Household food insecurity results when safe and nutritious food is not available, cannot be accessed in socially acceptable ways, or is not physiologically utilized completely. World Food Program’s (WFP) Purchase for Progress (P4P) is a pilot initiative that provides access to food markets and promotes agricultural productivity for over one million low-income smallholder farmers worldwide (>7,000 in Guatemala alone). P4P combines novel market development strategies with investments in capacity building in an effort to sustainably boost national food security and improve livelihoods. The objective was to characterize the main determinants of household food security and dietary diversity in the context of an agricultural and market development program in Guatemala. We compared food security and dietary diversity between P4P beneficiaries and a control group. We evaluated household conditions, food security (ELCSA), and dietary diversity (HDDS) in 372 households (271 P4P; 101 control) using a cross-sectional design and mixed-methods. Most Significant Change (MSC) methodology was used to characterize participants' experiences in a subset sample of 57 households (46 P4P; 11 control). Education level (EL), number of children (NC), household quality (HQS), food security (FSS), carotenoid-rich foods (VAS), and dietary diversity for households (HDDS), women (WDDS), children (IDDS), and normalized (HDDSn) were calculated from quantitative data. MSC interviews were transcribed verbatim. Interview transcripts were analyzed according to the principles of grounded theory, using open, axial and selective coding (NVivo ver. 9.2 and 10) which involved breaking down, examining, comparing, labeling, categorizing and integrating data into pre-determined and emerging categories. Connections among categories were established according to a coding paradigm comprising observed conditions, context, action/interactional strategies and consequences. We constructed and linked program impact pathways (PIP) based on a mixed-methods. Each pathway factor was laid out along the hypothesized PIP using as blueprints P4P’s program theory and current conceptual frameworks linking agriculture, food security and nutrition. Factor inclusion was supported by evidence from three sources: our study, P4P’s monitoring and evaluation, and current literature. A six-step process integrated information: data weighting, entry, preparation, analysis, interpretation and final integration. P4P participants were less food insecure (FSS=7.4±4.4 vs. 9.2±3.1; p<0.01), had increased VAS (p<0.01), overall and normalized dietary diversity (HDDS=8.9±1.8 vs. 7.0±1.8). HDDS was also higher among women and children (p<0.01). Among P4P participants, food security was associated (p<0.05) with education level (r=0.23).