Ben Schwencke, Business Psychologist, Test Partnership
How Accent Bias in Employment Interviews Worsens Social Inequality
Traditional employment interviews are considered the gold standard employee selection tool for the vast majority of hiring managers. Indeed, in the minds of many, it’s the only employee selection tool, with many screening processes relying solely on rounds of interviews.
Perplexingly, whenever I suggest alternatives or supplements to interviews—i.e., ability tests, personality questionnaires, situational judgment tests, etc.—I am met with knee-jerk skepticism. In particular, high levels of evidence regarding fairness and adverse impact are required to even begin the conversation.
But the fairness of interviews is never questioned or queried, despite a near century of evidence showing how biased and unfair they can be. People from disadvantaged groups, particularly neurodivergent people, are especially discriminated against in interviews, but organizations almost never raise this as a concern.
Another area of disadvantage caused by interviews is the issue of “accent bias,” discrimination that people face purely due to their accent. Research suggests that it is a major cause of bias in employment interviews, and that effect compounds massively with other characteristics.
What Is Accent Bias and How Does It Affect Interviews?
Accent bias in interviews occurs when people with non-standard accents (regional accents, foreign accents, etc.) are rated lower in interviews than those speaking standard accents. For example, those who speak General American would be more likely to receive higher ratings in interviews than those speaking Southern American.
You would think that something so trivial wouldn’t negatively affect someone’s likelihood of being hired, but the research suggests that accent bias can have a profound effect on someone’s chances. Overall, those with non-standard accents score around half a standard deviation lower on interview ratings than those with standard accents, which is considered a “moderate” effect (reference).
Although a difference of around half a standard deviation is considered moderate in absolute terms, in a practical hiring situation the effect can be very pronounced. For example, if an applicant with a non-standard accent is competing with 10 other candidates, the odds of them being chosen are extremely low, as selection ratio exacerbates the severity of adverse impact.
This effect is almost certainly entirely unconscious, as few people would discriminate actively based solely on someone’s accent. However, this effect is exacerbated by other variables, particularly demographics, suggesting an intersectional issue.
Who Experiences Accent Bias the Most?
From the research, it appears that women are by far the most likely to experience accent bias, with men seeing comparatively little discrimination (reference). Consequently, women with non-standard accents see an adverse impact of 0.83, almost a full standard deviation difference. This is considered a strong effect statistically and acts as a serious barrier to entry.
Moreover, the type of accent seems to matter too. Those with “racial/ethnic” accents saw the most discrimination, followed by “regional” accents, followed by those with “foreign” accents. In a U.S. hiring setting, a “racial/ethnic” accent could represent an Indian or African accent, “regional” could be Southern American, and “foreign” could represent German or French. This suggests an intersectionality between accent type and ethnicity, as well as one’s region. This likely compounds further with the impact of gender, making it especially hard for women with “racial/ethnic” accents.
Although the effect was statistically non-significant due to small sample sizes, it seems that audio-only interviews exacerbate the effect of accent bias, showing twice the accent bias of visual + audio interviews. This makes logical sense, as interviewers only have verbal communication available to them, making it harder to ignore the accent.
What Can We Do to Address This?
We know from the research that implicit bias training is not very effective, and is unlikely to correct the issue. Consequently, interventions should focus on minimizing the impact of interviews in general, rather than attempting to overcome the bias itself.
For example, organizations should diversify their selection processes, incorporating tools other than interviews in the selection process. Psychometric assessments, ability tests, knowledge tests, situational judgment tests, assessment centers, and work sample tests are all viable additions, reducing reliance on interviews.
By incorporating additional screening tools, especially during the early stages of the process, you limit the potential negative impact of interviews. Additionally, because these tools show incremental validity over and above interviews, you can expect a significantly greater quality of hire than an interview-only screening process.
Naturally, dropping interviews entirely is an unlikely outcome, given how popular employment interviews are as a screening tool. However, if other assessments do the bulk of the screening, then interviews are reserved for the final stage of the process as a final quality control measure, rather than a major obstacle for candidates.
Conclusions and Final Thoughts
It’s hard to imagine how something as trivial as someone’s accent could negatively impact their chance of employment, but unfortunately, this is a reality. Employers’ faith in interviews tends to be near absolute, and anything the candidate does that results in a negative perception is likely to hamper their chances. Apparently, this includes having the wrong accent.
Moreover, there seems to be a great deal of intersectionality, both for gender and ethnicity. Already, we can see a compounding problem here, with people from disadvantaged groups seeing even more unnecessary barriers put in place.
The other side of this issue which must be considered is the smooth ride that standard accent speakers enjoy. Generally speaking, those who speak the standard accent variant are more likely to be from majority and privileged groups, adding to unearned privilege. This likely results in a vicious cycle, whereby those with standard accents enjoy greater access to employment and all the advantages that entails, further segmenting that accent as the prestige variant.
Ultimately, organizations need to seriously reevaluate the role that interviews play in the selection process, especially if the entire selection process consists of interviews. Instead, a myriad of additional selection tools, many of which are tremendously complementary to interviews, could be considered in conjunction. This insulates the selection process from the impact of bias, giving candidates a fairer chance of success.
Ben Schwencke, Business Psychologist, Test Partnership