qgraph.sem: qgraph: Structural Equation Modelling

Description

This functions uses a "sem" object from the sem function (from the sem package; Fox, 2010) and outputs a multi-page pdf file containing goodness of fit indices, path diagram, graphs of the parameter estimates and graphs of the implied and observed covariance and correlation matrices.

Usage

qgraph.sem(res, layout="circle",  ..., vsize.man=3, vsize.lat=6, filename="qgraph",
	filetype="pdf", residuals=TRUE, panels=2, include=1:12, latres=TRUE, 
	curve=0, residSize=0.2, onefile=TRUE, width = 7, height = 7)

Arguments

res

The output from the 'sem' function. See details for extra details on specifying the model

layout

The layout used for the path diagram. Can be "tree", "spring", "circle" and "springtree"

...

arguments passed to qgraph. This is both for the path diagram and for the correlation/covariance plots.

vsize.man

Size of the manifest variables in the path diagram

vsize.lat

Size of the latent variables in the path diagram

filename

Name of the file (will be extended with the filetype)

filetype

The filetype to be used. Can be "pdf" to make a pdf (default) or anything else to plot in R. More filetypes will be supported in a future version.

residuals

Logical indicating if the residuals should be included in the path diagram. If this is FALSE then residual variances will be shown as colors on the nodes. Default is TRUE

panels

The number of panels to be used in the pdf file. Can be 1, 2, 4 or 8

include

A vector indicating which panels should be included in the output

latres

This is currently not supported, leave to TRUE

curve

Numerical value indicating the curve of edges that are on the same level in the "tree" layout, See details. This represent an offset from the middle of the straight edge through where the curved edge must be drawn. 0 indicates no curve, and any other valu

residSize

Size of the residual arrows

onefile

Logical indicating if a multi-page pdf should be produced. If FALSE each plot will be a new pdf. Use this only with panels=1 and filename="(Arbitrary name)%03d".

width

Width of each panel, in inches

height

Height of each panel, in inches

encoding

UTF8

Details

This function uses a "sem" object and outputs a multi-page pdf file. The function reads the 'sem' file and creates a residual variable for each variable present in the model. The default layout is a tree-layout (layout="tree") in which each variable is placed as a node on one of four vertical levels. At the bottom are the residuals of the manifest variables placed, Above that the manifest variables, above that the latent variables and at the top the residuals of the latent variables. The nodes are evenly spaced horizontally in order of appearance in the model (residuals are placed at the same horizontal position of their corresponding variable). So the order of specifying in the model defines the order of placement in the path diagram. If the 'residuals' argument is FALSE then residuals are omitted and this diagram will only contain two levels.

Alternatively the 'spring' layout can be used (layout="spring"). This will use the Fruchterman-reingold algorithm (Fruchterman & Reingold, 1991) is used, which has been ported from the 'sna' package (Butts, 2010). This is a force-directed algorithm. It is best to use this in combination with residuals=FALSE. Names for variables and edges used in the model specification are passed to the path diagram. To keep the model readable it is advised to limit these names to 3 characters. If a variable or edge name contains an asterisk, the asterisk is omitted and the name will be printed in the symbol font. This way Greek letters can be used (e.g. the edge name "l*" will make a lambda character). Because the symbol font conveniently uses Arabic numerals, parameter names like beta1 can be created with "*b1" in the model.

References

Sacha Epskamp, Angelique O. J. Cramer, Lourens J. Waldorp, Verena D. Schmittmann, Denny Borsboom (2012). qgraph: Network Visualizations of Relationships in Psychometric Data. Journal of Statistical Software, 48(4), 1-18. URL http://www.jstatsoft.org/v48/i04/.

Carter T. Butts (2010). sna: Tools for Social Network Analysis. R package version 2.2-0. http://CRAN.R-project.org/package=sna

John Fox with contributions from Adam Kramer and Michael Friendly (2010). sem: Structural Equation Models. R package version 0.9-21. http://CRAN.R-project.org/package=sem

Fruchterman, T. & Reingold, E. (1991). Graph drawing by force-directed placement. Software - Pract. Exp. 21, 1129?1164.

Examples

Run this code

require('sem')

# This example is taken from the examples of the sem function. 
# Only names were changed to better suit the path diagram.

# ----------------------- Thurstone data ---------------------------------------
#  Second-order confirmatory factor analysis, from the SAS manual for PROC CALIS

R.thur <- readMoments(diag=FALSE, names=c('Sen','Voc',
        'SC','FL','4LW','Suf',
        'LS','Ped', 'LG'))
    .828                                              
    .776   .779                                        
    .439   .493    .46                                 
    .432   .464    .425   .674                           
    .447   .489    .443   .59    .541                    
    .447   .432    .401   .381    .402   .288              
    .541   .537    .534   .35    .367   .32   .555        
    .38   .358    .359   .424    .446   .325   .598   .452  
            
model.thur <- specifyModel()
    F1 -> Sen,               *l11, NA
    F1 -> Voc,               *l21, NA
    F1 -> SC,                *l31, NA
    F2 -> FL,                *l41, NA
    F2 -> 4LW,               *l52, NA
    F2 -> Suf,               *l62, NA
    F3 -> LS,                *l73, NA
    F3 -> Ped,               *l83, NA
    F3 -> LG,                *l93, NA
    F4 -> F1,                *g1,  NA
    F4 -> F2,                *g2,  NA
    F4 -> F3,                *g3,  NA 
    Sen <-> Sen,             q*1,   NA
    Voc<-> Voc,              q*2,   NA
    SC <-> SC,               q*3,   NA
    FL <-> FL,               q*4,   NA
    4LW <-> 4LW,             q*5,   NA
    Suf<-> Suf,              q*6,   NA
    LS <-> LS,               q*7,   NA
    Ped<-> Ped,              q*8,   NA
    LG <-> LG,               q*9,   NA
    F1 <-> F1,               NA,     1
    F2 <-> F2,               NA,     1
    F3 <-> F3,               NA,     1
    F4 <-> F4,               NA,     1

sem.thur <- sem(model.thur, R.thur, 213)

# Run qgraph:
qgraph.sem(sem.thur,filename="Thurstone tree",layout="tree",edge.label.cex=0.6,
  curve=0.4,width=8,height=5)

# Spring layout:
qgraph.sem(sem.thur,filename="Thurstone spring",layout="spring",residuals=FALSE,
  width=5,height=5)

Run the code above in your browser using DataLab

Data engineering and BI courses are free!