This sentencing appeal concerns the procedural issue of the standard of review for decisions applying the Sentencing Guidelines to facts — in this case a decision applying the Guidelines' grouping rules — and the substantive issue of whether two episodes of unlawful sexual activity by a prison guard with an inmate, occurring on separate days, should be included in a single group. These issues arise on an appeal by Ricardo Vasquez, formerly a federal prison guard, from the November 6, 2003, judgment of the United States District Court for the District of Connecticut (Alvin W. Thompson, District Judge). The judgment sentenced Vasquez principally to 21 months' imprisonment after his plea of guilty to several counts charging sexual abuse of prisoners at the Federal Correctional Institution at Danbury, Connecticut ("FCI-Danbury") and one count of making a false statement.

Although opinions from this Circuit have used various formulations to describe the standard of review of decisions applying the Guidelines to the facts of a case, we conclude, in light of Buford v. United States, 532 U.S. 59, 121 S. Ct. 1276, 149 L. Ed. 2d 197 (2001), that the statutory requirement of "due deference to the district court's application of the guidelines to the facts," 18 U.S.C. § 3742(e) (Supp. II 2003), requires us to select either "de novo" review or "clearly erroneous" review depending on the primary nature of the application determination at issue. We further conclude that "de novo" review is appropriate in the circumstances of this case and, applying that standard, that the District Court's decision not to group the sexual offenses against the same inmate that occurred on different days was correct. We therefore affirm.

Vasquez, while serving as a guard1 at FCI-Danbury, engaged in sexual activity with four female prisoners. He had sexual intercourse with one inmate on one occasion; with a second inmate he had sexual intercourse on two consecutive days; with a third inmate he had sexual intercourse on one occasion and a few days later caused her to perform fellatio on him; and with a fourth inmate he engaged in repeated sexual touchings. The episode with the first inmate, each of the two episodes with the second inmate, and each of the two episodes with the third inmate were the subject of five separate counts of violating 18 U.S.C. § 2243(b), which is captioned "Sexual abuse of a minor or ward."2 The several episodes with the fourth inmate were the subject of one count of violating 18 U.S.C. § 2244(a)(4), which is captioned "Abusive sexual contact."3 During the trial, Vasquez pled guilty to the five counts of violating section 2243(b), the one count of violating section 2244(a)(4), and one count of making a false statement in violating of 18 U.S.C. § 1001(a)(2).

The Guidelines calculation concerned only the section 1001(a)(2) count and the five section 2243(b) counts because the section 2244(a)(4) offense is a Class B misdemeanor (maximum sentence of six months), see 18 U.S.C. § 3559(a)(7), which is not subject to the Guidelines, see U.S.S.G. § 1B1.9. The presentence report ("PSR") recommended the base offense level of 6 for the false statement count, see U.S.S.G. § 2B1.1(a)(2), and the base offense level of 9 for each of the five section 2243(b) counts, see id. § 2A3.3(a).

Then, pursuant to the "grouping" rules, id. §§ 3D1.1-.5, the PSR recommended that the section 1001 count and each of the five section 2243(b) counts should be a separate group, see id. § 3D1.2, for a total of six groups. The PSR further recommended a total offense level of 14, calculated by starting with 9, the highest offense level for any one of the six groups, see id. § 3D1.4, and adding one level for each of the five other groups, all of which were counted as a "unit" because they were equally serious and within four levels of the count with the highest offense level, see id. § 3D1.4(a).

The Defendant contended that the two section 2243(b) counts for sexual misconduct with the second inmate should have been included in one group, rather than two, and made the same argument with respect to the two section 2243(b) counts for sexual misconduct with the third inmate. The District Court rejected these claims, ruling that each episode of sexual misconduct with respect to the same victim on separate days was a separate harm for purposes of grouping.

The total offense level of 14 yielded a sentencing range of 15-21 months. The District Court selected a punishment of 21 months, and implemented the punishment by imposing 21 months on the false statement count, which carried a five-year maximum, 12 months concurrently on each of the section 2243(b) counts, which carried a one-year maximum, and six months concurrently on the section 2244(a)(4) count, which carried a six-months maximum. See id. § 5G1.2(c).

We consider first the appropriate approach to selecting a standard of review, then the standard of review applicable to the determination on this appeal, and then the merits of the Appellant's claim of improper grouping.

