Historically, orchestras have suffered from gender imbalance and females have been underrepresented as musicians. In the 1950s, orchestras started dealing with this problem through blind auditions. The “audition curtain” was introduced and was able to effectively combat the gender bias. This simple innovation made it 50% more likely that a female musician would make it to the final round of the audition process. Female musicians always had the talent to perform in top orchestras, but this simple technology allowed them to broadcast their potential in an objective way.
Think of Knack as “the curtain” for companies and other organizations that aim to combat bias and embrace inclusive hiring. This allows organizations to focus on objective information that is evaluated autonomously as opposed to the subjective interpretations of subjective processes like interviews. This means that Knack algorithms eliminate bias associated with race and ethnicity, age, and gender and level the playing field for everyone.
Knack algorithms remove age, gender, and race/ethnic bias, enabling employers to identify talent in a fair and inclusive manner. Knack’s ability to avoid bias is by design. We rely on machine-intelligent analytics of objective information (game play data), sidestepping subjective interpretations of low quality information (eg,resume, recommendations, biographical information, interviews).
We construct our models to be robust against bias, and Knack models have had no adverse impact on older adults, women, or racial / ethnic minorities.
The effective pass rate of protected groups has been equivalent to that of non-protected groups, going well beyond the 80% threshold (EEOC 4/5th rule). Our effective elimination of bias is impressive, considering that bias against minority job seekers based on names alone can be as high as 50% (Bertrand & Mullainathan, 2004, American Economic Review).
For example, working with a Fortune 100 Company we ensured that there was no adverse impact for each protected group. The test we used was the industry standard “two standard deviation rule” using Fisher’s exact test. Protected groups included minority race / ethnicity, female gender, and age 40 years and over. No protected group had a standard deviation (SD) below -2.0 (the critical cut-off). In fact, none were below -1.8 and in many instances protected groups were relatively advantaged (a positive SD).
Take a look at the following diversity and inclusion analysis we conducted when working with a Fortune 100 company. The light green indicates that the Knack area meets the Meets the Equal Employment Opportunity Commission (EEOC) “Pass Rate” for bias mitigation. The dark green indicates that the Knack area exceeds the EEOC “Pass Rate” meaning protected groups are relatively advantaged.
Achieving equity of opportunity is not easy, but we are trying to break down barriers and help unlock potential across all walks of life!