A Taxonomy of Classical Randomized Experiments
Lists
Only you see this
Abstract
In this chapter we introduce four specific examples of classical randomized assignment mechanisms, and we relate these examples to the general taxonomy of assignment mechanisms described in the previous chapter. The four examples, Bernoulli trials, completely randomized experiments, stratified randomized experiments (randomized blocks), and paired randomized experiments, all satisfy the four criteria necessary for assignment mechanisms to be classified as classical randomized experiments. These criteria, as discussed in more detail in Chapter 3, require that the assignment mechanism (i) is individualistic, with the dependence on values of covariates and potential outcomes for other units limited; (ii) is probabilistic – each experimental unit has a positive probability of being assigned to the active treatment and a positive probability of being assigned to the control treatment; (iii) is unconfounded – that is, given covariates, does not depend on potential outcomes; and (iv) has a known functional form that is controlled by the researcher.
BibTeX
@inproceedings{imbens2015taxonomy,
author = {Guido W. Imbens and Donald B. Rubin},
title = {{A Taxonomy of Classical Randomized Experiments}},
booktitle = {Cambridge},
year = {2015},
doi = {10.1017/cbo9781139025751.005},
}Loading related content...