Indeterminacy problems and the interpretation of factor analysis results

In: Statistica Neerlandica, Band 32, Heft 4, S. 181-199

Abstract

Abstract  This paper reviews indeterminacy problems for the factor analysis model and their consequences for the interpretation of the results. Two types of indeterminacy are discerned: indeterminacy of the parameters in the model (the number of factors, the specific variances and the factorloadings) and the indeterminacy of the factors, given the parameters in the model. It is argued that parameter indeterminacy is partly to be overcome, provided that a strong underlying theory for the subject matter under research is present. Factor indeterminacy remains a major stumbling‐block for the interpretation of results. The Guttman criterion is advocated as a measure of factor indeterminacy.