Members of the UMaine PERL are studying ways in which different modes of instruction can affect student learning. We report on three studies. In the first, we compared three published curricula and how they teach Newton's Third law. When studying student learning of Newton's Third Law after being taught using research-based curricula, we found that all students' understanding of the material improved. One class showed significant improvements over the others. In another class, improvement occurred with topics not explicitly taught in the reform-based setting. The other two studies took place in a new course, Intuitive Quantum Physics, for teaching quantum physics to non-science majors. The second study dealt with student learning of probability density. Even after repeated instruction on the concept solely in classical physics contexts, many students pattern matched to situations from the class or misinterpreted relevant graphs. Only when studying quantum situations that were explicitly contrasted to classical situations did their understanding truly improve. The third study concerned student learning of quantum tunneling. In the Intuitive Quantum Physics course, students responded very differently to seemingly similar instructional methods. After using instructional worksheets that asked for fewer explanations of reasoning but facilitated comparisons between similar tunneling situations, students demonstrated better understanding of the physics on specially designed post-tests. We describe the curricula, our physics education research results, and how they have affected our instruction.