Monday, January 18, 2016

Challenges in Study Replication

As a young researcher and graduate student, the issue of study replicability is one that has indeed come to mind over recent years. In agreement with Jeremy Berg’s position in the article The Reliability of Scientific Research, the scientific community must take ownership of this issue. The amount of studies that are reproduced does not seem to be enough; however, reproducing studies comes with a cost of not only time but also money. Therefore, regardless of how important this may be, it is impractical since both researchers and funding institutions are inevitably less likely to repeat a research study when there is opportunity to discover the new. However, one thing that did come to mind while reading this article was that there may be newer methods and technologies available, potentially making replicate studies just as fruitful as original studies.

Building off of Jared Horvath’s article, The Replication Myth: Shedding Light on One of Science’s Dirty Little Secrets, I also agree that irreproducibility is a common theme of scientific research, and has received increasing attention in recent years. However, many techniques require finesse and close attention, which makes study replication overwhelmingly challenging. To me, this would be similar to asking an artist to reproduce not only his or her own work, but also that of another artist. This is not to say that it is an impossible task, rather to point out that even the most experienced scientists may have trouble reproducing studies, regardless of the level of training and expertise.

Summarily, I think science has long reached the point where it is far behind the evolution of technology and demands of current society. Both of these factors seem to be pushing scientists to discover rather than to replicate their findings, and also places strain on financial institutions that are devoted to scientific discovery and advancement. To combat this issue I believe that funding institutions should, at the absolute least, select the most impactful studies and attempt replication. In addition, post-publication peer review is an excellent idea, and could potentially serve to unofficially suggest studies that those in the field believe should be reproduced.

No comments:

Post a Comment