Statistics 101

In United States v. Jones, a jury acquitted the defendant of all charges except for three counts of fraud based on improper billing, representing a total loss of $120.76.



The defendant was sentenced to 18-months imprisonment, and ordered to pay $224,133(!!) in restitution for the conduct associated with his acquitted counts. At sentencing, the government used a statistical extrapolation from some of the defendant’s billing records to establish loss.



Ready for a statistics refresher? The Sixth Circuit recognized that “a statistical estimate may provide a sufficient basis for calculating the amount of loss caused by a defendant.” Here, however, the statistical analysis was flawed because it did not form a representative sample of bills – over fifty patient billing records were missing. The district court did not “even realize that the fifty four files were missing and it definitely did not make a finding as to whether they were fraudulent.”



The lesson learned is to have a grasp of basic statistics. Because the government’s statistics were flawed and because the government had not proven the acquitted conduct by a preponderance, the sentence was vacated and remanded to determine the total amount of loss and restitution.