It is well settled that upon review of a Guidelines sentence the interpretation of a sentencing guideline is a question of law, subject to de novo review, see, e.g., United States v. Napoli, 179 F.3d 1, 6 (2d Cir.1999), and that a sentencing judge's finding of fact may not be disturbed unless clearly erroneous, see, e.g., United States v. Woodard, 239 F.3d 159, 161 (2d Cir.2001). Some uncertainty is encountered, however, in considering issues that are not easily categorized as issues of either fact or law. The most common problem concerns review of a sentencing judge's application of a guideline to the facts. Sometimes, what appears to be an "application" of a guideline is really an interpretation, as when a sentencing judge says, "I think that Guideline × means y, and I will therefore apply that meaning to facts a, b, and c." Difficulties arise when the sentencing judge is not so precise and simply says, "Based on my interpretation of Guideline × and the facts of this case, as I find them to be, I conclude that the Guideline applies [or is inapplicable]."

Congress endeavored to prescribe the standard of review for decisions applying the guidelines: "The court of appeals ... shall give due deference to the district court's application of the guidelines to the facts." 18 U.S.C. § 3742(e). The Supreme Court has instructed that "[t]he deference that is due depends on the nature of the question presented." Koon v. United States, 518 U.S. 81, 98, 116 S. Ct. 2035, 135 L. Ed. 2d 392 (1996) (ruling that "abuse of discretion" was proper standard for reviewing decision to depart from applicable sentencing range). That instruction is clear enough when the nature of the question is an issue of law, to which de novo review applies, and when the nature of the question is factual, to which the "clearly erroneous" standard applies, but the instruction does not provide guidance when the issue is whether the sentencing judge has properly applied a guideline to the facts.

Our Court has used five different formulations in articulating the standard of review of an application of a guideline to the facts.

(a) De novo. In a long series of decisions we have said that we review a district court's "interpretation and application of the [Sentencing] Guidelines de novo," United States v. Santiago, 384 F.3d 31, 33 (2d Cir.2004) ("Luis Santiago") (emphasis added). See, e.g., United States v. Smythe, 363 F.3d 127, 128 (2d Cir.2004) (same); United States v. Maaraki, 328 F.3d 73, 75 (2d Cir.2003) (same); United States v. Szur, 289 F.3d 200, 215 (2d Cir. 2002) (same). With one exception,4 all of our decisions stating the de novo standard can be traced back through a series of citations to a decision that did not state a de novo standard, but instead merely set forth the statutory standard of giving "due deference to the district court's application of the guidelines to the facts," 18 U.S.C. § 3742(e). See United States v. Deutsch, 987 F.2d 878, 884-85 (2d Cir.1993). The chain of citations back to Deutsch is set out in the margin.5

(b) Abuse of Discretion. Another line of decisions states that the standard is abuse of discretion. This standard was recently stated in United States v. Ravelo, 370 F.3d 266, 269 (2d Cir.2004) ("We review... the application of the Guidelines to [the district court's] findings for abuse of discretion unless the application presents a purely legal question, in which case we employ a de novo standard of review.") and in United States v. Coriaty, 300 F.3d 244, 249 (2d Cir.2002) ("[W]e ... will not overturn the court's application of the ... Guidelines to the facts before it unless we conclude that there has been an abuse of discretion."). Ravelo and Coriaty cite United States v. Deming, 269 F.3d 107, 109 (2d Cir.2001) (same), which cites United States v. Hernandez-Santiago, 92 F.3d 97, 100 (2d Cir.1996) (same), which cites United States v. Santiago, 906 F.2d 867, 871 (2d Cir.1990) ("Alvin Santiago") (same). Alvin Santiago provides no citation for its statement of an "abuse of discretion" standard.6

(c) Clearly erroneous. One decision stated the standard to be "clearly erroneous" review. See United States v. Lanese, 890 F.2d 1284, 1291 (2d Cir.1989) ("The determination of the application of a Sentencing Guideline is a question of fact, entitled to the `clearly erroneous' standard of review.").

