Conflict-reducing practices during evolution instruction have been recommended to increase students' perceived compatibility between evolution and religion, increase evolution acceptance, and decrease stereotypes about religious students in science. However, the efficacy of these practices has not been demonstrated in a randomized controlled design making it uncertain whether they are causing the effects reported in less controlled studies. Further, we do not know the extent to which the religious identities of instructors may impact their effectiveness. In this study, we randomly assigned 2623 undergraduate students in 19 biology courses across different states to receive an evolution video with 1) no conflict-reducing practices, 2) conflict-reducing practices implemented by a non-religious instructor, or 3) conflict-reducing practices implemented by a Christian instructor. We found that the evolution videos with conflict-reducing practices led to decreased conflict, increased compatibility, and increased acceptance of human evolution compared with the video without conflict-reducing practices. Further, the Christian and non-religious instructor conditions were equally effective at improving all student outcomes, except the non-religious instructor was more effective for increasing perceived compatibility between religion and evolution among atheist students. These results illustrate that conflict-reducing practices, implemented by either Christian or non-religious instructors, can be effective in a controlled study.