Ethnic diversity in STEM in the United States

Institute for Scientific Information Report

Diversity in the workplace has been found to have myriad of benefits, from accelerated innovation and discovery to better decision-making.

The Sustainable Development Goals from the UN stress the need for gender equality and reduced inequalities. As the scientific endeavor works towards solving these goals, it is particularly important that the workforce represents all countries, and that authors from diverse backgrounds are represented in scholarly publishing and journals too, in order to reduce inherent biases and enable a complete and global picture of the challenges we face.

In this ISI Insights paper, the authors combine self-identification data from the U.S. Census with the uniquely structured and curated Web of Science™ data to examine the issue of diversity in authorship of scientific publishing in the U.S.

Topics covered:

• defining a methodology for identifying the ethnicity of authors of STEM research articles
• identifying gaps in ethnic diversity in research
• discover the trends underpinning participation and inclusivity of authorship across disciplines
• understanding if and how the research landscape is changing, and whether there are changes in the levels of authorship of underrepresented minorities.

Research disciplines that were examined included Biochemistry, Computer Science, Mathematics, and Medical Research.

From the Introduction

In this Insights paper we sought to find a way to identify the ethnicity of authors of STEM research articles. The purpose was to not only be able to identify gaps in ethnic diversity in research but also to discover the trends underpinning participation and inclusivity of authorship across disciplines. In addition, we aimed to discover if and how the research landscape is changing, and whether there are changes in the levels of authorship of underrepresented minorities.

Top Level Findings

Most areas of study that we examined in the United States exhibit a sustained underrepresentation of specific minority groups, most notably Black Only and Hispanic authors. Moreover, the situation has changed little across the 10-year period we looked at. We see a need to develop mentorship, collaborations and partnerships that will encourage authorship from all underrepresented minorities. 

In all areas we looked at, publication rates from Asian/Pacific Islands Only authors are vastly higher than might be expected in the context of the U.S. population. For example, only 5% of the U.S. population self-identify with this category, but in the field of computer science – artificial intelligence they represent 33% of authorships, more than six times the greater population. This is an upward trending fraction in that field, where Asian/Pacific Islands Only authorship appears to have exceeded that of White Only (64% U.S. population) since 2017. No other minority group has a comparable upward trend. 

Of the fields we studied, medical research has the highest representation of Black Only authors (5.5%), but this is still less than half of the same group’s representation within the U.S. population of 12%. 

Math is dominated by White authorship but sees a steady growth of Asian/ Pacific Islands Only authors. However, there are low authorship levels of Black and Hispanic authorship in the field – and hardly any growth. 

Black Only and Hispanic authors do not show significant growth or decrease, and participation is mostly low and stagnant. 

Native American/Alaska Native authorship is very low across the studied disciplines and in some cases, such as computer science, sees decrease in participation.