Solutions
All Resources
Discover "Me" · Work Personality
.webp?width=559&height=292&name=2026.07.21%20Fireside%20KV%20-%201200x627%20(event%20featured).webp)
THE AI COACH THAT ACTUALLY GETS YOU.
Voice or text coaching built on psychology. For you, your team, or the candidates you place.
Hey Compono!
A coach that actually gets you.
Get 10 minutes free, then $15 a month. Cancel anytime.
Get Started ≫.png?width=383&height=200&name=team%20(1).png)
Gender decoder for job ads
Paste a job ad and see which way its language leans. The decoder checks every word against the masculine-coded and feminine-coded lists from peer-reviewed research, shows you the exact terms it found, and reports the lean. Your ad never leaves your browser.
Your job ad
What the decoder found
Every term below comes from the published Gaucher, Friesen and Kay (2011) word lists. Counts include repeats, and matching includes word endings.
Masculine-coded terms
Feminine-coded terms
Coded language is one of eight ways a job ad leaks candidates
Roast My Job Ad runs this same gender check plus seven more: salary transparency, clichés, vague duties, unicorn hunting, age-coded phrasing, readability and the candidate blind spot. Fair warning: the roast is blunt.
A gender decoder is a free tool that checks a job ad for gender-coded language: words that research links to gendered patterns in who applies. Paste an ad and it counts the masculine-coded terms (like competitive and determined) and the feminine-coded terms (like supportive and collaborative), then reports which way the ad leans overall. The lean is an advisory signal from peer-reviewed research, not a compliance verdict.
How the decoder works
The decoder compares your ad against the masculine-coded and feminine-coded word lists published in Gaucher, Friesen and Kay (2011), the peer-reviewed study behind most gender decoder tools. Matching works on word stems, so compet* catches competitive, competing and competition. The two counts are then compared, and the difference sets the verdict: equal counts read as balanced, a gap of one or two terms is a subtle lean, three to five is coded, and six or more is strongly coded.
The verdict is an advisory signal from published research, not a legal finding, and no word on either list is banned language. The point is to catch a strong lean you did not intend. Everything runs in your browser. Your ad is never uploaded, stored or seen by us.
New to the term? Read the plain-English definition of pay equity in the HR Glossary.
The word lists
These are the published stems the decoder matches, exactly as they appear in the research. The asterisk means any ending counts: compet* catches competitive, competing and competition.
Masculine-coded
Feminine-coded
Source: Gaucher, D., Friesen, J. and Kay, A. C. (2011). Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality. Journal of Personality and Social Psychology, 101(1), 109–128.
Want the whole ad pulled apart, not just the gender lean? Roast My Job Ad runs this check plus seven more.
- The lists are the ones published in the 2011 study, unchanged. Wording that codes in ways the study did not measure (Australian idiom, industry jargon, job-board boilerplate, newer template language) is invisible to the decoder, so a balanced verdict is not a clean bill of health.
- Stem matching over-catches. A word used in an unrelated sense (a "competency framework", a "support ticket") counts the same as the coded use. Read the matched terms and discount the ones that are not doing what the research describes.
- The lean thresholds are conventions carried over from the widely used decoder tools, not from the paper. The study measured the effect of the wording; it did not publish a scoring scale.
- The verdict is advisory. No word on either list is banned language, and a masculine-coded ad is not a discriminatory one. Explicitly discriminatory wording (a "salesman", an age limit) is a separate legal question the decoder does not check.
- The research found the effect on how appealing a role seemed to women and whether they felt they would belong, not on whether they believed they could do the job, and feminine-coded wording had much less effect on men. A masculine lean and a feminine lean are not equal risks.
- Gaucher, Friesen and Kay, "Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality", Journal of Personality and Social Psychology 101(1) (the word lists and the effect on perceived appeal and belonging), 2011
- Fair Work Ombudsman, Protection from discrimination at work (where the legal line sits, which the decoder does not test), accessed August 2026
What to do about it
A strongly coded ad narrows the applicant pool before the first CV arrives, and nobody on the hiring team finds out who scrolled past. The fix is smaller than a rewrite and bigger than a word swap. These steps are for whoever drafts and posts the ad.
