WEEK 1-2 Flashcards
Scientific Method
study politics identifying causal mechanisms
incremental, structural knowledge
apply methodology (guides principles of theory-building)
empiricism
prerequisite for reasoned analysis. recall: political science abt fact finding. statistical analysis tests empirically driven theories…
Empiricism is the belief that knowledge comes primarily from observable, measurable experience rather than theory or pure reasoning.
quantitative analysis benefits
gives high-speed way of collecting info and testing hyp
compliments qualitative/theoretical work
methodologies are used to collect large volumes of data
consistent criteria..
consistency guarantees reliability within principles of collection
uses of quantitative analysis
produces info (academic research, non-academic settings)
consumes info (reading academic research, reports, thinking of info in empirical terms and collection procedures)
hypothesis v. null hyp
Relationship between cause and effect.
NEVER 100% proven
null hyp: theory-based statement about what we would observe if there were no relationship between an IV and DV
Hypothesis testing
Process in which scientists evaluate systematically collected evidence to make a judgment of whether the evidence favors their hypothesis or favors the corresponding null hypothesis
When to reject null
rejecting null means you learned something about the dependent variable. cant reject null means didnt learn anything new
IV, DV, CONTROL, CONFOUNDING
IV = what you manipute (causal mechanism we want to observe)
DV = what you measure (outcome you want to observe
CONTROL = what you need to remain constant to ensure accurate results (Isolates true effect of IV
CONFOUNDING = variable that is correlated with both the IV and DV and somehow alters the relationship (Explains false correlations)
operationalize
when you are testing variables IRL
process of translation from theory to variables
PRE-HYP is…
theoretical model
principles for theory-building
- make theories causal
- avoid normative statements
- consider only empirical evidence
- pursue both generality and parsimony
- do not let data drive your theories
theory v. paradigm
theory = tentative conjecture about the causes of some phenomenon of interest
paradigm = shared assumptions and accepted theories in a field. Once a paradigm has been accepted, researchers can start conducting more technical work → paradigm shift
criticisms of political science
Is the scientific method itself prone to, or reflective of, ideological biases?
Is it possible to collect data independently of individual ideological concerns?
Can successful academics crowd out challenges even when their views are outdated?
Note the difference between closed and open-ended questions
empirical objects in poli sci
Institutions, Attitudes, Behaviours, Trends
Theory-building
building theories = self-generated based on interpretation. causal relationships within the ideas. parts of theory-building is translating something quantifiable
scientific method
building INCREMENTAL KNOWLEDGE by TESTING.
Theoretical propositions => evaluate those propositions empirically
research builds on previous research in increments
theory building process
- credible causal mechanism that connects X to Y?
(YES/NO - reformulate theory until the answer is yes)
- can we eliminate the possibility that Y causes X?
(YES/NO - proceed with caution to hurdle 3)
- Is there covariation between X and Y?
(YES/NO - think about confounding variable before moving to hurdle 4)
- Have we controlled for ALL confounding variables X taht might make the association between X and Y spurious?
YES (proceed with confidence and summarize findings)
MAYBE (control for confounding variables under answer is yes)
NO (stop and refolrmulate your causal explanation)
Endogeneity problem
situation in probabilistic statistics where the predictor variables and the dependent variable may be influenced by each other or by an unmeasured factor.
Process tracing:
Qualitative method using cases to identify pathways of causal mechanisms. Can use this to identify more specific aspects of a relationship with endogeneity. Flowcharts with multiple steps can help us with this.
Unit of analysis
parameters on which we are collecting data
Macro-level: Countries, provinces, institutions.
Micro-level: Individuals, surveys.
Time-series: Tracking variables over time.
Regression analysis
Coefficient: Measures the strength of the relationship between X and Y.
Standard Error (SE): The average deviation of observed values from predicted values.
p-value: Determines statistical significance (p < 0.05 = 95% confidence).
Confidence Interval: The range in which the true effect size likely falls.
PANEL data
also known as time series, collecting data on a unit over time
Structural data
sing a unit of analysis based on groups (countries, provinces, etc.)
