Working Memory and Question Framing: Why Simplifying the Question Changes the Answer

The leadership team has been debating the same decision for three meetings. Should they expand into the European market this year, given current resource constraints, the strength of two potential partnerships, the uncertain regulatory environment, the need to maintain domestic growth targets, and the CEO’s concern that waiting another year risks losing first-mover advantage to a competitor who may or may not be preparing a similar move?

Nobody can answer this question. Not because they lack information or intelligence, but because the question itself is unsolvable in its current form. It contains at least six distinct sub-decisions, three unresolved assumptions, and enough qualifiers to overload any individual’s ability to reason about it coherently. The team isn’t stuck because the answer is hard. They’re stuck because the question is bloated.

The research

Nelson Cowan, a cognitive psychologist at the University of Missouri, published an influential paper in Behavioral and Brain Sciences in 2001 that revised our understanding of working memory capacity. George Miller’s famous 1956 estimate of “seven plus or minus two” items had dominated for decades, but Cowan’s analysis, drawing on a broader set of experimental conditions, settled on a lower number: roughly four chunks of novel information. That’s the active workspace your brain has available for holding and manipulating the components of a problem in real time.

Four items. Not four words — four conceptual units. When a decision question contains six variables, three conditions, and two uncertainties, it exceeds working memory capacity before deliberation even begins. The result isn’t that you think slowly. It’s that you think poorly — dropping variables, oversimplifying in unpredictable ways, or cycling through subsets of the problem without ever holding the whole picture at once.

Amos Tversky and Daniel Kahneman demonstrated in their 1981 paper in Science that how a question is framed doesn’t just influence the answer — it can reverse it entirely. In their classic experiment, identical outcomes were presented as either lives saved or lives lost, and participants’ preferences flipped. The content was the same. The question was different. This finding established that question formulation is not a preliminary step before the real cognitive work begins. It is the cognitive work. The frame you choose determines which features of the problem become salient, which trade-offs become visible, and which options feel acceptable.

John Sweller, at the University of New South Wales, formalised this dynamic in his 1988 cognitive load theory, published in Cognitive Science. Sweller distinguished between intrinsic cognitive load (the inherent complexity of the material) and extraneous cognitive load (complexity introduced by poor presentation or framing). A badly framed question doesn’t make the decision harder because the decision is harder. It makes the decision harder because the framing itself consumes cognitive resources that should be directed at evaluation.

The mechanism

Herbert Simon, the Nobel laureate who coined the term “bounded rationality” in a 1956 paper in Psychological Review, identified the core constraint: human decision-making is not limited primarily by the quality of available information but by the capacity to process it. We don’t improve. We satisfice — we find the first option that meets our minimum criteria, because genuine improvement across all variables exceeds our processing capability.

This means that the structure of a decision question directly determines the quality of the decision. A question that tries to hold too many variables simultaneously forces choosing the first good-enough option at a lower level — you’ll grab the first option that resolves the most salient tension, which is often not the most important one. A question that has been compressed to its essential core allows your limited working memory to focus on what actually matters.

Michelene Chi, at the University of Pittsburgh, published a revealing study in Cognitive Science in 1981 comparing how experts and novices categorise physics problems. Novices sorted problems by surface features — the objects mentioned, the visual appearance of the diagram. Experts sorted by deep structure — the underlying principles that governed the solution. The expert advantage wasn’t more knowledge. It was better problem representation. They had learned to see through the surface complexity to the structural question underneath.

This is precisely what happens when you compress a bloated decision question into a single clear sentence. You’re not removing information. You’re removing extraneous cognitive load — the qualifiers, conditions, and tangential concerns that obscure the deep structure of the choice. The first draft of your question reveals what you think you’re deciding. The simplified version reveals what you’re actually deciding. These are frequently different questions.

When you can’t compress a decision into one clean sentence, that’s not a sign the decision is complex. It’s a sign you haven’t identified which decision you’re actually making.

The practical implications

The brain-dump-then-compress method works because it separates generation from evaluation. Writing everything out in a messy first draft externalises the problem, freeing working memory from the burden of holding all the pieces simultaneously. The compression step then forces prioritisation: which elements are essential to the decision, and which are context, conditions, or separate decisions masquerading as part of this one? The gap between the two drafts is itself diagnostic information.

If you can’t simplify, the problem is usually compound. A question that resists compression to a single sentence almost always contains multiple decisions bundled together. “Should we expand into Europe?” might actually be three questions: “Do we have the capacity?”, “Is the regulatory environment navigable?”, and “Does the timing advantage outweigh the resource cost?” Each of these is answerable. The bundle is not. Identifying the sub-questions and addressing them sequentially is almost always more effective than trying to solve the composite.

In groups, question clarity prevents circular discussion. When a team can’t converge, the most common cause isn’t disagreement about the answer — it’s misalignment about the question. Different people in the room are answering different questions. Writing the decision as a single sentence on a whiteboard and getting explicit agreement that this is what’s being decided eliminates one of the most common sources of meeting dysfunction.

The bigger picture

There’s a persistent assumption that complex problems require complex thinking. In some domains — scientific research, systems engineering, long-range strategy — this is true. But for the vast majority of decisions, the opposite holds. The problem isn’t insufficient complexity in your thinking. It’s insufficient clarity in your framing.

The most effective decision-makers aren’t those who can hold the most variables in their heads simultaneously. They’re those who have developed the discipline to strip a problem down to its load-bearing structure before attempting to solve it. They know that a question containing six conditions and three caveats isn’t a thorough question — it’s a question that hasn’t been thought through yet.

Simplifying the question is not a preliminary step. It’s the hardest cognitive work in the entire decision process. Everything downstream — generating options, evaluating trade-offs, committing to action — depends on having asked the right question in the first place. If that foundation is bloated or unclear, no amount of analytical rigour applied to the answer will compensate.