This paper addresses problems with performance on open-ended (i.e., not multiple choice) test questions administered on paper for students accustomed to writing with computers. The first section summarizes the results of three studies that examined differences between computer and paper-and-pencil test results. The second section identifies three reasons why the mismeasurement of students accustomed to writing with computers is problematic. The third section focuses on a short term solution of allowing students to select the mode in which open-ended responses are composed; reasons why state testing programs have resisted this option are listed. The fourth section addresses other areas of learning that computers can help students develop (e.g., problem solving, research, non-linear thinking, understanding of scientific concepts, collaboration, spatial reasoning, statistics, media, music theory, and modeling/simulation) and notes that these areas are not measured by current state-mandated testing programs. The fifth section describes the steps necessary for testing and accountability programs to develop and apply instruments and testing procedures that capture the types of learning impacted by computer use. (Contains 28 references.) (MES)