(d) Due deference. Several decisions have stated the standard simply by using the "due deference" language of the statute, without further explication. See, e.g., United States v. Lenoci, 377 F.3d 246, 250 (2d Cir.2004); United States v. Huerta, 371 F.3d 88, 91 (2d Cir.2004); United States v. Reis, 369 F.3d 143, 151 (2d Cir. 2004).

(e) Either de novo or clearly erroneous. In another line of decisions, the statutory "due deference" standard has been amplified by explaining that the standard of review should be either "de novo" or "clearly erroneous," depending on the nature of the application determination being reviewed: the "de novo" standard is used where the district court's application of a guideline "approaches a purely legal question," see Deming, 269 F.3d at 109; United States v. Medina, 74 F.3d 413, 417 (2d Cir.1996); United States v. Brumby, 23 F.3d 47, 49 (2d Cir.1994), and the "clearly erroneous" standard is used where the district court's determination "resembles a finding of fact," United States v. James, 998 F.2d 74, 82 (2d Cir.1993); United States v. Vazzano, 906 F.2d 879, 883 (2d Cir.1990); see also United States v. Stroud, 893 F.2d 504, 507 (2d Cir.1990) (applying "de novo" standard where the issue "turns primarily on the legal interpretation of a guideline term"). We will refer to this approach as the "either/or approach" to selecting an appropriate standard of review.

The other circuits have also used a variety of formulations to express the standard of review of a decision applying a sentencing guideline to the facts. See, e.g., United States v. Mariano, 983 F.2d 1150, 1158 (1st Cir.1993) ("We have interpreted [section 3742(e)] as requiring, in most instances, that the court of appeals review a trial court's application of a sentencing guideline to the facts for clear error."); United States v. Morrison, 983 F.2d 730, 732 (6th Cir.1993) ("Questions of law, however, such as the appropriate application of this guideline to a particular set of facts, are subject to de novo review."); United States v. Stokley, 881 F.2d 114, 116 (4th Cir.1989) ("On mixed questions of fact and law, there is no bright-line standard but rather a sliding scale depending on the `mix' of the mixed question.").

At oral argument, the Government advocated a distinct level of review for decisions applying a guideline — a fourth standard somewhere between "de novo" and "clearly erroneous" yet different from "abuse of discretion."7 The Defendant urges us to apply a "de novo" standard.

In trying to identify the meaning of "due deference" in section 3742(e), we start, of course, with the statute itself. The "due deference" provision was added in 1988 to the Sentencing Reform Act of 1984 ("SRA"), Pub.L. 98-473, Title II, Ch. II, §§ 211-238, 98 Stat.1987 (1984), by section 7103(a)(7) of the Anti-Drug Abuse Act of 1988 "(ADAA"), Pub.L. 100-690, 102 Stat. 4181, 4417 (1988); the ADAA also renumbered former subsection (d) as subsection (e), see id., § 7103(a)(8). The text of the "due deference" provision creates the ambiguity we have documented in the range of appellate opinions construing it, and we therefore look for any clues in the legislative history.

Unfortunately, we have not located any authoritative legislative history of the "due deference" provision. The provision originated in a draft bill presented by the Department of Justice at a hearing of the Senate Judiciary Committee on October 22, 1987. See Sentencing Commission Guidelines, Hearing before the Committee on the Judiciary, United States Senate, Serial No. J.-100-38, 100th Cong, 1st Sess. 135, 159 (Oct. 22, 1987) (Statement of William F. Weld, Asst. Atty. Gen. Criminal Division). The "due deference" provision was incorporated into S. 1822, 100th Cong., 1st Sess. (1987), which was passed by the Senate on October 28, 1987, 133 Cong. Rec. S15370 (daily ed. Oct. 28, 1987), apparently in an attempt to precede the November 1, 1987, effective date of the Guidelines. However, the appellate review section of S. 1822, including the "due deference" provision, was dropped in a Senate-House conference, see 133 Cong. Rec. H10016 (daily ed. Nov. 16, 1987), and was not included in the enacted bill, Pub.L. No. 100-182, 101 Stat. 1266 (1987).

