Quan 201: Experiments and Econometrics

Flashcards reviewing key concepts from the lecture notes on econometrics, causal inference, and randomized experiments.

What is econometrics?

A set of statistical tools that allow us to learn about the world using data.

What do empirical economists do?

Use econometrics to analyze data and can be considered data detectives.

What is a causal relationship?

How one particular factor (X) influences one particular outcome (Y).

Why are empirical estimates important?

They tell us about the size of an effect and can help improve theory.

What is the 'ceteris paribus' condition?

Other things equal; the assumption that all other relevant factors are held constant when examining the relationship between two variables.

What is a causal effect for a specific person?

The difference in outcome with treatment versus without treatment (e.g., health with insurance - health without insurance).

Why is it important to have a comparison group when making causal statements?

To avoid incomplete causal statements where the control group or comparison is not clear

Does correlation imply causation?

No, causal effects do not imply moral responsibility.

What is selection bias?

The difference in outcomes due to factors other than the treatment itself.

What is the difference in group means?

Average causal effect + selection bias.

What is positive selection?

People with insurance are healthier in the absence of insurance.

What is negative selection?

People with insurance are unhealthier in the absence of insurance.

What is a population in statistical analysis?

A complete set of items that is the subject of a statistical analysis.

What is a sample in statistical analysis?

A subset of items drawn from the population.

What is the goal of econometrics regarding populations and samples?

To learn about the population by looking at a sample.

What is mathematical expectation?

The population average of a variable.

What is conditional expectation?

The population average of Y given that X has the value x.

Why are randomized experiments considered the gold standard in econometrics?

They offer a way to eliminate selection bias and to make ceteris paribus comparisons.

How does random assignment eliminate selection bias?

Treatment and control groups are comparable, so the difference in outcome can be attributed to the causal effect of the treatment.

What is the Law of Large Numbers (LLN)?

As the sample size increases, the sample average becomes closer to the expected value.

What was the Oregon health experiment?

A study where eligible people were randomly given the opportunity to apply for free health insurance via a lottery system.

What were the findings of the Oregon healthcare experiment?

Access to health insurance has a small positive effect on self-reported health and mental health, a positive effect on financial situation, and no significant effect on physical health, however it increased emergency-department visits.

How did Bertrand and Mullainathan (2004) study racial discrimination in the labor market?

They sent out fake CVs with Black-sounding and White-sounding names and measured the call-back rates.

What were the results of Bertrand and Mullainathan's (2004) study?

Having a Black name on a CV seemed to cause a lower response rate from employers.

What is taste-based discrimination?

Discrimination based on a dislike of a particular group of people.

What is statistical discrimination?

Discrimination based on beliefs about the productivity of applicants from certain groups.

What are some limitations of experiments?

Long-term effects might be different, results might not be generalizable to other settings (external validity), people might behave differently because they know they are part of an experiment (Hawthorne effect), and the effect might be different if implemented at a large scale (general equilibrium effects).