How tutors model students: A study of personal constructs in adaptive tutoring
Derry & Potts (1998)

what was studied
In this study, Derry and Potts examined the constructs that tutors use to classify and discriminate among the students they work with in order to adapt their tutoring techniques to students' differences. Two sets of tutors participated: (a) a set of three tutors who worked with students using a computer-based instructional system designed to improve basic mathematics and problem-solving ability, and (b) a set of two tutors who worked in a remedial college support program – one in mathematics, the other in language arts.

how effectiveness was measured
Rather than measuring effectiveness per se, this study examined the fundamental concepts that tutors used to describe and differentiate among their students and how these constructs shaped the tutorial decision-making process. Repertory grid interviews and cluster analysis were used.

The interviews included two main phases: knowledge elicitation & rating grid. In the Knowledge Elicitation phase, each tutor was presented with a series of triads of their students' names and was asked to what was the most important attribute distinguishing the two most similar members from the third dissimilar member. The discriminating construct and its opposite were recorded and the process was continued until no new constructs emerged. Tutors were also asked to specify the observations on which they discriminated among the student triad and how such discriminations influenced their tutoring. In the Rating Grid phase, tutors scored their students (on a 5-point scale) on each construct they previously generated.

what the findings were
All tutors classified their students in terms of two fundamental dimensions: motivation and intellectual ability. Tutors from the first set (those who worked with a computer-based instructional system for improving mathematical skill) generated similar general models or categorical "types" of students: Problem Students, students with low motivation and ability, and Other Students, a category including Exceptional Students, who were high in both ability and motivation, and Eager Students, who were high in motivation but not ability. Analysis of tutors' responses indicated that tutors (1) gather evidence about students during the tutoring sessions, (2) maintain cumulative models of their students that characterize each individual in general terms of motivation and ability, and (c) adjust their tutoring to accommodate these general characterizations.