This study describes the development of a special education teacher observation protocol detailing the elements of effective decoding instruction. The psychometric properties of the protocol were investigated through many-facet Rasch measurement (MFRM). Video observations of classroom decoding instruction from 20 special education teachers across three states were collected. Twelve external raters were trained to observe and evaluate instruction using the protocol and assigned scores of "implemented," "partially implemented," or "not implemented" for each of the items. Analyses showed that the item, teacher, lesson, and rater facets achieved high levels of reliability. Teacher performance was consistent with what is reported in the literature. Implications for practice are discussed.