This paper addresses methodological and substantive issues regarding the use of portfolios of students' work in Chapter 1 evaluation and program improvement. In evaluations using portfolios desirable characteristics of the evaluation procedure are objectivity, reliability, and validity. The Chapter 1 Evaluation/Program Improvement System (CHEPIS) uses a local evaluation and an annual evaluation. Portfolios are useful sources of evidence about how effectively the Chapter 1 program meets its goals in performance and the duration of performance improvements. Major characteristics of portfolios include the following: (1) portfolios are a form of test and as such, theories and technologies of testing for reliability and validity can apply; (2) they can be useful and legitimate means of gathering data for the CHEPIS as they meet evaluation requirements, help participants draw situation-specific conclusions, and assist the participants in understanding students' learning better; (3) they have particular advantages for use in pre-kindergarten through first grade as they are specific and flexible for evaluating program progress toward desired outcomes; and (4) they have disadvantages as evaluation tools that must be recognized and controlled to the greatest extent possible. Statistical data are provided in two tables and one figure. Thirteen references are provided. (JB)