Many HR policies rely on the assumption that employee performance is distributed across a bell curve. But is that really the case? Not according to a recent paper by Ernest O’Boyle Jr. and Herman Aguinis, published inPersonnel Psychology.
In a study of over 600,000 people across five study groups and a number of professions, the authors found that performance was not normally distributed, but follows a Paretian distribution. Pareto curves produce fatter tails than those seen in a normal curve because they allow more extreme values to be present. This is often referred to as the 80/20 principle, and has been shown to apply in many contexts outside of HR. For example, market researchers state that about 80 per cent of a brand’s volume is purchased by about 20 per cent of its buyers. Paretian distributions typically have unstable means, infinite variance, and a greater proportion of extreme events.
Many HR policies and processes work on the assumption that individual performance contributes to the overall performance of the organisation and that this performance, when measured across big enough populations – like a department or business unit – clusters around a mean and then spreads out into symmetrical tails. That is, individual performance is assumed to follow a normal distribution. When this is not the case, HR either challenges the line managers involved, contesting their performance ratings, or questions the make-up of the sample. Then adjustments are made until the sample reflects the ‘normal’ performance.
The implications of these psychological studies for HR are pretty scary. Have we been ignoring high performers? Or have we been letting go an insufficient number of poor performers? Many organisational policies demand that, given normal distribution, if individuals’ performance is in the top 10-20 per cent of the curve, they should be paid more and motivated to stay with the company. If their performance is in the bottom 10-20 per cent of the curve, they should be paid less or may even be let go. I certainly presided over policies that did just this and routinely performance managed the bottom 10 per cent of employees each year.
But if performance is actually distributed on a Pareto curve, these actions are wrong or at least the percentages are. I am not a statistician but even I can see the issues.
Also because the Pareto curve demonstrates scale invariance, whether looking at the entire population or just the top percentile, the same distribution shape emerges. For selection, this means there are real performance differences between the best candidate and the second best candidate. Our ‘success profile’ research suggests those with the success factors to execute the future strategy have a disproportionate impact and this would seem to be supported by these studies. The ability to identify these exemplars will become even more important in the current economic environment. This study suggests that HR practitioners should focus on identification and differentiation at the tails of the distribution so as to best identify exemplars.
The researchers quote a theory called the Matthew effect; this states that those already in an advantageous position can use that position to gain disproportionate rewards because they are deemed to make a bigger contribution to the company. A similar argument can be made for both pay and other rewards such as training or high-potential type schemes that rely on measures of current performance. While this is already true in many organisations, HR needs to look at whether it has captured all the best performers and has policies to manage the Matthew effect without alienating others.
As with any study like this, there are questions about whether the data generalises to your own populations and job roles. But it is compelling enough to make you stop and consider whether that manager really does have an exceptional team of high performers and whether your exemplars are all recognised.