The purpose of this investigation was to undertake an objective and quantitative evaluation of how the degree of chin prominence influences perceived attractiveness. The chin prominence of an idealized profile image was altered in 2 mm increments from −24 to 12 mm, in order to represent retrusion and protrusion of the chin, respectively. These images were rated on a 7-point Likert scale by a pre-selected group of pre-treatment orthognathic patients, clinicians and laypeople. In treatment planning to alter the sagittal prominence of the chin in an individual with an otherwise normal soft tissue facial profile, an ‘ideal’ sagittal position with soft tissue pogonion on or just behind a true vertical line through subnasale may be used. Chin retrusion or protrusion up to 4 mm is essentially unnoticeable. Surgery is desired for chin protrusions greater than 6 mm and retrusions greater than 10 mm. The overall direction of aesthetic opinion appears to be the same for all the observer groups; the greater the retrusion or prominence of the chin, the less the rating of the perceived attractiveness and the greater the desire for surgical correction.
The chin is an important determinant of facial profile attractiveness. Its prominence is one of the facial characteristics that society tends to associate with an individual’s personality. Individuals, particularly men, with a deficient chin may be viewed as ‘weak’, whereas a prominent chin is often described as a ‘strong’ chin, implying strength of personality.
Each facial parameter, such as chin prominence, has an ‘average’ value or ‘norm’ for a given population, which is specific for age, sex and ethnicity. Each of these norms also has a range of variability, with the existence of a facial deformity often resulting from a significant deviation of one or more facial parameters from the accepted norm for a population. At what point does the deviation of a facial parameter move from the acceptable range of variability to being perceived as a facial deformity?
The magnitude of the deviation, whether it is due to an underlying dentoskeletal discrepancy, the overlying facial soft tissues or a combination of the two, is an important factor in deciding when orthognathic surgery may be required. If the magnitude of the discrepancy of a facial parameter is great (for example excessive chin prominence) the treatment planning decision may be relatively straightforward. However, a significant number of patients are regarded as ‘borderline’ in terms of need for surgical treatment. In such patients, the decision making process may be transferred from subjective clinical judgement to objective, evidence-based guidance based on data from studies investigating perceptions of facial attractiveness.
Chin prominence is a potentially important factor in the perception of facial attractiveness. The purpose of this study was to find objective evidence to aid clinicians in planning the treatment of patients requiring sagittal augmentation or reduction genioplasty.
The principal aim of this investigation was to evaluate quantitatively the influence of sagittal chin prominence on perceived attractiveness. The relationship between degree of chin prominence and attractiveness was recorded to ascertain the range of normal variability, in terms of observer acceptance, and to determine the clinically significant threshold value or cut-off point, beyond which the degree of chin prominence is perceived as unattractive and treatment is desired. The perceptions of orthognathic patients, clinicians and laypeople were compared for these different variables.
Materials and methods
Ethical approval was sought and granted for the study by the National Research Ethics Service (UK).
Two-dimensional facial profile silhouettes are used routinely to assess the perceptions of facial profile attractiveness. A facial profile silhouette image was created with computer software (Adobe ® Photoshop ® CS2 software; Adobe Systems Inc., San Jose, CA, USA). The image was manipulated using the same software to construct an ‘ideal’ facial profile image with proportions and soft tissue measurements based on currently accepted criteria ( Fig. 1 ). The chin prominence of the idealized profile image was altered in 2 mm increments from −24 to 12 mm, in order to represent retrusion and protrusion of the chin, respectively ( Fig. 2 ).
A pilot study was undertaken to perform a power calculation. Based on the results of the pilot study, the anticipated standard deviations of rating were 1.0 in all groups of observers. As such, this study aimed to recruit 75 pre-treatment orthognathic patients, 75 laypeople and 35 clinicians to guarantee 80% power to detect differences in the mean rating score of approximately 2.5 in the clinician group vs. 3.1 in the patient and laypeople groups (this corresponds to a standardized mean difference of 0.6).
One hundred and eighty-five observers took part in the study, separated into three groups (pre-treatment orthognathic patients, laypeople and clinicians) ( Table 1 ). Selection criteria for the orthognathic patients were: pre-treatment; primary concern was facial appearance; no previous orthodontic or facial surgical treatment; no history of facial trauma; and no severe psychological issues (e.g. body dysmorphic disorder). Selection criteria for the laypeople were: no previous orthodontic or facial surgical treatment; no facial deformities; no history of facial trauma; and non-healthcare employees. The group of clinicians (maxillofacial surgeons and orthodontists) were involved in the management of patients with facial deformities.
|Observer group||Number||Mean age (years)||95% C.I.||Age range||Sex (% male)||Ethnicity (% White)|
|Orthognathic patients||75||22||20, 24||13–60||42%||66%|
Each observer was given a questionnaire and asked to provide the following information: age, sex, ethnic origin, how would you rate the attractiveness of your facial appearance, and how important do you think it is to have an attractive facial appearance. An instruction sheet accompanied the questionnaire, asking the observers to rate each image in terms of facial attractiveness using the following rating scale: 1, extremely unattractive; 2, very unattractive; 3, slightly unattractive; 4, neither attractive or unattractive; 5, slightly attractive; 6, very attractive; 7, extremely attractive. Observers were also asked whether they would consider surgery to correct the appearance if this was their facial appearance (yes or no).
