I recently finished the book Everybody Lies – Big Data, New Data and what the internet can tell us about who we really are, by Seth Stephens-Davidowitz. A highly entertaining read, and it gives a fresh perspective of the field of Data science in the modern era of Internet behavior and social media filters. This is not meant as a book review. The book however starts with an interesting narrative of how Barrack Obama, during his time in office, fought the same underlying racist prejudice which later on brought Donald Trump to power, even though several official national studies concluded that the US had overcome harmful racial attitudes years earlier. The fact that Trump exploited and enforced these attitudes is another story. But how could the results from these studies be so wrong? According to Stephens-Davidowitz, it’s simply because people are not telling the truth when it comes to expressing ones public opinion, like for example in forms and surveys, much like we often lie in daily life.
On the contrary, people are not shy when it comes to using Google search. Big data therefore enables us to learn a lot about peoples truthful beliefs that they don’t talk about. For example that there is a widespread racist viewpoint in both democratic and republican constituencies, or surprising statistics of breastfeeding fetishes. People have probably always been hiding parts of their true mentality and inner fantasies. Even more so when it goes against the norm, morally just or not. With modern day tools like Google search, people can stay anonymous, well… relatively anonymous, and still explore and exercise these inner motives in a way that most would not dare to do outside of the digital realm.
All of this online activity can create quite an extensive data profile of a group of people and can furthermore make it possible to effectively interpret as well as influence the behavior of this group. Scary but evident.
The point is that Big data becomes a potent tool with which to measure and predict the truth. Modern data science can show us what people really think, instead of what they say they think. It could even show ourselves who we really are.
As a designer, this got me thinking. In UX research regarding digital product and service development, we are used to reflect about human behavior. In my field, it is a commonly accepted understanding that quantitative methods like Big data analysis can say what behavior that is caused and how much, from for example interacting with a product interface. However, it takes qualitative methods, like interviews or observations, to know why this behavior occurs.
Understanding the population of a city, country or the world and predict its actions can have a lot more benefit than that of someone winning a political campaign. It is undoubtedly important that leaders of the world can make the right decisions, as easy as possible. How leaders perceive the world and its population as well as predicting its behavior is furthermore very influential to what decisions they make. Isn’t it therefore consequently of great importance to also understand why a certain widespread belief or behavior exist and why people lie about it? so it can be properly understood or something can be effectively done about it?
With the development of more efficient data analysis techniques, *cough* machine learning, can Big data also make it possible to extract and understand the reasons behind certain human behavior and beliefs? Or are we, as the field of UX proclaims, forever dependant on complementary qualitative analysis?
Considering how increasingly difficult it becomes to find the signal in the noise, it feels almost impossible to also identify the construct in this way, since correlation does not imply causation. But how could we investigate the why if people are not honest about what they think?