<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=353110511707231&amp;ev=PageView&amp;noscript=1">

Ethical Considerations for Data Analytics in the Hiring Process

July 10, 2018

big data assessment ethics

It has been a couple of months since SIOP 2018, but I keep thinking about one of the sessions I attended. It was a debate surrounding the ethics of data use in future assessment models. With continually advancing technology, more and more sources of data are becoming available that could potentially be used to assess an individual’s personalities and other characteristics and perhaps even be used to make employment decisions. The goals of this session were to discuss the types of data available and if they should be explored or utilized.

Some examples of the data sources referred to include:

  • Social media usage (e.g., Facebook posts)

  • Online gaming experience

  • Verbal patterns

  • Facial recognition

  • DNA/genetic information

We know that some organizations, or at least individuals within organizations, will often look at a candidates Facebook or Pinterest profile to try to get a feel for their personality, interests, or experiences, even when we advise them not to. They just can’t help themselves and can’t resist exploring this easily available data that may shed additional insight into a candidate’s fit for a job. But is it ethical to hold a social media post someone wrote in college (or even high school) against them when they are applying for a job? Is it fair to conclude that someone is unable to handle stress because they wrote a single post about struggling with the pressures of graduate school during finals? Is it ethical to weigh someone’s online gaming habits or behaviors from middle school in the decision to hire them for a job in their thirties? How far back is it okay to go – data from two months ago? Data from two years ago? Data from 20 years ago? Can we refuse to hire someone because of the way they emphasize certain words in a sentence? It okay to select someone into a role based on their genetic makeup? These are the types of questions we are facing as this type of data is increasingly available and is sometimes even at our fingertips.

Further, what happens if we find that this type of data is actually predictive of job success? Is it okay to use simply because it is predictive without having any rationality or theory behind the findings? Should someone not be hired or promoted because, for example, their eyes are spaced closer together than average and this has been linked to lower job performance (this is a completely hypothetical example.)? Even if you don’t explicitly know the answer to these questions, I bet you can sense it…the answer is no.

However, the reality is that people are going to be collecting, analyzing, and using this data whether we think it is okay or not – or even whether we consent to it or not (just consider recent events!). As the presence of big data analytics and data scientists continue to increase in our field, we will likely find that use of these types of data (among many others) will be more prevalent for a variety of purposes, including hiring decisions. Some of the panelists were of the opinion that we (as I/O psychologists) should at least analyze and explore these types of data to better understand not only their predictive power, but also their likelihood to discriminate against some protected groups in an effort to ward off the use of prejudiced measures, to better answer some of the questions posed above, and to better educate others on these risky data sources. Yet, others felt that it was unethical to even begin to explore and analyze these sometimes controversial data.

Here are the overall takeaways to consider when it comes to analyzing or using these data sources:

  • We need to keep good science at the forefront.

  • We need to be wary of claims and results found by others who don’t have solid justification for using such data.

  • We need to establish the job-relatedness of any data that we use in decision making.

  • We need to ensure that such data is free of bias.

There was much food for thought presented and discussed at the SIOP conference this year, and while some of it was rather scary, if we continue to leverage our training, good science, and strong moral compass, we should feel confident in our conclusions.

New Call-to-action

Alli Besl, Ph.D. Alli Besl, Ph.D. was a Research Consultant based in the Pittsburgh office of PSI. Her areas of expertise include: employee turnover, selection and recruitment.