Investigating the association between mating-relevant self-concepts and mate preferences through a data-driven analysis of online personal descriptions

Research on mate preference have often taken a theory-driven approach; however, such an approach can constrain the range of possible predictions. As a result, the research community may inadvertently neglect traits that are potentially important for human mate choice if current theoretical models simply do not identify them. Here, we address this limitation by using a data-driven approach to investigate mating-relevant self-concepts (i.e., what individuals believe to be attractive about themselves). Using Latent Dirichlet Allocation (LDA; a clustering method developed in computer science) and a large sample of written descriptions from online personal advertisements (N = 7973), we identify 25 common topics that individuals use when advertising themselves. Men were more likely to advertise education/status, while women were more likely to discuss being honest/nurturing and caring for pets. We also assessed patterns of universal and compatible mate preferences for these 25 topics by collecting ratings of desirability from a separate group of 100 participants on a subset of these profiles (N = 468). Participants were also asked to write a personal description of themselves as if they were writing for a dating website. Overall, both male and female profiles that discussed outdoor activities, and music/art were rated as more desirable, while women that discussed a healthy lifestyle and friends/family were also rated as more desirable. Both men and women who discussed sex or mentioned being a parent were rated as less desirable. When comparing the topic probabilities between profiles collected online and those written by the raters, we found that raters preferred profiles that were more similar to their own, particularly for topics to do with being outgoing and agreeable.