Without question, discretion plays an important role in the criminal-sentencing process. Some of this discretion has resulted in racial and ethnic disparity in sentencing. Under the Sentencing Reform Act (1984), judicial discretion was constrained as a result of the Federal Sentencing Guidelines. For the 20 years that followed, for the most part, judges could only consider offense severity and offender criminal history when applying sentences. However, in 2005, the Federal Guidelines were made advisory following the United States v. Booker decision. This study evaluates the effects of this decision on sentence length. Using data from the Monitoring of Federal Criminal Sentences database, I estimate the effects of Booker, as well as race, ethnicity, and other variables on sentencing. Moreover, I use quantile regression models to examine whether the effects of those variables are consistent across the distribution of sentences. Implications are discussed.