ABSTRACT
Two categories of response format which are very commonly used in psychometric assessments are the 'normative' and the 'ipsative'. Normative response formats are designed to minimise the relationship between the items in an assessment. As part of the 'normative versus ipsative' debate, some people have raised objections about comparing ipsative responses between individuals. Ipsative scales, because of their mathematical interdependence, tend to force negative correlations in the data, whilst normative scales, because of various response sets such as central tendency bias, force up correlations between scales, thus flattening profiles. Original, normative and ipsative versions of the items were then separately factor-analysed using a Varimax-Promax rotated principal components analysis. Based on a relatively large number of personality scales, these data suggest that the normative and ipsative approaches correlate relatively well with each other and about equally well with external criteria. Ipsative scaling has often been criticised as an intrinsically inferior form of measurement.
