Writing teachers and educators can add to information from large-scale testing and teachers can strengthen classroom assessment by creating a tight fit between large-scale testing and classroom assessment. Across the years, large-scale testing programs have struggled with a difficult problem: how to evaluate student writing reliably and cost-effectively. Indirect measures, direct assessments, "holistic" scoring, and primary trait scoring (as used by the College Entrance Examination Board, the Educational Testing Service, and the National Assessment of Educational Progress) all have serious limitations. Even though not well defined, the portfolio movement provides a potential link between large-scale testing and classroom assessment and teaching. Several large-scale portfolio assessment programs are currently in place: (1) the Arts PROPEL program, a Pittsburgh school-district portfolio project in art, music, and imaginative writing; (2) the "Primary Language Record," a kind of portfolio introducing systematic record-keeping about language growth into all elementary classrooms in the United Kingdom; (3) a draft, state-wide plan for portfolio assessment in Vermont; and (4) the General Certificate of Secondary Education (GSCE) in language and literature, in which British students choose either a timed writing test plus a portfolio of coursework or simply a folder of coursework. However, just collecting and evaluating portfolios will solve neither the assessment problems nor the need to create a professional climate in schools. By coupling assessment and instruction in increasingly sophisticated ways, educators and teachers may be able to make a real difference in education. (Seventy-three references are attached.) (RS)