ORIGINAL ARTICLE
Facial attractiveness: A longitudinal study Egle Tatarunaite,a Rebecca Playle,b Kerry Hood,b William Shaw,c and Stephen Richmondd Cardiff, Wales, and Manchester, United Kingdom Background: The aim of the study was to investigate various factors that might affect facial attractiveness from ages 11 to 31 years. Methods: Sixty subjects were selected from a sample participating in the longitudinal Cardiff Survey. Three-quarter-view facial photographs were taken of the subjects smiling and not smiling. Photos were taken in 1981, when the subjects were aged 11 years, and in 2001, when they were aged 31. Twelve judges used a 9-point Likert scale to evaluate overall facial attractiveness and the attractiveness of various facial features. The judges were also asked to estimate the adults’ ages. Univariate and multivariate statistical tests and the generalizability theory were used. Results and conclusions: Overall facial attractiveness does not depend on any single feature; smiling and youthful facial appearance make women look more attractive; facial attractiveness tends to decrease over time from ages 11 to 31; people tend to retain their relative levels of attractiveness throughout their life spans; orthodontic treatment improves dental appearance, but it does not necessarily make a person more attractive in the long term. Nevertheless, the positive effect of orthodontic treatment could still be observed, especially in men with lower levels of facial attractiveness in childhood. (Am J Orthod Dentofacial Orthop 2005;127:676-82)
B
ody image appears to influence a person’s development.1-7 Mueser et al8 stated that the face is a slightly more important predictor of overall attractiveness than the body. In addition, it has been reported that the eyes, the oral region, and the complexion significantly contribute to overall facial attractiveness.5,9-13 However, it has been suggested that the rest of the face appears to be a more influential cue than dental esthetics in impression formation.6,15 Facial attractiveness tends to decrease as a person ages.7,16 A common belief is that beauty is in the eye of the beholder, suggesting that esthetic judgments are a matter of individual taste and purely subjective. Nevertheless, some persons are universally regarded as attractive, and there is evidence that people can agree on attractiveness.3,17-22 Thus, a concept of facial attractiveness as entirely subjective might not be correct. Very few studies have systematically evaluated the effects of orthodontic treatment on facial attractiveness in the long term. The aim of this project was to investigate various factors that might influence facial attractiveness over a
a
Visiting research scholar, Cardiff University. Medical statistician, Cardiff University. Professor of Orthodontics and Dentofacial Development, University of Manchester. d Professor of Orthodontics and head of Dental Health and Biological Sciences, Cardiff University. Reprint requests to: Prof Stephen Richmond, Department of Dental Health and Biological Sciences, Dental School, Heath Park, Cardiff CF14 4XY, United Kingdom; e-mail,
[email protected]. Submitted, October 2003; revised and accepted, January 2004. 0889-5406/$30.00 Copyright © 2005 by the American Association of Orthodontists. doi:10.1016/j.ajodo.2004.01.029 b c
676
20-year period, from ages 11 to 31. The objectives were to assess the facial features that contribute to overall facial attractiveness; the effects of gender, aging, and smiling on overall facial attractiveness; and the effect of orthodontic treatment on dental and overall facial attractiveness. SUBJECTS AND METHODS
The longitudinal Cardiff Survey,23 at Cardiff University Dental School in the United Kingdom, began in 1981 with an initial sample of 1018 white, 11-year-old schoolchildren; 331 of them attended the most recent examination in 2001, at age 31, and 60 were selected (by using the mean score obtained from a panel of 6 judges and derived from a 10-cm visual analogue scale), representing a broad range of facial attractiveness. The inclusion criteria were a complete set of photographs for each assessment year and groups equally divided by sex and orthodontic treatment. Three-quarter-view smiling and nonsmiling facial photographs for each of the 60 subjects at ages 11 and 31 were color corrected and cleaned. A questionnaire was devised with a 9-point rating scale for the various features of the face (cheeks, chin, eyes, hair, lips, nose, skin, and teeth) as well as overall facial attractiveness (Table I). The raters were also asked to decide whether the subject was attractive. For the adult group only, the raters were also asked to estimate each subject’s age. The 240 photos were randomly divided into 4 folders (smiling and nonsmiling children, smiling and nonsmiling adults) with 1 photograph and questionnaire per page.
