An article by James MacDonald (from the School of Natural Resources and Environment at the University of Michigan) describes an attempt to gather established emissions literature of various biofuels, and to identify discrepancies in the results. From a full fuel life cycle perspective, an analysis of the sources of the discrepancies, as well as potential areas for future research to reduce uncertainty, was also made. Emissions factors for biodiesel, ethanol, and cellulosic ethanol from five studies were taken, and emissions ranged from 0.52 kilograms of carbon dioxide per liter of ethanol, to 6.8 kilograms of carbon dioxide per liter of biodiesel. Feedstocks include three datasets for corn (ethanol), two for switchgrass (ethanol), one for corn stover (ethanol) and three for soybean oil (biodiesel). MacDonald observed "an incredible amount of variation that is difficult to explain, but some of the factors that provide a significant portion of the changes" were discussed. Some of the factors that might help explain the discrepancies were: (1) differences in process inclusion in the life cycle inventories, and (2) uncertainty in the inventory data/assumptions used, assessment boundaries, and allocation methodology for dealing co-products. According to the article, "life cycle assessments of biofuels are in need of some standard methodologies and a standard set of assumptions in order to create some consistency in the literature"..