From a young age we are taught not to jump to conclusions. Rather, we are encouraged to examine the data. At no time has this been easier than today—big data surrounds us, inviting us to dig deeper in order to arrive at correct conclusions and predictions.
The push for more data makes sense. If we look deeper, to more fundamental science, it seems intuitively plausible that we should arrive at the “true” cause of some phenomenon.
A teacher, when asked how to improve education, will respond based on experience and some experimentation. But then an education researcher will contend we need a deeper explanation, with more data collection and analysis. A cognitive scientist in turn insists that studying classrooms and schools will not help; we need to understand how the mind of both teacher and student operate. But then a chemist declares that brain chemistry and structure underlie cognitive science, and so a true cause can only be found in chemistry. Of course, the absurd conclusion to this search for “true causality” would be for a physicist to say that true causality lies in the interactions of the two fundamental constituents of all matter, quarks and leptons. This reductionism has a ring of truth: shouldn’t we be digging deeper for causal answers to questions that matter?
The problem with this reductive approach is that causality is not a free-floating thing, waiting for us to discover it. Rather, what counts for causation is determined by the questions we seek to answer. Accordingly, all of the above proposed approaches are legitimate insofar as they answer the question being asked—the “deeper” approach applied by the physicist has no more validity than that of the teacher.
Consider me sitting in my kitchen reading the news while a pot of water is boiling on the stove. My wife comes into the kitchen and asks, “Why is the water boiling?” Being scientifically inclined, I say, “The plasma you observe under the pot has high mean kinetic energy. Because entropy always increases, it transfers that energy to the pan and the pan to the water, which goes through a phase change….” Heather, ever patient with me, says, “No, doofus. Why is the water boiling!” My new tack, “There’s a handle on the stove that when turned opens a valve to release pressurized natural gas and closes a circuit that initiates a spark. Then….” Undeterred by my second causal explanation, she insists on knowing why the water is boiling. Finally, I answer, “I’m making mac and cheese for the kids.”
All of the answers that I offer my wife, as well as countless others, in a sense constitute legitimate causal explanations. However, the vast majority of them don’t work—they are incapable of answering the relatively straightforward causal question my wife was really asking: Why are you boiling that water? Causation is not out there waiting for discovery, but is inescapably circumscribed by the questions we pose, and the problems we seek to resolve. For this reason, some questions necessarily steer us to some answers over others.
The intuitively appealing notion that we should dig deeper and collect more data in order to arrive at a solution is partially right. Yet it’s equally important that we work hard at the outset to clarify the real question being asked, so that we understand how and where to search for our answer. This is at the core of how we approach complex problems at the Christensen Institute. Only when we’ve uncovered the real question that we’re asking or that others are asking us can we assess whether the Theories of Disruptive Innovation provide a genuine causal explanation and reliable predictions of outcomes.
For instance, when addressing how to encourage economic development in poor countries, we don’t ask, “How can we eradicate poverty?” but rather, “How can we create prosperity?” Our theories enable us to answer latter, but not the former. Similarly, regarding skyrocketing healthcare spending, we don’t ask, “How can we help Americans afford healthcare?” The theories don’t have an opinion on that. But we can ask, “”How can we make healthcare affordable?” to which the theories have a lot to say.
Determining causality, it turns out, is all about framing. Data and theory enable us to explain and affect the world around us. But in order for them to do so reliably, we must be sure to clarify the question being asked and choose the appropriate framework accordingly. Reductionism may be seductive, but sometimes the solution is as mundane as “I’m cooking mac and cheese”.
