APA Dictionary of Psychology

a data reduction approach in which a number of independent linear combinations of underlying explanatory variables are identified for a larger set of original observed variables. PCA reproduces all of the information in the original correlation matrix and does not assume that variables are measured with any degree of error. Thus, the result is a new set of variables (called principal components) that are uncorrelated with each other and ordered in terms of the percentage of the total variance for which they account. The technique is similar in its aims to factor analysis but has different technical features. Also called principal-components factor analysis.