Many chemistry departments have attempted to improve the representation of marginalized scientists in their schools and courses—for example, tackling “dude walls,” which honor historical or departmental science figures who are predominantly White men, and weed-out foundation courses, which are those with large enrollment and high fail rates that often disproportionately fail students who aren’t White males. But the science textbooks that thousands of students read each year still lag: White men predominate in the images and references to historical science figures. Many publishers are starting to implement changes to improve racial and gender diversity within their pages, but progress is slow.
When Melanie Nilsson, a professor of chemistry at McDaniel College, was looking through a college chemistry textbook, one image in particular stood out: a photograph of a Black athlete that accompanied a section on sports-related chemicals. Nilsson wondered how many individuals from marginalized groups are in the average chemistry textbook and how many of those individuals are depicted as scientists.
Nilsson and Mona Becker, a physics and chemistry instructor at Westminster High School, decided to quantify representation in US college chemistry textbooks. They thought that hard numbers on images and text citations would be most effective in persuading scientists to understand the scope of the problem and make changes. “We want [our findings] to speak directly to scientists,” Nilsson says.
Because ascertaining race seemed prone to errors, Becker and Nilsson focused on binary-gender representation: in 10 chemistry textbooks with editions published between 2016 and 2020, they categorized as male or female the images of people and the names cited in the indexes (J. Chem. Educ. 2021, DOI: 10.1021/acs.jchemed.0c01037).
The average percentage of images that contained only people the authors categorized as men (65%) was significantly higher than the percentage that contained only women (24%); the remainder (11%) contained men and women. In the text, stereotypically male names appeared on average 60 times as often as stereotypically female names. “I think the story line, if you look through these thousands of pages of text, is that women are undervalued. And that males, particularly White males, are the authority,” Nilsson says.
For four of the textbooks, Becker and Nilsson also compared gender representation in the first and most recent editions. “I still shake my head sometimes in disbelief that one of the textbooks actually decreased the percent of women in it from their first edition to their most recent edition,” Becker says. In another textbook they examined, the percentage of references to female scientists was unchanged.
Cissy Ballen, a professor of biological sciences at Auburn University, examined race and gender in a similar study of biology textbooks (Proc. R. Soc. B 2020, DOI: 10.1098/rspb.2020.0877). In seven widely used biology textbooks—popularity evidenced by books running to many editions—145 female scientists were mentioned, compared with 962 male scientists.
For the most recent textbook editions, Ballen’s team divided the references to historical scientists into the decades they worked, reasoning that the absence of scientists other than White men may reflect historical barriers to access within the sciences. But even when the pool of scientists in history became more diverse, people of color still weren’t included. Although Ballen saw some increase in the number of women mentioned from the 1970s to the present decade, “if you look at the intersection of race and gender, you really just see a large increase in [references to] White women over time,” she says.
Textbooks featured few scientists of color regardless of historical period discussed: of the references to research published between 1900 and 1999, just 3% included work by scientists of color; that number increased to only 8% for research published between 2000 and 2018, Ballen found.
The damage caused by these biases in textbooks is insidious. Students form a science identity—a perception of themselves as scientists—in part through the presence of role models who resemble them (Basic Appl. Soc. Psychol. 2016, DOI: 10.1080/01973533.2016.1209757; Front. Psychol. 2020, DOI: 10.3389/fpsyg.2020.02204). The absence of appropriate science role models in the classroom hinders the formation of science identities.
Images in textbooks can also have a short-term detrimental effect on academic performance. In a study, female students performed better on tests of their understanding of chemistry lessons when those lessons included pictures of women than when the lessons included pictures of men; the reverse was true for male students. (J. Soc. Psychol. 2010, DOI: 10.1080/00224540903366552). Textbooks that perpetuate the stereotype that women and other marginalized groups don’t belong in science could harm students’ academic performance and drive them from the science classroom and therefore careers in science.
Many publishers and textbook authors have read the studies on textbooks and are implementing policies they hope will address the problems. But people disagree on who ultimately bears the responsibility for reducing biases: Is it the author, publisher, peer reviewers, or consumers who choose which textbooks to use?
