Using two states' grades 3 through 8 state assessment databases, this study documents the extent to which students were assigned testing accommodations for ELA or mathematics in only one of two consecutive years. The percentage of students with disabilities who were assigned accommodations in the current year only or in the prior year only in a given grade statewide was not trivial, sometimes exceeding 25%. The relationship between inconsistent assignment to any accommodations and both students' prior proficiency level and aggregate growth is also documented. Group differences were observed at the state level. No practical differences were observed when covariates for inconsistent assignment were included in school value-added models, but very few schools had a substantial proportion of students with disabilities assigned accommodations inconsistently in a given grade. Implications for research and practice are discussed.