Tatarunaite et al 677
American Journal of Orthodontics and Dentofacial Orthopedics Volume 127, Number 6
Table I.
Facial attractiveness rating scale Very unattractive
Overall Cheeks Chin Eyes Hair Lips Nose Skin Teeth
1 1 1 1 1 1 1 1 1
Neither attractive nor unattractive 2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3
Twelve nondental-related white 31-year-olds, equally divided according to sex, were invited to rate the facial photographs. The raters were asked to complete the questionnaires at home and return them within a week. Each member of the panel was paid £60. The raters were unaware of the subjects’ages or that all the adults were the same age. Factor and linear regression analyses were used to study individual facial features. Factor analysis was used to identify possible underlying factors that might combine and contribute to overall facial attractiveness. Regression analysis was used to determine how well each facial feature predicted the overall attractiveness score. Univariate parametric and nonparametric tests were carried out on overall facial attractiveness score data and estimated ages when appropriate. As well as providing attractiveness scores, the raters designated subjects as either “attractive” or “unattractive.” A consensus was obtained from all 12 raters for each subject. When there was a tie, a subject was assigned to the “unattractive” group. Associations between designations at ages 11 and 31 were examined by using descriptive statistics (cross-tabulation). A crossed and nested analysis of variance (ANOVA)24 was used to investigate factors relating to facial attractiveness scores. Because the study had a mixed crossed and hierarchical design, a nested ANOVA model was used to test for significant factors while considering all other factors. Rater (R) and person (P) were the main effects. Smiling (S) (or nonsmiling) and age (A) were nested within persons (S[P] and A[P], respectively). Age and smiling (nested in persons) were crossed, because all subjects were photographed smiling and nonsmiling at ages 11 and 31. Interexaminer reliability estimation was undertaken by using the generalizability theory.25 Repeated measures analysis26 was used to determine the effect of sex and orthodontic treatment on
4 4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5 5
Very attractive 6 6 6 6 6 6 6 6 6
7 7 7 7 7 7 7 7 7
8 8 8 8 8 8 8 8 8
9 9 9 9 9 9 9 9 9
overall facial attractiveness. With the average attractiveness score over all 12 raters, a new model was examined that included just the smiling subjects, because only the smiling photographs showed teeth before and after orthodontic treatment. Age was the main within-subjects effect (repeated measure), and the 2-way interactions of treatment with age and sex and the 3-way interaction of treatment with age and sex were examined as between-subjects effects. Unless otherwise stated, a P value less than .05 was taken as statistically significant. Facial attractiveness scores were averaged over raters for all univariate statistical analyses. Multivariate ANOVA and interrater reliability were carried out over the 12 rater scores for each subject. RESULTS
The sample consisted of 29 men and 31 women; 28 had received orthodontic treatment between the ages of 11 and 31. The results of the factor analysis are shown in Table II. Only 1 factor was extracted with similar weightings for all features. This suggested that individual facial features contributed equally to overall facial attractiveness at both ages in the smiling and nonsmiling groups. According to the linear regression analysis results, in all 4 groups (smiling and nonsmiling children and adults), the facial feature most strongly associated with overall attractiveness was cheeks. The facial features least associated with the overall score were nose (nonsmiling photos) and teeth (smiling photos) (Table III). There were no significant differences between the smiling and nonsmiling subjects in both age groups (Table IV). The children were rated as significantly more attractive than the adults in the nonsmiling group only. Smiling women were rated significantly more attractive than their nonsmiling peers, and smiling men
678 Tatarunaite et al
Table II.