In the next session of the 100th Congress, the "due deference" provision was contained in section 312 of Title III of the proposed Minor and Technical Criminal Law Amendments Act of 1988, S. 2485 § 312, 100th Cong., 2d Sess. (1988). We have located no committee report on this Act. In explaining the Act to the Senate, Senator Byrd introduced a section-by-section analysis, which states that section 312 "provides that a district court's resolution of a mixed question of law and fact is entitled to due deference." 134 Cong.Rec. S7455 (daily ed. June 8, 1988). Although the analysis does not explain what was meant by "due deference," it makes clear that an intermediate standard, between "de novo" and "clearly erroneous" review was expected:

It is logical to establish different standards of appellate review for questions of law, questions of fact, and mixed questions of law and fact. Consider, for example, the case of a defendant who struck his victim, causing a laceration. The district court's finding that ten stitches were required to close the victim's wound is a purely factual determination that would not be disturbed by the court of appeals unless it were clearly erroneous.

The district court's finding that a serious bodily injury results in a four level increase in the defendant's offense level under the guidelines is a purely legal determination which, if incorrect, would be corrected by the court of appeals if it results in an inapplicable sentencing guideline range.

Mixed questions of law and fact should be subject to an intermediate standard of review. Thus, the district court's determination that ten stitches constitutes serious bodily injury under the particular circumstances of this case would be entitled to due deference from the court of appeals.

Id.

The "due deference" provision also appeared in section 7103(a)(7) of the Omnibus Anti-Drug Act of 1988, H.R. 5210, 100th Cong., 2d Sess. (1988), considered by the House later in 1988. When this Act was considered in the House, a section-by-section analysis was presented, apparently by Congressman Conyers, which differs slightly from the analysis presented to the Senate. The analysis of section 7103(a)(7) indicates not only that the "due deference" standard is between "de novo" and "clearly erroneous" review, but also suggests that the standard is a continuum of degrees of review:

When an aggrieved party appeals a sentence alleging a misapplication of the guidelines (i.e., that the sentence is the result of an incorrect application of the guidelines), the standard the court of appeals is to apply when reviewing that sentence depends on the nature of the alleged misapplication of the guidelines. The appellate court must accept a district court finding of fact unless the finding is clearly erroneous. The court of appeals will scrutinize more closely a district court determination of a legal question, applying a de novo standard.

Not every alleged misapplication of the guidelines will involve a finding of fact, however. Application of the guidelines involves a number of subjective determinations. For example, section 3A1.1 of the sentencing guidelines requires the district court to increase a defendant's offense level by 2 levels if the defendant knew or should have known that the victim of the offense was "unusually vulnerable due to age, [or] physical or mental condition." Although there may be a factual element to making such a determination, such as the victim's exact age of what (if any) physical ailment the victim had, the overall determination is not an objective finding of fact. Making the subjective determination required by section 3A1.1 (whether the victim was "unusually vulnerable"), unlike resolving purely factural [sic] questions, is not uniquely within the district court's expertise. The clearly erroneous standard, therefore, would be inappropriate.

On the other hand, because a determination under section 3A1.1 of the sentencing guidelines depends heavily on the unique factual pattern of the case, that determination cannot be considered simply a legal question. Section 3742 recognizes the importance of the district court's role in fact determinations by instructing the court of appeals to "give due regard to the opportunity of the district court to judge the credibility of the witnesses." The de novo standard, therefore, is also inappropriate.

Section 7103(a)(7) amends 18 U.S.C. 3742(d) to require the court of appeals to give "due deference to the district court's application of the guidelines to the facts." This standard is intended to give the court of appeals flexibility in reviewing an application of a guideline standard that involves some subjectivity. The deference due a district court's determination will depend upon the relationship of the facts found to the guideline standard being applied. If the particular determination involved closely resembles a finding of fact, the court of appeals would apply a clearly erroneous test. As the determination approaches a purely legal determination, however, the court of appeals would review the determination more closely.

134 Cong. Rec. H11257 (daily ed. Oct. 21, 1988) (footnotes omitted) (emphases added) (alteration in original).

Initially, we note our uncertainty as to what these excerpts from the two section-by-section analyses mean. They purport to identify an intermediate standard of review without describing it or relating it to any particular level of scrutiny, such as abuse of discretion. The second analysis recommends a "clearly erroneous" standard for determinations that are not exactly issues of fact but only "resemble[]" such issues. As to determinations that are not exactly issues of law but only "approach[]" such issues, it recommends review "more closely" than the "clearly erroneous" standard. We also note that these analyses are not authored by either the relevant House or Senate Committee, but appear to be staff analyses that, while sometimes helpful, especially when appended to the report of the pertinent committee, do not necessarily represent the collective view of the Congress that ultimately enacted the "due deference" provision as section 7103(a)(8) of the ADAA.

