<p>[eng] Is there such a thing as a “best scientific methodology” in regulatory</p><p>(decision-oriented) science? By examining cases from varying regulatory</p><p>processes, we argue that there is no best scientific method for generating</p><p>decision-relevant data. In addition, in regulatory science, the most suitable</p><p>methodologies often differ from what is considered best practice in</p><p>knowledge-oriented (academic) science. In data generation for regulatory</p><p>purposes, we are faced with a wide spectrum of preferred methodologies as</p><p>well as controversy as to methodological choice. What goes by the most</p><p>adequate scientific method can and will—justifiably and rationally—vary</p><p>significantly according to context and use. In order to make this argument,</p><p>we analyze four case studies, two from risk assessment and two from</p><p>benefit assessment. Our analysis shows that it is the noncognitive objectives</p><p>of a particular regulatory process that determine what counts as the most</p><p>appropriate scientific method. We use the concept of bounded rationality</p><p>to indicate that those methodological choices, despite being context-</p><p>dependent, can be interpreted as rational.</p>