American Journal of Orthodontics and Dentofacial Orthopedics June 2005
Factor loadings and communalities for factor analysis of facial features Nonsmiling 11-year-olds
Facial feature
Factor loading
Cheeks Chin Eyes Hair Lips Nose Skin Teeth Variance explained
.197 .189 .156 .174 .169 .138 .188 * 67.2%
Smiling 11-year-olds
Comm* .858 .795 .541 .671 .635 .424 .782 *
Factor loading
Comm
.177 .168 .153 .147 .156 .150 .166 .111 65.3%
.855 .772 .640 .592 .664 .610 .751 .338
Nonsmiling 31-year-olds Factor loading
Comm
.186 .169 .170 .180 .170 .145 .182 * 68.8%
.799 .660 .673 .748 .674 .491 .772 *
Smiling 31-year-olds Factor loading .174 .163 .159 .163 .163 .147 .160 .103 64.7%
Comm .812 .715 .678 .712 .714 .578 .683 .283
*Comm, Communality. Table III.
Results of linear regression for each facial feature with overall attractiveness score R2 values from univariate linear regressions
Facial feature Cheeks Chin Eyes Hair Lips Nose Skin Teeth
Nonsmiling 11-year-olds
Smiling 11-year-olds
Nonsmiling 31-year-olds
Smiling 31-year-olds
.796 .726 .574 .646 .483 .347 .648 —
.797 .728 .536 .599 .593 .484 .743 .358
.814 .645 .639 .748 .530 .394 .690 —
.797 .625 .602 .694 .571 .496 .639 .294
P ⬍.001. Table IV.
Median scores (range) for overall facial attractiveness in children and adults Age group
Nonsmiling
Smiling
P
11-year-olds (n ⫽ 60) 31-year-olds (n ⫽ 60) P
4.79 (3.4-6.3) 4.50 (2.4-6.3) 0.006
4.83 (3.0-6.4) 4.54 (2.9-6.7) NS
NS NS
NS, Not significant.
were significantly less attractive than smiling boys (Table V). Forty-nine (81.7%) and 43 (71.7%) subjects retained their facial attractiveness designation over the 20 years in the nonsmiling and smiling groups, respectively (Tables VI and VII). A few (8.9%) of the nonsmiling subjects who were rated unattractive as children were found to be attractive as adults. This contrasts with the 53.3% of attractive nonsmiling children who remained attractive as adults (Table VI). In the smiling group, 19.5% of the unattractive children became attractive adults, and
52.6% of the attractive children remained attractive as adults (Table VII). The median estimation of age in the smiling group was 32 years for the attractive group and 35 for the unattractive group; the difference was statistically significant. Similarly, a significant difference in the estimation of age was found for the nonsmiling group, with an age of 33 years estimated for the attractive group and 34 for the unattractive group. The estimated age ranged from 28 to 42 years (Table VIII). Table IX shows the median age for the various groups in relation to sex. The only statistically significant differences were found in women; attractive ones were rated significantly younger than their unattractive peers in both smiling and nonsmiling groups. Statistical analyses
Multivariate ANOVA showed that significant factors were rater (R), person (P), and age (A), and significant interactions were age by smiling (A[P]S[P]), rater by age (RA[P]), and rater by smiling (RS[P]) (Table X). A statistically significant rater effect showed
Tatarunaite et al 679
American Journal of Orthodontics and Dentofacial Orthopedics Volume 127, Number 6
Table V.
Median scores (range) for overall facial attractiveness by sex Women (n ⫽ 31)
Age group
Nonsmiling
Smiling
11-year-olds 31-year-olds P
4.9 (3.6-6.0) 4.8 (2.4-6.3) NS
4.9 (3.0-6.4) 5.0 (3.0-6.7) NS
Men (n ⫽ 29) P
Nonsmiling
Smiling
P
NS. .019
4.7 (3.4-6.3) 4.3 (3.1-5.8) NS
4.8 (3.1-6.2) 4.4 (2.9-5.8) .021
NS NS
NS, Not significant. Table VI.
Cross-tabulation for nonsmiling group Designation for 31-year-olds Nonsmiling group
Designation for 11-year-olds
Table VII.
