ABSTRACT
Educational research has observed widespread difficulties in scaling up technology that succeeded in the pilot phase, difficulties ascribed to a range of factors (Becker, 2000; Zhao & Cziko, 2001; Zhao & Frank, 2003). We intend this report not only to demonstrate methodological issues in assessing speech-interactive CALL, such as designing control conditions, but also to anticipate factors that influence whether and how well the benefits of systems like the Reading Tutor will scale up to practice.
