Although clinicians debate the relative importance of function and attractiveness in orthodontic treatment plan- ning, 1,2 there is no doubt that parents base their decisions primarily on the chance of improving their child’s appearance. Thus, this study by O’Neill et al provides a timely, well-controlled clinical trial of 2 procedures used to treat Class II Division 1 malocclusion. Consistent with the new philosophy of evidence-based dentistry, children were randomly assigned to the control (untreated), Fränkel, or Harvold groups. To minimize distracting influences of complexion, hair color and style, facial expression, etc, profile images were reduced to silhou- ettes. Panels of parents, art students, and dental students compared the attractiveness of pretreatment and post- treatment silhouettes with silhouettes of untreated con- trols produced over the same time period. Although the data suggest possible increases in attractiveness for the 2 treatment groups, the differ- ences are not statistically significant, primarily because of the extremely high variances (SD 2 ) in all respondent groups. Although the reliability of the judgments is extremely high, the validity of the end- points may be questioned. The visual analog scale was anchored with examples of very attractive, average, and very unattractive images selected by only 2 ortho- dontists. In agreement with other studies, the authors acknowledge that the physical bases of perceiving a face as attractive or unattractive differ widely among judges from orthodontic standards. 3 The issue is also complicated by the necessary assumption for paramet- ric statistical analysis that the intervals on the visual analog scale are subjectively equal. From that assump- tion, the authors might have considered the use of stan- dard scores or correcting individual raw scores to the (mean - raw score/SD) to help minimize the effects of the large interindividual differences. Although gender differences may not be a problem when judging 11-year-olds, the fact that neither the anchor nor the stimulus faces were gender-specific eliminates the possibility of determining if there are differences between perceived attractiveness in stimu- lus faces identified as male or female. Certainly there are in adults. Perhaps a different result would have been obtained if the silhouettes were all compared with each other in a forced choice task, such as paired comparisons. Other possibilities include requesting judgments of lips, chin, and/or mandible separately, or comparing pre- and posttreatment images with a standard such as the Rick- etts’ E-line. The conclusion of no differences in attrac- tiveness resulting from treatment with functional appli- ances, while courageous, may be premature. Donald B. Giddon, DMD, PhD Clinical Professor of Oral Health Policy and Epidemiology, Harvard University, School of Dental Medicine and Clinical Professor of Psychology in Orthodontics, University of Illinois at Chicago, College of Dentistry REFERENCES 1. Maxwell R, Kiyak HA. Dentofacial appearance: a comparison of patient self-assessment techniques. Int J Adult Orthod Orthog- nath Surg 1991;6:123-31. 2. Ostler S, Kiyak HA. Treatment expectations versus outcomes among orthognathic surgery patients. Int J Adult Orthod Orthog- nath Surg 1991;6:247-55. 3. Giddon DB. Orthodontic applications of psychological and per- ceptual studies of facial esthetics. Seminars in Orthod 1995; 1:82-93. 377 COMMENTARY On “Ratings of facial attractiveness after functional appliance treatment”