Please enable JavaScript.
Coggle requires JavaScript to display documents.
Research Methods in Political Science - Coggle Diagram
Research Methods in Political Science
philosophy
positivism (Comte)
logical positivism
empiricism + logical reasoning
retroduction
verification
deduction
falsification (Popper)
rejection induction
rejection verifiability
goal must be falsification
classical positivism
empiricism
induction
Carl Gustav Hempel
deductive nomological model
explanation through universal law-like generalization
hypothetico deductive model
tests ability of law to predict events
scientific realism
causal mechanisms (Tilly)
cognitive mechanism
relational mechanism
environmental mechanism
agent vs. structure (Coleman)
individualism
holism
reality includes observable elements with observable consequences
interpretivism
social world subjectively created
understanding by interpretation of meaning
hermeneutics: theory and methodology of interpretation
Imre Lakatos
scientific research programs
hard core: commonly accepted hypotheses
protective belt: auxiliary, adjustable hypotheses
Thomas Kuhn
normal science & scientific revolution
research design
research ethics & validity
validity & reliability
common threats
testing effects
instrumentation
maturation
selection bias
history
regression towards the mean
ceiling and floor effects
mortality/attrition
validity: whether the explanatory factor can be linked to the outcome (accuracy)
internal validity: whether the findings are true for the sample
content validity
construct/criterion validity
predictive
convergent
concurrent
discriminant validity
face validity
external validity: whether the findings are true for the population
generalizability
reliability: findings consistent when repeated (precision)
internal consistency
test-retest reliability
replicability
inter-coder reliability
research ethics
data analysis
ensure transparency
enable replication
presentation: do not plagiarize
data collection/storage
privacy
fraud prevention
3 ethical principles
confidentiality (privacy)
anonymous information
confidential information
public information
informed consent (voluntary)
do no harm (risk assessment)
formal review of research
pre-registration
data management plans
formal board approval
when science goes wrong
fake data
plagiarism
harm to subjects
Milgram study
Stanford prison experiment
data & measurement
concept: contract defined by mutual agreement (Gerring)
familiarity
resonance
parsimony
coherence
differentiation
depth
theoretical utility
field utility
measurement reliability & measurement validity
measurement reliability: the extent to which a measure continually obtains consistent results (precision)
error: difference between observation and true score
random error: expected, affects reliability
systematic error: indicative of bias, affects validity
reliability coefficient
Cronbach's a
split-half method
measurement validity: the extent to which a measure measures what is is intended to measure (accuracy)
units of analysis
levels of analysis
characteristic
measurement
unit
fallacies
ecological fallacy (macro -> micro)
individualistic fallacy (micro -> macro)
triangulation
types
investigator triangulation
methodological triangulation
data triangulation
outcomes
convergence: results in accurate data
inconsistency: results in nuanced data
contradiction: corrects inaccuracies
types of data
quantitative vs. qualitative
sources
observation
documents
people
secondary data
primary vs. secondary
causality & overview of designs
limitations
nomothetic causality unlikely
zero-order relationship
partial relationship
intervening variables
mediator variable
moderator variable
reinforcer
distorter
suppressor
probablistic (instead of deterministic)
antecedent variables
confound variable
research design: strategy for investigation of RQ
cross-sectional & longitudinal design
cross-sectional: analysis at a single point in time
longitudinal study: analysis over time
cohort study: same group over time
panel study: random sample over time
comparative design
single-N
small-N
large-N
experimental design
field experiment
natural experiment
laboratory experiment
historical design: studies events & processes
conditions of causality (Mill)
temporal ordering
spatial & temporal contiguity
covariance/correlation
non-spurious connection
methods of data collection
questionnaires/surveys
interviewing
semi-structured
focus group
unstructured
ethnography
structured
discourse/content analysis
research questions & theories
characteristics
relevance
scientific relevance
social relevance
use
guide
researchability
novelty
question types
descriptive
explanatory
prescriptive
predictive
normative
unanswerable questions
false dichotomy
fictional event
incomplete information
metaphysical question
tautological question
theory: simplified version of reality
types
process
inductive reasoning
deductive reasoning
nature
empirical
normative
scope
grand theory
middle range theory
grounded theory
sorting of codes (categories)
memo writing (describe categories)
coding collected data
hypotheses: expected answer to the research question
variable
independent variable
dependent variable
relationship
null relationship
association
covariance
correlation
causation
reciprocal causation
data collection
interviews & focus groups
interviews
forms of interview
semi-structured
unstructured
structured
special types
elite interview
recruitment approaches
direct recruitment
gatekeepers
public advertising
expert interview
expert surveys
modes of interview
telephone interview
postal/internet survey
comprehension problems
face-to-face interview
interviewer effect
methodological issues
social desirability bias
interview guide
synchronous
asynchronous
researcher bias
neutrality
no influence/guiding/reacting/interrupting
take notes/record in a discrete manner
informed consent
focus groups
characteristics
6-10 members
purposive selection
interviewer role: moderator/facilitator
semi-structured, 60-90 minutes
analysis process
quantification
coding
data retrieving (ordering)
data reduction
data analysis
create transcript
experiments
types
field experiment
natural experiment
survey experiment
laboratory experiment
methodological issues
randomization
random sampling (external validity)
random assignment (internal validity)
experimental assumptions
conditional independence (exogeneity)
no selection bias
counterfactual assumption
no omitted variable bias
no publication bias
experimental designs
solomon 4 group design
delayed effects design
posttest-pretest experiment
factorial design
posttest experiment
key characteristics
group assignment (random/control-treatment matching basis)
design (between/within)
intervention
analysis
randomized allocation (control/treatment)
validity & ethics
reactivity
ethical issues
concealment
deception
ecological validity
surveys & sampling
survey
ensuring accurate measurement
pitfalls
measurement validity
measurement reliability
ensuring accurate representation
coverage of sampling frame
sampling error
non-response
cross-level inferences
types
cross-sectional survey
longitudinal survey
rolling cross-section
trend study
cohort study
panel study
cross-lagged causal analysis
non-scientific/unethical surveys
push polls
frugging & sugging
purpose: prediction
methodological issues
close-ended or open-ended questions?
