The paper advises that with a good quality assurance (QA) program, errors can be minimized and used to improve laboratory and personnel management. Hypothetical error rates add little value to the evaluation of the strength of the evidence in a specific case. Errors with significant consequences due to mistakes or bias are identified and addressed most effectively through peer review by retesting, reanalysis and/or blind verification. Consensus guidelines and standards for quality, education, and training must be developed for all disciplines; assumptions, inferences, and significance of all interpretations should be documented. In addition, research needs and direction should be documented, so that all are adequately informed about these practices. Research and development in methodologies and technologies from collection to interpretation of the analyses should be advocated and aggressively pursued. Greater outreach to the scientific community should be supported. New methods and techniques require validation, which involves assessing the ability of defined procedures to obtain reliable results, to define conditions that are required to obtain the result, to determine the limitations of the analytical procedure, and to identify aspects that must be monitored and controlled. For some analytical results only qualitative assessment of the significance are conveyed. In order to foster communication, coordinate efforts, leverage resources, and facilitate decisions about future directions, this paper recommends the convening of an International Summit on Forensic Research, to be continued periodically. The attendees should include expert practitioners, researchers, representatives of professional organizations, representatives from private industry, academicians, and other interested persons who can help develop a comprehensive strategic plan for forensic science research. 2 tables and 66 references