Analyst’s Pendulum (Thinking)
A Concept From Both Psychology & Biology Important To All Analysts
Holistic and reductionist thinking are not immensely common terms around the corporate water cooler. They are popular terms in other fields. Biology, psychology, health, and wellness to name a few. They describe a very simple dichotomy of thinking. One quite important to the field of data science and analytics. And so our pendulum swings…
Reductionist and holistic thinking represent two extremes. Or perhaps they simply describe varying degrees of perspective. After all, it isn’t exactly a mistake that our pendulum images swing in two dimensions. They are two methodologies for divining deeper insight and impact. Our metaphor works well here.
Holistic thinking addresses a system as a whole entity. Reductionist thinking breaks it down to the sum of its parts. Which is superior, which is more effective, is an intriguing question. They are both unique and useful perspectives, though for completely opposite reasons.
Holistic Models
Models which look at the greater picture, which focus on connected relationships, which consider concepts like complexity and emergence are holistic. They represent things more broadly, as a whole. They recognize ecosystems and interconnection. They are often top down. They see the whole as more than the sum of the parts.
Reductionist Models
The counter to holistic models are reductionist ones. They are the essence of analysis. They break things down and break things up. Reductionist models see all things as the sum of their parts. They focus on those parts. They are the very epitome of analysis.
For analysts, the pendulum swings. Well, at least, it should. Unfortunately, many analyst’s pendulums don’t swing when it comes to this sort of thinking… errr, these. Many analysts stick to one type of thinking or the other. They are simply more comfortable dividing or thinking big picture. They lack either the perspective or ability to work in both paradigms.
Not at once, mind you. Rarely does an analyst need to think in both directions at once. They do need to have a general willingness to alternate as situations call for it. It is likely true that more analysts have the willingness than are able to demonstrate it.
Beyond the willingness and ability of the analyst lies the limitations of their data. Although, impacts do vary. Gaps in data force many analysts to pursue reductionist models. If they can’t capture the whole, how can they model it? Almost as common, similar gaps drive other analysts to pursue only holistic models. It seems easier to overcome gaps in the data by rising above them. Holistic models are often ripe with allocations and assumptions. What better way to compensate for lost data?
In the end, let the pendulum swing. Attack problems from both angles. One at a time is fine, but be willing to try both. Recognize that bad or missing data is a limitation no matter which methodology you embrace. And thanks for reading!
For more series like this: