Citing Baddeley (2003)*, the Oxford Handbook of Thinking and Reasoning defines working memory as ‘the ability to maintain and manipulate information in short-term memory while resisting interference.’
A recent report by the National Academy of Sciences partially dedicated to helping students with problems in science and engineering points out that ‘complex problems contain a considerable amount of relevant information, which often must be held in students’ working memory. […] As the working memory load of a problem increases, success rate decreases.’ The report also mentions that research shows ‘that chemistry students who have a larger working memory capacity […] perform better on problem solving.’ This is not limited to chemistry; in fact, the report points out that ‘a large body of research in cognitive psychology documenting enhanced performance (e.g., language comprehension, problem solving, etc.) when more working memory resources are available to complete the task at hand.’
So let’s look at we can do to have more working memory resources available.
Define your problem effectively in a PIC
Asking the right question is hard. That’s why most people don’t do it. To start with, it’s not just the question, it’s the entire introductory flow, with the situation and the complication (see more here). Since it’s so difficult to decide which facts should be included in the introductory flow and which shouldn’t, people go for the safe bet: they include everything, thinking that the reader will choose for himself. That is, they delegate the hard work. But they also overload unnecessarily the reader’s working memory.
As a result you end up with vague introductions called ‘background’ with anywhere between two and 300 pages of material that is all relevant to the subject at hand but only variable related to the problem you want to solve.
People operate with the motto ‘more is more’, but doing so is counterproductive. So more isn’t more.
But neither is less, because if you forget a single key piece of information, your introduction won’t hold water.
So here, really, the right quantity is more. It’s a tough job, but summarizing your introduction in a problem identification card will help. (Learn more about this here.)
Use issue trees as road maps
Another moment where you risk to overload your working memory is when you are gathering evidence. Consider having to answer the question ‘how do we improve learning abilities in Costa Rica?’ A typical way to approach this is to start with finding out about current learning abilities in Costa Rica. And maybe historic trends. And then maybe you’ll want to break down this information by age groups, or ethnicities, or geographical location, or wealth. Great! Now time to get the info, but that’s when you find out that this data doesn’t really exist—at least not at the level of reliability that you need—so you need to get it first. So you hire a team of sociologists, they spend eight months on the job and give you a beautiful, 900-page report exhaustively listing all the information you needed in a reliable way. Read the report and you will know about learning abilities in Costa Rica. But you will have little insight into answering your key question: how do you improve those.
That’s because you’re boiling the ocean: you are spending an immense effort to move forward by gathering info about the subject. Of course, you are moving forward, but it is really inefficient because you are gathering a lot of material that isn’t really relevant. But, even more problematic, is that you won’t know what to do with this info, because you will have saturated your working memory.
What you need to do instead, is work forward in solving your problems: don’t start with the evidence, start with the questions. This is where issue trees are critical.
By breaking down your key question in an issue tree, you look at all aspects of it, formulate hypotheses, identify which analysis you need to conduct to test your hypotheses, and what data will fuel your analysis. Only then are you ready to get the data, because only then will you be able to answer the ‘so what?’ associated with each piece. Chances are that the data you are looking for will be in the 900-page report you would have commissioned, but chances are that a large number of these pages wouldn’t help you directly in solving your problem. So you’re a lot more efficient in your solution process—and you’re also keeping your working memory, well, working.
Issue trees are particularly useful here because they allow you to relate bits of evidence straight to the questions that you are trying to answer. Take the example of the dog who disappeared from your friend’s house and you don’t know if he has been kidnapped or if he just escaped. So you define your key question, you break it down in an issue tree, and you formulate your hypotheses. Then you go fish for information (speak with neighbors, the police, the vet’s office, etc.) and gather this evidence straight into the tree, identifying on the spot which evidence is relevant to the hypothesis you’re testing and what it means for the hypothesis.
The issue tree becomes your road map, helping you structure and navigate the analysis that you need to conduct. It also becomes a central repository for your evidence: the one place where you capture what each bit of information means for each of your hypotheses. Add a color code (say red to signal evidence that contradict a hypothesis and green for evidence that is compatible with it) and, at a glance, you get to see where you stand.
How’s that for a working memory boost?
* Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience, 4(10), 829-839.