Imagine this: you’re creating a job ad for your company. You’ve been swapping out words like “competitive” and “dominant” with “supportive” and “collaborative”. You’ve read that reducing gendered language in job ads is more inclusive. Similarly, you feel confident that this change will attract a more diverse pool of candidates. But, what if those changes don’t make a difference?
For years, hiring managers have been told that language in job ads can subtly signal who belongs and who doesn’t. Studies, like the influential 2011 research1 by Gaucher et al., suggested that masculine-coded words, such as “assertive” or “dominant” discouraged women from applying for roles. This has led to the rise of tools designed to remove bias from job descriptions.
But is this advice as reliable as we believe?
Recent research tells us otherwise. A study by Castilla & Rho (2023)2 challenges this narrative. It suggests that gendered language in job ads has little impact on hiring in real world settings. These findings could completely change how we think about gendered language.
The research that started it all
In 2011, researchers at the University of Waterloo (Gaucher et al.) asked university students to evaluate job ads written in either masculine or feminine-coded language. Researchers then asked about their feelings of belongingness, and their likelihood of applying for the jobs.
The results showed that masculine-coded terms, such as “assertive” or “dominant”, made women less likely to feel belonging in a workspace. As a result, women felt less likely to consider applying for those roles.
This groundbreaking study led to one clear implication: the words in job ads influence who applies. This study spread like wildfire. Hiring managers redesigned their job descriptions. Tech companies built tools to identify and remove gendered language. I’ve made these recommendations too, before I dived into the research myself.
However, this study wasn’t perfect. It had limitations:
- The study lacked diversity – The participants were all university students from the same university, not real job seekers.
- The scenario was artificial – The experiments were performed in a lab, not using real data.
- The study measured intention only – They asked participants if they felt that they would apply for the job. They didn’t measure actual actions.
- The study didn’t account for online job platforms.
In essence, this study showed that behaviours could be influenced in a controlled setting. Unfortunately, we might have taken a leap, and assumed it works in the real world as well.
While Gaucher et al.’s findings led early conversations about gendered language, more recent research by Castilla & Rho paints a different picture.
Looking at the real world
In contrast, Castilla & Rho from MIT analysed real-life behaviour of job seekers. They took real data from a popular North American job board, including:
- 296,000 job ads posted by 32,000 recruiters
- 487,000 unique job seekers
- 2 years of data
Instead of just asking people how the job ads made them feel, they analysed real behaviour. Did job seekers click on, inquire about or apply for jobs with masculine or feminine coded language? Did the recruiter’s gender (displayed prominently on the job ad) influence these decisions?
The findings? In the real world, language made little to no difference in whether women applied for a job. The researchers said:
Contrary to the currently popular recommendation, organizational efforts to tweak the language in job postings may not be the most fruitful avenue for organizations to
diversify the gender of their applicant pools.
Revisiting the original experiment
Not entirely satisfied with their analysis, the MIT researchers created a follow up experiment.
They followed a similar approach to the original study. They created three different versions of the same job advertisement. One masculine, one feminine, and one completely neutral. To keep the experiment as fair as possible, the job ad was completely the same except seven different words.
Our research team at [OUR SCHOOL NAME] is looking for a [FIRST WORD] and [SECOND WORD] candidate willing to gain experience as a research assistant. An ideal candidate should be [THIRD WORD] and care about the success of our projects.
They then replaced the blanks with gendered words. For the neutral version, they didn’t add any words at all.
| Feminine Wording | Masculine Wording | Neutral Wording | |
| First word | Committed | Determined | No word |
| Second word | Supportive | Assertive | No word |
| Third word | Sensitive | Outspoken | No word |
They then created a real job ad which had almost 1,500 applications. Each of the applicants were randomly assigned to one of these three options. The researchers then measured how many people applied for the job.
Like before, they found that changing the gendered words had very little impact on real world job applications. To conclude their study, Castilla & Rho wrote:
Our research, however, shows that differences in the masculinity/femininity of the ad language or the gender of the recruiter do not matter much in the real world, particularly in online job postings in the United States.
What does this all mean?
The 2011 study showed that in an artificial setting, changing gendered language has an effect. A similar study in 2023, aimed at real job applicants showed only a very minor effect. When analysing historic large-scale data from hiring platforms, it doesn’t have an effect.
