A normalization step is widely exercised in life cycle assessment (LCA) studies in order to better understand the relative significance of impact category results. In the normalization stage, normalization references (NRs) are the characterized results of a reference system, typically a national or regional economy. Normalization is widely practiced in LCA-based decision support and policy analysis (e.g., LCA cases in municipal solid waste treatment technologies, renewable energy technologies, and environmentally preferable purchasing programs, etc.). The compilation of NRs demands significant effort and time as well as an intimate knowledge of data availability and quality. Consequently only one set of published NRs is available for the United States, and has been adopted by various studies. In this study, the completeness of the previous NRs was evaluated and significant data gaps were identified. One of the reasons for the significant data gaps was that the toxic release inventory (TRI) data significantly underestimate the potential impact of toxic releases for some sectors. Also the previous NRs did not consider the soil emissions and nitrogen (N) and phosphorus (P) runoffs to water and chemical emissions to soils. Filling in these data gaps increased the magnitude of NRs for “human health cancer,” “human health noncancer,” “ecotoxicity,” and “eutrophication” significantly. Such significant changes can alter or even reverse the outcome of an LCA study. We applied the previous and updated NRs to conventional gasoline and corn ethanol LCAs. The results demonstrate that NRs play a decisive role in the interpretation of LCA results that use a normalization step.