psych (version 1.0-23)

poly.mat: Find polychoric correlations of item data

Description

Uses John Fox's hetcor function (from polychor package) to find a matrix of polychoric correlations for integer data. Essentially a wrapper for hetcor to convert integer item data into factor (categorical) data and then use hetcor. Just a useful shortcut for subsequent factor analysis.

Usage

poly.mat(x, short = TRUE, std.err = FALSE, ML = FALSE)

Arguments

x

A matrix or data frame of integer data

short

short=TRUE, just show the correlations, short=FALSE give the full hetcor output

std.err

std.err=FALSE does not report the standard errors (faster)

ML

ML=FALSE do a quick two step procedure, ML=TRUE, do longer maximum likelihood

Value

• A matrix of polychoric correlations (if short=TRUE), otherwise a list of various estimates (see hetcor).

Details

Typical personality and item data are integer values (0,1 for ability; 1,2, 3, 4 for attitude scales). The normal correlation procedures will find Pearson correlations (cor). The polycor and hetcor functions from John Fox's polychor package will find polychoric correlations for categorical data. This wrapper function converts integer data to categorical data and then calls hetcor.

Examples

## The function is currently defined as
function(x,short=TRUE,std.err=FALSE,ML=FALSE) {
require(polycor)  #John Fox's Polycor package
xm <- as.matrix(x)   
xm.cat <- matrix(as.factor(xm),ncol=dim(xm)[2])
colnames(xm.cat) <- colnames(xm)
r.het <- hetcor(xm.cat,std.err=std.err,ML=ML)
colnames(r.het$correlations) <- colnames(xm)
if(short) {return(r.het$correlations)} else {return(r.het)}
  }