Case Studies (Ref: Ch. 7/8) 個案研究 13 Case Studies (Ref: Ch. 7/8) 個案研究 中央大學.資訊管理系 范錚強 mailto: ckfarn@mgt.ncu.edu.tw 2013.05 updated

個案的研究策略 為研究目的所進行的個案研究設計和執行 如何定義被研究的個案 如何決定需要蒐集的相關資料 在資料蒐集後，應該做些什麼 中央資管：范錚強

比較研究策略 Strategy Form Control Contemporary events Experiment how, why yes yes Survey who, what no yes where, how many, how much Archival who, what no yes/no analysis where, how History how, why no no Case study how, why no yes 中央資管：范錚強

對個案研究的偏見 欠缺嚴謹性 「教學個案」與「研究個案」 欠缺科學性概化(scientific generalization)的基礎 推展和概化理論 (analytical generalization) 頻率 frequencies (statistical generalization) 花太久時間和太多大量的資料 中央資管：范錚強

技術性定義（一） 個案研究是一項實徵的探索，它 調查一個現時的現象，並在其現實情境中(real-life context)執行 這個現象與情境並沒有清楚的界線 中央資管：范錚強

技術性定義（二） 個案研究探討 情境中有太多可能相關的變數與資料項 倚賴多重資料、證據的來源，而這些資料必須要被加以交叉檢定(triangulation) 盡量運用以前研究發展出來的理論命題(theoretical propositions)來導引資料的收集和分析 中央資管：范錚強

個案研究的變化 單一與多重個案研究 個案研究  定性研究 並非資料類型所決定，而是哲學的主張 但，強烈、本質上的共通之處 如民俗誌 ethnographic methods 並非資料類型所決定，而是哲學的主張 但，強烈、本質上的共通之處 中央資管：范錚強

個案研究與評估 解釋真實世界的因果關係，而這些關係過於複雜無法以調查或實驗加以探討 描述一個現象（突發事件）及其所發生的現實情境 以敘事或描繪的方式展現對特定主題的評估 探索一些被評估事件所發生的情況，而這些事件並沒有清楚、單一的結果 中央資管：范錚強

個案研究的設計 研究設計：連結所蒐集的資料（及所取得的結論）和原始研究問題的邏輯（如何經由適當研究設計以回答研究問題） 所有的實徵研究都有（內隱或外顯）研究設計 求取任何設計在下列品質方面的最大化： 構念效度 construct validity 內部效度 internal validity 外部效度 external validity 信度 reliability 中央資管：范錚強

研究設計 一項行動計畫 一項「藍圖」 避免：資料所提供的證據無法佐證原來的研究問題（誰這麼笨？很多人！） 原始問題  一些結論 研討什麼問題？ 哪些資料是相關的？ 收集哪些資料？ 如何分析結果？ 避免：資料所提供的證據無法佐證原來的研究問題（誰這麼笨？很多人！） 中央資管：范錚強

研究設計的組成 一項研究的問題 (how & why) 它的命題（如果有的話） 它的分析單元 (unit of analysis) 連結資料到命題的邏輯 解譯研究發現的準則 中央資管：范錚強

研究命題 導引注意力到研究範疇內應被檢視的標的上 當只有你被強迫陳述出一些研究命題時，你才有可能走向正確的方向 探索性研究雖然沒有命題，但 反應重要的理論議題 告訴你在哪裡找尋相關的資料 探索性研究雖然沒有命題，但 應陳述研究目的 state a purpose 需要判定探索成功的準則 中央資管：范錚強

分析的單元 Unit of Analysis 最基本的問題：定義一個「個案」 原始的研究問題分析的單元 與同儕討論一下可能的個案 可能是一個個人或一些較難以清楚定義的事件、個體（例如：某項計畫） 命題：避免蒐集「所有的東西」 原始的研究問題分析的單元 國家、產業、政策、特定資金流等 與同儕討論一下可能的個案 比較研究發現 中央資管：范錚強

連結資料到命題 有許多不同的方式 其中具潛力的一種：型態配合 (pattern-matching) several pieces of information from the same case may be related to some theoretical propositions time-series pattern systematic pattern matched with a proposition two potential patterns  two rival propositions 中央資管：范錚強

詮釋所獲得資料的準則 Data match one pattern better than they matched the other how close does a match have to be so as to be considered a match? Most of the time: no statistical analysis Hope: the different patterns are sufficiently contrasting that the findings can be interpreted in terms of comparing at least two rival propositions 中央資管：范錚強

研究設計階段的「理論」 Theory development prior to the conduct of any data collection differ from ethnography or grounded theory No guidance can be more misleading make their field contacts as quickly as possible Theory development as part of the design phase is essential for case studies 中央資管：范錚強

