The argument is advanced that research could be more effective for use by teachers and simultaneously more scientific, in the sense that it would emulate the historical process of the development of quantitative science, if it were (1) of the form referred to as naturalistic case studies and (2) oriented toward discovering the cognitive and social interaction mechanisms that underlie the learning processes in classrooms. The limitations of individual case studies of children with difficulties learning in school are examined and some practical approaches to interventionist case studies done in classrooms are considered. It is proposed that the standard view of quantitative research in the social sciences derives from the physics curriculum rather than from the history of physics. This view also mistakenly places higher priority on finding predictive linear relationships among measurements rather than qualitative understanding of mechanisms, which could guide researchers toward useful quantitative research.

The researcher who has not yet penetrated the world of the individuals being studies is in no firm position to begin developing predictions, explanations, and theories about that world.