There has been much discussion recently about why the percentage of students scoring Proficient or above varies as much as it does on state assessments across the country. However, most of these discussions center on the leniency or rigor of the cut score. Yet, the cut score is developed in a standard-setting process that depends heavily on the definition for each level of performance. Good performance-level descriptors "(PLDs)" can be the foundation of an assessment program, driving everything from item development to cut scores to reporting. "PLDs" should be written using a multistep process. First, policymakers determine the number and names of the levels. Next, they develop policy definitions specifying the level of rigor intended by each level, regardless of the grade or subject to which it is applied. Finally, content experts and education leaders should supplement these policy definitions with specific statements related to the content standards for each assessment. This article describes a process for developing "PLDs", contrasts that with current state practice, and discusses the implication for interpreting the word "proficient," which is the keystone of No Child Left Behind.