The best selection processes are those that have articulated what characteristics make a great employee and have effective ways to evaluate those characteristics.
Even if you didn't care about fairness to the candidates themselves, structuring interviews and mitigating biases is proven to be the best way to choose the best candidates, rather than choosing the person who is good at interviewing or "looks the part".
Unconscious bias weighs most heavily when we don't have adequate structure or evaluation criteria. In addition to addressing those two issues, other straightforward steps can also be very helpful: finding diverse candidate pools and appointing diverse hiring boards.
Facts and Data
Orchestras in the 70s began adding way more women after conducting "blind" auditions. Researchers collected audition records from 8 major symphony orchestras that consisted of 14,121 person-rounds, 7,065 individuals and 588 audition-rounds. Top orchestras went from 6% to 21% female after implementing a new process in which musicians auditioned behind a curtain. The curtains ensured the judges could not be unconsciously biased by seeing the musician’s gender. According to Harvard's analysis, blind auditions explain 25% of the increase in women in orchestras between the 1970s and 1990s. [Harvard Kennedy School]
An analysis of medical school selection processes found that interviews provided no predictive value in student success. In the late 80s, Texas legislators required the UT Houston to increase the medical school class size from 150 to 200 — after they had already chosen their 150 students. This turned into an eye-opening field study that showed the value of their application process. The results? There turned out to be no performance difference between initially accepted and initially rejected students. Furthermore, the researchers found that about 75% of the difference in initial ratings between initially accepted and initially rejected students was due to interviewers’ perceptions of the candidates in unstructured interviews, not more objective measures [HBR]
Elite jobs often are awarded based on "cultural fit" not ability. A study in 2012 done of elite investment banking, consulting, and law firms found that their interview processes heavily relied on "cultural matching" (i.e. similarity to the interviewer) rather than ability. The researcher found that 75% of Law firms, 65% of Investment Banks, and 40% of consulting firm hirers used cultural matching as their top evaluative criteria [Kellogg Northwestern]
I was on an interview panel once for a job opening in our department. One applicant used only male pronouns when talking about ANYONE. All hypothetical employees were "he" and managers were "him," etc. It was awful. I cringed every time the applicant opened his mouth. At the end of the interview, the man who was on the panel with me, made a polite comment that our organization is very diverse and we can't blanket everyone with he/him pronouns. When the applicant left, my coworker acknowledged how bad that interview may have felt for me and mentioned that he himself was uncomfortable with the applicant's use of pronouns.
What Not to Do
It seems like we’re always at work. We work nights; we work weekends; we are pretty much in the office or traveling. It’s way more fun if the people around you are your friends. So, when I’m interviewing, I look for people . . . I’d want to get to know and want to spend time with, even outside of work . . . people I can be buddies with. ~ Consultant (source)
What Not to Do
A lot of this job is attitude, not aptitude . . . fit is really important. You know, you will see more of your co-workers than your wife, your kids, your friends, and even your family. So you can be the smartest guy ever, but I don’t care. I need to be comfortable working everyday with you, then getting stuck in an airport with you, and then going for a beer after. You need chemistry. Not only that the person is smart, but that you like him. ~ Banker (source)
"Cultural fit" obviously matters, but assess it in a considered and structured way. Culture fit interviews, like all interviews, can be less biased when they are designed to assess for specific, predetermined factors that are relevant to the job: e.g., the candidate is collaborative, likes taking on challenges, is comfortable with a quickly changing environment [Paradigm]
Use structured interviews. Structured interviews standardize the process among candidates, eliminating much subjectivity. These interviews pose the same set of questions in the same order to all candidates, allowing clearer comparisons between them. This may seem obvious approach, but is very underused. The flow of conversation during the interview will be slightly more awkward, but the payoff is worth it. You can do a "cultural fit" interview after doing the main structured part [HBR]
Develop a rubric to assess interview answers in advance. Developing a rubric to define what great answers, decent answers, and poor answers look like for each question ensures that all candidates are held to the same standard. In addition to minimizing bias, this type of structured process has been shown to produce the most effective hiring decisions [Paradigm]