Past research has identified social and environmental causes and correlates of behaviors thought to be associated with obesity and weight gain among children and adolescents. Much less research has documented the efficacy of interventions designed to manipulate those presumed causes and correlates. These latter efforts have been inhibited by the predominant biomedical and social science problem-oriented research paradigm, emphasizing reductionist approaches to understanding etiologic mechanisms of diseases and risk factors. The implications of this problem-oriented approach are responsible for leaving many of the most important applied research questions unanswered, and for slowing efforts to prevent obesity and improve individual and population health. An alternative, and complementary, solution-oriented research paradigm is proposed, emphasizing experimental research to identify the causes of improved health. This subtle conceptual shift has significant implications for phrasing research questions, generating hypotheses, designing research studies, and making research results more relevant to policy and practice. The solution-oriented research paradigm encourages research with more immediate relevance to human health and a shortened cycle of discovery from the laboratory to the patient and population. Finally, a "litmus test" for evaluating research studies is proposed, to maximize the efficiency of the research enterprise and contributions to the promotion of health and the prevention and treatment of disease. A research study should only be performed if (1) you know what you will conclude from each possible result (whether positive, negative, or null); and (2) the result may change how you would intervene to address a clinical, policy, or public health problem.