In order to obtain concrete information to improve test writing, this study analyzes 13 achievement test drafts written by graduate teaching assistants, lecturers, and adjunct faculty teaching introductory and intermediate college Spanish. The analysis focuses on factors of test method and content likely to create variance in test performance unrelated to learners' abilities. Bachman and Palmer's (1996) rubric for describing target language use and test tasks functions as an organizational tool to describe and analyze the drafts. Results reveal patterns of inappropriate input and inadequately specified procedures, tasks, scoring criteria, and expected responses. Findings suggest common pitfalls for the novice test writer to avoid in test development, and provide a starting point for teachers of any language who want to improve their own achievement tests.