Many psychologists are embracing a set of open research practices designed to produce more replicable results. They include:
Preregistering experiments. Preregistration involves creating a time-stamped document with your hypotheses, research design and analysis plans (Nosek, B.A., PNAS, Vol. 115, No. 11, 2018opens in new window). Projects can be preregistered on platforms such as AsPredicted.org and the Open Science Framework, which is run by the Center for Open Science. When the paper is published, a link to the preregistration can be included in the published article, enabling reviewers, editors and readers to see that you haven’t engaged in questionable practices like p-hacking (trying lots of analyses until you get a significant effect) or HARKing (hypothesizing after results are known). “Basically, you have evidence that you called your shots ahead of time,” says Julia Strand, PhD, an assistant professor of psychology at Carleton College. The Center for Open Science partnered with APA in 2017 to offer open science badges to authors, create an APA data repository and designate a preferred preprint server for APA journal articles. APA also recently updated its Journal Article Reporting Standards to include guidance on preregistering studies (Appelbaum, M., American Psychologist, Vol. 73, No. 1, 2018opens in new window).
Submitting registered reports. Registered reports do everything that preregistration does and more. You submit your detailed research plan to a journal before collecting any data, the submission undergoes peer review and, if approved, the paper is accepted in principle. “The reviewer basically asks, ‘Is this an interesting question? Is it methodologically sound?’” says Benjamin Le, PhD, a professor of psychology at Haverford College. “And if so, it doesn’t matter whether the results support the hypothesis or not.” This agnostic attitude toward outcomes has made way for more studies with negative results; null findings account for more than half of registered report publications, compared with just 5% to 20% of general literature publications across all fields (Allen, C., PLOS Biology, Vol. 17, No. 5, 2019opens in new window).
Sharing data, code and materials. Transparency in your research extends beyond saying in advance what you’re going to do. It also means showing your work. The best way to do that is to upload your data to one of many open access platforms. You can also share custom code you wrote to analyze the data as well as other materials you used in the experiment that can help others replicate your findings.
