Answers are grounded in this paper and cite their sources.
-When a set of variables cannot be regarded as being drawn from a well defined population, the estimation of communalities, population parameters, and universe factors is not a rational undertaking. A descriptive factor analysis can be performed by weighting test vectors inversely as the components of total test variance unable to determine a test battery common factor. Such components are well defined and a descriptive factor analysis based on weighting by such components requires no assumptions about the population and no estimations of communalities ; it requires only that regressions be reasonably linear.
No takes yet. Share an insight, caveat, or question.
John M. Butler (1968) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: