We examine the effects of, and interplay among, several proven research-based reforms implemented in an introductory large enrollment (500+) calculus-based physics course. These interventions included Peer Instruction with student response system in lecture, Tutorials with trained undergrad learning assistants in recitations, and personalized computer assignments. We collected extensive informal online survey data along with validated pre/post content- and attitude-surveys, and long answer pre/post content questions designed to assess learning gains and near transfer, to investigate complementary effects of these multiple reforms, and to begin to understand which features are necessary and effective for high fidelity replication. Our average [median] normalized gain was .62 [0.67] on the FCI, .66 [0.77] on the FMCE, yet we find we cannot uniquely associate these gains with any individual isolated course components. We also investigate the population of students with low final conceptual scores, with an emphasis on the roles played by demographics, preparation, and self-reported attitudes and beliefs about learning.