In M-E-Rademaker/cSEM.DGP: Generate Data for Structural Equation Models

cSEM.DGP: Generate data for structural equation models

Installation:

The package requires the development version of cSEM (version 0.1.0:9000 as Feb. 2020). To install the development version use:

# install.packages("devtools")
devtools::install_github("M-E-Rademaker/cSEM")

To install the development version of cSEM.DGP use:

# install.packages("devtools")
devtools::install_github("M-E-Rademaker/cSEM.DGP")

Getting started

The best place to get started is the cSEM.DGP website.

Purpose

Generate data for structural equation models including up to eight constructs. Data generation is based on parameter values given in lavaan model syntax.

In addition to supplying numeric values, variable values (symbolic names) for parameters are allowed. To achieve this, the package makes use of lavaan's labeling capabilities. Users may replace a given parameter in, i.e. the structural model by a symbolic name and assign a vector of values to that name. These values will be used to generate data for all possible combinations of these values with the remaining fixed parameters.

The package works nicely in combination with the cSEM package.

Examples

Without variable parameters

Simply write your model in lavaan model syntax. Add a fixed numeric value for each parameter. Note, currently you must either set all parameters or none. The type of output can be chosen: a data.frame (return_type = "data.frame", the default), a numeric matrix (return.type = "matrix"), or a correlation matrix (return.type = "cor").

require(cSEM.DGP)

model <- "
# Structural model
eta3 ~ 0.6*eta1 + 0.4*eta2
eta4 ~ 0.4*eta1 + 0.35*eta3

# Measurement model
eta1 =~ 0.8*y11 + 0.9*y12 + 0.8*y13
eta2 =~ 0.7*y21 + 0.7*y22 
eta3 =~ 0.9*y31 + 0.8*y32
eta4 =~ 0.9*y41 + 0.8*y42

# Measurment error correlation
y11 ~~ 0.2*y12

# Correlation between exogenous constructs
eta1 ~~ 0.5*eta2
"

dat <- generateData(model, .return_type = "cor")
dat

Including variable parameters

To include variable parameters:

1. Replace a numeric value by a symbolic name (i.e., a character string) of your choice.
2. Supply a named vector of values to generateData() where the name corresponds to the chosen character string.

Now, data for all possible combinations of these values with the remaining fixed parameters will be generated.

# We want to vary the path coefficient between eta2 and eta1 and 
# the first loading of eta1.
model <- "
# Structural model
eta2 ~ gamma1*eta1
eta3 ~ gamma2*eta1 + 0.35*eta2

# Measurement model
eta1 =~ lambda*y11 + 0.9*y12 + 0.8*y13
eta2 =~ 0.7*y21 + 0.7*y22 + 0.9*y23
eta3 =~ 0.9*y31 + 0.8*y32 + 0.7*y33
"

dat <- generateData(model, .return_type = "cor", 
                    lambda = c(0.8, 0.9),
                    gamma1 = c(0.2, 0.3),
                    gamma2 = c(0, 0.2, 0.3),
                    .handle_negative_definite = "drop"
                    )
dat

The return type in this case is a nested tibble. To access e.g. the first data set type:

dat$dgp[[1]]

M-E-Rademaker/cSEM.DGP documentation built on Feb. 12, 2020, 6:36 a.m.