Review by: Holm, Chandra (2005-01-11)

In 1970s, Dick Snow and Lee Cronbach published results of their investigations on how differences in attitudes and aptitudes of individuals affect their learning, and how learning environments need to be adapted to these individual differences in aptitudes in order to optimize learning outcomes. In this paper, Valerie Shute, a principal research scientist in the Research & Development Division at Educational Testing Service and Brendon Towle of the learning company, Thomson Netg, describe how this concept of ATIs (Aptitude Treatment Interactions) can be used to create adaptive eLearning methods instead of ‘one-size-fits-all’ learning solutions. The authors develop the concept of an ‘adaptive engine’ that can be used to build such eLearning environments in which eLearning need no longer be confined to making learning material accessible but can focus on improving learning by adapting instruction and content to suit individual learners.

Shute and Towle’s paper is aimed primarily at designers of eLearning courses and researchers. It is based on a study involving 300 paid participants in which the interaction between exploratory behavior, one particular trait of learning style that is learner specific, and two kinds of learning environments, the rule-application and rule-induction methods provided by an ITS (intelligent tutoring system) were tested. The learning outcome was determined by four tests given after the learning sequence was completed. The authors explore what the knowledge gained by this research means for creating suitable eLearning environments that are adapted to learners’ needs instead of being a depository of just lecture notes and links. Three models based on three important components of eLearning – content model, learner model, and instructional model are described not only with regard to their basic ideas but also with regard to extending them to suit the concept of adaptive eLearning environments in which context is necessary to go beyond taking a systematic approach for presenting content. Two methods – SMART (Student Modeling Approach to Responsive Tutoring) and BIN (Bayesian Inference Network) – are presented as possible ways of modeling student behaviour and selecting appropriate learning objects.

At the beginning of any learning sequence, the adaptive engine assesses the learner’s knowledge level and presents a suitable topic by selecting learning objects appropriate to the learner’s needs. The learning outcome is again assesssed, and new topics made of LOs specifically selected to meet the student’s needs are presented. This chain continues till the learning outcome meets the set goal. The speciality of Shute and Towle’s adaptive engine is that pretests can be generated as and when needed, assessments can be based on a sequencing algorithm, and LOs are selected based on given rules which make use of the principle of ATIs.

In conclusion, in this paper Shute and Towle develop the idea of an adaptive engine to create an effective eLearning environment. This publication addresses an important problem in the field of eLearning. Its only drawback – if such an expectation from a publication in Educational Psychologist is valid – is that it does not describe in enough detail the technical aspects of implementing such an adaptive engine.

Source:
http://www.elearning-reviews.org/topics/pedagogy/learning-design/2003-shute-towle-adaptive-elearning/