Randomized Controlled Trials Flashcards
(13 cards)
Assignment mechanism
The mechanism by which individuals came to receive the treatment they actually received
Randomized experiments
Units are randomly assigned by the researcher into a control and treatment group. There is no selection and you can estimate the ATE simply by comparing the average observed outcomes in the treatment and control group
Natural experiment
Units are randomly assigned but not by the researcher into control and treatment group
Measuring Compliance
Assessing to what extent do participants take up treatment
Random assignment
The treatment assignment is independent of the potential outcomes and the difference in means estimator is unbiased for the ATE
Balance Test
Need to assess that randomization worked and can test whether the treatment and control group are balance before treatment by regressing pre-treatment covariates on the treatment. Basically, this means testing the null hypothesis that the difference between treatment and control = 0
Sampling Variation
Occurs as a result of the procedure that selected units into the sample
Estimand SATE
The Sample average treatment effect. Here uncertainty occurs only from hypothetical randomization and inferences are limited to the sample in the study
Estimand PATE
The Population average treatment effect, in which variation from the sampling process and treatment assignment occurs.
Stratified Randomized Experiments
Can also block sample units into groups with similar observable characteristics and then run experiments within each group
Internal Validity
the ability of a study to estimate causal effects within the study population
External validity
the ability to extrapolate and generalize the estimates to other samples and populations
Threats to internal validity
Failure of randomization, spilover effects, non-compliance, and attrition or when the outcomes cannot be measure for some participants
