Kernels vs. ears, and other questions for a science of treatment dissemination