In a series of large-scale (N>100) studies of analogy use in college physics, we have explored how, when, and why analogies affect student reasoning. In the first of these studies, we demonstrated that analogies affect student reasoning when taught in a large enrollment physics course. In the present follow-up study, we demonstrate that teaching EM waves concepts implicitly via analogy leads to greater conceptual change compared to teaching explicitly without analogies. Students were divided into two groups, one taught using analogies (string and sound waves) and the other taught without analogies (EM waves only). On a targeted concept question given before and after instruction, students who were taught with analogies outperformed those taught without analogies demonstrating that analogies can affect student reasoning in productive ways, even when taught implicitly. We propose framing as a mechanism to begin to explain why analogies can be productive when used implicitly.