例：資訊系統建置 The case study will show why implementation only succeeded when the organization was able to re-structure itself, and not just overly the new MIS on the old organizational structure. (Markus, 1983) a nutshell of a theory of MIS implementation The case will also show why the simple replacement of key persons was not sufficient for successful implementation. (Markus, 1983) the nutshell of a rival theory 中央資管：范錚強

理論的角色 The complete research design embodies a “theory” of what is being studied not grand theory in social science Goal: to have a sufficient blueprint for your study (and this requires theoretical propositions) Theory development takes time & can be difficult existing works may provide a rich theoretical framework for designing a specific case study 中央資管：范錚強

描述性理論 the purpose of the descriptive effort the full but realistic range of topics that might be considered a “complete” description of what is to be studied the likely topic(s) that will be the essence of the description answering these questions help you toward developing the needed theoretical base 中央資管：范錚強

探索性個案研究 Literature provides no conceptual framework or hypotheses Should be preceded by statements about: what is to be explored the purpose of the exploration the criteria by which the exploration will be judged successful Types (or levels) of theories theory of the program vs. the theory of program implementation 中央資管：范錚強

從個案概化到理論 Theory  Case  Generalization Cases are not “sampling units” analytic generalization statistical generalization Cases are not “sampling units” more like experiments a previously developed theory is used as a template with which to compare the empirical results of the case study replication: two or more cases support the theory 中央資管：范錚強

研究設計的品質 Logical tests of any given research design Construct validity Internal validity External validity Reliability 中央資管：范錚強

構念效度 Establishing correct operational measures for the concepts being studied tactic phase of research use multiple data collection sources of evidence establish chain of data collection evidence have key informants composition review draft case study report 中央資管：范錚強

內部效度 tactic phase of research do pattern-matching data analysis Establishing a causal relationship, whereby certain conditions are shown to lead to other conditions, as distinguished from spurious relationships tactic phase of research do pattern-matching data analysis do explanation- data analysis building do time-series data analysis analysis 中央資管：范錚強

外部效度 Establishing the domain to which a study’s findings can be generalized tactic phase of research use replication research design logic in multiple- case studies 中央資管：范錚強

信度 Demonstrating that the operations of a study - such as the data collection procedures can be repeated, which the same results tactic phase of research use case study data collection protocol develop case study data collection data base 中央資管：范錚強

基本個案設計 single-case designs multiple-case designs holistic (single unit of analysis embedded (multiple units of analysis) TYPE 1 TYPE 4 TYPE 2 TYPE 3 中央資管：范錚強

單一個案設計 The critical case in testing a well formulated theory The case represents an extreme or unique case The revelatory case Holistic versus Embedded study 中央資管：范錚強

多重個案設計 Can be expensive Consider as multiple experiments Every case should serve a specific purpose within the overall scope of inquiry Consider as multiple experiments following the replication logic, not the sampling logic Each case selected so that it either predicts similar results (a literal replication) produces contrasting results but for predictable reasons (a theoretical replication) 中央資管：范錚強

為何不採用一般的抽樣邏輯？ Case studies should not generally be used to assess the incidence of phenomena A case study would have to cover both the phenomenon of interest and its context, yielding a large number of potentially relevant variables this would require an impossibly large number of cases - too large to allow any statistical consideration of the relevant variables 中央資管：范錚強

個案方法 develop theory design data collection protocol write individual case report write individual case report write cross-case report develop policy implications modify theory draw cross-case conclusions Conduct remaining case study conduct2nd case study conduct 1st case study select cases 中央資管：范錚強

Conducting Case Studies: Preparation for Data Collection Prior skills of the investigators The training and preparation for the specific case study The development of a case study protocol The conduct of a pilot case study Remember: case study research is among the hardest types of research to do 中央資管：范錚強

Difficulties The data collection procedures are not routinized Little room for traditional research assistant The continuous interaction between the theoretical issues being studied and the data being collected Take advantage of expected opportunities Care against potentially biased procedures 中央資管：范錚強

Basic Required Skills Ask good questions - and to interpret the answers Be a good listener and not be trapped by his/her own ideologies or preconception Be adaptive and flexible Have firm grasp of the issues being studied Should be unbiased by preconceived notions, including those derived from theory 中央資管：范錚強

Case Study Protocol An overview of the case study project objectives, issues, relevant readings Field procedure credentials & access to “sites,” sources of information, procedure reminders Case study questions the specific questions, “table shells” A guide for the case study report outline, format for narrative etc. 中央資管：范錚強

Case Study Questions The heart of the protocol Differ from those in a survey interview the questions are posed to you, the investigator each question accompanied by a list of probable sources of evidence Should reflect the full set of concerns from the initial design (only at the single-case level) 中央資管：范錚強