Blog post

Why We Need a Data Model to Understand Pay Equity

By Debra Logan | April 12, 2021 | 0 Comments

CIO Leadership, Culture and PeopleHR Function Strategy and ManagementTalent Analytics

All humans are biased.  We are limited in what we can perceive by our senses.  Our brains and central nervous systems limit what we can process.  We are shaped by our genetics (nature), our experiences (nurture) and our preferences.  We can overcome these biases with rigorous training, discipline and experimentation.  We can never fully eliminate them.  Humans share a set of species specific characteristics.  In that sense, we truly are one human family: to be human is to be ‘biased’.

Our Biases Are Like Icebergs

Our biases can be good things.  They allow us to make lightening quick judgements and take appropriate action with zero conscious thought. They are also bad things.  We automatically categorize people as ‘like us’ or ‘not like us’ in ways that limit our true understanding of other people.  Some biases are conscious. We can, with effort, train ourselves to overlook them.  Some biases are unconscious.  That means there is nothing we can do about them because we do not even know they are there.  Many of our biases would seem to be  matters of taste, style or preference.  Think about something as simple as the kind of pets we like:  some  prefer cats, some dogs.  And we can logically (we think) explain why.  But imagine a cat curled up and purring as you stroke it or a half-grown puppy looking up at you with absolute trust, love and enthusiasm as you prepare to throw the ball ‘one more time, ok, buddy, then time to go in!’  Maybe neither of those scenarios appeal and you are biased toward pet-hair-free furniture.

One bias that we all share is that of a preference for ‘people like us’.  There are thousands of reasons but it boils down to this:  we are simply more comfortable with ‘people like us’.  This is an iceberg bias:  a small part of it is perfectly visible but a lot of it is not and never will be until you hit it in an obvious way.  When I meet someone from Pittsburgh for example, I know we have things in common on the surface:  no need to ask, ‘Are you a Steelers’ fan?’   If you’re from Pittsburgh, you ARE a Steelers’ fan. If you’re not, you would never, never admit it!  I have a colleague who says he enjoys talking to me because I ‘sound like his mother, sisters, aunts and cousins’.  We remind each other of home and family.  I think that is a good example of a benign bias which is not completely obvious, but figures into our friendship.

Our Biases Effect Our Judgment

So why does any of this matter?  Bias affects our judgement when we are evaluating people.  As much as we like to think we ‘picked the best person for the job/this person deserves it’ whether recruiting, developing or promoting, forming teams, deciding on bonuses and pay rises.  This is not deliberate discrimination.  This is the submerged part of the iceberg bias. There is way more under the waterline than we know, or ever will, unless we hit it.  It cannot be trained away, not completely.  There is only one way to navigate around the iceberg.  We need to create data models that represent people plus algorithms & methods that use it to make unbiased comparisons.   The person-as-job-related-data-model is crucial to figuring out the pay equity challenge.  We need to bring data driven rigor to our hiring, development, retention and performance evaluations.  We need a set of data elements that reduce bias with rigorous & shared definitions of substantially job related characteristics.

The creation of that model, along with the fundamental fixes and algorithmic elements that my colleague Christie Stuckman is blogging about, is the next step in the EPIC work.  Christie has already written about this here:  https://blogs.gartner.com/christie-struckman/2021/03/22/what-is-relevant-information-when-assessing-pay-equity/?_ga=2.172994791.593004584.1618219424-892691859.1530692391

This is a good list to start.  Our next series of blogs will build this out in parallel with the creation of methods and a discussion of the math.

Comments are closed