Focusing on the Problem – Part One

In Part One of John Schmitt's exploration of problem solving, he discusses the importance of adequately understanding a problem before coming up with a solution.

Albert Einstein is credited with the saying, “If I had an hour to solve a problem, I’d spend 55 minutes thinking about the problem and five minutes thinking about solutions.”  The message is clear:  Focus on understanding the problem.  If you truly understand the problem, the solution will become obvious.  We sense a certain intuitive truth to that statement, but there is also research to back it up.

A 2000 study by D.A. Kobus, S. Proctor, T.E. Bank and S.T. Holste for the Space and Naval Warfare (SPAWAR) Systems Center looked at decision response times among Marine Corps infantry officers in dynamic tactical scenarios.  (Shameless plug:  They used tactical decision games from my TDG workbook, Mastering Tactics, available for free download here.) The final report, titled Decision-Making in a Dynamic Environment:  The Effects of Experience and Information Uncertainty, is available here

The study found that experienced officers spent much more of their time examining the situation than did the less experienced officers — but that once they felt they understood the situation, they were much quicker to develop a course of action.  (Actually, I think it is more nuanced than simply that experienced decision makers spend longer considering the problem. They have an appreciation of how much total time they have and are willing to spend most of that time on understanding the problem.) Conversely, less experienced officers spent less of their time considering the situation and proportionally more of their time generating and considering courses of action.

This finding might seem counterintuitive at first.  Wouldn’t it take inexperienced officers longer to understand the problem? My theory is that experienced decision makers spend longer considering the problem because they appreciate its complexity and realize there is much to consider. They know the right questions to ask, and it takes time to answer those questions—which usually leads to more questions. Novices, by contrast, don’t appreciate the complexity of the problem and don’t know the right questions to ask. They don’t know where to begin, so they resort quickly to generating courses of action in the hope that something will inspire them.  

This research is consistent with Gary Klein’s Recognition-Primed Decision (RPD) model, which suggests that when experts recognize a problem through experience, they don’t have to reason their way through one or more courses of action. They know what to do. The solution emerges naturally from the recognition of the problem. When experts generate a single course of action, they do it based on their specific understanding of the problem.  

In contrast, as the SPAWAR study suggests, novices will tend to exhaust their analysis of the problem quickly and start generating multiple courses of action. This is consistent with Rational Choice Theory (RCT), which posits that decision making is a process of comparing multiple courses of action in an effort to find the optimal solution. But when novices, lacking a solid understanding of the problem, begin generating multiple courses of action, they do it more or less randomly. They are throwing things at the wall hoping something will stick, cognitively speaking. Maybe we should start calling Rational Choice Theory DMD—Decision Making for Dummies—because it really does seem like a process designed for beginners. (Interestingly, most of the research that led to RCT used naïve subjects in the form of college undergraduates performing unfamiliar tasks, so there was no opportunity to examine the effect of experience.)

So, if we know that focusing on the problem is key, and if we know that once experts understand the problem they know what to do without having to generate and compare multiple courses of action, why is it that senior leaders continue to ask their staffs for multiple courses of action? “Give me some options” is a refrain we hear from senior executives in practically every field. Part of the answer is that we still have not succeeded in finally killing off the Zombie stragglers of DMD. RCT is still embedded in a lot of people’s minds as the “right” way to make decisions, even if they don’t actually do it that way. More importantly though, when I hear senior leaders ask for options, I take it as a sure indicator that (1) they do not understand the problem and (2) they realize they cannot afford to spend the time it will take to understand the problem.  So (3), they are looking for a shortcut.

On the surface, that seems to be the same thing the novices are doing.  But there is another possibility. I believe the senior leaders are doing something more sophisticated. They are asking for options not because they actually expect to choose one but as a way of exploring the problem space quickly. The options serve as hypotheses about the problem type. Understanding the problem and solving the problem are the same cognitive process. Per the RPD, they will use their experience to run a mental simulation of each option to see how it turns out. If Option X looks like it has some merit, that tells them something about the problem:  it must be an X-Type of problem in some important way. That’s a valuable insight. If Option Y looks like it will not work, they can eliminate Y-Type problems from consideration. This allows them to take directed excursions through the problem space and quickly eliminate bad options rather than trying to analyze the full problem space from Square One—which they don’t have to time for. That may be even more valuable. It is a clever way of repurposing an RCT tool designed for other purposes to their own needs. In fact, proponents of RCT, when confronted with its consistent failure to improve decision making, retreated to this same position—that RCT is primarily useful in helping people understand the situation. That may be, but if the objective is to accelerate learning about the problem, I bet we can build naturalistic decision-making tools better suited to that purpose.

More about that in a later post.