How do pairs matched in physical attractiveness form if people are unaware of their own attractiveness?

The correlation of physical attractiveness in romantic partners has been widely documented. However, it has also repeatedly been demonstrated that people are largely unaware of their own attractiveness, which raises the question about the mechanism responsible for the within-pair matching. One hitherto unexplored possibility is that low accuracy in attractiveness self-assessments results from methodological drawbacks. Participants were usually asked to rate their attractiveness on a numeric scale, and independent judges evaluated them on the basis of facial photographs. We hypothesized that the accuracy of self-assessment may be increased if (1) participants and judges evaluate the same characteristic, e.g., both groups assess facial attractiveness, (2) own attractiveness is estimated in a comparative manner (with reference to attractiveness of other individuals) rather than by abstract numbers, (3) judges rate attractiveness of people as seen in video clips rather than in photographs. To test these hypotheses we photographed and videotaped faces of 96 women and 78 men. Independent judges rated attractiveness from these photographs and video clips, and the participants assessed own attractiveness in several ways. None of the above hypotheses was confirmed by statistical analysis. We discuss how the within-pair matching in attractiveness can arise, given such poor awareness of own appeal.