This article reviews the literature of hypermedia research. Several perceived problems are identified, including: irrelevant assessment of outcome variables; weak methodological design in conducting experimental research; lack of theoretical foundations in designing hypertext systems; lack of specification of precise learning outcomes; and lack of learners' prior knowledge in the content area. Suggestions are made to reform the future hypermedia research agenda, including: future research should focus on assessing relevant student learning outcomes (knowledge acquisition, knowledge integration, and knowledge generation); researchers should develop reliable and valid instruments and methodology of experimental design (reporting reliability of criterion measures); hypermedia design should reflect hypotheses derived from established learning and instructional design theories; and future studies should consider learners' prior knowledge and varied learning styles. (Contains 23 references.) (Author/MES)