The scientific community can be
quick to vilify an experiment that is deemed “irreproducible.” People assume
that the experimental design was poor or that the research was conducted sloppily
when two seemingly comparable studies don’t result in the same conclusion. In reality, immediately discounting science
that is not consistently reproduced causes us to lose out on the ability to
understand the nuances in whatever we are studying. If we get inconsistent
results or a result that does not match up with the current literature, there
is pressure to ditch the project for something more fruitful (i.e., publishable).
However, identifying factors that cause an outcome to be different than
expected can have real-world applications. For instance, failed efforts to replicate results
showing Sildenafil to be an ideal treatment for heart disease resulted in the implementation
of Viagra for erectile dysfunction. Irreproducibility should be acknowledged, but it should not be a death sentence for a study.
A
potentially more serious problem in science today is the lack of
generalizability in many studies. Do study results hold true across age, race,
and sex? Are the clinical assessments developed in the United States still
useful in less industrialized countries? These problems are being addressed, to
some extent, with things like the addition of the NIH requirement to consider
sex as a biological variable in all grant applications. However, we still have a long way to go. As an
example, a study published in January of 2018 indicates that the progression of
labor in Nigerian and Ugandan women does not necessarily reflect the “Friedman
Curve,” a timecourse developed in the United States. Additionally, even within the United States, research
studies tend to include cohorts of participants that are not reflective of the
diverse population that we have. These are discrepancies that need to be addressed in
order to maximize the benefits of the research we conduct.
No comments:
Post a Comment