Educational outcomes often result from two (or more) clearly hierarchical sampling dimensions, such as when students not only represent themselves but also represent some larger group. In multilevel modeling analyses, the students might be considered “Level 1” but be nested within schools at “Level 2” and would be modeled appropriately using the traditional multilevel model. However, when sampling dimensions are not clearly hierarchical, such as if students at Level 1 are simultaneously nested within more than one Level 2 variable (e.g., both schools and neighborhoods), the traditional multilevel model must be abandoned in favor of a cross-classified model. An understanding of the cross-classified model is critical given the potential curricular and fiscal policy ramifications that could result from incorrectly analyzing nonhierarchical, multilevel educational data. Therefore, ...