countEffects: Average and conditional treatment effects for count outcomes

In chkiefer/countEffects: Estimation of Average and Conditional Treatment Effects on Count outcomes

Description

A general function to estimate average and conditional treatment effects for a count outcome.

Usage

countEffects(
  y,
  x,
  k = NULL,
  z = NULL,
  data,
  method = "nb",
  distribution = "condNormal",
  control = "default",
  measurement = list(),
  na.rm = TRUE
)

Arguments

y	Dependent variable (character string). Has to be the name of a count variable (i.e., non-negative integer).
x	Treatment variable (character string) currently treated as binary variable.
k	Categorical covariate (character string).
z	Continuous covariate (character string).
data	A data frame.
method	By default a negative binomial regression model is fitted (i.e., "nb"). Alternatively, with "poisson" a standard Poisson regression is fitted.
distribution	Distribution of the covariate. By default, normal distribution is assumed (i.e., "normal"). Currently, a uniform distribution (i.e., "uniform"), a Poisson distribution (i.e., "poisson"), a chisquare distribution (i.e., "chisquare"), and a negative binomial distribution (i.e., "negbin") are possible.
control	Name or value of the control group
measurement	List containing information for a measurement model (i.e., name of the latent variable and its indicators)
na.rm	Should rows with missing values be removed?

Value

Object of class countEffects.

Examples

## Example with normally distributed covariate:
m1 <- countEffects(y="dv", x="treat", z="pre", data=example01,
                   method="poisson", distribution="condNormal")
summary(m1)

chkiefer/countEffects documentation built on Jan. 24, 2021, 5:04 a.m.