A national conversation is continuing to brew over whether Science, Technology, Engineering, and Math (STEM) skills should be emphasized in schools and universities over humanities and liberal arts education. Several recent studies have found that the perceived higher marketplace value of STEM skills has led to a drop in the percentage of students pursuing humanities college degrees. This trend has led to significant and sometimes heated debate over whether STEM is inherently more valuable than the humanities, both at the individual level in terms of career prospects, and for society overall.
This debate is particularly pertinent given the growing importance of data in nearly all aspects of business and life. As IBM CEO Virginia Rometty told graduates at Northwestern University’s commencement ceremony this month, data is a “new natural resource” that will be just as vital to mankind in the 21st century as hydrocarbons and steam were in past centuries. The increasing importance of data would suggest that STEM skills such as statistics and computer programming are more important than ever, particularly in analytics, but do humanities play an important role in understanding data as well?
In my view, STEM and humanities are not mutually exclusive, and in fact, the combination of both STEM and humanities skills can lead to the most powerful and effective use of data. For people interested in analytics, the combination of these skills is not only helpful, but crucial to understanding data and meaningfully communicating business results.
My field of interest, HR analytics, is defined by the effective combination of both STEM skills and an understanding of the humanities. STEM, including the ability to use statistical software and data visualization technology, enables HR analysts to evaluate the statistical effects of employment practices, draw conclusions about the workforce using data, and build predictive equations to forecast future human capital trends. Likewise, a humanities-centered understanding of history, HR laws and regulations, and a deep knowledge of how society and the labor force interact are needed to determine if statistical results make logical sense, and to communicate findings in an understandable manner and a proper context.
As a prime example, HR analytics can be very helpful in advancing equal employment opportunity and workplace diversity. One way to study the pay gap between men and women in a workforce is to use multiple linear regression, which produces an equation that evaluates how different demographic variables, including gender, combine to affect employees’ pay. Such a statistical model can help identify if gender is a statistically significant predictor of pay above the effects of experience, education and other merit-based factors. But a regression equation alone does not fully explain why a certain result may be observed, or if the observed result makes sense. Understanding the history of how women have sought equal opportunity in the workplace, societal causes of the pay gap, and the laws that have been enacted to advance pay equity, is essential to interpreting the results of such a statistical model, and to developing better solutions to address the pay gap.
Another growing area of analytics that requires both STEM and humanities expertise is analysis of unstructured data, which generally speaking, is data that cannot be easily counted and sorted into spreadsheets. Examples of unstructured data in the workforce include employee reviews of a company on websites such as Glassdoor.com, and written narratives in employee performance evaluations and surveys. Using text analytics on free response survey questions can reveal frequently repeated patterns in employee sentiment, which can help identify workplace issues that may have otherwise been unknown.
However, a major challenge of text analytics software packages is that the resource dictionaries used to identify patterns are often limited, in that they sometimes fail to identify synonyms or conditional logic (where the meaning of a word is dependent on a preceding or subsequent word). This gap is where humanities expertise, in particular grammar and writing skills, becomes crucial. One can edit a resource dictionary to better group synonyms unique to an industry and more efficiently identify repeated phrases or terms. Likewise, grammar skills can be incorporated into a text analytics dictionary to help the software learn when the combination of certain words produces a unique meaning, and to better understand sarcasm and tone based on the surrounding words and phrases. In this way, writing and grammar skills can be combined with analytical skills to reach data-based conclusions.
Since analytics is such a widely employed discipline for many fields, and often incorporates both STEM and humanities skills, a major opportunity exists in our nation to combine these two skillsets. Some educational institutions are already seeking to move beyond the traditional STEM vs. humanities debate and meet this combined skill need. As just one example, Emory University now offers a unique dual major in math and political science that equips students with advanced knowledge of social science and history along with skills in predictive modeling, statistics and calculus. Combining humanities and STEM skills enables us to make full use of the new natural resource of data in the 21st century.
The opinions reflected in this column are solely those of the author.
Matthew Albucher is part of the GovLoop Featured Blogger program, where we feature blog posts by government voices from all across the country (and world!). To see more Featured Blogger posts, click here.