POLS285 Test 1 Flashcards
WForms of inference?
-Descriptive inference: Using observations from the world to learn about characteristics of populations.
-Casual Inference: Using observations from the world to learn about causal relationships.
What are the different types of studies?
-Cross sectional: We know it is measured cross sectionally because it varies across spatial units.
-Time series: Spatial Unit is the same across all observations.
What is the fundamental problem of causal inference?
Says that we can never know for certain whether a change in X caused a change in Y.
What are the four causal hurdles?
1.Correlation without causation: Correlation without underlying consideration of other factors (e.g. Mexican lemon imports prevent highway deaths).
- Causal mechanism: A credible and logical story about why or how X affects Y (and does this story make sense as well as the presence of direct evidence).
3.Reverse causation: Can we rule out the possibility that the covariation we observe is due to Y causing X?
- Confounding variable: A confounding or omitted variable Z influences both X and Y causing a spurious relationship between X and Y.
What are the different types of relationships?
-Deterministic relationships: Relationships that if some cause occurs, then the effect will occur with certainty.
-Probabilistic relationships: Such as that increases in X are associated with increases (or decreases) in the probability of Y occurring, but those probabilities are not certainties. One case alone does not decrease our confidence in the theory, let alone disprove it entirely.
What is an observational study?
-Research designs in which the researcher does not have control over values of the independent variable, which occur naturally; it is necessary that there be one degree of variability in the independent variable across cases, as well as variation in the dependent variable.
What are the different types of experiments?
-Lab experiment: Intervention in setting created and controlled by researcher (e.g. computer terminals with interface asking questions).
-Survey experiments: Individuals fill out survey questions, different subjects see different surveys.
-Field experiments: take place in real world settings. Researchers randomly assign treatment and observe outcome (examining real world causal relationships but difficult to measure).
-Natural experiments: Exploit real-world situations in which the values of IV arise “as if” a researcher had assigned them randomly (not a real controlled experiment).
What are the different types of validity?
-Content validity: The degree to which a measure contains all the critical elements that, as a group, define the concept we wish to measure.
-Face validity: Whether, not its face, the measure appears to be measuring what it purports to be measuring.
-Construct validity: The degree to which the measure is related to other measures that theory requires them to be related to.
What are the strengths and weaknesses of natural experiments?
-Strengths: Observe actual processes in real-world context. Leverage treatment that would be impossible or unethical to introduce intentionally.
-Weaknesses: Without true control one must assume assignment is random (this is why natural experiments are not pure experiments). The risk of confounding variables in natural settings is high.
What are the strength and weaknesses of field experiments?
-Strengths: Real world causal relationships.
-Weaknesses: Difficult to measure some outcomes, difficult to manipulate behavior of certain actors (e.g., elites). Potentially expensive and raises ethical concerns. Harder to control for confounders.
What are the strengths and weaknesses of lab experiments?
-Strengths: Easy to randomly assign, high degree of control (with respect to confounders), easy to carefully observe stimulus and response and is often useful for testing causal mechanisms, especially psychological ones.
-Weaknesses: Lab conditions may no mirror real world, convenience samples don’t represent broader population
What are the strengths and weaknesses of survey experiments?
-Strengths: Can ask specific questions about factors of interest. Can randomly survey large populations (good for external validity).
-Weaknesses: Answers to questions are hypothetical; people aren’t taking real-world action. Non-response may be high compared to others.
Types of validity of to assess an experiment?
-Internal validity: Extent to which study produces high levels of confidence about whether IV causes DV.
-External validity: Degree to which we are confident results apply beyond participants and circumstances of study to broader population.
What are ratio measures?
-Features: All features of interval measures plus true 0 point; can use values to calculate ratio as a result.
-Dataset form: Age of students
What are interval measures?
-Features: attributes are rank ordered, have quantifiable distance between them.
-Examples: Fahrenheit or Celsius thermometer.
-Dataset form: Average monthly temperatures for Kingston.
