How Occam’s Razor is Dulled: Seeing the Science For What It Is

Cathy Xiang reflects on the process of crushing "beautiful hypothesis" as discussed in the seminar given by Dr. Steve Nowicki.

As aspiring researchers, we often hear and are inspired by the successful stories in research. From hearing about the data that seemingly fits so perfectly with the hypothesis to learning about groundbreaking research findings, it becomes easy to get caught up in the glamorized image of research. But what happens to the hypotheses that fail to produce the desirable results? What happens when you discover that the logical and straightforward hypothesis that you have crafted for twenty years doesn’t align with your data? How do we ensure that our biases don’t cloud the science and the data that is in front of us?

Such critical questions were answered by Dr. Steve Nowicki in the seminar “How Occam’s Razor Is Sometimes Dulled.” Dr. Nowicki, a professor of Biology, Psychology & Neuroscience who studies evolutionary biology in the context of animal behavior, details the story of how his own “beautiful hypothesis” becomes crushed by the reality of the data.

“How do we ensure that our biases don’t cloud the science and the data that is in front of us?”

Many scientists, along with Dr. Nowicki, prescribe to Occam’s razor: the notion that the simplest explanation in science is typically the correct explanation. So, when Dr. Nowicki and his lab studied the mating habits and birdsongs, they gravitated towards the simple and intuitive development stress hypothesis. According to Dr. Nowicki, the hypothesis was developed on the basis that female songbirds choose their mate based on their songs. Additionally, early stress in birds impacts the birds’ abilities to learn songs so songs might be an indicator of a male songbird’s cognitive abilities. Thus, females prefer well learned songs because males that learn well have built better brains and are “smarter” in general. And, males that sing well should be smarter, overall.

This development stress hypothesis among songbirds was well-celebrated and well-accepted, with Dr. Nowicki and his lab receiving great praise for his findings.

However, as Dr. Nowicki continued his research, he found that the link between male songbirds’ song and cognitive ability was weak and when replicated in a natural setting, the original relationship was not found. In approaching challenges to his previous findings, Dr. Nowicki took the crucial and courageous step in applying for a grant to investigate his own intuitive hypothesis using the new findings that he had. In Dr. Nowicki’s final findings, he found that how well a male songbird learns to sing is not a reliable indicator of cognitive ability in general, rejecting his original hypothesis in a paper published in 2018 in Animal Behavior.

During the seminar, Dr. Nowicki brought up a key message in the importance of challenging old ideas, despite how intuitive and “beautiful” the concept can be. With the pressures of publishing and acquiring grant funding, it could have been easier to let the issue lie and continue with such a well-established hypothesis. But, in conducting research with integrity and objectivity, we take an important step forward in advancing the field and acknowledging that our mistakes can often guide our research towards the right direction.

We see the implications everywhere when we choose to believe what we want to rather than what is in front of us. From falsifying data to other acts of research misconduct for the sake of pristine results and glory, buying into the glamorized image that research is often painted in overshadows the true purpose of research to develop and present new knowledge in a field of study.

Just as we are inspired by the “beautiful” data that fits into the “beautiful” hypothesis, we can be inspired by the null hypothesis and the data that challenges scientists’ views on existing theories. The beauty of science and data is that no matter where our biases lie, the purpose of science is to eventually reveal the truth. And, in recognizing the negative results and even the statistically insignificant ones, we still take a vital step forward in understanding a more full picture of the field we conduct research in.

“The beauty of science and data is that no matter where our biases lie, the purpose of science is to eventually reveal the truth.”

Cathy Xiang, Huang Fellow ’25

Cathy XiangCathy is a rising sophomore from Plainsboro, New Jersey hoping to major in Biology and Global Health.