estimates_hierarchical: Estimate average causal effects under a two-stage...
In szonszein/interference: Design-Based Estimation of Spillover Effects

Description Usage Arguments Details Value References Examples

Estimate average causal effects and their variance under two-stage hierarchical treatment assignment, assuming partial and stratified interference.

1	estimates_hierarchical(estimator_data)

estimator_data

data frame of N observations and the variables:

group:: Integer vector specifying the group label.
group_tr:: Numeric indicator of group assignment to high saturation (psi) in the first stage.
indiv_tr:: Numeric indicator of individual assignment to treatment in the second stage.
obs_outcome:: Numeric vector of observed outcomes.

estimates produces values of the estimators proposed by Hudgens and Halloran (2008) of the population average direct causal effect under high treatment saturation (psi), the population average direct causal effect under low treatment saturation (phi), the population average indirect causal effect, the population average total causal effect, and the population average overall causal effect. It also produces values of the variance estimators which assume stratified interference (i.e. potential outcomes of a unit are affected by its own treatment assignment and only the treated proportion of its group; the precise set of treated group members does not matter).

A list of 2 lists:

A list of 5 scalars corresponding to the values of the estimators of the average direct causal effect under high treatment saturation (psi), the average direct causal effect under low treatment saturation (phi), the average indirect causal effect, the average total causal effect, and the average overall causal effect.
A list of 5 scalars corresponding to the values of the variance of the estimator of the average direct causal effect under high treatment saturation (psi), the average direct causal effect under low treatment saturation (phi), the average indirect causal effect, the average total causal effect, and the average overall causal effect.

Hudgens, M.G. & Halloran M.E. (2008). Toward causal inference with interference. Journal of the American Statistical Association, 103(482), 832–842.

Aronow, P.M. et al. (2020). Spillover effects in experimental data. arXiv preprint, arXiv:2001.05444.

# Simulate first and second stage treatment
# assignment and outcomes:

group <- rep(1:6, each = 30/6)
c <- 1/2
k <- c(2/5, 3/5)

tr_assignment <-
make_tr_vec_permutation_hierarchical(group, c, k, R = 1,
                                     seed = 357)[[1]]

potential_outcomes <- make_dilated_out_hh(N = 30,
                                          seed = 357)

# Create data frame with group label, first and second stage
# treatment assignment and outcomes:

estimator_data <- make_estimator_data(tr_assignment,
                                      potential_outcomes)

# Estimate average causal effects and their variance:
estimates_hierarchical(estimator_data)