The relationships between use of specific tutorial skills and pupil outcomes is the domain of the present paper, the third of three studies of mathematics tutoring designed and conducted using three populations of tutors--teachers, paraprofessionals, and students. The other two studies individually address the effects of tutoring on the achievement and attitudes of fourth and/or fifth grade pupils, and the effects of training in tutorial skills on the achievement and attitudes of fourth and/or fifth grade pupils. Minicourse 5, Individualizing Instruction in Mathematics, is used as the training vehicle for each of the three studies, although the materials are adapted for use with the student tutors. Regression analysis procedures are proposed in order to investigate the relationship of the tutoring skills and interpersonal behaviors to pupil outcomes. The data which forms the basis of the regression analyses consists of frequency counts indicating the number of times that a particular skill is used in the tutorial setting. A more appropriate question in these analyses might have concerned itself not with the quantity of skill usage, but with the quality or appropriateness of a particular skill being used. (Author/AM)