So, should you stop caring about gendered language in your job ads? Not quite. What this research has shown is tweaking an existing job ad to be more inclusive is ineffective. As we’ll see in the next study, a completely different approach can be successful.
Improving the use of gendered language in job ads
In 2021, the British telecoms giant Openreach3 committed to recruit a minimum of 20% of females into entry-level engineering roles. This goal was 10 times higher than historic levels. They were aware that engineering roles are considered masculine roles, but when they presented a gender-inclusive job ad, they saw women’s interest increase by more than 200%. This is significantly higher than what Gaucher and Castilla & Rho saw in their studies.
Despite 80% of women saying that they wouldn’t be interested in an engineering role, 56% were interested when the role was reimagined, and the word ‘engineer’ was removed.
So what led to this impressive change in behaviour? The job description was designed to be inclusive from the ground-up.
Let’s find out what they changed.
1) They renamed the role
They renamed “Openreach Engineer – Trainee” to “Trainee Network Coordinator”. This eliminated the traditional associations of “engineer” with a male-dominated role. The new title felt more approachable and inclusive, making it easier for women to imagine themselves in the position.
2) They focused on human-centric descriptions
The job description emphasised the human impact of the role “keeping families in touch, making sure businesses continue working, helping kids learn and play”. These statements resonated with values like connection and community, which research shows are more appealing to women.
3) They highlighted personal traits over technical skills
Rather than focusing on rigid technical qualifications, the job requirements were reimagined to emphasise personal traits and transferable skills that candidates could easily identify with.
Instead of requiring prior experience or specific technical skills, the advert described qualities:
- Multitasking and time management: “I’m great at multitasking and managing my time”.
- Reliability and problem-solving: “I get things done, and people know they can rely on me”.
- Curiosity and a willingness to learn: “I like finding out how things work and learning how to fix them if they’re broken”.
4) They avoided requirements which sounded overly masculine
The original advert leaned heavily on masculine-coded language, particularly around physical demands. For example, it stated:
“I’d love a job that keeps me fit and active, and I don’t mind if it sometimes involves getting a bit dirty.”
While this won’t significantly shift perceptions by itself, it reinforced the broader inclusivity efforts.
5) They removed unnecessary barriers
The ad explicitly stated that “no qualifications or previous experience are needed to apply.” By eliminating intimidating language, Openreach reduced the likelihood of self-selection out of the hiring process.
Self-selection is the phenomenon where candidates opt out of applying because they feel unqualified. This is a common challenge for women who feel they must meet all the listed criteria before applying.
6) They emphasised training and support
The ad reassured candidates that they would receive comprehensive training and ongoing support.
“You will get all the training you need from us, with a manager and specialists at the end of the phone to support you throughout.”
7) They explicitly stated that both men and women worked this role
“Both men and women, from all walks of life, work in this role […]. We are really looking for people who feel they might have the skills and attitudes above so, even if you are a little uncertain, we would love to hear from you and help you find out more.”
This reassured women that they wouldn’t be the first or only woman in this role. By explicitly highlighting gender diversity in their workforce, Openreach addressed the fear of isolation in a male-dominated role.
Conclusion
The evidence is clear: tweaking gendered language in isolation won’t revolutionise your hiring practices. While small changes like swapping “assertive” for “supportive” may show promise in controlled settings, the real world is much more complex. Salary transparency, role design and inclusive messaging can be far more impactful.
What’s the lesson here? Inclusivity isn’t a last minute edit, it needs to be the foundation. Openreach’s success shows us that reimagining job ads from the ground up – emphasising values, potential and accessibility, can have a far more significant impact.
Interested in learning more about ethical hiring? See our other articles!
References
- Gaucher, Danielle, Justin Friesen, and Aaron C. Key. “Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality.” Journal of Personality and Social Psychology 101.1 (2011): 109-28. ↩︎
- Emilio J. Castilla, Hye Jin Rho (2023) “The Gendering of Job Postings in the Online Recruitment Process“.
Management Science 69(11):6912-6939. https://doi.org/10.1287/mnsc.2023.4674
↩︎ - New research reveals hidden bias in job adverts deters 50% of female applicants for engineering roles, Openreach https://www.openreach.com/news/new-research-reveals-hidden-bias-in-job-adverts-deters-50-of-female-applicants-for-roles-at-openreach ↩︎
