People in Japan consider a face more attractive when the person performs a traditional bow, and this occurs regardless of the model's race. Observers in the US, by contrast, show almost no reaction to bowing. This suggests the effect comes down to cultural habits rather than a universal body language. The results were published in the journal Japanese Psychological Research.
A bow is a gesture in which a person inclines the head and upper body forward. In Japan it is customary to greet, express gratitude or apologize this way. The inclination can be slight for a passing greeting or deeper for an apology.
The work was led by Takayuki Osugi of Yamagata University. The scientists drew on earlier data showing that tilting an image forward, imitating a bow, increases the perceived attractiveness of a face among Japanese raters. Psychologists explain this by the fact that a bow triggers cognitive schemas - mental templates that organize knowledge and expectations. In this case the gesture is linked to politeness and respect, and this creates an overall favorable impression that carries over to the evaluation of appearance. This carryover is called the halo effect.
The first experiment involved Japanese people recruited through an online platform. In six sub-experiments, in which the sex and race of the models varied, between 45 and 49 people rated faces under three conditions: a long bow, a short bow and a static position. About half of the participants took part in more than one sub-experiment, which were conducted with breaks over roughly three years. A re-analysis accounting for repeat participation produced the same results.
Japanese raters gave higher attractiveness scores to bowing models than to those standing still. For example, when rating European women, the mean score was 43.2 in the static condition, 51.1 for a short bow and 59.0 for a long bow. Overall, across the different model types, scores rose by roughly 6-12 points for a short bow and by 15-28 points for a long bow compared with the static condition. The longer the bow lasted, the higher the rating.
The same videos were then shown to Americans recruited through Amazon Mechanical Turk. Each sub-experiment involved between 30 and 40 people. Americans barely changed their ratings when the models bowed. Some differences reached statistical significance but were small - usually less than three points on a 100-point scale, whereas for the Japanese the gain was 15 points or more. A direct comparison of the two samples confirmed that the bow effect is strong among the Japanese and practically absent among the Americans.
Additional experiments tested whether the effect works for faces of differing baseline attractiveness. Japanese participants rated attractive and less attractive models. Bowing raised scores in both groups. For example, among less attractive women the score rose from 33.3 in the static condition to 42.0 for a short bow and 49.0 for a long bow. Among attractive women it rose from 48.0 to 54.3 and 62.7 respectively. Americans, meanwhile, easily distinguished levels of attractiveness, but bowing had almost no effect on their ratings.
The authors note limitations. They did not record how familiar the American participants were with Japanese culture, anime or travel. Racial data on the American sample were also not collected, so future work will need to establish whether the observer's race affects perception. In addition, participants completed the tasks on their own computers, so screen size and viewing distance were not standardized. Live interaction includes intonations, postures and eye contact that digital avatars lack.
The study is titled "Effect of Bowing on Facial Attractiveness of Other Races". The authors are Takayuki Osugi and Jun I. Kawahara.