This study examines the psychometric properties of the Writing Attitude Scale for Teachers (WAST) using scores of preservice teachers to gather evidence of content, internal structure, response processes, reliability/precision, and correlations with external variables. Four contending models were compared using CFA, but the unidimensional model was championed. The WAST was further refined using IRT to 10-items and shown to have high precision across most of the latent continuum. The WAST was shown to have incremental evidence relative to the Writing Apprehension Test when predicting attitudes towards attending a professional development workshop on writing instruction and time devoted to writing instruction.