間有邊幾個 concrete step？講唔到全條 chain = 未識做。

Mixed methods 幾時值得做

Undergraduate FYP 常見誤區：為咗 show sophistication 而做 mixed methods。CUHK 教育學院 EDM6402 對呢個 tendency 有一個非常直接嘅提醒：

Researchers should carefully select a single design that best matches the research problem. This will make the study more manageable and simpler to implement and describe.

CUHK Faculty of Education — EDM6402 Mixed Methods

意思係如果 single design 已經 match 到 research problem，就 stick with single design——mixed 唔係更 impressive，係更難做。Undergraduate FYP timeline 好多時做唔到 mixed methods 兩邊都 rigorous。

Mixed methods 真正值得做嘅 3 個 scenario：

• Triangulation：同一個現象用兩種 data 互相驗證。例：問卷量到 usage frequency 同 anxiety score correlation，interview 深入了解為咩會有呢個 correlation。
• Explanatory sequential：先 quantitative 揾出 pattern，再 qualitative 解釋 pattern 點解出現。例：先 survey 揾出 usage 高嘅 group 得分低，再 interview 呢班人了解具體 mechanism。
• Exploratory sequential：先 qualitative 揾出 relevant variable／theme，再 quantitative 量度呢啲 variable 喺 population 嘅分佈。例：先 interview 揾出「FYP burnout」嘅 3 個 dimension，再 build survey 量呢 3 個 dimension 喺全體 Year 3 學生嘅分佈。

如果你嘅 research question 唔屬呢 3 個 scenario 任何一個，single design 更合適。EDM6402 對 mixed methods 嘅 advantage 亦有 balanced 講法：

A major advantages of mixed methods research is that it enables the researcher to simultaneously answer confirmatory and exploratory questions, and therefore verify and generate theory in the same study.

CUHK Faculty of Education — EDM6402 Mixed Methods

Key 係「同時 answer confirmatory 同 exploratory questions」——你條 research question 有冇同時兩層？冇 = 唔需要 mixed。有 = 值得考慮，但 verify 你 timeline 做唔做到兩邊。

如果 4 個 axis 走完你仲係唔肯定 method 應該點揀，或者 supervisor 已經 push back 你 method choice 但你唔清楚點改，可以搵 UniWriterPro 由 human writer 起草嘅 FYP 代寫——我哋見過幾百份 methodology 段，可以直接指出邊個 axis 嘅 mismatch 令 design 唔 hold water。如果只想現有 methodology draft 由第三隻眼 tighten wording 同 logic，可以睇 academic proofreading 服務。發 assignment brief 或現有 methodology draft 過嚟就可以 quote。

另外，如果你打算用 GenAI 輔助 draft methodology 段（例如 brainstorm sample recruitment strategy、check analysis technique 嘅 name），記得對返你 department 嘅 GenAI 使用政策同 declaration 要求，詳細睇 香港八大 GenAI 使用政策同 declaration 指引。

FAQ

相關文章：FYP 題目點揀？交 proposal 前 10 分鐘可行性檢查、FYP Proposal 點寫：4 段 workflow 由 background 推到 research question、香港八大 GenAI 使用政策同 declaration 指引。