It carried out a review of the empirical literature examining the nutrition implications of agriculture in Bangladesh. Search Criteria and Protocols: It was searched 10 databases (eg, WorldCat, PubMed, EMBASE, and Google Scholar) between November 2013 and March 2014 using search terms pertaining to agriculture, nutrition, and food. The search terms included different combinations of the key words Bangladesh, nutrition, food security, agriculture, farm, and smallholder. Further searched 10 different Web sites associated with the Government of Bangladesh and international research and development organizations, such as CGIAR, World Bank, United Nations Standing Committee on Nutrition, and the Food and Agriculture Organization and searched the bibliographies of relevant studies for additional, non duplicate references. This initial search yielded 2400 articles. It included full-text publications in indexed journal articles, books, grey, or unpublished sources linking nutrition outcomes to elements of agriculture in Bangladesh, published or released between January 1994 and March 2014. Examples of nutrition relevant outcomes included the intake of calories, macronutrients, micronutrients, or specific foods; changes in anthropometry; dietary diversity; food- and nonfood expenditures; and women’s status and empowerment as linked to nutrition outcomes. The dietary intake of environmental contaminants, most notably arsenic, was included due to its direct impact on health status and in many cases related nutritional outcomes. This time period was chosen because the large majority of studies published earlier did not include any type of impact evaluations. It excluded opinion pieces, conceptual papers, research released prior to 1994, publications that were not available in English, research that did not relate elements of agriculture to nutrition-relevant outcomes, animal studies, research that did not contain Bangladesh specific results, research that solely focused on the consumption of foods produced outside the homestead instead of by the household itself, meeting abstracts, literature reviews that summed up articles that were already included in the results, and research that could not be retrieved through more than 5 additional databases and search functions. All sources were entered into RefWorks, and duplicate, irrelevant, and inaccessible studies were removed. Analytic Approach: The final included studies were mapped to one or more of 6 agriculture–nutrition pathways and the relevant nutrition-relevant outcomes they measured. Each study was rated separately by 2 independent reviewers according to research quality using an adapted version of quality review protocols developed by the UK’s Department for International Development.17 The protocols were adapted by assigning more weight to internal validity and assigning actual point values to the criteria. The research quality rating system comprised 15 tests on conceptual framing, transparency, appropriateness and rigor, internal and external validity, reliability, and clarity. Fourteen indicators received 1 point each, with the final indicator, internal validity, receiving more weight with 4 points. Internal validity scores were assigned based on the study design used: randomized controlled trials (RCTs) received 4 points; quasiexperimental studies received 3 points; longitudinal studies received 2 points; and descriptive and cross-sectional studies received 1 point. Based on their cumulative performance on this 18-point system, the studies were then graded as high quality (13 to 18 points), moderate quality (9 to 12 points), or low quality (0 to 8 points).