Despite increasing calls for transparent integration, the mixed methods field lacks replicable approaches that center theory to answer multilevel-systems questions. We describe the Deductive-Inductive Dual Pillar Integration Process (DID-PIP), a five-phase integration technique for multilevel research. The DID-PIP centers a theoretical framework to generate meta-inferential understandings of multilevel systems that moves beyond confirmation or rejection of theoretical codes. Drawing from sociological, qualitative paradigms, and mixed methods reporting standards, we outline the theoretical rationale, define the DID-PIP, and illustrate its application through a literature synthesis of racial-equity monitoring in special education policy. This contribution is novel because it provides researchers with a structured, theory-driven integration approach that is methodologically rigorous and builds understandings of phenomena within and across multilevel systems. [This is the online first version of an article published in "Journal of Mixed Methods Research."]