This study investigates the effects of two factors: the mode of problem-solving instruction (i.e. Web-based versus non-Web-based) and the level of academic achievement (i.e. high achievers versus low achievers) on students' problem-solving ability and biology achievement. A quasi-experimental design was used, in which the experimental group received six weeks of Web-based problem-solving instruction in biology and the control group received non-Web-based problem-solving instruction for the same content and for the same period of time. Pre-, post- and retention tests of problem-solving and biology achievement were administered before and at two different time intervals after the instruction. With the pretest scores as a covariate, the results of MANCOVA followed by protected univariate "F" tests suggest that Web-based problem-solving instruction has the potential to enhance and sustain the learner's problem-solving skills over an extended period of time. (Contains 5 tables.)