HR analytics should combat illusory correlation with a bit of fun

When I see people describe HR analytics as people who are “just collecting data”, I often wonder if they know how wrong they are. HR analytics have tremendous amount of power – perhaps more than it realizes. It’s real power lies not in obtaining data or facts and passing it on. Instead, the real power lies in converting facts to information and knowledge and presenting it.  Question is, do people employed within HR analytics know how to mange this power?

How you present your information is important. Jeremy Shapiro has highlight how important it is to keep the information down to the bare essentials due to cognitive load. I have also shown how  cognitive dissonance will create a bias for a certain decision despite of facts and evidence may favor the alternative. A third important concept to be aware of when presenting your information is illusory correlations.

Illusory correlation is the belief that two variables are associated with one another when little or no actual association exists. This was probably best illustrated by Hamilton & Gifford in 1976 in the following experiment (from Stephen Franzoi’s ‘Social Psychology’) . Hamilton & Gifford asked participants to read information about people from two different groups, “Group A” and “Group B”. Twice as much information was provided about Group A than about Group B which made Group B a kind of “minority group” in this study. In addition, twice as much of the information give about both groups involved desirable behaviors rather than undesirable ones. Desirable information included statements such as, “John, a member of Group A, visited a sick friend in the hospital”. An example of an undesirable statement was, “Bob, a member of Group B, dropped litter in the subway station”.

Even though there was no correlation between group membership and the proportion of positive and negative information, participants perceived a correlation. As the figure below shows, participants overestimated the frequency with which Group B, the “minority group” (who where only described has a much as the “majority group”) and the undesirable actions (which occurred only half as much as desirable behaviors) were both distinctive aspects of participants’ social perceptions.

This is an important study for several reasons. It concludes that humans make wrong correlations/judgments even when they are faced with correct data and that this often goes against a minority group.

There are many reasons why humans (and therefore also HR analytics people) fall prey to illusory correlation. One reason is the psychological term ‘heuristics‘, which essentially mean that we use cognitive shortcuts to make judgments fast. We leave out a lot of relevant information in order to make a judgment. Stereotyping is one such heuristic. Whatever the reason is – we all do it all the time!

How does this apply to HR analytics? First, we need to understand that what and how we present “objective data” is hugely influential to how judgments and conclusions are made. Therein lie the power of HR analytics. Secondly, we must be careful that we do not make false and illusory correlations. We can do this in several ways; 1) constantly be aware of this phenomenon (studies show that education, reevaluation and reinterpretations help), 2) be in a better mood (Steven Stroessner (1992) found that people in a positive mood are less likely to perceive an illusory correlation). Perhaps all we need to do is to have more fun in HR analytics 🙂

5 thoughts on “HR analytics should combat illusory correlation with a bit of fun

Leave a Reply