School district performance ratings are presented on intuitive grading scales. These ratings are widely publicized by local news outlets and can potentially influence voter behavior. Based on behavioral research applied to the local public finance context, we hypothesize that voters will anchor their votes in initial performance ratings even if the performance rating subsequently changes. This article investigates this hypothesis with a panel of all Texas school district bond referenda from 1998 to 2015. Estimates of probit models with careful consideration of selection bias provide support for our anchoring hypothesis. Evidence of anchoring votes, however, is nuanced and varies across two different Texas school district rating systems.