Methods circle 27 November 2015 QUALITATIVE COMPARATIVE ANALYSIS A VERY SHORT INTRODUCTION Laura Elena Sibinescu 27.11.2015 1
TOPICS Set-theoretic approaches in social sciences What is QCA? What kind of research is it useful for? Basics and types of QCA QCA workflow Resources 27.11.2015 2
1. SET-THEORETIC APPROACHES Method of analyzing social reality where: Data: set membership scores Ex: Finland is a full member in the set of EU countries, Russia is fully excluded Relationships between social phenomena: set relations Ex: All EU members are democratic countries Results: necessity, sufficiency, causal complexity Ex: Democracy is a necessary condition for EU membership (but not sufficient) 27.11.2015 3
2. WHAT IS QCA? Developed by Charles Ragin in the late 80s Dynamic, innovative community, applications in many fields 4 characteristics: Bridges qualitative and quantitative analysis Provides tools for analyzing causal complexity Good for small/intermediate-n research Brings set-theoretic methods into social enquiry 27.11.2015 4
2. WHAT IS QCA? Bridges qualitative and quantitative analysis: Requires familiarity with cases prior to analysis Can identify cross-case patterns Respects diversity of cases Respects heterogeneity of conditions 27.11.2015 5
2. WHAT IS QCA? Analysis of causal complexity: Equifinality: different, mutually non-exclusive paths to the same outcome Conjunctural causation: the effect of a condition occurs in combination with other precisely specified conditions Asymmetry: the presence of a set/condition/outcome and its absence denote two different phenomena Ex: if A*B*C X, it does not mean that A* B* C X 27.11.2015 6
2. WHAT IS QCA? Good for small/intermediate-n research design: Too many cases for researchers to keep track of case-specific knowledge Too few for conventional statistical methods How many cases are suitable for QCA analysis: Consider the number of conditions & their logical combinations Total number of possible combinations is 2 k, where k = number of conditions If cases are too few many combinations of conditions have no corresponding cases, affects results Ex: 5 conditions 32 possible combinations; with 10 cases it leaves 22 combinations with no empirical equivalent 27.11.2015 7
2. WHAT IS QCA? Brings set-theoretic approaches into social science Social theory is mostly verbal translates well into set-theoretical formulations Indicates causal relations in terms of set-relations necessity and sufficiency 27.11.2015 8
2. WHAT IS QCA? What kind of research: Most important can be designed in terms of set relations (research problem) and membership scores (data) Solid theoretical foundation + case knowledge needed in advance Adequate number of cases and conditions In combination with other methods (e.g. case study, process tracing) 27.11.2015 9
3. BASICS AND TYPES OF QCA Language of QCA: Dependent variable outcome Independent variable condition Causal explanation causal complexity Causal explanation expressed as necessity and sufficiency Solution formula combination of conditions for the overall cases based on logical minimization (Boolean algebra) 27.11.2015 10
3. BASICS AND TYPES OF QCA Explanatory expressed as logical operators translated into everyday language: logical AND, OR and negation Ex: A + B* C X reads The presence of condition A OR the presence of condition B combined with (AND) the absence of condition C leads to outcome X Necessity and sufficiency: A X, A is sufficient for X (if the condition is present, the outcome is also present; A is a subset of X) A X, A is necessary for X (if the outcome is present the condition is also present; A is a superset of X) 27.11.2015 11
3. BASICS AND TYPES OF QCA Types of QCA: Crisp set (csqca) Fuzzy set (fsqca) Multi-value (mvqca) Temporal (tqca) 27.11.2015 12
3. BASICS AND TYPES OF QCA csqca and fsqca Crisp sets membership scores are either 1 (fully in) or 0 (fully out) Ex: Canada s membership score in the set of OECD countries is 1 Membership scores for outcome also expressed either as 1 or 0 Fuzzy sets membership scores on a scale between 1 and 0 scale depends on researcher s reasoning Ex: Norway s membership score in the set of rich countries by GDP/capita is 0.9, North Korea s is 0.2 Careful with calibration best if theoretically grounded, otherwise needs to be very transparent Scales including 0.5 are tricky cut-off point used for calculation BUT: Country X is neither rich nor poor ) 27.11.2015 13
3. BASICS AND TYPES OF QCA mvqca Extenstion of csqca, shares most features Designed for multinomial concepts that are not dichotomous in nature but the outcome should be a crisp set De-emphasizes set relations in terms of necessity and sufficiency tqca When the order of events is important Explanatory sequences Still in development 27.11.2015 14
4. QCA WORKFLOW 1. Case knowledge and data 2. Theoretical basis for finding conditions 3. Calibration of set membership scores for data 4. Data matrix 5. Truth table + logical minimization 6. Analysis of necessity 7. Analysis of sufficiency 8. Presenting/reiterating results 27.11.2015 15
4. QCA WORKFLOW Calibration: turning data into membership scores For fsqca Choose scale (ideally theoretically grounded; if not choose cut-off points but explain why!) Direct and indirect methods of calibration most software can do it automatically 27.11.2015 16
4. QCA WORKFLOW Data matrix CASES CONDITIONS OUTCOME E C I D BG 0 0 0 0 CZ 1 1 1 1 EE 1 1 1 1 HU 0 1 1 0 LV 1 1 0 0 LT 1 0 0 0 PL 1 1 1 0 RO 0 0 0 0 SK 1 1 1 1 SI 1 1 1 1 27.11.2015 17
4. QCA WORKFLOW Truth table CASES CONDITIONS OUTCOME E C I D BG, RO 0 0 0 0 CZ, EE, SK, SI 1 1 1 1 HU 0 1 1 0 LV 1 1 0 0 LT 1 0 0 0 PL 1 1 1 0 27.11.2015 18
4. QCA WORKFLOW Truth table + logical minimization: Logical minimization summary of the information in the truth table using the rules of Boolean algebra (see Resources) Gives solution formula for sufficiency but not necessity 27.11.2015 19
4. QCA WORKFLOW Analysis of necessity and sufficiency: Analysis of necessary condition should be done first Very high consistency (what % of empirical data fits inside the postulated set relation) threshold around 0.9 Analysis of sufficiency Examine each row on the truth table & decide if it can be interpreted as sufficient (see Logical minimization) Logical remainders researcher must decide what to do with them 27.11.2015 20
4. QCA WORKFLOW Presentation of results: Which conditions explain the outcome Which cases are covered (or not) by the solution How well does the solution fit with the underlying empirical evidence Forms of presentation: tables, Venn diagrams, solution term Interpretation of results: Focus on cases (combine with other methods) Focus on (part of) the solution 27.11.2015 21
5. RESOURCES 27.11.2015 22
5. RESOURCES compasss.org Publications (articles and working papers updated frequently) Books Software Training and events Some recommended courses: ECPR Methods school IPSA Summer school 27.11.2015 23