Figure 1: Recommendations to improve reproducibility and rigor of biomedical research organized across the 4 domains: teaching computational skills to produce reproducible research (“Teach”), development and distribution of data and software (“Develop and Distribute”), implementation of reproducible research (“Implement”), and incentivizing reproducible research (“Incentivize”).
Gigascience, Volume 9, Issue 6, June 2020, giaa056, https://doi.org/10.1093/gigascience/giaa056
Create, implement, and enforce standard operating procedures for collecting, reporting, storing, sharing data and for onboarding new students
Store everything electronically and make a back-up; take photos if need-be
Work with a statistician consultant or company (ex: Experimental Design Assistant)
Train students on ideal reproducibility and and rigor practices and statistics
Publish and pre-register your protocols and clinical trials
Use version control software such as github, Box
Use LabArchives, an electronic lab notebook that helps with data sharing, storage and collaboration
Design your analyses so that they can be run automatically as a script
Celebrate no or negative results in addition to positive results
Include full transparency and compete methodological detail in lab notebooks and manuscripts
Attend talks or conferences, enroll in courses, and view tutorials on enhancing reproducibility
Use open source code, software, and tools and make your data open
Run periodic tests!
Keep a clean and carefully labelled lab
Request the code for papers you are reviewing
Provide public access to scripts, runs, and results
Ask professional associations and societies to promote the importance of reproducible research (ex: editorials, official statements and guidelines,...)
Publish protocols and research outcomes, including negative and null results and replications studies https://plos.org/reproducibility-assessment/
Pre-register study design(s), primary outcome(s) and analysis plan(s)
Pressure publishers to require more detailed methodology sections
Take reproducibility into consideration for promotion/tenure
Support incentives from publishers and funders (ex: badges or credits)
Make replication studies and testing common place