Unattractive n (%) Attractive n (%) Total
Unattractive n (%)
Attractive n (%)
Total
41 (91.1) 7 (46.7) 48 (80.0)
4 (8.9) 8 (53.3) 12 (20.0)
45 (100) 15 (100) 60 (100)
Cross-tabulation for smiling group Designation for 31-year-olds Smiling group
Designation for 11-year-olds
Unattractive n (%) Attractive n (%) Total
Table VIII.
Median age estimates (age range) for attractive and unattractive adults
Smiling (n ⫽ 60) Nonsmiling (n ⫽ 60)
Attractive
Unattractive
P
31.9 (28.2-40.3) n ⫽ 18 32.5 (29.1-38.3) n ⫽ 12
35.2 (30.4-41.1) n ⫽ 42 34.4 (29.4- 41.6) n ⫽ 48
.001 .010
that the same people received different scores from different raters. A significant person effect reflected that inherent differences in persons’ facial attractiveness were perceived as such by the panel. A significant age effect agreed with the univariate analysis, indicating that children were generally rated as more attractive than adults. A statistically nonsignificant smiling effect was consistent with the univariate analysis results and suggested that a change in facial expression (smiling and nonsmiling) did not influence overall facial attractiveness scores significantly in either age group. However, it seemed that smiling had an effect when the person’s age was considered. This finding was illustrated by both significant interaction between age and smiling as well as the univariate analysis results,
Unattractive n (%)
Attractive n (%)
Total
33 (80.5) 9 (47.4) 42 (70.0)
8 (19.5) 10 (52.6) 18 (30.0)
41 (100) 19 (100) 60 (100%)
indicating that, although facial attractiveness generally decreased over time, the statistically significant difference was observed only in the nonsmiling group. Moreover, the significant interactions between rater and smiling, and between rater and age, suggested that not all raters agreed while making a decision on the effects of smiling and aging (maturation) on a person’s facial attractiveness. The generalizability theory was used to derive an expression for determining interrater reliability in the mixed model. Variance components were calculated from the ANOVA model mean squares and expected mean squares terms (by using Splus 6.0 for Windows; Insightful, Seattle, Wash) (Table XI). The low interrater reliability score (0.37) reflects the significant rater main effect and interactions of rater with age and smiling in the ANOVA model. However, although different raters were scoring the same subjects differently, the nonsignificant interaction between rater and person indicates that raters ordered the photographs in a similar way (from least to most attractive). There were no significant differences in median dental and facial attractiveness scores for treated and untreated smiling adults (Table XII). The teeth and overall facial attractiveness changes over time were
680 Tatarunaite et al
Table IX.
American Journal of Orthodontics and Dentofacial Orthopedics June 2005
Median age estimates (age range) for attractive and unattractive adults by sex Men (n ⫽ 29)
Smiling Nonsmiling
Women (n ⫽ 31)
Attractive
Unattractive
P
Attractive
Unattractive
P
37.8 (36.2-39.4) n⫽2 32.9 (32.9-38.3) n⫽3
35.1 (31.4-40.3) n ⫽ 27 34.5 (31.2-40.0) n ⫽ 26
NS
31.6 (28.2-40.3) n ⫽16 31.3 (29.1-34.8) n⫽9
35.6 (30.4-41.1) n ⫽ 15 34.3 (29.4-41.6) n ⫽ 22
.003
NS
.016
NS, Not significant. Table X.
Factors and factor interactions explaining sources of variation in the overall facial attractiveness scores (evaluated by ANOVA) Factor
F-value
Degrees of freedom
F-value from F tables
P
R P A(P) S(P) A(P)S(P) RP RA(P) RS(P)
28.747 2.08 4.618 0.845 2.499 0.893 2.341 1.378
(11, 649) (59, 49) (60, 60) (60, 60) (60, 660) (649, 660) (660, 660) (660, 660)
1.99 1.7 1.7 1.7 1.4 1.0 1.0 1.0
⬍.05 ⬍.05 ⬍.05 NS ⬍.05 NS ⬍.05 ⬍.05
Table XII. Median scores (range) for teeth and overall facial attractiveness in smiling adults as to orthodontic treatment
Attractiveness score
Table XI.
