mlmd: Maximum Likelihood Model Dependence Simulations

In pampakid/amor: Assessing Model Robustness in the Computation of Treatment Effects

Description

Takes in a binary dependent data set and outputs a data.frame of the estimated average treatment effects.

Usage

mlmd(data, dep, treat, predictors, level, n.sims, method = "odds",
  matched.data = NULL)

Arguments

data	A binary dependent data set.
dep	The dependent outcome binary variable that will be used to measure the effect (e.g. "success").
treat	The treatment variable (e.g. "treatment").
predictors	Vector of predictor variables.
level	Level of allowed interactions. Level = (1, 2, 3) and correspond to one-way, two-way and three-way interactions respectively.
n.sims	Number of desired simulated models.
method	Format method to present results. The default is set to predictions of log-odds ="odds". For results on the scale of the predictors ="coeffs".
matched.data	Additional data set of pre-processed data to compare the same random models on both data sets.

Value

Returns a data.frame composed by the list of estimated treatment effects and the model used in each iteration. In the case of including an additional data set, the data.frame will have to columns with the respective estimates.

Examples

## Not run: 
peacekeeping <- read.table("peacekeeping.csv")
p <- c("untype4", "wartype", "logcost", "wardur", "factnum", "trnsfcap", "decade")

mlmd(peacekeeping, "pbs2s3", "untype4", predictors=p level=2, method="odds")
## End(Not run)

pampakid/amor documentation built on Nov. 7, 2019, 3:15 p.m.