Can machines eliminate prejudice? Using People Analytics to overcome bias

There’s a fundamental puzzle at the core of the gender diversity issue. By now, it is well established that businesses led by women and teams that include women tend to outperform their all-male competitors. Large, well-designed, global studies have supported this finding again and again.

So, if you’re a business leader who wants to see your company thrive, why aren’t you rushing to hire women into senior roles as fast as possible?

Part of the problem is that leaders they are often swayed by the “looks like me” syndrome when they go to select the best talent. When top talent comes in a package that looks different, the selection committee often just doesn’t recognize it. If you’re female — or African-American or young or old or disabled, for example — you are more likely to be passed over because you don’t look like the folks who are making the selection.

Years ago, the world of symphony musicians had the same problem. Back then, if you went to see a performance by a world-class orchestra, the musicians were male, almost without exception. This was because everyone “knew” that male players were better — stronger, more musical, whatever.

A very persistent woman applicant, eager to defeat the myth, pestered the audition committee at a major orchestra into listening to candidates play behind a screen. A clear winner emerged — and when the screens came down, the committee members were stunned to see that the woman was the winner. To the credit of the music world, many orchestras accepted the facts and learned the lesson. Since then, many auditions are held behind a screen, and the proportion of women in major orchestras has increased significantly.

I don’t think it’s possible, or even desirable, for job interviews to be conducted behind a screen. For one thing, voices would have to be disguised as well as faces. But there is another powerful tool that can help business leaders get past their biases — People Analytics.

Using People Analytics to hire smarter

People Analytics uses data and analytics tools to make informed decisions about people. Through the internet, and especially through social media, employers have access to a wide variety of data about job candidates. For external job candidates, these data might include:

• “Digital exhaust” — chat, email, contacts, social media, documents, calendar, search, collaboration, cloud, etc.

• Work samples

• Cognitive tests

• Structured or unstructured interviews

• Tests of job knowledge

• Personality tests

• Reference checks

• Rate of promotion

• Résumé

• Job descriptions

• Academic record — GPA, university attended

• Document analysis

For internal candidates, all of the above data might be available, as well as:

• Surveys — employees and customers

• Performance evaluations

• Network analysis — surveys, email and phone patterns, bulletin boards, social media, archival data bases (shared projects, work histories, event attendance), public data bases (e.g. co-authorship, field work, observations, diaries, electronic tags)

• 360 feedback

• Compensation history

• Videos — movement in building, meetings, calls, etc.

Powerful new analytics tools are available to make sense out of all these data and create a picture of the candidate. For example, increasingly sophisticated text analysis enables us to combine qualitative with quantitative data in new and powerful ways. Instead of scouring every tweet a person has shared over 10 years, you can run their Twitter account through an algorithm to determine whether the person is optimistic or pessimistic, antagonistic or helpful.

As the same time, business leaders can also use People Analytics to identify the characteristics that differentiated previous successful hires. This is a really important part of the process, because the qualities that predict success in one role and one company may be quite different from the predictors in a different context.

Defeating gender bias using People Analytics

So back to hiring women. The traditional triad of resumes, reference checks, and interviews — perhaps with a couple of online assessments added in — provides massive opportunities for gender bias to creep in. Take reference checks. When I was in graduate school, one of my fellow students saw a reference letter that had been sent in by one of her professors. It said something to the effect of, “Amy is a very strong student, but she is so attractive that she’ll probably get married and drop out of school.” That kind of data is not going to help anyone get admitted to an elite program or hired to a top job!

Imagine instead that job candidates provide a variety of the types of data listed above, and top candidates are identified using People Analytics tools. Only then would interviews be scheduled, preventing the hiring team from selecting for gender, race, class, or other biased markers in the initial screen.

The experience of symphony orchestras suggests that a lot more women would be hired this way. No system is perfect, and some of the data used in the analysis might already reflect gender bias. But bringing in more data points reduces our reliance on any one indicator, making us more likely to see the whole candidate, instead of the part likely to set off our biases.

People Analytics is one of the current best approaches to help businesses leaders hire top talent, not just people who look the part. The research is clear —  when women win, businesses win. Time to make it happen!

If you would like to know more about using People Analytics in your hiring and talent management, contact us at ggolden@gailgoldenconsulting.com.