In this experimental study, 44 undergraduates completed five computer-based instructional lessons and either two multiplechoice tests or two fill-in-the-blank tests. Color-coded borders were displayed during the lesson, adjacent to the screen text and illustrations. In the experimental condition, corresponding border colors were shown at posttest. In the control condition, border colors in the post test were mismatched with those used in the lesson. Participants were not informed of this color manipulation. Based on Tulving and Thomson's (1973) encoding specificity principle, the experimental group should remember significantly more content. Conversely, Murnane, Phelps, and Malmberg's (1999) ICE theory predicted that color-coding will have little effect on memory retrieval, since participants did not explicitly and actively integrate the topics with the color-coding scheme. Results generally favor ICE theory, however, the effects of practice/feedback varied by test format and color condition. A topic-color test measured the number of topic-color associations that were implicitly encoded and remembered. Gender and format differences were detected. (Contains 4 tables and 3 figures.)