Recent research has demonstrated that worked example based instructional designs can effectively foster learning of engineering concepts and are supported by contemporary educational theories, including cognitive load theory. However, a number of interrelated fundamental questions, which have neither been addressed in the educational psychology nor in the engineering education literature, remain open including: (A) What is the impact of means-ends practice? (B) What is the effect of backward vs. forward fading of worked example steps? and (C) What is the effect of adaptivity to learner performance? The goal of the present study was to answer these questions by comparing the learning and perceptions about learning of engineering college freshman who learned how to solve electrical circuit problems in five different computer-based learning conditions: (1) problem solving with step-by-step feedback, (2) means-ends problem solving with total feedback, (3) backward fading, (4) forward fading, and (5) adaptive feedback. Forward fading and adaptive feedback practice promoted more students' near problem solving transfer ability than backward fading practice. Furthermore, the adaptive feedback practice group outperformed students in the backward fading practice group on measures of far problem solving transfer.