The Environmental Protection Agency (EPA) is required by law to assure that the use of pesticides does not cause unreasonable risks to humans or the environment when risks are compared with benefits. Weed scientists conduct hundreds of comparative efficacy tests each year, but the results are often of little use to the Agency in benefit assessments because the tests are unpublished or otherwise unavailable to the Agency, the tests are conducted in a manner unusable for regulatory purposes, or there are inconsistencies between tests conducted year to year or at different sites. Despite the lack of high quality data, the Agency is compelled to make the best regulatory decision possible with the information at hand, and it may appear to some that decisions are based more on policy than science. EPA is looking for experimental methods that will improve the quality of benefits data available to the Agency.