Data are said to be ipsative when the sum of the measures obtained over the variables is a constant for each individual. In this article, three types of ipsative data that are commonly encountered in social research are discussed. Each of them can be used to control the effect of different response set biases. Using an artificial example in the context of factor analysis, however, it is showed that the three types of ipsative data cannot simply be analyzed as if they were normative because this will give us misleading and improper results. Correct methods for factor-analyzing the additive and the ordinal ipsative data are discussed. Empirical results from our examples indicate that the suggested methods recover the normative factor structure successfully. It is concluded that ipsative data, if analyzed appropriately, can provide us useful information in psychological research.