Randomized controlled trials are a special case of designs using an unbiased instrument to take care of confounders even if they are unmeasured or unknown. Another example of studies using instrumental variables is the Mendelian experiment and Directed Acyclic Graphs show the power of such designs to enhance the internal validity. It is argued that internal validity is a necessary condition of ethically sound application of research findings in (clinical) practice and policy. Internal validity can only gradually and slowly be approximated through a spiraling process of ever more exact and varied replications, secondary analyses, meta-analyses, and umbrella reviews. This advocacy of slow science argues for distinct roles and responsibilities of practitioners and researchers in translational research.