Researchers have regarded inductive reasoning as one of the seven primary mental abilities that account for human intelligent behaviours. Researchers have also shown that inductive reasoning ability is one of the best predictors for academic performance. Modelling of inductive reasoning is therefore an important issue for providing adaptivity in virtual learning environments (VLEs). Research on Cognitive Trait Models (CTM) is currently underway to model such mental abilities by inferring learners' behaviours in VLEs. Despite its recognized importance underlying the learning process of human beings, little effort is spent by the research community to support learners' inductive reasoning process in such computer-based learning environments. This paper provides a conceptual basis which addresses this issue by asking the question of how to model the inductive reasoning ability of a learner. A structural overview of CTM is first presented to give contextual information about modelling learner's inductive reasoning ability, followed by examination of the characteristics of inductive reasoning ability in five different aspects: domain knowledge, generalization, working memory capacity, analogy, and hypothesis generation. On the basis of knowing the characteristics of inductive reasoning, extraction of the manifestations of inductive reasoning ability can then be made. The enumerated manifestations are then discussed. (Contains 1 table and 1 figure.)