Have you looked beyond publishing your paper? A perspective on reproducible research
“A scientific publication is not the scholarship itself, it is merely advertising of the scholarship.”
Reproducible Research for Scientific Computing: Tools and Strategies for Changing the Culture
I really like this quote. It makes me think about what a journal article represents. A published paper is not the end, but rather a means to the larger picture – the actual research itself – which is broader in scope than the paper.
What is the larger picture?
The data and methodology used in research play a critical role in ensuring that a published paper is a true representation of the research it reports. They serve to ensure that the propositions presented in a publication stand up to scrutiny. They do not, and cannot, cease to exist once a theory is supported by the evidence at hand, once a study receives funding, or even when a paper is accepted by a journal.
Given the intense competition today, researchers often think that publishing a paper is the ultimate step in the ladder of success. But the fact is that scientific inquiry is much more than a journal article. It is about the science itself. In recent years, the increasing emphasis on the need for replicating scientific studies as well as the alarming rate of retractions indicate the need for us to think of the reproducibility of our scientific findings, so that they can be confirmed, disputed, or developed by other scientists and practitioners anywhere and at any time. This in turn highlights the importance of the data and methodology. In this post I’d like to talk about the role both data and methodology play in ensuring that research is reproducible.
Data: Show it so they know it
Data sharing is a great way of ensuring the replicability of the scientific method. The insistence for research data repositories (such as ClinicalTrials.gov or Open Science Framework) and the popularity of open access databases are on the rise in academia. By sharing your data, you not only help others understand your work in the larger context, but you also indicate your willingness to let other researchers validate your findings by replicating your study. A transparent and comprehensive dataset enables an investigation to be reproduced, and subsequently validated, with the same information in a different setting. Keeping your data under wraps defeats this cause.
Methodology: What goes around comes around
The journey from proposition to proof is not a one-way street. Any enquiry is iterative; a scientific author thinks and rethinks, and there’s no way to circumvent the feedback loop of research. As a researcher, you MUST be open to adopting an approach that accommodates the iterations in research. A linear approach that simply aims to arrive at a conclusion is a disservice to “development,” a term used most often in conjunction with research. Only by researching and developing will you be able to find answers to some of the most pressing problems faced by society, and one of the ways to find solutions to these problems is to make sure that the research methodology enables reproducibility.
There is no easy way out here. An intuitive paradigm is to empirically observe, generate an idea, hypothesize that the idea holds true, and test the premise by, well, observing again. While a sufficient number of contradictory observations can lead to a rejection of your hypothesis, no number of supportive observations can establish its universal truthfulness. A greater amount of support garnered through multiple observations can, however, strengthen the concept of a research. I say this because “establish,” “universal,” and “truthfulness” are weighted words in scientific research. They indicate absolutes; to ensure the efficacy of your findings, it’s best to avoid such absolutes and stick to facts that can be tested over and over again.
Do you see the larger picture now?
This brings me back to the quote I shared at the outset. A published paper is not the end of a scientific journey. It merely represents the actual science, which is much larger. If we agree that the main purpose of scientific research is to enable human development and help address critical problems, then we must give replication the importance it deserves and support it in every way we can.
How can you do your bit to support it?
Share your data, adopt a flexible methodology that can be adapted to other contexts, identify the limitations of your data and methodology, and clarify the scope for future studies that could build on your findings. A scientific publication may be authored by you alone, but its theoretical and practical contributions are co-created by you and the academic community. By allowing theorists to repeat your study and practitioners to observe the applicability of your findings, you’re letting in the academic community – which might just bear greater fruit than being let in by this community.
References
- Control, treat, repeat: Replicating an experiment on uncertainty
- Video - John Ioannidis: "Reproducible Research: True or False?" | Talks at Google
You might also be interested in reading the following posts:
Published on: May 04, 2018
Comments
You're looking to give wings to your academic career and publication journey. We like that!
Why don't we give you complete access! Create a free account and get unlimited access to all resources & a vibrant researcher community.
