As applied linguists begin to accept the naturalistic paradigm as a legitimate and viable approach to research in their field, they can draw upon various models that have been developed in the educational literature on naturalistic program evaluation. In this literature, model is sometimes used interchangeably with design. When a distinction is made, model tends to be seen as the more grandiose of the two terms and refers not only to the plan for collecting the information on which the evaluation will be based, but also to its theoretical basis and purpose. I will continue with the terminology of the preceding chapter, using design to refer to specific plans for collecting and interpreting data to evaluate language education programs. These designs have, in most cases, been given names or labels and, in addition to specifying steps to take and considerations to make, represent particular motivations, both theoretical and practical, for doing program evaluation.

The responsive model

As early as 1967, Stake was advocating an approach to program evaluation that has been characterized as transactional (Patton 1980). Originally referred to as the countenance model (Stake 1967), it was refined into the responsive model (Stake 1975). The model is transactional in the sense that it responds to the requirements that various audiences have for information concerning the program.