Scientific Bias Across Different Fields

As a scientist, it is increasingly more common to hear about papers being retracted after being published and about how journals are becoming more demanding in terms of what they require for submission. Many postdoctoral fellows and professors joke about how it is notoriously impossible to repeat experiments published in high-impact journals that were conducted at other research institutions. One explanation regarding the issue of reproducibility could be that the conditions at every institution are different, leading some researchers to obtain positive results, while others obtain negative results. However, the issue of reproducibility raises a more concerning issue: how much bias is introduced into experiments by researchers?

Science is driven by the underlying need to obtain significant results. Publications are crucial to advancing a scientific career and it is difficult to publish without having statistically significant data. Can the need for obtaining significant results cause a researcher to introduce bias into their experiments? According to Daniele Fanelli, a researcher at Stanford who studies publication bias, bias is greater in the social sciences, such as psychology, economics, and sociology, when compared to the physical sciences, such as chemistry and biology. 

Some fields, such as psychology, seem to be inherently more prone to experimenter bias. Although Fanelli suggests that the social sciences tend to have more bias than the physical sciences, that doesn’t mean that the physical sciences are bias free. In some cases, the physical sciences can introduce bias through experimental design, such as testing enough animals until significance is reached, or through the presentation of the data. No matter what field a researcher is working in, it is important to keep in mind how bias can be introduced into experiments and an effort must be made to reduce bias as much as possible.

https://www.washingtonpost.com/news/in-theory/wp/2017/03/31/how-biased-is-science-really/?utm_term=.018e6cf3b4b5