The purpose of this study was to investigate evidence of reliability, criterion validity, and grade-level differences of curriculum-based measures of writing (CBM-W) with 612 students in grades 1-3. Four scoring procedures (words written, words spelled correctly, correct word sequences, and correct minus incorrect word sequences) were used with two CBM-W tasks (picture-word and story prompt) during fall, winter, and spring of one academic year. A subsample of participants (n=244) were given a criterion measure in spring of the academic year. Pearson's r coefficients were calculated to determine evidence of alternate form reliability and criterion validity, and a MANOVA was used to detect significant growth within and across grade levels. Results indicated that scores on both CBM-W tasks had adequate reliability and validity coefficients in grades 2-3 and mixed results in grade 1. Significant growth was detected within and across all grades at each time point on each task. Implications for research and practice are discussed.