Sort the matched terms into two piles
Words that describe the work, and words that arrived with the template. Keep the first pile even if it leans; swap the second for the plain description of what the person will do. Balance is the goal, not a flip to the other list.
Rewrite the requirements list before the rest
That is where the agentic filler clusters (driven, competitive, a proven leader, dominant in the market) and where a candidate decides whether the ad is talking to them. Requirements written as things the person will do rather than things they must be tend to decode cleaner on their own.
Fix the template as well as the ad
The lean came from wherever the last ten ads came from. Change the master, add the decoder to the posting checklist, and the next hiring manager starts from a balanced page.
Track the applicant mix over the next few ads
If your ATS records gender where candidates choose to share it, compare the share of women applying for the same role before and after the rewrite. Two or three ads is a signal, not proof, but it is your data rather than a 2011 study's.
Agree what the role needs before the next ad is drafted
Most coded language is a stand-in for a definition nobody wrote: what the work is like, how fast it moves, how the team makes decisions, how much autonomy there is. Write that down and the ad describes a real job instead of a template's idea of one.
Decode the ad. Then
define the real job.
Compono Hire is an applicant tracking system (ATS) that measures fit before you make the offer, starting with a definition of what the role actually needs and matching every candidate against it on work personality and culture fit. Most ATSs start at the ad and manage what follows; the definition behind the ad is left to whoever wrote it.
Step 5 asks you to write that definition down. Hire makes it the first step in the process, so the ad describes the real work, the coded filler has nothing to fill, and the people who apply are measured against the same definition the ad came from.
Common questions
Where do the word lists come from?
From Gaucher, Friesen and Kay (2011), "Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality", published in the Journal of Personality and Social Psychology. The researchers compiled masculine-themed and feminine-themed word lists from the psychological literature, then showed two things: job ads in male-dominated industries used more of the masculine wording, and the wording itself changed how appealing a role seemed. Nearly every gender decoder online, including this one, uses those published lists.
Does gender-coded language actually change who applies?
In the Gaucher, Friesen and Kay experiments, women rated masculine-worded ads as less appealing and reported a weaker sense that they would belong in the role. The mechanism is the interesting part: the wording did not change whether women believed they could do the job, it changed whether the workplace felt like somewhere they would fit. Feminine-coded wording had much less effect on men. A masculine lean quietly shrinks one side of the applicant pool while leaving the other side untouched.
Is masculine-coded language in a job ad illegal?
No. Words like driven or determined are ordinary language, and using them is not discrimination. That is a different question from explicitly discriminatory wording, such as advertising for a "salesman" or setting an age requirement, which can breach discrimination law in most jurisdictions. The decoder deals only in the subtle lean, and its output is general information, not legal advice.
Should I rewrite every job ad to be feminine-coded?
No. The goal is balance and accuracy, not flipping the lean. If the work genuinely is competitive, say so. The problem the research points at is unintended lean: template language that codes heavily one way without describing anything real about the job. Keep the coded words that carry information and swap the ones that are filler.
How is this different from Roast My Job Ad?
The Gender Decoder does one thing deeply: it decodes gendered language and shows you every matched term. Roast My Job Ad runs the same gender check as one of eight, alongside salary transparency, clichés, vague duties, unicorn hunting, age-coded phrasing, readability and the candidate blind spot, and delivers the verdict as a roast. Decode here when language lean is the question. Roast there when you want the whole ad pulled apart. Both are free and both run entirely in your browser.
Next step
Build the business case
Turn this number into a document your CFO can question line by line.
Open the builder →Related toolRoast my job ad
Keep the same numbers moving. The next calculator picks up where this one stops.
Run it next →HR GlossaryPay equity
The plain-English definition, with the calculators and rules that sit around it.
Read the definition →Talk to usBring us your number
A 30-minute conversation about what it means for you, not a demo script.
Talk to us →
