AAMC Award Focuses on Improving Science Writing and Coding
Biomedical researchers often find it difficult to reproduce laboratory and clinical studies – a problem that raises costs, reduces the credibility of investigators and undermines the public trust in science, according to two University of Miami faculty members taking a fresh approach to this national problem.
“We believe that poorly written prose and software code are under-recognized contributors to failures to replicate other researchers’ results,” said Joanna Johnson, Ph.D., director of writing and director of scientific writing programs in the College of Arts and Sciences.
Johnson and her colleague, Kenneth W. Goodman, Ph.D., professor and director of the Miller School of Medicine Institute for Bioethics and Health Policy, and co-director of UM Ethics Programs, have teamed up to address these issues, while incorporating writing and coding into the University’s responsible conduct of research (RCR) training.
Recognizing the importance of reproducibility, the Association of American Medical Colleges (AAMC) recently presented a third-place Innovation in Research and Research Education Award to Johnson and Goodman for their collaborative project, “Writing Prose and Writing Code.” Johnson accepted the UM award at the AAMC’s “Learn Serve Lead 2017” annual meeting Nov. 3-7 in Boston.
“Scientific prose is often of low quality, suffers from hedging, boasting, passive voice and other tricks, and is rarely among the skills included in the training of researchers,” said Goodman. “Also, investigators who write, modify and use computer software are untutored in the role of data analysis and its tools.”
Johnson noted that the too-frequent failure to replicate the results of scientific experiments is not only relevant to the advancement of knowledge but also costly.
“An analysis of past studies indicates that more than 50 percent of preclinical research – about $28 billion per year – is not reproducible,” she said. “This problem also contributes to the delays and overall costs of therapeutic drug development.”
In their RCR training, Goodman and Johnson are teaching UM biomedical, nursing, engineering, psychology and other science students about the importance of reproducibility of their work. “When writing up studies, clarity and simplicity are crucial,” Johnson said. “We all have a professional and moral responsibility to write prose that is accurate and presented so that others can clearly understand what we are saying.”
Researchers who modify other developers’ code or write their own applications to analyze data need to document their actions and decisions, added Goodman. “Be as transparent as possible, so that other researchers can reproduce and validate your work.”
The 2017 AAMC Innovation in Research and Research Education Award winning projects highlighted institutional biomedical research and research training initiatives enhancing research data rigor and reproducibility in innovative and more effective ways.