F1 Introduction and course overview Flashcards
Explain what exogenous and endogenous variables are.
Exogenous variables:
Variables whose values are determined outside the system and are not affected by other variables within the system. Considered an independent variable.
Endogenous variables:
Variables whose values are determined within the system and are influenced by other variables. Considered a dependent variable.
What is a quasi-exogenous shock and variables?
Shock: An external event or change that influences a system. Mostly independent of the endogenous factors within that system but not completely random or fully outside the system’s influence.
Variables: Strong plausible case that units did not influence the assignment of treatment. Variables are treated as exogenous under certain conditions (even though they are not strictly exogenous).
What is an identification strategy?
A strategy to identify a causal relation between some variables
What does randomization mean?
The researcher have control of the assignment of treatment and units are randomly assigned to treatment (treatment is exogenous).
What does quasi-randomized/quasi-experimental mean?
Some process not controlled by the researcher. Treatment is assigned somewhat as if random.
What is temporal precedence?
Correct chronological order. Criterion for establishing causality, where the cause (x) must occur before the effect (y) in time.
What is time series data?
Observations of a unit over time which yields variation across time. Time series analysis is one unit over several time periods.
What is cross-section data?
A single point in time with multiple units. Variation across units.
What is panel data?
Multiple units over several time periods. Variation across units and time.
What are hierarchical models?
Models where data is organized according to a group-level variable (e.g. students within schools). Also known as multilevel models.
The model account for relationships within and between groups.
What are nested data models?
Data have nested structures. One level is contained within the other (e.g. students within schools)
What are mixed models
A multilevel model with at least one an independent variable estimated with random slope and one without.
What are random effects models?
?
Mention strengths of quantitative designs
1) Rigorously test theories with a large set of observations
2) Approximate randomness through causal inference
3) Better separation of theory and test
Mention strengths of qualitative designs
1) Achieve intimate familiarity with topic/setting/context/case
2) Detailed knowledge on phenomena of interest and causal relationships
3) Build theories based on this knowledge
