The MplusAutomation package leverages the flexibility of the R
language to automate latent variable model estimation and interpretation
using Mplus, a powerful latent variable modeling program developed by
Muthén and Muthén (www.statmodel.com). Specifically, MplusAutomation
provides routines for creating related groups of models, running batches
of models, and extracting and tabulating model parameters and fit
statistics.
You can install the latest release of MplusAutomation directly from
CRAN by running
install.packages("MplusAutomation")
Alternately, if you want to try out the latest development
MplusAutomation code, you can install it straight from github using
Hadley Wickham's devtools package. If you do not have devtools
installed, first install it and then install MplusAutomation.
#install.packages("devtools")
library(devtools)
install_github("michaelhallquist/MplusAutomation")
For questions, answers, and updates on the status of the
MplusAutomation package, email or subscribe to the Google group
list.
You can find a detailed example of how to use the MplusAutomation
package in the
vignette
.
Here is an example of using the package to run a simple path model using
the mtcars dataset built into R.
library(MplusAutomation)
pathmodel <- mplusObject(
TITLE = "MplusAutomation Example - Path Model;",
MODEL = "
mpg ON hp;
wt ON disp;",
OUTPUT = "CINTERVAL;",
rdata = mtcars)
## R variables selected automatically as any variable name that occurs in the MODEL, VARIABLE, or DEFINE section.
## If any issues, suggest explicitly specifying USEVARIABLES.
## A starting point may be:
## USEVARIABLES = mpg disp hp wt;
fit <- mplusModeler(pathmodel, modelout = "model1.inp", run = 1L)
That is all it takes to run Mplus! MplusAutomation takes care of
figuring out which variables from your R dataset are used in the model
and which are not (if it get's confused, you can also specify
usevariables). It creates a dataset suitable for Mplus, calls Mplus to
run the model on the dataset, and reads it back into R.
There is even pretty printing now. To see the results:
library(texreg)
screenreg(fit, summaries = c("Observations", "CFI", "SRMR"), single.row=TRUE)
==================================
Model 1
----------------------------------
MPG<-HP -0.06 (0.01) ***
WT<-DISP 0.01 (0.00) ***
WT<->MPG -1.02 (0.38) **
MPG<-Intercepts 29.59 (1.53) ***
WT<-Intercepts 1.82 (0.18) ***
MPG<->MPG 14.04 (3.52) ***
WT<->WT 0.21 (0.06) ***
----------------------------------
Observations 32
CFI 0.87
SRMR 0.14
==================================
*** p < 0.001, ** p < 0.01, * p < 0.05
The fit is not great, to add some extra paths we can update the model.
pathmodel2 <- update(pathmodel, MODEL = ~ . + "
mpg ON disp;
wt ON hp;")
fit2 <- mplusModeler(pathmodel2, modelout = "model2.inp", run = 1L)
We can make some pretty output of both models:
screenreg(list(
extract(fit, summaries = c("Observations", "CFI", "SRMR")),
extract(fit2, summaries = c("Observations", "CFI", "SRMR"))),
single.row=TRUE)
====================================================
Model 1 Model 2
----------------------------------------------------
MPG<-HP -0.06 (0.01) *** -0.02 (0.01)
WT<-DISP 0.01 (0.00) *** 0.01 (0.00) ***
WT<->MPG -1.02 (0.38) ** -0.73 (0.26) **
MPG<-Intercepts 29.59 (1.53) *** 30.74 (1.27) ***
WT<-Intercepts 1.82 (0.18) *** 1.68 (0.19) ***
MPG<->MPG 14.04 (3.52) *** 8.86 (2.21) ***
WT<->WT 0.21 (0.06) *** 0.19 (0.05) ***
MPG<-DISP -0.03 (0.01) ***
WT<-HP 0.00 (0.00)
----------------------------------------------------
Observations 32 32
CFI 0.87 1.00
SRMR 0.14 0.00
====================================================
*** p < 0.001, ** p < 0.01, * p < 0.05
If you want confidence intervals, those can also be printed, so long as they were requested as part of the output (we did in the initial model, which propogates to later models that were updated()ed based on the original model):
screenreg(list(
extract(fit, cis=TRUE, summaries = c("Observations", "CFI", "SRMR")),
extract(fit2, cis=TRUE, summaries = c("Observations", "CFI", "SRMR"))),
single.row=TRUE)
================================================================
Model 1 Model 2
----------------------------------------------------------------
MPG<-HP -0.06 [-0.08; -0.05] * -0.02 [-0.05; 0.00]
WT<-DISP 0.01 [ 0.00; 0.01] * 0.01 [ 0.01; 0.01] *
WT<->MPG -1.02 [-1.77; -0.27] * -0.73 [-1.25; -0.21] *
MPG<-Intercepts 29.59 [26.59; 32.58] * 30.74 [28.25; 33.22] *
WT<-Intercepts 1.82 [ 1.46; 2.17] * 1.68 [ 1.31; 2.04] *
MPG<->MPG 14.04 [ 7.14; 20.95] * 8.86 [ 4.52; 13.20] *
WT<->WT 0.21 [ 0.10; 0.32] * 0.19 [ 0.10; 0.28] *
MPG<-DISP -0.03 [-0.04; -0.02] *
WT<-HP -0.00 [-0.00; 0.00]
----------------------------------------------------------------
Observations 32 32
CFI 0.87 1.00
SRMR 0.14 0.00
================================================================
* 0 outside the confidence interval
If you have a tutorial or examples using MplusAutomation, please add
them to the github
Wiki.
In addition, on the
Wiki, is a
list of publications that cite or use MplusAutomation. If you use
MplusAutomation in your own work --- papers, posters, presentations,
etc. --- please add a citation to the list, and if possible, include an
abstract or link to the full text. This helps us get to know our users
and how MplusAutomation is being used.
Finally, if you find bugs or have suggestions for new features or ways
to enhance MplusAutomation, please let us know! Just click the
'Issues' button at the top of the github page or go
here
and open a New Issue.
Lastly, if you use MplusAutomation and have space, we greatly
appreciating citations. In addition to being easier to track, the
recognition and credit help make it easier for us to continue putting
our time into developing and sharing this software!