The public's confusion around what constitutes a healthy diet is related in part to nutrition studies that do not standardize the foods consumed by research participants or participants' adherence to their diet programs, argues Kevin Hall in this Perspective. According to Hall, domiciled feeding studies - which involve subjects being housed and fed in comfortable yet controlled facilities - could greatly improve the understanding of dietary influences on health. While consensus exists regarding fundamental aspects of what constitutes a healthy diet, consensus quickly erodes when discussing more detailed questions of optimal human nutrition. Recently, whether modern nutrition science is up to the task of answering related questions has been questioned. Exacerbating the issue, says Hall, is the lack of carefully controlled studies to explore it. "...imagine trying to develop a new drug without being confident that investigators could administer quantities of the drug to subjects or objectively measure its dose-response," he writes. Many randomized diet studies - often hinged upon subjective, participant-reported metrics, and which rarely verify or directly monitor the consumption of on- or off-study foods - do not reveal the effects of consuming different diets, but rather, the effects of people following differing diet advice. The results of such studies "conflate diet adherence with the effects of the diet," he says. Hall argues that we need to facilitate more human nutrition studies where subjects can comfortably reside continuously at a research facility, thereby allowing investigators to control and objectively measure their food intake. Well-designed domiciled feeding studies could help elucidate the fundamental mechanisms by which known diet changes affect us all, providing a bevy of information on the complex interactions between diet changes, the microbiota and host physiology, for example, as well as providing standardized diet biomarkers and test technology that can be used in longitudinal diet studies to assess diet-disease associations. Hall does also note limitations of these studies, such that "long-term nutrition studies in free-living people will always be required."