Practice Recommendations or Not? The LoGeT Model as Empirical Approach to Generate Localized, Generalized, and Transferable Evidence

In educational research, there is the general trade-off that empirical evidence should be generalizable to be applicable across contexts; at the same time, empirical evidence should be as specific as possible to be localizable in subject-specific educational interventions to successfully transfer the empirical evidence to educational practice. This trade-off is further increased by the fact that the diverse instructional contexts, such as school or student characteristics constrain the applicability of empirical evidence. Several approaches have been proposed to address this issue, however, emphasized the different problems (i.e., localization, generalization, transferability) rather in an isolated manner. To this end, in this article, we introduce a synergistic approach, the LoGeT (localize, generalize, transfer) model, which systematically integrates co-design (localization strategies) and ManyClasses principles (generalization strategies) with co-constructive transfer activities, to generate empirical evidence that may be applicable in educational practice. To illustrate the LoGeT model, we present three long-term projects, covering different granularities and durations of educational interventions across different fields of education (teacher education, adaptive teaching, non-interactive teaching) that successfully applied the LoGeT approach. Finally, we outline further directions for future iterations of the LoGeT model. We hope that the LoGeT approach may be a stimulus to guide researchers as well as practitioners alike to design generalizable and evidence-based educational interventions that are rooted in localized instructional contexts.