The success of an intervention depends not only upon its theoretical soundness, but also on proper implementation that reflects the guidelines derived from its theoretical conception. Debates surrounding the effectiveness of problem-based learning (PBL) have focused on its theoretical conception and students' learning outcomes, but implementation is seemingly absent from the picture. This paper attempts to describe what research evidence is needed to fill in this missing information and provide a clearer picture of PBL. The author examines current PBL implementation practices and identifies potentially confounding variables that may play a role in inconsistent or conflicting research results in PBL. For example, various models of PBL have been developed and implemented to afford the specific instructional needs of the institution or learner population. These PBL models are in fact quite different in terms of the nature of problem solving and the degrees of selfdirected learning, which theoretically, should result in different types of learning outcomes. Without distinguishing the models used, the results of comparative PBL research could have been confounded. Furthermore, human factors are another set of confounding variables that could influence the students' learning processes and consequently affect PBL implementations and research results. To remedy these problems and reach PBL's full potential, as well as obtain a more accurate picture of PBL as an instructional method and its effects on students' learning, some fundamental changes are needed.Keywords Problem-based learning Á Implementation Á Problem solving Á Self-directed learning Á Student learning Despite its popularity in educational settings, problem-based learning (PBL) generates a great deal of skepticism and speculation among theorists. More than half a dozen metaanalyses and systematic reviews of PBL studies have been conducted to answer the question, These meta-analyses examined the effect of PBL on various aspects of students' learning outcomes, such as domain knowledge acquisition, problem solving skills, self-directed learning, group processing, and social and psychological soft skills. However, the results from the meta-analyses were not conclusive, even yet, conflicting.More recently, Kirschner et al. (2006) brought this issue back to the surface by openly criticizing PBL as an ineffective instructional method because it ignores or contradicts human cognitive architecture and cognitive load principles. Hmelo-Silver et al. (2007) and Schmidt et al. (2007) countered the argument by presenting the related cognitive science principles and empirical evidence that support the soundness of PBLs theoretical foundation. Specifically, some researchers suggested that PBL may not be effective in all aspects of student learning, but is especially effective in certain aspects of student learning because of its instructional emphases and characteristics. For example, there has been a general agreement that PBL is effective in promoting students' problem sol...