IT IS ORDERED:
1. Petitioner Peter Michael was convicted of first-degree sexual assault based on an encounter with L.R., age eighteen, who was a close friend of Michael’s daughter. The night before she left Anchorage for college, L.R. and Michael’s daughter got together with some friends, where L.R. drank some alcohol. The two returned to the Michael house early in the morning of August 16, 1998. L.R. went to bed in a downstairs guestroom. Before she fell asleep, Michael got into the bed with her. He then reached under her shirt and touched her breasts. L.R. said no and “nudg[ed] him away,” but Michael persisted. He then pulled down L.R.’s shorts and underwear and penetrated her with his fingers. The penetration lasted, according to L.R.’s testimony, “[t]en seconds, five seconds,” and involved “[o]ne, maybe two” fingers. L.R. continued to tell Michael to stop and elbowed him. He stopped the penetration, but returned to touching her breasts. L.R. continued to protest and then “nudged him pretty hard with [her] elbow.” Michael then stopped, kissed her on the cheek again, apologized, and left. There was no allegation that physical force was used.
2. For a first-time felony offender, the first-degree sexual assault conviction carries a presumptive sentence of eight years.1 At sentencing, Michael attempted to establish two statutory mitigating factors: that his conduct was among the least serious conduct within the definition of his offense2 and that the harm caused by his conduct was consistently minor.3 He also attempted to prove that he had extraordinary potential for rehabilitation and that his sentence therefore should be referred to a three-judge panel for adjustment.4 The superior court refused to adjust the sentence, finding that the age difference and relationship of trust between Michael and L.R. precluded a finding that Michael’s conduct was among the least serious included in the definition of the offense. Despite its finding that Michael, a veteran with a steady employment record and no previous criminal history, was an excellent candidate for rehabilitation, the superior court also declined to refer the case to a three-judge panel. The trial court imposed the presumptive eight-year term for the first-degree sexual assault conviction and a *519concurrent term of three years with two years suspended for the second-degree conviction. Michael appealed. Applying the clearly erroneous standard of review, the court of appeals affirmed the superior court’s rejection of Michael’s proposed mitigating factors. The court of appeals also remanded the case to the superior court with instructions to clarify its findings on rehabilitation and to decide again whether to refer the case to a three-judge panel.5 The superior court clarified its findings and made the referral, apparently under the impression that the court of appeals’s remand order changed the standard for referral. The court of appeals provided some additional clarification after receiving those findings, stating that it had not changed the standard for referral, and, based on the superior court’s clarified findings, affirmed that court’s initial refusal to refer the ease to a three-judge panel. We agreed to review the sentencing issues raised in Michael’s petition for review.
3. We have never determined the standard of review of a superior court’s decision regarding the presence or absence of aggravating or mitigating factors. The parties appeal’ to have assumed that the court of appeals’s decisions applying the clearly erroneous standard of review resolved this question.6 We conclude that the court of appeals erred in reviewing the superior court’s rejection of Michael’s proposed mitigating factors under a clearly erroneous standard. We hold that the correct standard of review of a superior court’s application of statutory aggravating and mitigating factors to a given set of facts is de novo review.
4. The existence or non-existence of an aggravating or mitigating factor is a mixed question of law and fact. The determination of whether the defendant’s conduct is among the least serious conduct within the definition of the offense involves a two-step process: the court must (1) assess the nature of the defendant’s conduct, a factual finding, and then (2) make the legal determination of whether that conduct falls within the statutory standard of “among the least serious conduct within the definition of the offense.” Any factual findings made by the court regarding the nature of the defendant’s conduct are reviewed for clear error, but whether those facts establish that the conduct “is among the least serious” under AS 12.55.155(d)(9) is a legal question.7 We note that this analysis mirrors the approach the federal courts have taken in reviewing the application of the Federal Sentencing Guidelines to particular facts.8
5. De novo review of the application of statutory aggravating and mitigating factors furthers the goal of reducing disparity that underlies our presumptive sentencing scheme.9 The “clearly erroneous” deferential standard of review creates a danger that because of differences in superior court sentencing practices some defendants will receive the “least serious conduct” mitigator while other defendants who have engaged in *520similar or even less serious conduct will not.10 Whether particular conduct is “among the least serious conduct included within the definition of the offense” should not vary from case to case or judge to judge.11 While the need for more exacting scrutiny is particularly clear in the case of the “least serious conduct” mitigating factor, de novo review of the application of the other statutory aggravating and mitigating factors to particular facts will also promote uniformity in sentencing by ensuring that those statutory factors are applied in a consistent way. We therefore conclude that the superior court’s determination of whether a statutory aggravating or mitigating factor applies to a particular set of facts is reviewed under the de novo standard.12 To the extent that the court of appeals’s decisions are to the contrary, those decisions are overruled.
6. The court of appeals has not passed on the question whether the superior court’s rejection of the “least serious conduct” statutory mitigating factor would survive de novo review. Prudence dictates that we allow the court of appeals to consider this question in the first instance.13
7. Under ordinary circumstances, we would remand the case directly to the court of appeals for a determination of this question. But in a recent request for notification of ease status, Michael has indicated that as of May 26, 2005 he had only seventy-eight days left to serve on his sentence. We believe that either through a least-serious conduct finding or through the three-judge panel’s consideration of the non-statutory mitigator of extraordinary rehabilitative potential, it would be clearly mistaken not to suspend at least some portion of the presumptive term. Since Michael has already nearly completed serving the full presumptive term, any adjustment will certainly place him in jeopardy of serving more time than his sentence on remand. Given these circumstances, fundamental fairness requires Michael to be treated as if he has already completed serving the unsuspended portion of his sentence.14 We therefore REMAND the case first to the trial court with directions to release Michael from state custody, subject to appropriate parole/probation conditions, pending remand proceedings in the court of appeals.15
8. The superior court should then refer the case back to the court of appeals to determine whether under a de novo standard of review, the superior court in its initial decision erred in rejecting Michael’s proposed mitigating factor under AS 12.55.155(d)(9). If the court of appeals decides, using a de novo standard of review, that Michael’s conduct is among the least-serious conduct within the definition of the *521offense, it will remand the case to the trial court for resentencing.
