The issue with most studies concerned with mate selection preferences in humans is that they rely on declarations and rational actions of experimental subjects, which are affected by their pre-conceived opinions and prejudices. Moreover, current research suggests that subcortical structures and processes, rather than the neocortex, play the principal role in actual partner choice behaviour. Consequently, we have only limited information on how relevant our current knowledge is in relation to real-life human ethology. To address these issues, we surveyed 2,718 women and 4,073 men between the ages of 16-50 and compared their declared and observed preferences for various properties in short-term and long-term partners. We found differences between what the subjects declared to prefer and what they preferred in reality: for example, men declared that wealth was the second least desirable property out of eleven in short-term partners, but we observed that in reality, they considered wealth the third most important factor after charisma and sense of humour. Similarly, while women declared that dominance and masculinity were desirable properties in short-term partners, in the observational part of the study, they showed little preference for these traits. Furthermore, we investigated the effects of the phase of the menstrual cycle, hormonal contraception, pregnancy, and partnership and parenthood status on these preferences. We found some support for the good parents hypothesis and no support for the good genes and the immunocompetence handicap hypotheses when observed, rather than declared preferences were considered. We also detected that hormonal contraception, and parenthood and partnership status influenced declared preferences in considerable ways, but had only a small, if any, impact on observed preferences. We suggest interpreting the results of studies reliant on declarations and rational actions of experimental subjects with great caution.