Analysis of Qualitative Data Social Research Methods 2113

Analysis of Qualitative Data (質性資料分析) Social Research Methods 2113 & 6151 Spring, 2007 May 15~22, 2007 1

Four parts • • Comparing Methods of Data Analysis Coding and Concept Formation Analytic Strategies for Qualitative Data Other Techniques 2

Comparing data analysis methods Similarities of quantitative and qualitative data analysis (質性與量化資料分析的相似處) • inferring from the empirical details of social life – To infer (推論): based on evidence • Involving a public method or process (公開的方 法或過程蒐集資料) • Comparison is central (資料分析的重點在於比較) – Compare evidence to locate patterns (similarities and differences) • Striving to avoid errors, false conclusions, and misleading inferences (避免各項錯誤) 3

Comparing data analysis methods Quantitative data analysis Qualitative data analysis • A specialized, standardized set of techniques • Begin to analyze data after data collection and processing • Test hypotheses • Social life measured by using numbers, then using statistics • less standardized, often inductive • Begin analysis while collecting data • Create new concepts and theory • Data relatively imprecise, diffuse, and context-based 4

Explanations and Qualitative Data • Developing explanations that are close to concrete data and contexts (解釋接近具體的資料 與脈絡) – Less abstract theory grounded in concrete details, sensitive to context, could be causal • Either highly unlikely or plausible – Can eliminate an explanation b showing contradictory evidence • Best to make theories and concepts explicit (最 好讓理論和概念明確) 5

Coding and Concept Formation • Conceptualization: grounded in data – Concept formation begins during data collection – Data analysis: organizing data into categories on the basis of concepts/themes, i. e. , coding (根據概念或是主題，把資料分成不同類別來分析) – Ideas and evidence mutually interdependence 6

Coding in Qualitative Data Analysis • Coding: organizing the raw data into conceptual categories and create themes/concepts (編碼: 將 原始資料依概念類別整哩，以發展出新的主題或 概念) – Difficult for novice researchers • Two simultaneous activities: – Mechanical data reduction (資料縮減) • Reducing large amount of data into manageable piles • Quickly retrieve parts of data – Analytic categorization of data (資料依分析類別整理) 7

Three types of coding Open coding (開放式編碼) Axial coding (主軸式編碼) Selective coding (選擇性編碼) 8

Open Coding • Open coding: performed during a first pass through collected data (剛開始蒐集資料的 第一階段使用開放式編碼) – Reads all field notes or other data – Writes a preliminary code label on the edge of a record • Brings themes to the surface from deep inside the data (將主題從資料內部浮現出 來) 9

Axial Coding • A “second pass” through the data, already have an organized sets of codes/concepts • To review and examine initial codes: think about linkages between concepts (回顧檢視先前的編碼: 思考概念間的連結) – Organize themes and identify the axis of key concepts – Ask about causes and consequences, conditions and interactions, strategies, and processes, – Cluster categories/concepts (將概念或類別群聚) 10

Selective Coding • Scan all the data and previous codes, look selectively for cases for comparisons and contrasting (再一次地掃描資料與先前的編 碼，選擇性地檢視能進行比較或對比的個 案) 11

Coding and Concept Formation • Analytic memo writing (撰寫分析備忘錄): discussion of thoughts and ideas about the coding – A link between the concrete data or raw evidence and theoretical thinking • Outcropping (表面事實): events on the surface – Events as presenting deeper structural relations 12

Analytic memo writing 13

14

15

Analytic Strategies for Qualitative Data • NOTE: data analysis means a search for patterns (類型) in data • Ideal types (理念型): models or mental abstractions of social relations/processes – Pure standards used for comparison – To contrast the impact of contexts and as analogy 16

Analytic Strategies for Qualitative Data • Successive approximation (連續的近似): repeated iterations between the empirical data and the abstract concepts, adjusting theory and refining data collection each time • The illustrative method (舉例法): taking theoretical concepts and treating them as empty boxes to be filled with specific empirical examples and descriptions 17

Analytic Strategies for Qualitative Data • Analytical comparison (分析比較): developed by John Stuart Mill; need multiple cases – Identifying many factors for a set of cases, sorts through logical combinations of factors, and compares them across cases – Nominal comparison because factors most often are nominal – A few cases & intensive data analysis 18

Analytical comparison (分析比較 ) 19

Analytic Strategies for Qualitative Data • Analytical comparison: method of agreement and method of difference • Method of agreement (一致法) – What is common across cases, using a process of elimination • Method of difference (差異法): 1) locating cases similar in many respects but different in a few crucial ways, 2) focusing on the differences among cases 20

Analytic Strategies for Qualitative Data • Narrative Analysis (敘事分析): a historical writing as well as a types of qualitative data analysis • Narrative: multiple meanings – A narrative (as raw data): referring to the condition of social life – Narrative text: a story-like format people apply to organize and express meaning in social life – Narrative inquiry: method of investigation and data collection 21

Narrative Analysis • Narrative: multiple meanings – A narrative style: “storytelling” ; blending description, empathetic understanding, and interpretation – Narrative a a method of analyzing data – Narrative analysis: path dependency, periodization, & historical contingency 22

Analytic Strategies for Qualitative Data • Negative Case Method (反面個案方法): what is not explicit in the data; what did not happen – Systematically examining the absence of what is expected – A single negative case can tell a lot about what theories did not take into account 23

Other Techniques • Use maps or diagrams to summarize your points and illustrate social relations – Network analysis: use maps to show connections among a set of people, events, places – Diagrams/charts: organize ideas and investigate relations in the data – Maps: see spatial relations and supplement/reinforce results from other data 24

Software for Qualitative Data Used more in recent years several choices, select wisely 25

Qualitative Data Analysis • Qualitative data more difficult to deal with than numerical data • Need careful reading of your data (mostly texts): coding and writing analytic memos • Learn some generic and particular techniques • Writing skills are critical in presenting qualitative data 26