The methodology and problems of research on cross-age tutoring are reported. Using a phenomenological framework, the two researchers involved in the study worked toward a research method and design that would account for the backgrounds of both; one was trained in the quantifiable methods of natural science, the other in qualitative methods of human science. The study focused on the relationship that developed between eighth graders and first graders paired in a tutoring program. The researchers' graduate assistants observed and interviewed children involved in the program throughout the semester. Differences in approach between the principal investigators surfaced in decisions about the need for hypotheses, objectivity in a situation requiring involvement, reliability, and generalizability. Data analysis has so far only dealt with the observations. More problems may arise in analyzing the interviews and presenting final results. Further issues arising from a phenomenological approach include the need to consider observer reactions, the need for the principal investigators to be directly involved in data collection, and the need to study the teachers' reactions to the tutoring. Despite all the differences the researchers conclude that this collaboration has helped them to confront their own biases. (IS)