To obtain through the use of discriminant analysis a relatively bias-free sample of extraction and nonextraction, Class I, equally susceptible to both treatments' malocclusions and to analyze retrospectively the soft tissue changes between the two different treatment groups. The cephalometric, model, and demographic data of 215 patients (females and males) fueled a stepwise discriminant analysis that provided the borderline homogenous subsample (30 extraction and 32 nonextraction cases). The pretreatment and postreatment cephalograms of the borderline sample were then subjected to a thorough soft tissue cephalometric analysis. The results indicated that the three variables that played the most important role in the clinician's treatment decision were indicators of lower crowding, soft tissue convexity, and lower incisor protrusion. Significant differences (P < .001) regarding upper and lower lip protrusion, upper lip thickness (P < .05), and the nasiolabial angle (P < .05) occurred. Extraction treatment of Class I borderline malocclusions led to significant soft tissue changes regarding the upper and lower lip position and thickness as well as the nasiolabial angle, whereas the nonextraction treatment resulted in significant upper lip retraction and lower lip protraction.