treatment_model: Enhanced Estimation of Treatment Effects of Binary Data from...

In JobNmadu/Dyn4cast: Dynamic Modeling and Machine Learning Environment

Enhanced Estimation of Treatment Effects of Binary Data from Randomized Experiments

Description

Observational study involves the evaluation of outcomes of participants not randomly assigned treatments or exposures. To be able to assess the effects of the outcome, the participants are matched using propensity scores (PSM). This then enables the determination of the effects of the treatments on those treated against those who were not treated. Most of the earlier functions available for this analysis only enables the determination of the average treatments effects on the treated (ATT) while the other treatment effects are optional. This is where this functions is unique because five different average treatment effects are estimated simultaneously, in spite of the one line code arguments. The five treatment effects are:

1. Average treatment effect for the entire (ATE) population

2. Average treatment effect for the treated (ATT) population

3. Average treatment effect for the controlled (ATC) population

4. Average treatment effect for the evenly matched (ATM) population

5. Average treatment effect for the overlap (ATO) population.

There are excellent materials dealing with each of the treatment effects, please see Understanding propensity score weighting

Usage

treatment_model(Treatment, x_data)

Arguments

Treatment	Vector of binary data (0 = control population, 1 = treated population) LHS for the treatment effects estimation
x_data	Data frame of explanatory variables for the RHS of the estimation

Value

A list with the following components:

Model	Estimated treatment effects model.
Effect	Data frame of the estimated various treatment effects.
P_score	Vector of estimated propensity scores from the model
Fitted_estimate	Vector of fitted values from the model
Residuals	Residuals of the estimated model
`Experiment plot`	Plot of the propensity scores from the model faceted into Treated and control populations
`ATE plot`	Plot of the average treatment effect for the entire population
`ATT plot`	Plot of the average treatment effect for the treated population
`ATC plot`	Plot of the average treatment effect for the controlled population
`ATM plot`	Plot of the average Treatment effect for the evenly population
`ATO plot`	Plot of the average Treatment effect for the overlap population
weights	Estimated weights for each of the treatment effects

Examples

Treatment = treatments$treatment
data = treatments[, c(2:3)]
treatment_model(Treatment, data)

JobNmadu/Dyn4cast documentation built on March 5, 2025, 9:56 p.m.