Variance components and interrater reliability expression (generalizability theory) Factor
Term
Variance component
R P A(P) S(P) A(P)S(P) RP RA(P) RS(P) RA(P)S(P)
R p A(P) S(P) A(P)S(P) RP RA(P) RS(P) e
0.1604505 0.2167225 0.33167 ⫺0.0000187 0.07362 ⫺0.04129 0.384 0.108755 0.57365
2 2 2 ⫹ S(P) ⫹ A(P)S(P) 2P ⫹ A(P) 2 2 2 2 2 ⫹ S(P) ⫹ RA(P) ⫹ RS(P) ⫹ A(P)S(P) ⫹ 2RP ⫹ e2 2P ⫹ A(P) 0.6220125 ⫽ ⫽ 0.368 1.6884175 Reliability from variance components (Splus) 0.6186606 ⫽ ⫽ 0.367 1.6856756
assessed by comparing the scores for children and adults in both groups as to treatment. There was a significant improvement in tooth attractiveness (not facial attractiveness) over time in the group that received orthodontic treatment (Table XIII). Regression analysis showed that the attractiveness score at age 11 predicted the score at age 31 in both the
Facial (n ⫽ 60) Teeth (n ⫽ 59)*
Treated group (n ⫽ 28)
Untreated group (n ⫽ 32)
P
4.7 (2.9-6.2) 5.7 (1.7-6.8)
4.5 (3.0-6.7) 5.1 (3.0-6.7)
NS NS
NS, Not significant. *One subject’s tooth-attractiveness data were missing.
treated and untreated groups. However, there was a significant interaction between orthodontic treatment and facial attractiveness score in childhood. This indicates that orthodontic treatment had a positive effect on facial attractiveness, especially in the persons with lower attractiveness scores at age 11 (Fig). Repeated measures analysis was used to study the combined effects of orthodontic treatment, age, and sex. The only significant term in the model was the interaction of age by sex by treatment (P ⫽ .013). This can be interpreted with the aid of the univariate analysis and descriptive statistics. Generally, average facial attractiveness scores decreased significantly in the smiling male group over the 20 years (Table V). Nevertheless, a different pattern of change in the attractiveness scores over time emerged in the male groups as to orthodontic treatment. The average facial attractiveness score for the treated smiling men decreased to a lesser extent (average scores were 4.30 and 4.27 for children and adults, respectively [n ⫽ 11]), when compared with their untreated counterparts (average scores were 5.01 and 4.30 for children and adults, respectively [n ⫽ 18]). DISCUSSION
Facial photos were used as a means of stimulus presentation because it has been shown that photographic records provide valid, reproducible, and representative ratings of dental and facial appearance.27 Raters of the same age as the subjects were chosen, because it was assumed that lay people were
Tatarunaite et al 681
American Journal of Orthodontics and Dentofacial Orthopedics Volume 127, Number 6
Table XIII.
Median scores (range) for teeth and overall facial attractiveness in smiling children and adults by treatment
group Treated group (n ⫽ 28)
Untreated group (n ⫽ 32)
Attractiveness score
11-year-olds
31-year-olds
P
11-year-olds
31-year-olds
P
Facial (n ⫽ 60) Teeth (n ⫽ 59)*
4.8 (3.0-6.4) 4.5 (2.8-6.4)
4.7 (2.9-6.2) 5.7 (1.7-6.8)
NS .007
4.9 (3.5-6.2) 4.8 (3.0-6.4)
4.5 (3.0-6.7) 5.1 (3.0-6.7)
NS NS
NS, Not significant. *One subject’s tooth-attractiveness data were missing.
