Tailoring of instructional methods to learner levels of expertise may reduce extraneous cognitive load and improve learning. Contemporary technology-based learning environments have the potential to substantially enable learner-adapted instruction. This paper investigates the effects of adaptive instruction based on using the isolated-interactive elements effect that occurs when learners who are initially presented with elements of information in an isolated, non-interactive form (followed by a fully interactive form) outperform those who are presented with the same information only in a fully interactive form. Cognitive load theory explains the effect for novice learners by their potential cognitive overload when dealing with a fully interactive form of instruction from the beginning. However, according to the expertise reversal effect in cognitive load theory, the effect may reverse for relatively more knowledgeable learners. Experiment 1 found that more knowledgeable accounting students performed better with interactive rather than isolated presentations. For less knowledgeable learners, there was no statistically significant performance difference between the presentation formats. Thus, there was a significant interaction between the instructional procedures and levels of learner prior knowledge as an indicator of an expertise reversal effect. In one of the two conditions used in Experiment 2, information was adaptively presented in isolated form to less knowledgeable learners but in interactive form to more knowledgeable learners (based on the pre-tests of learner prior knowledge). In another (control) group, students were randomly allocated to isolated and interactive instructional formats irrespective of levels of their prior knowledge. As expected, the adaptive instruction group was superior to the non-adaptive group. The paper concludes with implications for the technology enabled design of learner-tailored instructional presentations.