Suggests that traditional measurement-based approaches to evaluation are theoretically inappropriate in whole-language classrooms. Argues that responsive evaluation can be applied at the classroom level and that the data generated will tell more about children's developing control of language than standardized tests do. (RJC)