Qualitative Research Methods

Qualitative Methods

• Focuses on intro, case studies, and process tracing based on Gerring 2017, “Qualitative Methods,” Annual Review of Political Science.

Qualitative vs. Quantitative Approaches

• Quantitative (Quat) views science as a unified endeavor with shared rules, assumptions, and standards for measurement/analysis.
• Quantitative goals include replication, knowledge cumulation, and researcher consensus
• Qualitative (Qual) sees knowledge as embedded in theoretical, epistemological, or ontological frameworks.
• Epistemology tackles the science of knowing (how we know what we know).
• Methodology is within epistemology, determining how to find something out.
• Ontology explores the philosophical study of being and existence.
• Qualitative researchers believe social science inquiry is determined by how questions are approached
• Observation and scientific pursuit are grounded in human experience, including values and beliefs
• Methodological pluralism is the need for different ways to approach and answer research questions.

Conventional Definitions: Qualitative vs. Quantitative

• Qualitative research is expressed in natural language, uses small samples, and selects cases in an opportunistic or purposive way
• It focuses on particular individuals, events, or contexts, and concentrates on concepts or ideas (idiographic)
• Quantitative research is expressed in numbers/statistical models, uses large samples, and selects cases via systematic (random) sampling
• It focuses on features generalized across populations and general laws or patterns (nomothetic)

Gerring's Example of Qualitative Research

• Example is a detective story, focusing on suspects motives, location at the time of the crime, and motives of other suspects
• Each clue involves different people, so they can't be in a matrix data set and are dealt with in prose (narrative analysis)

Contrasting Quantitative and Qualitative Work

• Quantitative work examines covariation patterns in a matrix of observations, analyzed within a formal model (e.g., statistics)
• Qualitative inferences are based on noncomparable observations addressing different problem aspects, analyzed informally
• Qualitative researchers use unstructured interviews, focus groups, ethnography and archival research, looking for clues from sources

Converting Words to Numbers

• Qualitative data can be converted to quantitative data

Value of Qualitative Analyses

• Use when little is known about a subject or there is little theory
• Can follow quantitative analyses when a correlation is found and needs more understanding
• Example: interviewing leaders of democratic regimes about decisions on the use of force
• Useful when interested in just one or a few cases as opposed to gathering data on many cases

Case Selection

• Case selection depends on research goals/strategies.
• Descriptive studies aim to describe a single variable and exemplifying a common pattern

Causal Analysis Types

• Causal quasi-experimental analysis identifies causal relationships and isolates effects of X on Y Causal exploratory analysis seeks to build or identify hypotheses starting with an outcome (Y) and identifying potential causes (Xs)
• Causal estimating analysis tests a hypothesis by estimating a causal effect to see if X could have caused Y (positive, negative, or null)
• Causal diagnostic analysis confirms, refines a hypothesis based on literature, and identify mechanisms at work

Causal Inference with Qualitative Research

• Quantitative research uses correlation, time order and elimination of alternative explanations for causal inferences
• Qualitative data inferences are "looser and less precise" due to fewer observations and less formal tests

General Rules for Causal Inference with Qual Data

• Analyze sources based on relevance, proximity, authenticity, validity, and diversity of viewpoints
• Look for sources that are generalizable, neglected by research, enhance outcome probability, and are exogenous
• Canvas widely for rival explanations and pick cases where alternative explanations can be eliminated or easily tested
• Construct as many testable hypotheses as possible for each explanation
• Use counterfactual thought experiments, considering if X may have mattered under different external conditions
• Utilize chronologies/diagrams to clarify temporal and causal interrelationships, mapping how events led to others

Qualitative Tests for Causal Relationships

• Hoop test: If X can jump, it's a possible explanation; if not, eliminate it (necessary but not sufficient).
• Smoking gun test: Observing a smoking gun shows that they did it (sufficient but not necessary).
• Doubly-decisive test: observation is necessary and sufficient for demonstrating X caused Y.
• Straw-in-the-wind test: neither necessary nor sufficient for establishing X caused Y (weak evidence)

A Crowd-Based Approach

• Making a qualitative causal argument more persuasive can be attained with a crowd-based approach that relies on a number of sources

Multimethod Research

• Using a variety of methods with different advantages and disadvantages to test theories via "triangulation"

Case Study Methods

• Can be used to review one or very few cases to examine a research question
• Seeks depth and complexity, while wanting to generalize. Balancing both can be difficult

Defining Terms

• Positivism is the belief in using scientific methods to find truth and laws, and tell us about a reality in a population i.e generalizability
• Post-positivism and interpretivism is the belief in cannot describing/explaining phenomena in casual relations.