9. Alternatively, if the court of appeals decides that Michael’s conduct does not merit a “least-serious” characterization, the court of appeals should direct that this case be referred to the three-judge sentencing panel for consideration of the non-statutory mitigator found by the superior court. We reach this conclusion for two reasons: (1) the superior court found that Michael proved his excellent potential for rehabilitation and (2) this finding of rehabilitative potential was not clearly erroneous. In these circumstances, the case should be referred to the three-judge panel. Considering Michael’s extraordinary potential for rehabilitation and the totality of the circumstances in light of the goals of sentencing as laid out in State v. Chaney,16 it is apparent to us that the range of sentences that could justly be imposed must necessarily stop short of the presumptive sentence, which is aimed at more serious conduct by a more intractable criminal defendant. Most of the purposes of sentencing— rehabilitation, preventing criminal conduct during confinement, and deterring future offenses by Michael17 — will be served by a sentence less than eight years. And the remaining sentencing goals, deterrence of others and the expression of community condemnation, do not on their own justify the length of the term, in light of the nature of Michael’s crime.18 It would therefore be manifestly unjust to sentence him to that term without accounting for his extraordinary potential for rehabilitation.
Entered by direction of the court.
. AS 12.55.125(1X1).
. AS 12.55.155(d)(9).
. Id. (d)(13).
. AS 12.55.165(a).
. Michael characterizes the court of appeals's decision to remand rather than reverse as lacking "integrity.” We assume that he is not actually making the serious charge that his language implies, and we remind counsel that questioning a court’s ethics is not an acceptable, or an effective, rhetorical tactic.
. See, e.g., Lewandowski v. State, 18 P.3d 1220, 1222 (Alaska App.2001).
. Cf. Ornelas v. United States, 517 U.S. 690, 116 S.Ct. 1657, 134 L.Ed.2d 911 (1996) (once the "historical facts” are established, probable cause determinations, including whether the facts suffice to generate a reasonable suspicion in a reasonable officer, are to be determined de novo on appeal); In re J.A., 962 P.2d 173, 175-76 (Alaska 1998) (in determining whether a child is in need of aid, the court gives deference to factual findings but reviews the legal determination of whether probable cause exists under a de novo standard).
. See, e.g., United States v. Yazzie, 407 F.3d 1139, 1148 (10th Cir.2005) (Briscoe, J., concurring) (describing how appellate review of guidelines involves both factual and legal issues, and that the determination of whether "a given set of facts" establishes the existence of a particular statutory enhancement is a legal question reviewed de novo); United States v. Hoskins, 282 F.3d 772, 778 (9th Cir.2002) (application of the abuse of trust enhancement involves a mixed question of law and fact requiring de novo review).
. Senate Bill (S.B.) 56, 24th Leg., 1st Sess. (2005) (Alaska's presumptive sentencing scheme "is designed to avoid disparate sentences").
. Cf. State v. Armstrong, 2002 WL 31185806 (Alaska App., Oct.2, 2002), (affirming the superi- or court's finding that the least serious conduct mitigator applied even though the conduct at issue appears to be more serious than Michael's conduct). The defendant in Armstrong took the thirteen-year-old best friend of his daughter shopping and provided her with alcohol until she "passed out." Id. at *1. The girl woke up to discover Armstrong having sexual intercourse with her in his bed. Id. We assume that the court of appeals's determinations that the defendant in Armstrong was entitled to the least serious conduct mitigator while Michael was not is attributable to the deferential standard of review used by the court of appeals.
. Cf. In re J.A., 962 P.2d at 175 (probable cause determination in CIÑA case is reviewed de novo in part to avoid permitting different trial judges from drawing different conclusions as to whether the facts are sufficient to constitute probable cause).
. We recognize that the legislature has recently amended AS 12.55.155 to provide that a jury determine whether many of the statutory aggravating factors are present. S.B. 56. We need not determine in this case the proper standard of review for a determination made by a jury under the new statute.
. Cf. Austin v. United States, 509 U.S. 602, 623, 113 S.Ct. 2801, 125 L.Ed.2d 488 (1993) (declining to decide when a forfeiture is constitutionally excessive until the lower courts have had the opportunity to consider the question).
. The unique procedural circumstances at issue here raise a question whether due process would be violated if AS 12.30.040(b) and AS 11.41.410(b) were interpreted to bar Michael's release pending proceedings on remand.
. Notwithstanding Appellate Rule 512(a), the provisions of this order take effect immediately. See Appellate Rule 507(b).
. See Bossie v. State, 835 P.2d 1257, 1258 (Alaska App.1992) (citing State v. Chaney, 477 P.2d 441, 444 (Alaska 1970)); see also Smith v. State, 711 P.2d 561, 569-70 (Alaska App.1985).
. See Chaney, 477 P.2d at 444; see also AS 12.55.005.
. See Chaney, 477 P.2d at 444; see also AS 12.55.005.