HR Data Intelligence: Learning from the A-level results debacle

How did they get it so wrong?

It cannot be denied that algorithms are extremely intelligent. We only need to look at the work of tech giants such as Google and Amazon to see the potential of mathematically driven AI. However, the recent A-level results fiasco has highlighted that, although this software is super complex, it isn’t always entirely fair. This was felt by the thousands of students who were left disappointed by their exam results. This is especially evident when you consider the majority of victims were from less privileged backgrounds. Many across the nation were left shocked by the injustices caused by a data-led process that was supposed to be reasonable and objective. Yet, rather than pondering over past errors, we can look to the future and view this as a learning curve. Certainly, it is an opportunity to review our own HR data intelligence and discover if improvements can be made.

Artificial intelligence or human intelligence?

If only an individual had taken a glimpse at the A-level results across the years, they may have foreseen potential issues beforehand. Certainly, they may have seen that those in private schools achieved rather favourable results in comparison to state-funded institutes. Arguably, common sense dictates that a student predicted to receive a C doesn’t deserve to be downgraded to a U. Perhaps if Ofqual had investigated a little more closely beforehand, they’d have seen that the algorithm had not taken into account the individuals academic potential but rather, used data stemming from previous school or regional averages. A logical response would have been to apply common sense to the overall data, instead of blindly depending on a rating system that didn’t take into account factors on an individual basis.

What does this mean for the future of HR?

So, what does this mean for HR? What can we learn from the results fiasco? Well, there is increasing pressure for the HR sector to become more data-driven, especially in light of the innovative technology now available to us. Graphical data can be incredibly valuable in discovering potential correlations that may be growing engagement or attrition; monitoring the gender pay gap or distribution of reward to BAME staff. Or perhaps, when considering trends or evidence to support people strategies. Yet, the whole fiasco should be treated as an educational breakthrough about the dangers of using data carelessly and without context.

If we become too reliant on technology, we run the risk of losing touch with our natural human judgment. Certainly, we should be incorporating data into our approach. However, it also is our responsibility as HR professionals to not allow this to overshadow the bigger picture. To be constantly alert to unconscious bias, we must have the confidence to look deeper into the HR data intelligence and ensure that it is consistently representative of our inclusive values.

The A-level debacle has taught us that a balance between technology and traditional human touch is the way forward. Our human experience and common sense can see pieces of a puzzle that may otherwise be missing from data-led conclusions. Therefore, we can allow HR data intelligence to encourage decisions, but we cannot let it be the deciding factor.

Performance Management Resources

Part of our ethos is to help build a better workplace for people. We achieve this through great performance management software, the HR Uprising Podcast, and a variety of free thought leadership resources. Therefore, if you would like to learn more about other topics, you can find some additional resources using the buttons below.

We also have a team of organisational development consultants who are on-hand to help meet your needs for culture change. If you would like to find out more about this service, why not get in touch by contacting us here.

Upcoming Webinar Schedule

The HR Uprising Podcast: Evidence-Based HR with Rob Briner

The HR Uprising Podcast: HR Tech Trends with Mervyn Dinnen

White Paper: Managing Underperformance in a virtual world