March 1, 1977Journal of Educational Statistics
Assignment to Treatment Group on the Basis of a Covariate
Authors
Discussion
Loading...
Member takes
Implication
Randomized trial analyzes treatment group assignment based on a covariate, emphasizing evaluation techniques for accuracy.
Key Points
- This research focuses on how to assign treatment groups based on a covariate and estimate outcomes accordingly.
- Employed conditional expectation estimation for dependent variable Y based on covariate X for treatment and control groups.
- Utilized regression adjustment techniques and assessed the nonparallelism and nonlinearity of conditional expectations.
- Considered blocking strategies on covariate values with sensitivity to blocking coarseness.
- Found that proper regression adjustment requires linear and parallel conditional expectations for reliable estimates.
- Identified sensitivity in estimates based on the distribution overlap of covariate X across treatment groups and blocking nuances.
Cite This Study
Donald B. Rubin (1977) studied this question.