We deem the explanations in these analyses too metaphysical to be useful. A sentencing judge's application of a guideline might involve primarily a determination of an issue of fact or primarily a determination of an issue of law, but we see little utility in inquiring whether the determination "resembles" an issue of fact or "approaches" an issue of law, and even less utility in calibrating review standards so finely that review of a determination that "approaches" an issue of law will be done "more closely" than review under the "clearly erroneous" standard. Fortunately the Supreme Court has provided us with a sounder basis for understanding the meaning of "due deference" in section 3742(e).

The Supreme Court has thus far considered the "due deference" language of section 3742(e) on only two occasions. In the Court's first encounter with section 3742(e) in Koon, as we have already noted, the Court merely instructed that the deference that is due to a district court's application of a guideline to the facts "depends on the nature of the question presented." Koon, 518 U.S. at 98, 116 S. Ct. 2035. The question presented in Koon concerned a departure, a discretionary determination peculiar to the Sentencing Guidelines and not an example of the recurring issue of applying a guideline to facts. The Court concluded, not surprisingly, that the appropriate standard of review of a discretionary departure determination at that time was abuse of discretion. See id. at 98-100, 116 S. Ct. 2035.8

The Court's second consideration of section 3742(e) is far more pertinent to our pending appeal. In Buford v. United States, 532 U.S. 59, 121 S. Ct. 1276, 149 L. Ed. 2d 197 (2001), the ultimate issue was whether the defendant was subject to the enhanced punishment that the Guidelines specify for a "career offender," see U.S.S.G. § 4B1.1, a classification applicable to a defendant with two or more convictions for violent or drug-related felonies, id. § 4B1.1(a). The Guidelines specify that "related" felony convictions count as one for purposes of the career offender guideline. See id. § 4B1.2(c) (incorporating "related" test of subsection 4A1.2(a)(2)). Application note 3 to subsection 4A1.2 further specifies that convictions are "related" if they "were consolidated for ... sentencing." Id. § 4A1.2, comment., n.3. The Seventh Circuit, in which Buford arose, had previously ruled that prior convictions not formally consolidated for sentencing are nonetheless to be considered "related" for purposes of subsection 4B1.2(c) if they have been "functionally consolidated," United States v. Joseph, 50 F.3d 401, 404 (7th Cir.1995). The precise issue in Buford was whether two prior convictions were "functionally consolidated" under the Seventh Circuit's gloss on application note 3. The district court had ruled that they were. See Buford, 532 U.S. at 62-63, 121 S. Ct. 1276.

Viewing this issue as an example of the sort of question to which section 3742(e) requires due deference to the district court's application of a guideline to the facts, Buford, 532 U.S. at 63, 121 S. Ct. 1276, the Supreme Court ruled that the Seventh Circuit "was right to review this trial court decision deferentially rather than de novo," id. at 64, 121 S. Ct. 1276. Although the quoted words and the balance of the Court's opinion make clear that the proper standard is not the "de novo" standard, the opinion does not explicitly state what is meant by "deferential[]" review. Nevertheless, there can be no doubt that the Court meant deferential review to be review under the "clearly erroneous" standard.

First, the Court explained that it had granted review in Buford because of the different approaches the courts of appeals had taken to the proper standard for reviewing a determination of whether convictions were "related" and described the two approaches as "review[] for clear error" and "review[] de novo." Buford, 532 U.S. at 63, 121 S. Ct. 1276 (collecting courts of appeals cases). Second, and more significantly, the decision of the Seventh Circuit, which the Supreme Court affirmed, had applied the "clearly erroneous" standard, see Buford, 201 F.3d at 940 ("[T]he district court's conclusion that Buford's state cases were not functionally consolidated for sentencing cannot be called clearly erroneous.").9

In affirming the use of the "clearly erroneous" standard to the determination of "relatedness" in Buford, however, the Supreme Court made it clear that it was not endorsing that standard for review of all determinations applying a guideline to facts. After noting that the facts concerning "relatedness" were not disputed and that the issue was, in a sense, legal, Justice Breyer, writing for a unanimous Court, provided this critical explanation:

The legal question at issue is a minor, detailed, interstitial question of sentencing law, buried in a judicial interpretation of an application note to a Sentencing Guideline. That question is not a generally recurring, purely legal matter, such as interpreting a set of legal words, say, those of an individual guideline, in order to determine their basic intent. Nor is that question readily resolved by reference to general legal principles and standards alone. Rather, the question at issue grows out of, and is bounded by, case-specific detailed factual circumstances.

