Is a sentence that falls within the advisory guideline range categorically one that does not create unwarranted disparities among defendants with similar records who have been convicted of similar conduct? As a corollary, what evidence must a court consider in reviewing whether a sentence creates unwarranted disparities among similarly situated defendants, and may a court categorically reject circuit-wide or nationwide statistics?

Title 18 U.S.C. § 3553(a) mandates that in imposing sentence, a district court “shall” impose a sentence that is “sufficient but not greater than necessary” to comply with identified sentencing purposes, and that in determining the particular sentence, the court must consider certain enumerated factors including, as relevant here “the need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct.” See 18 U.S.C. § 3553(a)(6). In assessing whether such unwarranted disparities exist, however, courts have largely rejected comparisons grounded in the only evidence readily available to defendants, that is, statistics compiled by the United States Sentencing Commission and written opinions detailing the offense conduct and sentence received by other defendants convicted of similar conduct. Further, several courts have held that if a sentence falls within the advisory guideline range, it by definition avoids unwarranted sentencing disparities. As a result, § 3553(a)(6)’s directive to sentencing courts to avoid unwarranted disparities offers defendants an illusory guarantee of fairness and appellate oversight. The questions presented here are: Is a sentence that falls within the advisory guideline range categorically one that does not create unwarranted disparities among defendants with similar records who have been convicted of similar conduct? As a corollary, what evidence must a court consider in reviewing whether a sentence creates unwarranted ii disparities among similarly situated defendants, and may a court categorically reject circuit-wide or nationwide statistics?