Attractiveness of Anthropometrically Average Facial Anatomy: Is the Whole a Sum of Its Parts?
Literature suggests that faces that possess average anatomy for the population are considered more attractive. When patients seek facial cosmetic surgery, they often have a certain feature or features that they desire improvement on. With the subjective and evolving nature of beauty, it can be difficult for plastic surgeons to differentiate between relative change and definite improvement. Thus, identifying ideal dimensions of facial features would provide evidence-based guidance for aesthetic surgery. This study aims to investigate the independent relationship between measurements of anthropometrically average facial features and perceived beauty, as well as their relative importance to each other and the overall perception.
Methods/Technique:
Ratings
To assess perceived attractiveness, crowdsourcing, a well-founded online method of studying aesthetic and reconstructive outcomes, was utilized. The photographs were uploaded to Google Forms with multiple-choice responses ranging from 1-7, with 1 and 7 being the least and most attractive, respectively. The images were presented in a random order to each respondent. Monochrome photos were used to negate the biases of hair, skin, and eye color. Amazon Mechanical Turk, a widely used crowdsourcing platform, was used to receive ratings of the images. Respondents were allowed a maximum of one hour to complete the survey. Respondents were given a survey with all images of the cohort to complete.
Measurements
Our female and male cohorts each consisted of standardized frontal-view photos of 41 individuals from DeBruine 2019, with 1 composite derived from the other 40 (Figure 1 and Figure 2). Composite images were generated using Webmorph.org, which is a web-based software that specializes in averaging and transforming faces. Each face had the same number of standardized points placed manually to delineate the position of facial landmarks (Figure S1). The corresponding points allowed for averaging each facial landmark across all the faces to develop the composite.
All photos were calibrated by interpupillary distance. Next, linear facial features (including measurements of the upper face, middle face, lower face, and facial height and width) were measured by pixels with WebMorph.org (Figures 1-3). Lastly, angular facial features were measured in degrees using Mirror software.
Statistical Analysis
A t-test was used to verify that the composite with anthropometrically average facial features was the most attractive in each cohort. After verifying that the anthropometrically average faces were most attractive, a Spearman correlation test calculated correlation coefficients between anthropometrically facial measurements and perceived attractiveness. A larger correlation coefficient indicates a stronger association between anthropometrically average facial measurements and perceived attractiveness. p<0.05 was considered significant.
Results/Complications:
For the male and female cohorts, the respondents (n=870 and 876, respectively) found the composites (derived from average facial anatomy) significantly more attractive than the rest of the cohort (both p<0.0001). Upon verification of the greater attractiveness of average facial anatomy, each anthropometrically average feature was independently analyzed.
For the male cohort, only anthropometrically average upper lip height had a statistically significant correlation with attractiveness (Table 1) (R=0.376; p=0.017) (Table 1). Anthropometrically average upper face features, including periocular and brow measurements, did not have a significant independent correlation with perceived attractiveness (all p>0.05). Anthropometrically average middle face features, including ear and nose measurements, did not have a significant independent correlation with perceived attractiveness (all p>0.05). All other anthropometrically average lower face features, including chin and jaw measurements, did not have a significant independent correlation with perceived attractiveness (all p>0.05). Anthropometrically average facial widths and heights did not have a significant independent correlation with perceived attractiveness (all p>0.05).
For the female cohort, only anthropometrically average bigonial width had a statistically significant correlation with attractiveness (R= 0.352; p=0.026) (Table 2). Anthropometrically average upper face features, including periocular and brow measurements, did not have a significant independent correlation with perceived attractiveness (all p>0.05). Anthropometrically average middle face features, including ear and nose measurements, did not have a significant independent correlation with perceived attractiveness (all p>0.05). All other anthropometrically average lower face features, including chin and jaw measurements, did not have a significant independent correlation with perceived attractiveness (all p>0.05). Anthropometrically average facial widths and heights did not have a significant independent correlation with perceived attractiveness (all p>0.05).
Conclusion:
Our study identifies anthropometrically average male lip height and anthropometrically average female bigonial features significantly correlated with perceived beauty. Though overall anthropometrically average faces were found to be significantly more attractive, the majority of facial features when independently analyzed for their correlation with beauty were not found to be significantly associated. Our findings suggest that the attractiveness of the anthropometrically average face is not largely due to the summative attractiveness of its individual facial components.