Buford, 532 U.S. at 65, 121 S. Ct. 1276 (emphasis added).

This passage confirms the correctness of our use of an "either/or approach" to the standard of review of application determinations, as exemplified in our prior decisions in Deming, Medina, Brumby James, Vazzano, and Stroud. Although one of these decisions cited the section-by-section analysis of the ADAA presented to the House, see Stroud, 893 F.2d at 506-07, all of them wisely avoided the vagaries of the staff analyses and ended up, in a straightforward manner, interpreting "due deference" to mean "de novo" review of determinations that primarily involve issues of law and to mean "clearly erroneous" review of determinations that primarily involve issues of fact. Buford refines this approach by making clear that even a nominally legal issue might be so "bounded by, case-specific detailed factual circumstances," 532 U.S. at 65, 121 S. Ct. 1276, as to warrant "clearly erroneous" review.

The "either/or approach" to review of guideline application determinations calls to mind the thoughtful analysis of former Judge William Norris, explicating for the in banc Ninth Circuit the appropriate standard of review for determination of a so-called mixed question of fact and law.10 See United States v. McConney, 728 F.2d 1195, 1200-04 (9th Cir.1984) (in banc). Recognizing that a mixed question of fact and law concerns the application of a legal standard to facts, see id. at 1200 (citing Pullman-Standard v. Swint, 456 U.S. 273, 289 n. 19, 102 S. Ct. 1781, 72 L. Ed. 2d 66 (1982)), Judge Norris outlined the following approach:

If application of the rule of law to the facts requires an inquiry that is "essentially factual," [Pullman-Standard] at 288, 102 S. Ct. 1781 — one that is founded "on the application of the fact-finding tribunal's experience with the mainsprings of human conduct," Commissioner v. Duberstein, 363 U.S. 278, 289, 80 S. Ct. 1190, 4 L. Ed. 2d 1218 (1960) — the concerns of judicial administration will favor the district court, and the district court's determination should be classified as one of fact reviewable under the clearly erroneous standard. If, on the other hand, the question requires us to consider legal concepts in the mix of fact and law and to exercise judgment about the values that animate legal principles, then the concerns of judicial administration will favor the appellate court, and the question should be classified as one of law and reviewed de novo.

Id. at 1202. Judge Norris added that "usually the application of law to fact will require the consideration of legal concepts and involve the exercise of judgment about the values underlying legal principles." Id.

The application of a sentencing guideline to facts is a classic illustration of the so-called mixed question of fact and law, or more precisely an instance of the general task of applying a rule of law to facts. The Ninth Circuit's "either/or approach" to that general task foreshadowed the Supreme Court's similar approach in Buford to the more particularized task of applying a sentencing guideline to facts. Although the approach will sometimes involve a close call as to whether a particular determination of a sentencing judge should be considered to involve primarily an issue of fact warranting "clearly erroneous" review or primarily an issue of law warranting "de novo" review, that task seems entirely manageable, especially if doubtful calls are usually resolved in favor of the plenary standard. The approach avoids the vagaries of an imprecise intermediate standard, and it focuses on the reasonably clear inquiry as to whether the sentencing judge's application of a guideline to facts primarily (or essentially) involves an issue of fact or law, rather than on the vague inquiry as to whether the application determination "resembles" an issue of fact or "approaches" an issue of law.

2. Selecting the Standard of Review Appropriate for the Appellant's Challenge to the Grouping Determination

To apply the "either/or approach," we must determine whether the District Court's grouping decision involves primarily an issue of fact or primarily an issue of law. That task requires some understanding of the Guidelines' "grouping" rules.