Perspective on Case Studies

• Case studies sometimes are critiqued by positivists and scholars because they view them as post-positivist or lacking in rigor
• Can be valuable for making generalizations to do what larger studies cannot in providing a detailed view of one case
• Allows for a high internal validity, measures accurately represent concept

Defining "Case"

• Case is defined as an instance of a broader phenomena that is being studied
• Focus on revolutions, wars, decisions by leaders, countries, or time periods
• The case may be applicable to a broader number of similar cases

Generalization

• When trying to generalize, researchers must be clear on what they want to because the case may be applicable to different phenomena
• Most case studies want to generalize, just as large "n" studies do, the main difference is the number of cases examined

Qualitative Methods in Case Studies

• Can also use numbers and statistics to measure the aspects of key variables

What a Case isn't

• Is not an illustrative example
• Not quantitative studies

What a Case is

• Can have one or more goals
• Single-outcome studies do not see phenomenon to generalize other cases but only seeking to understand
• All other case studies look at theory testing and development

Case-Selection Techniques

• Convenience selection selects a case that they are familiar with
• Random selection is randomly selecting a case from cases
• Strategic selection is strategically selecting a case based on their hypothesized characteristics in relations to broader cases

Single Case Studies

• The case is chosen in order to explore a theory for the causal mechanisms in a particular context
• Advantage: Higher levels of validity because the studies can go deeper into the case
• Disadvantage: They are unable to control confounding variables and are subject to selection bias

Empowered Single Case Studies

Kinds of studies:

• Selecting an extreme case with extreme values that helps to generate a hypothesis
• Selecting a deviant case that doesn't fit the theory and can be useful for new hypotheses
• Selecting a crucial case, which is most likely to exhibiting the outcome that plays an important role in testing

Selecting a crucial case

• Based on an assessment of real crucial-ness
• Do not tell much about representativeness but if the theory holds in a hard case implies it holds
• Least Likely: If I can make it there: the inference that if it can be observed it can be anywhere
• Most Likely: If I cannot make it there: If we cannot observe Y in the case of this X, then unlikely that X effects Y

Select a pathway

• Which is through to embody a distinct casual path from X to Y
• Useful to probe mechanism, embody typical relation, lend itself to testing

Advantages of Single Case Studies

• High on conceptual/internal validity
• Allows the opportunity to develop new hypothesis by exploring the causal mechanism

Disadvantages of Single Case Studies

• Selection bias/weak on external validity, might select case already exhibiting certain characteristics.
• Weak on internal validity because cannot control potential confounding factors that are being observed

Comparative Case Studies

• Most similar design: two cases are selected on the basis of presenting similar characteristics, setting aside or X and Y
• Most different design: two cases being selected on the basis of varying in every possible aspect except X and Y
• Strengths of comparative case studies show they have a great external validity than single studies alone

Weaknesses of Comparative Case Studies

• Because you are examining more than one can, cannot get to full extent of detail as one case
• Run the risk of comparing apples and oranges with the two cases

Options for How to do a Case Study

• Process Tracing: examination of diagnostic evidence selected, also known as "casual pathways"
• Structured focused comparison: Same questions are asked across cases when focusing on questions relevant to understanding the theory
• Method of congruence: see that outcomes are the same in expectations of a theory

Process tracing/Causal pathways

• What is process tracing: qualitative design involving with-in case analysis.
• Attempts to identify processes between and establish causal dynamics
• Process tracing can establish specific events, different events, and the former caused latter

Tracing the processes

• Seeking to trace leading from X to Y and explain the mechanisms
• Mechanisms are the triggered by caused to link outcomes

Three ways to think of nature of mechanisms

• Variables
• understanding when unpacked in detail
• unpacking in detail

Minimalist Understanding of Mechanisms

• does not seek to detail how X causes Z, which then causes Y

System-Understanding of Mechanisms

• The goal is to unpack the explicit process but to trace exactly from the processes
• Evaluate hypotheses about the presence, involves connections and description is essential

Another way to think processes

• Casual pathways that contains:
• Step from casual pathway, the relevant events
• And the reasons why actors in this case are linked together

Using Within-Case Evidence to Causal Inferences about Mechanisms

• Four types of evidence studying casual mechanisms:
• Statistical in empiracle
• chronology of events from mechanism
• Whose mere existence provides
• That support

Selecting cases

• Cases were tested
• Deviant cases: Detect omitted conditions
• Quadrant cases can be ignored

Variations of Process Tracing

• Process: observe between Xs and Y to map the process

• Theory-building process tracers. Analysis is used to ask "how did we get here"

Causal mechanisms can be built

• This is focused on historical
• Testing and building

Components of research design

• Select cases to studying, conceptualization
• What the statistical methodology

Difference between correlation and causation.

• Correlation that's not but

Criteria for

• To establish is that as changes so does
• Show
• Make sure and
• Are must be

Types of research designs

• ,

• ,

• that are measured at same time

• With measures

Panel the time. element the

That is, we can effect change

Experimental research

• Can correlate, to eliminate

Characteristics of Experiments

• Researchers assign: condition and condition.
• To randomize assigned
• To test: the how
• To have for treatment

Difference kinds of experimental research

• With test exam
• Random test
• The. the we have

A Pre-test

• To the .
• We do know the Specifics:

Met the requirement

• We can.

Advantages and Diadvantages

• Advantages

• : the demontrated

• the of for:

• from who

• from ,

• of coverage

• : to and experimental signi

• Trying to use but

• variable

• The Still

Survey Research

• That to
• The.
• Closed is want

Closed type of

• Easy And
• Everyone Can

A way to respond

• This sort To

Disadvantages

• Force Represent
• Wording May
• Changes -Welfare To
• Or ,

Questions Of what

• To About

• -Like

• To that

• Incompetent willing:

• Can To What

• To

• Explain

• A About

And And

What

• Should That

• And: What

• To:

To What

• It can

• What

Methods for that

• Will To And

Of where

To get: From We and What Be of