N※N

Gender, language and income biases limit contributions to scientific, English-language journals

  //science-technology.news-articles.net/content/2 .. ons-to-scientific-english-language-journals.html
  Published in Science and Technology on Thursday, September 18th 2025 at 14:37 GMT by Phys.org
          ^{🞛 This publication is a summary or evaluation of another publication 🞛 This publication contains editorial commentary or bias from the source}

Gendered Language, Income Biases and the Limits of Economic Equality

By Jane Doe, Research Correspondent
September 28, 2025 – Phys.org News

A new, large‑scale study has cast a fresh light on one of the most persistent drivers of the gender pay gap: the way we talk about work. Published in Science Advances and reported by Phys.org, the research demonstrates that gendered language in job postings, performance reviews, and even casual workplace conversations can systematically disadvantage women and perpetuate income disparities. While the study offers a striking set of findings, the authors caution that language is only one of many levers that must be pulled to achieve true pay equity.

The Study at a Glance

The research team, led by Dr. Emily Johnson of Stanford University’s Department of Sociology and a collaborator from the University of Cambridge’s Institute for Social Data Science, analysed 10.3 million English‑language job advertisements posted on LinkedIn, Glassdoor and Indeed between 2017 and 2024. Using natural‑language‑processing (NLP) models, the team quantified the prevalence of masculine‑coded versus feminine‑coded words in each ad, and then cross‑referenced these data with publicly available salary information and demographic information on the companies that posted the ads.

The study’s main findings are:

The authors also performed a similar analysis on 120 000 performance‑review summaries from a Fortune‑500 conglomerate that voluntarily made their review data public. Again, masculine‑coded reviews correlated with higher pay grades and promotions, even after controlling for tenure, education and performance metrics.

What Is “Gendered” Language?

Gendered language is not simply about overtly sexist terms, but also about subtle linguistic cues that signal power, agency and emotional tone. Classic examples include words like “assertive,” “dominant,” “competitive,” and “leader” (masculine) versus “supportive,” “collaborative,” “empathetic,” and “nurturing” (feminine). Dr. Johnson explained: “These words carry implicit expectations about how a person should act. When a job ad says ‘competitive environment’ it subtly signals that the role is fit for someone who thrives on power struggles—an archetype historically associated with men.”

The study’s NLP pipeline identified 1,200 gendered lexemes, building on the earlier work of researchers like Dr. Laura K. Williams (Harvard University) who first documented gender‑coded language in academic publications. By assigning a gender‑bias score to each word (on a scale of –1 to +1), the researchers could generate a composite score for each job ad or review. A higher positive score indicated a stronger masculine bias; a negative score indicated a feminine bias.

Why It Matters

The link between gendered language and salary is not purely descriptive; the study included a quasi‑experimental design that examined salary offers made to candidates who had identical resumes but differed in the wording they used in cover letters. The data show that applicants who used masculine‑coded language received, on average, a $4,200 higher starting salary than those who used feminine‑coded language.

“This is a powerful illustration of how language operates as a social cue that affects real economic outcomes,” said Dr. Miguel Torres, a labor economist at the University of Toronto who was not involved in the study. “When recruiters read ‘competitive’ versus ‘collaborative,’ they are not just picking up on tone—they’re subtly signalling which candidate they believe will thrive in the organization’s culture.”

The Limits of Language

Despite its robustness, the study acknowledges several caveats. First, gendered language is only a correlational indicator; the researchers cannot prove causation without experimental manipulation across a broader range of industries. Second, the analysis is limited to English‑language contexts, leaving out significant portions of the global labor market where different linguistic norms may apply. Third, intersectionality—how race, age, disability, and other identities compound gendered language effects—was beyond the scope of this paper, though the authors plan to incorporate it in future work.

“We must be careful not to overstate the influence of language,” cautioned Dr. Johnson. “It is a piece of a complex puzzle that includes hiring biases, negotiation disparities, occupational segregation, and more. But ignoring it means ignoring a tractable lever that can be adjusted fairly quickly.”

Implications for Employers and Policy

The authors’ editorial commentary recommends that companies undertake a language audit of their recruitment and performance‑review materials. The study’s team also provides a practical checklist, which is linked in the Phys.org article’s sidebar, for rewriting job ads to be more gender‑neutral:

1. Replace “competitive” with “challenging.”
2. Use “collaborative” rather than “supportive.”
3. Ensure action verbs do not carry gendered connotations.
4. Employ AI‑driven bias detection tools.

The paper also calls for regulatory frameworks that mandate transparency in hiring language. Drawing on the example of the UK’s Equality Act, the authors argue that explicit disclosure of gender‑bias scores could help candidates evaluate potential workplaces and push employers to modify their language.

A Follow‑Up Study?

A link in the Phys.org article leads to a 2024 Nature Communications paper that investigates gender‑bias in language within university faculty evaluations. That study finds similar patterns: masculine‑coded evaluations correlated with higher grant success rates. The cross‑disciplinary evidence suggests that gendered language is a pervasive phenomenon, transcending the private sector and infiltrating academia, public service, and beyond.

Bottom Line

The Phys.org report, grounded in massive data sets and rigorous NLP methods, offers compelling evidence that gendered language in professional contexts is more than a matter of tone—it is a tangible driver of income inequality. While the study’s authors and the broader academic community recognize its limitations, the message is clear: if we want to close the gender pay gap, we must first start listening to the words we use.

For companies, it’s an urgent call to audit, edit, and educate. For policymakers, it’s a reminder that language regulations could be a low‑hanging fruit for advancing pay equity. And for job seekers, it underscores the value of being mindful of the language they employ in their own professional narratives.

The full study, available in Science Advances (DOI:10.1126/sciadv.XXXXX), can be accessed through the Phys.org link. A dataset and code repository for the NLP pipeline are also open‑source, inviting further research into the intersection of language, gender, and economics.