The images were placed in random order into the software application Microsoft PowerPoint ® . Each image was identified by a randomly assigned double letter in the top right corner of the screen (e.g. BD, CE, DG). A duplicate of one of the images was used to assess intra-examiner reliability (images DG and EF). Each observer sat undisturbed in the same room in front of the same computer and 17 in. flat screen monitor. The presentation and the images were created in such a way that each of the profile silhouette images, when viewed on the monitor, had the same dimensions as a normal human head, based on an average lower anterior facial height. This helped to reduce the potential effect of image magnification or size reduction on the observer’s perception. Each observer examined the images in the PowerPoint ® presentation by pressing the ‘Page Down’ button on the keyboard, in their own time.
The Likert-type rating scale used is largely accepted in the psychology literature as the most useful rating method. The seven-point Likert scale described above was used by each observer to rate each image in terms of attractiveness.
The observer’s ratings were recorded in a Likert scale from 1 to 7. Mixed regression was used to assess the differences in ratings for the three groups (pre-treatment orthognathic patients, laypeople, and clinicians) whilst adjusting for the concurrent effects of age, sex, ethnicity, self-rating for facial attractiveness, the importance given to an attractive facial appearance, the observer’s anteroposterior jaw relationship (Class I, II or III), the observer’s vertical face height (average, increased or decreased), observer’s facial asymmetry (yes/no) and the degree of sagittal chin prominence of the images. The multivariate regression models are fitted in a stepwise manner, including all those variables that reach a significance below P = 0.25 univariately. Given the recognized low power of the relevant test, the benchmark for a significant interaction was set at the 10% level. The mixed regression uses a multi-level approach to take into account the clustering effect by observer. The model was validated using a logarithmic transformation for the rating scale to assess the effect of departure from normality.
The variable ‘self-rating of attractiveness’ had very few observers in its border values (very unattractive, seven patients; very attractive, seven laypeople) and was recoded into two levels: attractive ( n = 123) vs. unattractive ( n = 62). The variable ‘vertical face height’ was dichotomized into two levels: normal ( n = 166) vs. non-increased/decreased ( n = 19). Only 11 observers responded that attractiveness was very unimportant (one patient) or unimportant (six patients and four laypeople). As a consequence this variable was dichotomized as very important vs. not very important.
All the laypeople and the clinicians were skeletal Class I whilst 96% of the patients were skeletal Class II or III ( Table 2 ). There was no significant difference in perceptions of attractiveness between observers with skeletal Class II and III jaw relationships ( P = 0.91) but they appeared to differ significantly from those with skeletal Class I. When skeletal Class was fitted on the patient group alone no difference was detected between skeletal Classes II and III ( P = 0.86).
|Jaw relationship (skeletal class) no. (%)||Orthognathic patients||Laypeople||Clinicians||Total|
|Class I||3 (4%)||75 (100%)||35 (100%)||113 (61%)|
|Class II||37 (49.3%)||0||0||37 (20%)|
|Class III||35 (46.7%)||0||0||35 (19%)|
|Total||75 (100%)||75 (100%)||35 (100%)||185 (100%)|
A duplicate of one of the images was used to assess intra-examiner reliability (images DG and EF). On long one-way analysis of variance, the variability between observers, for replicated images DG and EF, was highly significant ( P < 0.0001): the value of the F (184,185) statistic was 3.03. These results indicate that there was little variation in the intra-observer ratings for these images. The intra-class correlations was ICC = 0.50 (95% C.I. 0.40–0.61) (moderate reliability).
Perceived attractiveness of images
The univariate and multivariate mixed linear regressions for rating are given in Tables 3 and 4 . The most important factor influencing rating is the degree of sagittal chin prominence. The effect of chin prominence is more marked when the chin is protrusive (forward, positive) than when it is retrusive (backward, negative). Ratings decrease, on average, by 0.31 of a level of the Likert scale (95% C.I. 0.28–0.31) for each unit increase in chin protrusion and by 0.15 of a level of the Likert scale (95% C.I. 0.14–0.16) for each unit chin retrusion. In both cases, the difference begins only after 2 units (i.e. 4 mm) of change.
|Description||Coef. – interval||P value|
|Age||0.01 (0.00, 0.02)||0.004|
|Sex||0.05 (−0.16, 0.27)||0.64|
|Ethnicity||0.07 (−0.14, 0.28)||0.51|
|Self-rating of appearance||−0.08 (−0.24, 0.10)||0.38|
|Importance of an attractive appearance||−0.11 (−0.34, 0.11)||0.33|
|Observer’s skeletal class (anteroposterior jaw relationship)||0.06|
|II vs. I||−0.18 (−0.44, 0.08)||0.19|
|III vs. I||−0.30 (−0.57, −0.04)||0.03|
|II vs. III||0.13 (−0.20, 0.45)||0.45|
|Observer’s skeletal class (II and III vs. I)||−0.24 (−0.45, −0.03)||0.03|
|Observer’s vertical lower anterior face height||0.28 (−0.05, 0.62)||0.10|
|Chin prominence (deviation in mm)||−0.11 (−0.12, −0.11)||0.0000|
|Chin prominence (protrusion vs. 0 vs. retrusion)||0.0000|
|Normal vs. chin retrusion||2.39 (2.20, 2.58)||0.0000|
|Chin protrusion vs. chin retrusion||−0.60 (−0.69, −0.52)||0.0000|
|Normal vs. chin protrusion||2.99 (2.80, 3.19)||0.0000|
|Laypeople vs. patients||0.29 (0.07, 0.52)||0.01|
|Clinicians vs. patients||0.05 (−0.23, 0.33)||0.71|
|Laypeople vs. clinicians||0.24 (−0.04, 0.52)||0.10|