Who’s soft? Social sciences are in fact scientific

One day last semester, I was sitting at Hillman Library with some friends, looking over the requirements to apply for the Graduate Research Fellowship Program through the National Science Foundation (NSF).

While I was at the table, one of the people sitting with me asked me what I was working on. As an ecomomics major, I informed her that I was applying for a NSF grant to pursue graduate study in economics. She replied, “but it’s the National ‘Science’ Foundation.”

Unfortunately, the attitude that the social sciences don’t constitute legitimate scientific work is widespread, even in an academic setting. This disdain has manifested itself in academia through the usage of the term “hard science,” to describe the natural sciences, whereas the social sciences are merely referred to as “soft sciences.”

Social science departments are really not helping the matter. At Pitt, students can pursue a Bachelor of Arts in Anthropology, Economics, Linguistics, Political Science and Sociology — all of which are considered social sciences, broadly construed. 

Additionally, the claim that social sciences are somehow “less scientific” than their natural counterparts is naive and misguided. In particular, I will examine and refute two sentiments regarding the social sciences.

First, many students and academics believe that natural scientists have a much better understanding of the underlying mechanisms in their field of study. This is particularly pronounced in economics — following the financial crisis in 2007, even prominent economists were wondering how confident they should be in the field’s predictions.

Nevertheless, this attitude conflates macroeconomics with the discipline as a whole, which is understandable, as the news is usually dominated with coverage of macroeconomic fluctuations. However, this false identification misses the large strides made in other subfields of economics, such as applied microeconomics and behavioral economics.

Further, in addition to overstating the problems in the social sciences, this belief understates the problems in the natural sciences.

Take physics for example. Academics and intellectuals hold physics as the purest of the sciences, having a rich history of successfully explaining physical phenomena. Yet, quantum mechanics — a well-known subfield of the discipline — has incredibly important problems that have remained open for the last half-century.

Additionally, the mathematical development of chaos theory in the 1960s and 1970s exposed our shallow understanding of seemingly simple phenomena in the natural world — such as turbulence in water flow or even weather patterns.

Second, the results in the natural sciences are generally seen as more definite than those in the social sciences, as conflicting evidence seems to appear more often in the latter.

This is especially prominent in psychological research, as many of the findings from the field over the past couple decades are, in light of replication problems, being called into question by modern researchers. Take, for example, the open letter written by psychologist Daniel Kahneman warning of a “train wreck looming” over the “robustness of priming results,” referring to recent work on social priming. As the journal Nature reports, “this skepticism has been fed by failed attempts to replicate classic priming studies, increasing concerns about replicability in psychology more broadly.”

However, this “replication crisis” has been taking place throughout the sciences, including medical and biological science. In fact, the 2005 article “Why Most Published Research Findings Are False,” argues from a probabilistic standpoint that the push for novel results and lack of replication has grave implications for scientific work.

Further, it is actually the case that the social sciences are in a better position to tackle such statistical challenges. This stems from an increased emphasis on learning the intricacies of statistical tools.

This emphasis is immediately evident following an examination of the requirements for majoring in the natural and social sciences at Pitt. The mathematics-economics major, designed to prepare students for graduate study in economics, requires students to take an introductory class in statistical methods, as well as probability theory, mathematical statistics and econometrics, or statistics for economists.

On the other hand, biology majors are only required to take an introductory statistics course, while chemistry and physics students have the opportunity to take probability theory and mathematical statistics as electives. As students of the social sciences are better trained in statistical methods, it follows that they will be more capable of implementing them without running into statistical pitfalls.

Additionally, this emphasis is reflected at the professional level. The field of economics has a robust subfield, econometrics, dedicated to developing statistical techniques for economists to use in their work.

In fact, econometricians have actually developed tools that have been accepted into the statistical canon. You’d be hard pressed to find work like that being done in a chemistry department.

In light of this, it might be more accurate to say that the social sciences are “more scientific” than the natural sciences — whatever that means. Perhaps we should all simply be content recognizing scientific work as, well, scientific.

Write to Thomas at teh18@pitt.edu.