On the factoring and interpretation of ipsative data

The debate in this journal on the validity of factoring ipsative data has involved the use of synthetic data and artificial ipsatization of scores, processes which can often obscure rather than clarify the issues in dispute. This paper presents a more direct illustration of these issues by having subjects make both ipsative and non-ipsative responses to the same test items. The results show clearly the dangers of factoring correlation matrices derived from ipsative data. It is also argued that conventional norms should not be used for the interpretation of scores from ipsative tests.