Is Statistics Making Us Stupid?

How a brainless use of statistics lead to wrong decisions and wastage of effort

Statistics and data science in action from ancient Indian epic Mahabharata.

“The king lost his way in a jungle and was required to spend the night in a tree. The next day he told some fellow travelers that the total number of leaves on the tree were “so many” (an actual number was stated). On being challenged as to whether he counted all the leaves, he replied, “No, but I counted the leaves on a few branches of the tree, and I know the science of die throwing.”

Photo by jesse orrico on Unsplash

Data is the fuel of the digital economy, and statistics or data science is the tool that converts the fuel into energy for business growth.

Teams who make judicious use of the tool build delightful, compelling products. Teams who misuse the tool — statistics — waste effort and time for the company. Believe it or not, the misuse of statistics is quite rampant. In this post, we’ll list out several misuses of statistics.

First, let’s get into

How statistics is making us stupid?

The title ‘Is Statistics Making Us Stupid’ is a play on ‘Is Google Making Us Stupid.’

Statistics can make us stupid because it takes away our ability of critical thinking. We start relying on it and stop using our brains for decision making.

Please find some thoughts on why teams become numb on the face of statistics.

1. Laziness is the engine of progress. The desire to look for an easier way to do tasks drives innovation. Naturally, we want to offload decision making to machines and mathematical formulas. For example, a measure statistics, like confidence level, which is a result of an unaccessible complicated mathematical formula, may give us reason to stop thinking.
2. Cognitive load and System 2 thinking: Complex statistical formula evokes system 2 thinking. It is effortful and laborious. The cognitive load takes away the ability of basic logic, and folks end up making wrong decisions.
3. Social Currency and Signaling: Data science and statistics have become a cool buzzword. For product teams and data analysts, it has become a signaling mechanism to be a cutting edge. We, our team or our company is a dinosaur if we don’t use data science — even when we don’t need it. When we need it, we choose a complicated statistical method over the simpler one. Heard of the expression — building space shuttle to cross Mississippi

Examples of misuse and abuse of statistics

Marketing campaign and product launch (AB test world)

Data science (WIP)

Closing thoughts: What we need to do?

We now know more about the world, but the world has become more complex. Making sense of the data, and how do we use the data to make decisions and build products is the most critical skill of all. Statistics and data science is a tool to help make decisions. Unfortunately, it is a tool. No more!

Here is what we can do to make sure we are not using/abusing the beautiful science.

1. Awareness that data interpretation is a skill. Applying statistics as a skill does not have perfect overlap with bookish statistical depth. Statistics help, but we need human intelligence to operate the tool. On the other hard, there is no shortage of people who looks at analytics as a function to pull data. They come from the product, BI, and in some cases, from analytics. Please, do yourself a favor, and don’t take pride in pulling data. With time, all aspects of it will get automated. The job of a PM or an analyst is to choose the right analytical framework and do data interpretation for decision-making.
2. Holistic view and an eye on the goal always helps. It’ll stop you from wandering into paths that lead nowhere. It is advisable to develop a growth model for your product. It will help you in making decisions, it will help you in prioritizing the product built.
3. Don’t be in the aura of a super-sophisticated statistical method. Try to develop system thinking about how it fits into solving the problem. If it helps, use it, and if it doesn’t, don’t think twice before abandoning it. Don’t use a statistical method because it is cool. Don’t ignore an analytical approach because it looks too simple, and everyone uses it.
4. Team: Don’t hire people who err towards complicated methodologies. Please read more on this here.