The L.l. Thurstone Psychometric Laboratory University of North Carolina

This chapter presents ViSta-PrnCmp, the module for Principal Components Analysis (PCA) in ViSta. This procedure is capable of analyzing numerical variables so they can be represented on a lower dimensionality space. The visualization for ViSta-PrnCmp includes a scatterplot-matrix of component scores; a bi-dimensional biplot; a tri-dimensional (spin-plot) version of the biplot; a box-diamond-dot plot and a scree plot of the relative fit of the components. The report includes the eigen-values and eigenvectors of the correlation or covariance matrix; the correlations of the variables with the dimensions and other optional output. ViSta-PrnCmp also allows creating data objects such as component scores that can be submitted for further analysis.