Fig. Average overall facial attractiveness scores by treatment.
in a better position to assess the facial attractiveness of their peers. The oblique view of the face was used, because people are generally observed by others at a slight angle in daily interaction,28,29 and it provides the best impression of one’s facial appearance.30 Three-quarter facial photographs have been reported to be reliable and valid in the assessment of facial attractiveness.27 It seems that recent findings are consistent with previous studies in many ways. This study also highlights weaknesses in previous studies. Our study design tested for significant factors while considering all other factors. For instance, the findings in this study are consistent with those of Shaw6 and Shaw et al,15 in which dental attractiveness seemed to be less influential on overall facial attractiveness than the rest of the face. Nevertheless, smiling improved facial attractiveness in women. Smiling men were rated as less attractive than smiling boys. This agrees with other evidence,7,31 suggesting that a positive facial expression improves facial attractiveness in women more than in men. In regard to the absolute stability of one’s facial
attractiveness over the life span, recent findings are in line with the evidence reported in the relevant literature.7,16 Generally, there was a tendency for a person’s facial attractiveness to diminish over the 20 years as judged by lay raters. Other than that, Zebrowitz et al7 suggested that there appeared to be relative stability of facial attractiveness across the life span, ie, people tended to retain their level of attractiveness when compared with those of the same age. Similarly, our findings indicate that subjects tended to retain their level of attractiveness over the 20 years of observation period; this was especially valid for the unattractive children. Moreover, it seems that highly attractive children had a greater chance of remaining attractive as adults than their less attractive peers. Youthful facial appearance is known to enhance overall facial attractiveness7 considerably, especially in women.32-34 This is supported by our results; attractive people were generally perceived as significantly younger than their unattractive counterparts. Nevertheless, this was mainly due to the estimated age differences between the attractive and unattractive women, but not the men. Although the raters did not necessarily agree on the same facial attractiveness score, they tended to rank the photographs of the each person similarly. This is in line with the results of other relevant studies3,17-22 and supports the idea that facial attractiveness is less subjective than is generally thought. Orthodontic treatment made teeth look significantly better, but it did not generally improve a person’s overall facial attractiveness. Nevertheless, the positive effect of orthodontic treatment on overall facial attractiveness was still observed, especially in the men with lower initial facial attractiveness scores. CONCLUSIONS
1. Overall facial attractiveness does not depend on any single facial feature. 2. Smiling and youthful facial appearance make a woman look more attractive, but facial attractive-
682 Tatarunaite et al
ness is not significantly affected by these factors in men. 3. Generally, a person’s facial attractiveness tended to decrease over the 20 years from ages 11 to 31. Nevertheless, people tend to retain their relative level of face attractiveness throughout their life spans (especially if they were unattractive as children). 4. Generally, orthodontic treatment improves the appearance of the teeth but does not necessarily make a person look more attractive in the long term. However, the positive effect of orthodontic treatment can still be observed, especially in men with lower facial attractiveness scores in childhood.14 We thank Anne Kingdon for her determination in maintaining contact with the subjects in the Cardiff Study and organizing the recall program and database. REFERENCES 1. Adams GR. Physical attractiveness research: toward a developmental social psychology of beauty. Hum Dev 1977;20:217-39. 2. Baldwin DC. Appearance and aesthetics in oral health. Community Dent Oral Epidemiol 1980;8:244-56. 3. Bull R, Rumsey N. The social psychology of facial appearance. New York: Springer-Verlag; 1988. 4. Alley TR, editor. Social and applied aspects of perceiving faces. Hillsdale (NJ): Lawrence Erlbaum Associates; 1988. 5. Franzoi SL, Herzog ME. Judging physical attractiveness: what body aspects do we use? Pers Soc Psychol Bull 1987;13:19-33. 6. Shaw WC. The influence of children’s dentofacial appearance on their social attractiveness as judged by peers and lay adults. Am J Orthod 1981;79:399-415. 7. Zebrowitz LA, Olson K, Hoffman K. Stability of babyfaceness and attractiveness across the life span. J Pers Soc Psychol 1993;64:453-66. 8. Mueser KT, Grau BW, Sussman S, Rosen AJ. You’re only as pretty as you feel: facial expression as a determinant of physical attractiveness. J Pers Soc Psychol 1984;46:469-78. 9. Lerner RM, Karabenick SA, Stuart JL. Relations among physical attractiveness, body attitudes, and self-concept in male and female college students. J Psychol 1973;85:119-29. 10. Terry RL, Davis JS. Components of facial attractiveness. Percept Mot Skills 1976;42:918. 11. Terry RL. Further evidence on components of facial attractiveness. Percept Mot Skills 1977;45:130. 12. Helm S, Kreiborg S, Solow B. Psychosocial implications of malocclusion: a 15-year follow-up study in 30-year-old Danes. Am J Orthod 1985;87:110-8. 13. Sergl HG, Stodt W. Experimental investigation of the aesthetic effect of various tooth positions after loss of an incisor tooth. Trans Eur Orthod Soc 1970;497-507. 14. Kerosuo H, Hausen H, Laine T, Shaw WC. The influence of incisal malocclusion on the social attractiveness of young adults in Finland. Eur J Orthod 1995;17:505-12.