When it comes to deciding which scientists to cite or profile in a textbook, “we typically leave it to the authors,” says Anthony Palmiotto, editorial director at OpenStax, a creator of open-source textbooks. That means diversity, equity, and inclusion (DEI) issues can be flagged during the editing and peer review process but aren’t set out in guidelines for the author. Becker and Nilsson, who have both reviewed textbook chapters, say some other publishers don’t explicitly ask reviewers to comment on representation in the text and images.
Other textbook publishers, such as Cengage and Macmillan Learning, provide training for authors and outline their DEI guidelines at the beginning of the writing project. “We’ve realized it needed to start way back at the inception of the idea,” says Susan Winslow, president of Macmillan Learning.
In addition to the technical experts reviewing the textbook chapters, OpenStax has a diversity review panel featuring academics from the diversity committees of science organizations such as the American Chemical Society and the American Physical Society (ACS publishes C&EN). Other publishers have hired DEI consultants to assist with the development of internal guidelines and training and respond to DEI issues flagged within the textbooks.
Several publishers have renegotiated with their vendors to obtain stock images that are more representative of gender and racial diversity among the student population. Winslow sees it as a simple issue of supply and demand: if publishers ask vendors for more-diverse images, the vendors will provide them.
The most recent editions of several science textbooks offer better gender, racial, and disability representation, with introductory textbooks for nonmajors, such as What Is Life? (published by Macmillan Learning) and Chemistry in Context (written by ACS and published by McGraw Hill), showing the best representation in studies by Nilsson and Becker and by Ballen. In the same studies, more specialized chemistry textbooks do not show as much improvement.
In one organic chemistry textbook published in 2018 by Cengage, a reference to the chemical composition of sunscreen was accompanied by the Coppertone sunscreen logo: an image of a dog pulling off the swimsuit bottoms of a topless little girl. “Based on our study, this textbook contained 35 total depictions of females and 8 images that contained only females,” Nilsson says. “What message are we sending when that’s one of the few females in the entire textbook?”
C&EN asked Cengage about the Coppertone logo before the interview. Thais Alencar, vice president of product management and learning experiences at Cengage, tells C&EN the company conducted an internal review and decided to immediately remove the image from the digital textbook and remove it from the next paper edition. “There’s no pedagogical reason for it to exist there,” Alencar says. All the publishers interviewed for this story accept feedback from students and instructors and encourage readers to flag issues with textbook content.
Fixing several decades’ worth of underrepresentation will take more than simply adding photos that show more women and people of color, according to Daniel Domin, director of assessment, evaluation, and achievement at Dominican University.
Publishers may think that adding any pictures of people of color to science textbooks will help students of color identify as scientists, Domin says. “But when you start doing the research, you begin to realize some images are just laden with bias.” Domin says a careless attempt to rectify bias could create greater biases.
As publishing houses continue to try to improve representation in their textbooks, they should examine the kinds of images they use and the way they present information. In a 2007 study, Domin found that too often, images of people of color perpetuated biases (J. Chem. Educ. 2007, DOI: 10.1021/ed084p342). One type of bias was fragmentation bias, in which contributions from prominent scientists in a marginalized group are presented in boxes outside the main text, othering their achievements as separate from the main scientific narrative. Other biases included stereotypes, such as depicting Black people as athletes instead of scientists, and tokenism—making only the minimum effort to improve representation.
Although science publishers have established diversity committees and amended author and review guidelines to improve the gender and racial representation in their textbooks, research indicates that any improvement is incremental. Ballen’s team estimated that if textbook citations of Black scientists continue to increase at the same rate, it would take nearly 500 years for the percentage of Black scientists in biology textbooks to reflect the current percentage among biology students (7.7%) in the US. The percentage of Asian scientists mentioned in biology textbooks reflects Asian populations within the US (6.0%) but will not reflect their makeup of today’s biology students (15.2%) for approximately 50 years. “It’s 2021! Come on, you know we can do better than this,” Westminster High School’s Becker says.
Claire L. Jarvis is a freelance writer based in Atlanta.