What HR can learn from the exam debacle and using data blindly
Lucinda Carney, August 20, 2020
Few of us can feel anything other than shock at the dreadful inequities caused by the methods used to moderate this year’s exam results. How could a seemingly fair and objective process get things so wrong?
Of course, there are many lessons to be learned from this politically, and hindsight is a wonderful thing, but so much damage has been done to students everywhere and particularly those from less privileged backgrounds.
Algorithms are often considered to be clever, mathematical devices that smart companies like Google use. They are meant to be fair, yet this is sadly often not the case.
We know that there is an increasing pressure for HR to be numeric and data-led, and there’s no doubt that people's data can be incredibly valuable for us when considering trends or looking for correlations or evidence to support people strategies.
However, could we be running the risk of over-relying on data and forgetting to bring in that human factor or gut instinct?
It is hard to imagine that if someone at Ofqual or within government had taken a quick glance at the year-on-year changes in A-level results for a handful of public schools, and then compared them to a handful of sixth form colleges, that they wouldn’t have thought that something might be awry.
Think of a typical 18-year-old who has had their teacher-assessed C grade downgraded to a U by the algorithm. What does common sense say about a C grade candidate ending up with a U as opposed to a D, E or even a B grade?
The truth is that this terrible mess wouldn’t have happened had someone applied common sense to the mathematics, rather than blindly following the requirements of a forced distribution.
As HR professionals, data can definitely be our friend, but it is only part of the picture and it is our role to consider it in the context of other forms of evidence.
We need to be alert to inherent bias that could be built into AI interviewing tools or forced distribution of performance-related pay. This doesn’t mean we reject the data, but we can’t afford to be intimidated by it either.
Let this awful situation in education be an inspirational lesson to us as HR professionals about the power and the danger of using data blindly. We are in increasingly data-driven organisations and we should use this to be more evidence based.
Graphical data can be incredibly valuable in plotting potential patterns that may be increasing engagement or attrition; monitoring the gender pay gap or distribution of reward to BAME staff.
However, we would do well to remember the adage ‘lies, damned lies and statistics’ and ensure that we have the confidence to dig into the data to see what sits below it, to ensure that it is representative and isn’t introducing unconscious bias into the process.
If there is one thing that HR should learn from the exam fiasco, it is that we need to embrace technology, but not let it be all-consuming.
We must build the confidence to be able to stand up to the data and apply common sense, bring the human touch and keep it in its place if we want to ensure that it brings fairness rather than furthering inequity within our organisations.
Lucinda Carney, CEO of Actus