Abstract
Biomedical journals continue to be the single most important conduit for disseminating biomedical knowledge. Unlike clinical medicine, where evidence is considered fundamental to practice, journals still operate largely in a 'black box' mode without sufficient evidence to drive their practice. We believe there is an immediate need to substantially increase the amount and quality of research by journals to ensure their practice is as evidence based as possible. To achieve this goal, we are proposing the development of an international 'best practice journal research network'. We invite journals and others to join the network. Such a network is likely to improve the quality of journals. It is also likely to address many unanswered questions in publication science, including peer review, which can provide robust and generalizable answers.