Annually, states are required to monitor whether school districts disproportionately refer certain student groups into special education services. However, these monitoring methods do not account for student factors theorized to be associated with the presence of a learning disability. I explore whether conclusions drawn from current measures of disproportionality remain unchanged once relevant student-level factors are accounted for. Using parameters estimated from logit models, I construct a weighted adjusted risk ratio (WARR) that accounts for student-level characteristics. Comparing the WARR with the commonly used weighted risk ratio (WRR), I find that the unconditional WRR tends to overstate overrepresentation for some minority groups. However, the differences are not large enough to reverse any conclusions about the direction of disproportionality in districts.