psych (version 1.0-17)

VSS.plot: Plot VSS fits

Description

The Very Simple Structure criterion ( VSS) for estimating the optimal number of factors is plotted as a function of the increasing complexity and increasing number of factors.

Usage

VSS.plot(x, plottitle = "Very Simple Structure", line = FALSE)

Arguments

x

output from VSS

plottitle

any title

line

connect different complexities

Value

• A plot window showing the VSS criterion varying as the number of factors and the complexity of the items.

Details

Item-factor models differ in their "complexity". Complexity 1 means that all except the greatest (absolute) loading for an item are ignored. Basically a cluster model (e.g., ICLUST). Complexity 2 implies all except the greatest two, etc.

Different complexities can suggest different number of optimal number of factors to extract. For personality items, complexity 1 and 2 are probably the most meaningful.

The Very Simple Structure criterion will tend to peak at the number of factors that are most interpretable for a given level of complexity. Note that some problems, the most interpretable number of factors will differ as a function of complexity. For instance, when doing the Harman 24 psychological variable problems, an unrotated solution of complexity one suggests one factor (g), while a complexity two solution suggests that a four factor solution is most appropriate. This latter probably reflects a bi-factor structure.

For examples of VSS.plot output, see http://personality-project.org/r/r.vss.html

References

http://personality-project.org/r/r.vss.html

See Also

VSS, ICLUST, omega

Examples

#test.data <- Harman74.cor$cov
#my.vss <- VSS(test.data)         #suggests that 4 factor complexity two solution is optimal
#VSS.plot(my.vss)                 #see the graphics window

## The function is currently defined as
function(x,plottitle="Very Simple Structure",line=FALSE)
  {
  	n=dim(x)
  	symb=c(49,50,51,52)              #plotting sym
	plot(x$cfit.1,ylim=c(0,1),type="b",ylab="Very Simple Structure Fit",xlab="Number of Factors",pch=49)
    if (line) lines(x$fit)
	 

	title(main=plottitle)
	x$cfit.2[1]<-NA
	x$cfit.3[1]<-NA
	x$cfit.3[2]<-NA
	x$cfit.4[1]<-NA
	x$cfit.4[2]<-NA
	x$cfit.4[3]<-NA
	lines(x$cfit.2)
	points(x$cfit.2,pch=50)
	lines(x$cfit.3)
	points(x$cfit.3,pch=symb[3])
	lines(x$cfit.4)
	points(x$cfit.4,pch=symb[4])
  }