Peer review has been around since journals were first published. Peer review organized by journal editors is, however, relatively recent, having become popular in the mid-1900s. Prior to that time, editors decided what to publish. The change of approach has not been good for science. Mandatory journal peer review is biased against the proper scientific study of important problems. It is also unreliable, slow, expensive, and has led to the mindless publication of many incorrect and useless studies. We suggest that journal peer review should be replaced by assessment of whether a paper follows proper scientific procedures. We describe the development of checklist software to aid in this process. Using the software would reduce the time and cost of reviewing research papers and help to avoid biased reviews. We pretested the software using a convenience sample of published papers and compared preliminary findings with those from software designed to assess the conformity of advertisements with evidence-based persuasion principles. The online journal PLoS employ a criterion they call “soundness,” which is akin to assessing conformance to science, for choosing articles to publish. Since that journal was founded ten years ago, PLoS has become the world’s largest publisher of research articles. We suggest that journal editors, PhD programs, universities, law courts, and research funders including governments adopt the Conformance to Science checklist to efficiently identify research worthy of support and use, and to thereby encourage the growth of scientific knowledge.