Caveats of Non-Ipsatization of Basic Values: A Review of Conceptual and Statistical Issues
Ipsatization, or a correction of variables by their common component, has been routinely applied to measures of basic values. However, it was rarely discussed in terms of the consequences it may entail and especially in contrast to the consequences of non-ipsatization. The current paper aimed at clarifying conceptual and statistical issues as well reviewing the existing critiques of ipsatization. Conceptually, ipsatization of values is intertwined with the definition of values which emphasizes their ordered nature. A common factor that is removed by ipsatization was discussed to represent either a nuisance variable to turn ratings into preferences, a response style, social desirability, or an overall level of motivation. None of these, though, were proved to be a single determinant of a common value factor. The conceptualizations were not mutually exclusive and partly overlapped with the content of values. Statistically, both applying and failing to apply ipsatization can cause bias. A small simulation study illustrated that the use of a popular ipsatization method of within-person centering tended to increase bias when the specific factors were positively intercorrelated or correlated to a common factor. In contrast, non-centered scores were biased when the common factor had relatively large variance compared to specific value factors and when the factors were negatively intercorrelated. Revisiting previous overarching recommendations, we concluded that no general advice in regard to ipsatization can be provided. Instead, each measurement situation requires its own evaluation. An explicit theory of a common factor in given settings can provide arguments for or against ipsatization.