Short-term studies often do not capture the full story of the effect that a researcher is trying to describe. If one wishes to elucidate long-term educational processes and outcomes, the only direct way of doing so is to stretch the duration of research projects and to study the same subjects over time. The limited availability of valid longitudinal data seems to be largely due to methodological problems and logistical difficulties. The focus of this article is on true longitudinal studies in science education and on other studies which aimed at studying science education over the long term. Topics discussed include: (1) long-term inferences; (2) defining and finding longitudinal studies; (3) the realities of longitudinal studies; (4) research-based knowledge in science education; and (5) longitudinal thinking. Thirty-four longitudinal studies in science education are identified and discussed in terms of focus and design features. A 124-item reference list is appended. (CW)