Imagine that you are a Londoner during the latter stages of 2WW when London came under bombardment by German "vengeance weapons" – the V-1 buzz bombs and V-2 rockets. Based on your personal experience and personal experience of other Londoners you have a strong suspicion that these bombs are landing in definite clusters, with an unusual number of bombs landing on the poorer parts of the city, and thus making these areas of the city more dangerous than others. Now imagine that after the war I would present you with a map showing points of impact of 67 V-l bombs in Central London. The map would look this way:
What would be your impression? Would it confirm your hunch that bombs fell in clusters? My bet is that you would think so. Even in a case that you are familiar with the true “data story” behind, when looking at the map you probably have very strong impression that the points are not dispersed randomly - the lower right quadrant and also upper left quadrant look rather devastated; the upper right and lower left quadrants, however, appear to be relatively intact. Nevertheless, despite this strong impression, proper statistical analysis would reveal that the opposite is true. In fact, after the war, R. D. Clarke has applied statistical test to discover whether there is any evidence in favor of belief popular among the Londoners in latter phase of WW2 that German bombs were landing in clusters and unproportionatelly on the poorer parts of the city. Clarke examined 144 square miles of south London and 535 bombs that had fallen in this area. He divided south London into 576 small squares and counted the number of bombs falling in each square. In a case that bombs fell uniformly over this area, then these counts should be well described by the Poisson distribution. Clarke found that this was the case, and concluded that there is no evidence in favor of the clustering hypothesis, implying that the points of impact of German bombs were actually randomly dispersed throughout London and that people had been misled by their intuitions.
How is this story relevant for us, people working in HR industry? Well, we also tend to have plenty of personal experiences with area of our expertise and we also tend to form strong impressions about meaning of these experiences in terms of what they tell us about rules and regularities governing the “HR universe”. Based on our experiences we can be for example strongly convinced that performance of business units is determined by the quality and longevity of staff working there, that best managers are charismatic, that attending sales trainings increases sales, that current remuneration system stimulates better performance and improves retention etc.
Nevertheless, being just humans whose mind supposes a greater degree of order and equality in things than it really finds we can see signal where only random noise reigns, exactly as Londoner did during WW2. Being in a such situation it would be very handful to have some simple tools that would prevent us to fool ourselves and that would enable us to identify more reliable pieces of knowledge on which we could base our decisions that would bring us with higher probability closer to our goals, being it better job performance of employees, higher employee engagement or higher retention rates. In the next post we will look in more detail on a first such tool and using one concrete example I will demonstrate how it can be used for answering HR-related questions producing answers that may have real impact on the business results of the company. So stay tuned!
PhDr. Luděk Stehlík
Consultant, R&D Manager
Assessment Systems Czech republic