EffectLiteR

This R package can be used to estimate average and conditional effects of a treatment variable on an outcome variable, taking into account any number of continuous and categorical covariates. It automatically generates lavaan (Rosseel, 2012) syntax for a multi-group structural equation model, runs the model in lavaan, and extracts various average and conditional effects of interest.

EffectLiteR is a new program based on the ideas (not the code) of EffectLite (Steyer & Partchev, 2008; www.causal-effects.de) and many (former) researchers at the Department of Methodology and Evaluation Research at Friedrich Schiller University Jena.

Installation

EffectLiteR is on CRAN (https://cran.r-project.org/web/packages/EffectLiteR) and the latest stable version can be installed as usual:

install.packages("EffectLiteR")

For testing purposes, the development version of EffectLiteR can be installed directly from this GitHub repository using the additional package devtools. Under Windows, please make sure Rtools (http://cran.r-project.org/bin/windows/Rtools/) are installed and no older version of EffectLiteR is currently loaded):

install.packages("devtools")
library(devtools)

install_github("amayer2010/EffectLiteR")

Alternatively, the subfolder /old contains tar.gz files (for installing from source) and Windows binaries. However, these may be outdated. When devtools is used, dependencies are installed automatically. Otherwise, you need to install the following packages yourself:

install.packages(c("lavaan","methods"))
install.packages(c("shiny","foreign","ggplot2"))

Run EffectLiteR

The main function of the package is effectLite(). Type example(effectLite) for some examples. The shiny interface can be called by effectLiteGUI() after having loaded the package EffectLiteR:

library(EffectLiteR)

example(effectLite)
effectLiteGUI()