The Guidelines provide a set of grouping rules to guard against the risk that technically distinct but related forms of criminal conduct, capable of being charged in separate counts, do not result in excessive punishment. See U.S.S.G. §§ 3D1.1-.5. When counts alleging instances of criminal conduct are placed in the same group, the number of such counts does not result in more punishment than would result from a single count.11 By contrast, when counts alleging instances of criminal conduct are placed in separate groups, increased punishment results, although the degree of increase is less than would occur if separate punishments for each count were imposed consecutively. The grouping rules modulate the degree of increased punishment by a formula that increases the adjusted offense level by small increments depending primarily on the number of groups. See id. § 3D1.4.

The grouping rule for closely related counts provides: "All counts involving substantially the same harm shall be grouped together into a single Group." Id. § 3D1.2. The issue in the pending case is whether two counts charging a prison guard's sexual misconduct with one prisoner on two separate occasions should be placed in one or two groups.

The grouping rules provide guidance for determining when separate counts "involve substantially the same harm" within the meaning of section 3D1.2. The parties agree that the relevant portion of this guidance is as follows:

Counts involve substantially the same harm within the meaning of this rule:

. . . . .

(b) When counts involve the same victim and two or more acts or transactions connected by a common criminal objective or constituting part of a common scheme or plan.

Id. § 3D1.2(b). Application note 4 adds the further explanation that counts are to be grouped when they are "part of a single course of conduct with a single criminal objective and represent essentially one composite harm to the same victim." Id. § 3D1.2, comment., n.4. As an example of grouped counts, note 4 offers "two counts of mail fraud and one count of wire fraud, each in furtherance of a single fraudulent scheme ... even if the mailings and telephone call occurred on different days." Id. However, note 4 continues, "two counts of rape for raping the same person on different days ... are not to be grouped together." Id. (emphasis in original). Based on these examples, the Appellant contends that grouping of the same crimes involving the same person is appropriate whenever the crimes do not involve the use of force. The Government disagrees that force is a requirement for placing crimes against the same person in separate groups, and points out that the reason why the monetary crimes illustrated in note 4 are grouped is because punishment for such offenses is already increased by reason of the aggregate amount of loss involved, see id. § 2B1.1(b)(1) (loss table).

A grouping determination bears some similarity to the "relatedness" determination that Buford ruled was subject to "clearly erroneous" review. And an application of the grouping rules might, in some circumstances, present primarily an issue of fact, warranting "clearly erroneous" review. Moreover, in this case, the particular grouping dispute "grows out of, and is bounded by, case-specific detailed factual circumstances," as in Buford, 532 U.S. at 65, 121 S. Ct. 1276. Nevertheless, we conclude that the grouping determination in this case is appropriate for "de novo" review.

The narrow issue in dispute is whether grouping of offenses involving the same person on different days is appropriate only when force is used. That is a legal issue. Ultimately the grouping determination turns on the meaning of "substantially the same harm" in section 3D1.2, an issue of guideline interpretation for which Buford deems "de novo" review appropriate. Perhaps some aspect of a factual issue, warranting "clearly erroneous" review, would arise if application of the grouping guideline in a case like ours turned on whether the inmate's "consent," which the Appellant alleges, was more significantly influenced by sexual desire or by submission to authority. We conclude, however, that such a factual issue need not be considered in this case in order properly to interpret the "same harm" language of section 3D1.2.

3. Applying the "de Novo" Standard of Review to the Grouping Determination

Having concluded that the "either/or approach" is appropriate for selecting the standard of review for application of a guideline to the facts and having selected the "de novo" standard to review the particular determination made in this case, we readily conclude that the District Court was correct not to group the separate acts of sexual misconduct occurring with the same inmate on different days.

First, we agree with the District Court's interpretation of section 3D1.2 that use of force is not a requirement for placing the same crimes against the same person in separate groups. Crimes do not necessarily "involve substantially the same harm," which is the test for grouping under section 3D1.2, just because force is not used. As the Government points out, grouping is appropriate for application note 4's example of counts of mail and wire fraud against the same victim, not because of the lack of force but because the punishment for these grouped counts will be increased by aggregating the dollar amounts involved in each count.

Second, whether or not some non-monetary crimes might be grouped, two episodes of sexual misconduct that society has legitimately criminalized occurring with the same person on different days are not "substantially the same harm" for purposes of section 3D1.2.

The judgment of the District Court is affirmed.