Improving Reproducibility and Reliability of Scientific Research

            Since beginning graduate school and embarking on my early career as an independent researcher, I was astounded by the degree of difficulty in recapitulating experimental results from peer-reviewed primary literature. Even if the protocol and experimental design is clearly laid out, reproducing data can be a daunting task. While many may attribute this lack of reproducibility and reliability of scientific data to the ‘publish-or-peril’ culture of contemporary science with researchers not taking the proper steps to ensure their data is statistically significant and controlling for their inherent biases, Horvath et al., 2013, ScientificAmerican, states that “unreliable research and irreproducible data have been the status quo since the inception of modern science”. Horvath also states that several seminal studies such as Millikan’s oil drop experiment and Galileo’s postulation of his law of motion, which have laid the foundations for our understanding of our physical world, even proved to be difficult to reproduce. While it is true that researchers may be susceptible to biases and prone to overlooking details that may deviate from their hypothesis, designing thorough and sound experiments is an iterative process that relies on many rounds of critical assessment by peers. Peer-review offers the opportunity to obtain different perspectives that may have been overlooked and ultimately culminate in improved experimental design.

While irreproducibility of scientific data is not a foreign concept to the scientific community, it is the job of the scientist to be completely transparent about their experimental design and what they deem as potential pitfalls to their experimental setup. Transparency is also critical for educating the public of scientific research, as many individuals do not have the training to critically assess scientific literature. Articulating the implications of a research study as well as areas of contention or conflicting results in a way that is translatable to the public is paramount for allowing individuals to make informed decisions on their own. For instance, this is especially important in biomedical research, where individuals can formulate logical decisions that can impact their livelihood (e.g. decision to get vaccinations). With much skepticism of scientific research and its importance as of lately, it is therefore instrumental that scientists can maintain the trust of the public, and reliable studies can be published that can facilitate further discussion among the scientific community as well as equip individuals with the knowledge they need to make informed decisions.