ROUTINELY COMPUTABLE CONFIDENCE INTERVALS FOR FACTOR LOADINGS USING THE ‘JACK-KNIFE’

The problem of the statistical significance of factor loadings is attacked, using the ‘jack‐knife’. The technique, while not new, has curiously not been applied to factor analysis. The procedure is outlined and applied to two classes of data: (1) real data from which a clear pattern of significant factor loadings emerges, and (2) random data which do not fare well under...