Bad news, older tech workers: Job advert language works against you
And there's a study now to prove it – science!
Economists have crafted online fake job ads and found that those incorporating age-related stereotypes discourage older workers from applying – another cause for concern among tech workers as the industry faces multiple lawsuits over ageism.
The research – conducted by Ian Burn (University of Liverpool, UK), Daniel Firoozi, Daniel Ladd and David Neumark (UC Irvine, US) – is described in a paper titled, "Help Really Wanted? The Impact of Age Stereotypes in Job Ads on Applications from Older Workers," distributed via America's National Bureau of Economic Research.
"We find strong evidence that ageist language related to communication skills, physical ability, and technology skills, even when it is not blatantly or specifically age-related, deters older workers from applying for jobs," the paper argues. "Job ads that feature ageist language attract younger applicants on average than job ads that do not feature ageist language."
The authors conclude that job-ad language has about the same effect as direct discrimination in hiring – such as rejecting job applications from anyone over a certain age.
However, the researchers observe that the effect of discouraging older people to apply for jobs doesn't show up in the analysis of corporate hiring practices – non-applicants don't get counted. This has implications for how companies write job ads and how policymakers assess discriminatory behavior.
An age-old tech problem
Claims of age discrimination are common among technology companies. IBM is alleged to have laid off as many as 100,000 employees from 2015 through 2019 as part of a global effort to de-age its workforce. Google paid $11 million in 2019 to resolve an age discrimination lawsuit brought on behalf of 227 people, and then faced another age discrimination claim two months later.
According to the American Association of Retired Persons (AARP), a nonprofit advocacy group, 78 percent of older workers say they've experienced discrimination in the workplace. That represents the highest level since the organization began tracking that data in 2003.
In a phone interview with The Register, David Neumark, professor of economics at UC Irvine, co-director of the Center for Population, Inequality, and Policy, and a co-author of the paper, explained that this study flips a previous study [PDF] from 2019 on its head.
In that project, Neumark, Burn, and Patrick Button (Tulane, RAND Corporation) sent out 40,000 job applications and found "robust evidence of age discrimination in hiring against older women" in terms of job applicants who received responses.
- Intel laid me off for being too old, engineer claims in lawsuit
- Google age discrimination case: Supervisor called me 'grandpa', engineer claims
- HP Inc, HPE both slapped with racism, ageism lawsuit
- How a Facebook post by blabbermouth daughter cost her parents $80,000
The latest research takes the opposite approach by creating fake jobs ads derived from actual job ad text and documenting the demographics of those who respond.
"The question this paper is about is why?" explained Neumark. "Why would an employer run an ad with ageist language? The most obvious reason would be that when you discriminate, you can get caught by people like me who do this kind of work, both as an academic and as a consultant on discrimination cases, by treating people disproportionately. If you can actually drive down the number of older applicants from the group you hire from, you don't have to do that."
The study does not examine actual employer behavior, despite its reliance on job-ad language from real employers. The fake job ads presented and evaluated were assembled with the help of a machine learning algorithm designed to generate phrases that are semantically similar to ageist stereotypes but aren't obvious.
"What our evidence does address is whether age stereotypes expressed in job ads affect the likelihood that older job seekers apply for jobs, likely by signaling to job applicants that older workers are less likely to be hired," the paper states.
Truth in data?
Neumark acknowledges that you cannot see intent in data.
"But we know three things," he said. "We know from the first study [in 2019] that [employers] actually discriminate – those were rigorous experimental measures of discrimination, because they see résumés that have nothing different about them except age, and they prefer the younger folks."
"The same employers who behave that way actually used more stereotyped language. And here we see that that language discourages older workers from applying."
It's possible to construct an explanation for these findings that don't involve discriminatory intent, Neumark says, but he cites the maxim of Occam's Razor to support his contention that they're related.
Neumark believes machine learning will prove useful for flagging discriminatory language in job ads by offering a neutral measurement mechanism. And others appear to be thinking along similar lines. Neumark said he presented his findings in a seminar in Brussels and was told that there's already a tool called the gender decoder being used to evaluate job ads for gender bias.
The paper argues that the research findings should encourage the US Equal Employment Opportunity Commission (EEOC) to expand its guidance on the wording of job ads.
"Using language that explicitly deters older workers from applying is already illegal under the ADEA, but the subtler usage of ageist language that we study suggests that job-ad language that would not be flagged as explicitly illegal can still have pernicious effects on older workers in the labor market, and possibly facilitate age discrimination," the paper concludes. ®