There is an extensive need for school systems to reliably assess the data literacy and data use skills of their educators. To address this need, the current study seeks to refine the NU Data Knowledge Scale (NUDKS) for assessing teacher data literacy for classroom data. A data-based decision-making framework provides the theoretical underpinnings for the instrument. The study's objective is to refine the NUDKS such that items are located at various points along the data literacy continuum. In this fashion, the NUDKS should be able to measure teacher data literacy throughout the data literacy continuum. To this end, item response theory is used to provide the estimates of the items' locations and teacher data literacy. Analyses revealed that the NUDKS conformed to the Rasch model. To facilitate the future use of the NUDKS, concordance tables were created to provide a quick determination of teacher data literacy.