Data is very ambiguous. It often doesn’t give you a clear answer to the question; “What should I do?”, which is really the question you hope that Big HR Data will help you answer. To get to that point, you must take your data and convert it into knowledge. The question is; what will you see when you stare at the data? The answer is clear but not very encouraging.
I have previously argued that psychology has a lot to offer us in our understanding of how we work with data. This is also the case when we need to understand what we look at Big Data. Big Data is a much hyped term which essentially just refers to a lot of data – a lot of data in terms of volume, variety and velocity. But a lot of data it is and you therefore need to cut and slice it. You need to choose to pay attention to some data over other.
The psychological concept of confirmation bias can help us understand what we see in Big Data. Confirmation bias refers to the fact that people will search for, pay attention to and interpret data in a way which supports their own preexisting view. And conversely, people has a tendency to undervalue or look away from data which conflicts with their existing view.
In short, when you are looking at a lot of data, you will focus on the data which supports your view and ignore the data which contradicts your view. Hence, you will find data that shows that your project has been a success, that you do add value and that your decisions were right. Funny that.
Confirmation bias has been used to explain other bias effects such as cognitive dissonance (where people continue to hold their beliefs despite evidence showing the opposite) and illusory correlations (the belief that two variables are associated with one another when little or no actual association exists).
Several studies supports the confirmation bias theory. An interesting one [described here] took place during the 2004 US presidential election and involved a group of people who had described themselves as having strong feelings about one of the candidates. They were shown apparently contradictory pairs of statements, either from Republican candidate (George W. Bush), the Democratic candidate (John Kerry) or a politically neutral public figure. They were also given further statements that made the apparent contradiction seem reasonable. From these three pieces of information, they had to decide whether or not each individual’s statements were inconsistent. There were strong differences in these evaluations, with subjects much more likely to interpret statements by the candidate they opposed as contradictory.
In this experiment, the subjects made their judgments while in a magnetic resonance imaging (MRI) scanner which monitored their brain activity. As subjects evaluated contradictory statements by their favored candidate, emotional centers of their brains were aroused. This did not happen with the statements by the other figures. The experimenters inferred that the different responses to the statements were not due to passive reasoning errors. Instead, the subjects were actively reducing the cognitive dissonance induced by reading about their favored candidate’s irrational or hypocritical behavior.
This is important for HR analytics and HR Data (or Big Data) for several reasons. First, if analytics is supposed to help us make better HR decisions, confirmation bias will instead make us make the same decisions just with better arguments. That’s not good. Second, if analytics is a way for HR to argue its case more credibly up in the organization the credibility will be shattered if there is a sense that confirmation bias has been prevalent in the analytics stage.
While psychologist have been good at identifying this behavior and making studies to prove it they still have to come up with a great way to overcome it. But in practice what you need to do is:
- Be aware of your existing bias, hypothesis and view.
- Seek information which challenges specific biases. For example, assign a person to play the devils advocate during meetings. See as many objective data to contradict questionable biases.
- Reflect on the information which challenges your view.
- Incorporate the contradictory view into your existing hypothesis and test it
Like any four-step program this sounds really easy but is difficult to practice. Especially because it is a lot of hassle and easy to ignore. But I fear that confirmation bias may stand in the way for unleashing the true value of Big Data in HR.