account for reactivity
maintain logical question order/grouping
wording
avoid vague terms or unknown acronyms
avoid leading, negative, or double-barreled questions
avoid biased/loaded/sensitive items
maintain operationalisation
respondent issues
solutions
balanced questions
pre-testing
randomization
monitoring/verifying
interviewing modes
telephone interview
postal/internet survey
comprehension problems
face-to-face interview
interviewer effect
pitfalls
recall problems
social desirability bias
misunderstanding
acquiescence
samples
accounting for bias
selection bias
response bias
self-selection bias
nonresponse
sampling techniques
probability sampling
stratified random sampling (subgroups)
reweighting
systematic random sample (interval)
cluster sampling (large -> small)
simple random sample
with replacement
without replacement
non-probability sampling
convenience sampling
purposive sampling
quota sampling
snowball sampling
theoretical sampling
sample: manageable subset of population
sampling issues
disproportionality
a priori weighting
post hoc weighting
sampling error
decreases with sample size
desired sample size depends on
homogeneity
desired precision
response rate
cooperation rate
surveyed rate (participate too often)
contact rate
response rate
population
finite vs. infinite
known vs. unknown
ethnography & participant observation
advantages
unpicking of multifaceted phenomena
first hand in depth observation
useful for sensitive topics
methodological issues
case selection: single case study
access: open or closed?
role of observer
participant or observer?
overt of covert?
observation: structured or unstructured
documentation: field notes/recording
analysis: systematic qualitative summary/interpretation
thick description
validity & ethics
replicability
trustworthiness
transferability (external validity)
dependability (reliability)
credibility (internal validity)
confirmability (objectivity)
reduce bias
ethical issues
informed consent: voluntary participation
protection of privacy: hide personal details
participant observation
comparative & historical research
historical research
focus on time as context
cross-sectional comparative research
historical process research: longitudinal
historical events research: cross-sectional
comparative historical analysis (CHA): longitudinal, many cases
parallel theory demonstration
contrast of contexts
macro-causal analysis
focus on timing of events
narrative case-studies
events structure analysis (ESA)
construction of overarching narrative account
series of short, individual statements of events
order into diagram/flowchart
historical institutionalism
critical juncture
positive feedback
self-enforcing mechanisms
process-tracing
domino-theory: must include all intermediate steps
case selection & sampling
case selection approaches
MSSD
MDSD
case selection challenges
outliers
heterogeneity
selection bias
historical contingency
path dependency
selection by outcome
case selection process
select sample
identify population
case selection techniques
diverse case selection
extreme case selection
representative case selection
deviant case selection
crucial case selection
influential case selection
pathway case selection
MSSD/MDSD case selection
comparative method
comparison: must serve theoretically justified research
uses of comparison
theory testing
theory generation
theory application
types
small-N (2-5)
large-N
concept stretching: measure cannot be applied universally
intermediate-N: qualitative comparative analysis
truth table
crisp set
fuzzy set
analysis: paired comparison of all possible factors
single-N
thick description
pitfalls
false uniqueness
false universalism
textual/content analysis & big data
textual analysis
foundational theoretical considerations
interpretivism
constructivism
speech act theory (language serves action and function)
post-structuralism (discourse creates)
advantages
no bias
allows access to sensitive topics
methodological issues
determine categories
intensity or valence (tone)
manifest or latent
case selection: which documents should be analyzed?
coding
closed (a priori, codebook)
open (grounded, protocol)
analysis
quantitative (use codes to identify relationships)
qualitative (presenting quotes & trends)
determine recording unit
syntactical (explicit)
referential (implicit)
thematic (broad theme)
types
discourse analysis
content analysis
reliability
coder stability
reproducibility
objectivity
plausibility
humans vs. computers
manual content analysis
inter-coder reliability
computer assisted analysis
reduced ability causal inferences (low internal validity)
systematic analysis (perfect reliability)
big data
analysis techniques
computer assistancr (big datasets)
data mining (looking for correlations)
ethical issues
privacy (public/private info)
purpose (private/commercial)
informed consent (voluntary participation)
confidentiality (protection of privacy)
3 V's
velocity
variety
volume
recommended measures
anonymization
data minimization
data encryption
data analysis
data analysis
data management
record data in spreadsheet
save backup of original data
check values, scale, direction
types of analysis
univariate analysis
bivariate analysis
multivariate analysis
using secondary data
proxy indicator (close enough)
validity & reliability accounted for
drawbacks: no control
ethical principles assumed
inferential statistics
distributions
population distribution
sample distribution
sampling distribution
estimation
Central Limit Theorem
standard error
confidence interval
null hypothesis
critical values
statistical significance
assumptions
simple random sample with perfect response
no non-sampling error
complete coverage