Research on juvenile and adult sentencing has been characterized by theoretical, methodological, and empirical limitations that preclude adequate description, prediction, or assessment of decisionmaking processes and outcomes. Five limitations are prominent: emphasis on atheoretical, empirical attempts, generally unsuccessful, to increase predictive accuracy; limited conceptualizations of dependent variables (e.g., incarceration versus nonincarceration); overreliance on individual, offender-level data with minimal reference to victims, practitioners, or contextual factors; failure to incorporate analytically multiple research methods; and inattention to intended and unintended uses and effects of sentencing. These limitations can be highlighted by focusing on a context—-juvenile justice—in which the goals of sentencing are varied, conflicting, and, due to recent reforms, changing. Using interview and survey data, the present research examines juvenile sentencing reform in Texas to highlight these limitations and to oudine an analytical framework for improved description, modeling, and assessment of sentencing.