American Journal of Orthodontics and Dentofacial Orthopedics June 2005
15. Shaw WC, Rees G, Dawe M, Charles CR. The influence of dentofacial appearance on the social attractiveness of young adults. Am J Orthod 1985;87:21-6. 16. Alley TR. The developmental stability of facial attractiveness: new longitudinal data and a review. Merrill Palmer Q 1993;39: 265-78. 17. Alley TR, Hildebrandt KA. Determinants and consequences of facial aesthetics. In: Alley TR, editor. Social and applied aspects of perceiving faces. Hillsdale: Lawrence Erlbaum Associates; 1988. p.101-40. 18. Iliffe A. A study of preferences in feminine beauty. Br J Psychol 1960;51:267-73. 19. Udry J . Structural correlates of feminine beauty preferences in Britain and the U.S.: a comparison. Sociol Soc Res 1965;49:33042. 20. Edler RJ. Background considerations to facial aesthetics. J Orthod 2001;28:159-68. 21. Thornhill R, Gangestad SW. Facial attractiveness. Trends Cogn Sci 1999;3:452-60. 22. Patzer GL. The physical attractiveness phenomena. New York: Plenum; 1985. 23. Shaw WC, Addy M, Dummer PMH, Ray C, Frude N. Dental and social effects of malocclusion and effectiveness of orthodontic treatment: a strategy for investigation. Community Dent Oral Epidemiol 1986;14:60-4. 24. Armitage P, Berry G. Statistical methods in medical research. 3rd ed. Oxford: Blackwell Scientific; 1994. 25. Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. 2nd ed. Oxford: Oxford University Press; 1995. 26. Altman DG. Practical statistics for medical research. London: Chapman and Hall; 1991. 27. Howells DJ, Shaw WC. The validity and reliability of ratings of dental and facial attractiveness for epidemiologic use. Am J Orthod 1985;88:402-8. 28. Peerlings RHJ, Kuijpers-Jagtman AM, Hoeksma JB. A photographic scale to measure facial aesthetics. Eur J Orthod 1995; 17:101-9. 29. Sarver DM. The face as the determinant of the treatment choice. In: McNamara JA, Kely KA, editors. Frontiers of dental and facial esthetics. Ann Arbor: Center for Human Growth and Development; University of Michigan; 2001. p. 19-54. 30. Van der Linden FPGM, Boersma H. Diagnostic aids. In: Van der Linden FPGM, Boersma H, editors. Diagnosis and treatment planning in dentofacial orthopedics. London: Quintessence; 1987. p. 91-110. 31. Schulman GI, Hoskins M. Perceiving the male versus the female face. Psychol Women Q 1986;10:141-54. 32. Milord JT. Aesthetic aspects of faces: a (somewhat) phenomenological analysis using multidimensional scaling methods. J Pers Soc Psychol 1978;36:205-16. 33. Alley TR. The effects of growth and aging on facial aesthetics. In: Alley TR, editor. Social and applied aspects of perceiving faces. Hillsdale: Lawrence Erlbaum Associates; 1988. p.51-62. 34. Berry DS, Zebrowitz-McArthur L. The impact of age-related craniofacial changes on social perception. In: Alley TR, editor. Social and applied aspects of perceiving faces. Hillsdale: Lawrence Erlbaum Associates; 1988. p. 63-87.
