Psychology › Section G · Thinking and intelligence › Topic G.3
Problem solving
From fixing a jammed printer to choosing a university, life is a stream of problems and decisions. This topic sets out the main strategies people use, the mental shortcuts that make thinking fast, and the predictable traps those shortcuts set.
Strategies for solving problems
Solving a problem starts with defining it clearly, then choosing a problem-solving strategy, a plan for reaching a solution. Trial and error tries one possible solution after another until something works, as when you restart a router, then check the cable, then try another device; it is common but can be slow. An algorithm is a step-by-step procedure that guarantees a correct answer if followed exactly, such as long division or testing every combination on a three-digit padlock. Algorithms are reliable but can be impractical: the padlock has 1,000 combinations. A heuristic is a mental shortcut, a rule of thumb that is usually good enough and much faster, but carries no guarantee.
Useful heuristics
Working backwards starts from the goal and reasons back to the present. Breaking a problem into sub-goals turns a daunting task, such as writing a long essay, into manageable steps: choose a question, research, outline, draft, revise. Means-end analysis repeatedly asks how far you are from the goal and which step will most reduce that gap. People lean on heuristics most when they face too much information, too little time, too little information, or a decision that matters little. The psychologist Gerd Gigerenzer argues that such fast and frugal shortcuts often perform very well in real environments, so heuristics are not simply errors.
Worked example: working backwards
Zara's flight leaves at 14:10 and bag drop closes 60 minutes before departure. The train to the airport takes 75 minutes, and she wants a 15-minute buffer for delays. When must she leave home? 1. Start at the goal: bag drop closes at 14:10 minus 60 minutes, which is 13:10. 2. Add the buffer: she should aim to arrive by 12:55. 3. Subtract the journey: 12:55 minus 75 minutes is 11:40. 4. Answer: leave by 11:40. 5. Check forwards: leave at 11:40, arrive at 12:55, wait 15 minutes until 13:10, when bag drop closes, an hour before the 14:10 departure. Every constraint is met.
Two kinds of thinking
Daniel Kahneman, who shared the 2002 Nobel Memorial Prize in Economic Sciences for research on judgement and decision making, much of it done with the late Amos Tversky, popularised a distinction between two modes of thought. System 1 is fast, automatic and intuitive; it recognises a face or completes "bread and ..." without effort. System 2 is slow, deliberate and effortful; it checks a calculation or weighs a contract. Most daily thinking runs on System 1, which is efficient but relies on heuristics. Many of the errors below appear when System 1 answers a question that needed System 2, and the slower system fails to step in.
Mental set and functional fixedness
A mental set is a tendency to keep using an approach that worked before, even when a simpler or better one is available. In Abraham Luchins's water-jar problems of the 1940s, people who had solved several problems with a long formula kept using it on later problems that a short formula solved, and some failed problems the old formula could not solve at all. Functional fixedness is a kind of mental set in which you can see an object only in its usual use. In Karl Duncker's candle problem, people asked to fix a candle to a wall with a box of drawing pins and some matches often failed to see that the emptied box could be pinned up as a shelf.
Confirmation bias
Confirmation bias is the tendency to look for, notice and remember evidence that supports what you already believe, and to neglect evidence against it. In Peter Wason's 1960 task, people were told that the sequence 2, 4, 6 fitted a rule and asked to discover it by proposing other sequences. Most proposed sequences such as 8, 10, 12, received a yes, and confidently announced "even numbers rising by two". The real rule was simply "any three rising numbers". They rarely tested sequences that could prove their idea wrong, such as 1, 5, 20. The same pattern appears when people search only for news that agrees with them.
Shortcuts in judgement
Tversky and Kahneman described several heuristics that distort judgements. Availability: we judge how likely something is by how easily examples come to mind, so vivid, recent or heavily reported events seem more common than they are. Representativeness: we judge by how well something matches a stereotype or prototype, often ignoring base rates, how common each category actually is. Anchoring: an initial number pulls later estimates towards it, even when it is plainly irrelevant; in one study, people who saw a wheel of fortune stop on a high number gave higher estimates of the percentage of United Nations members that were African countries.
Framing, hindsight and overconfidence
Framing is the effect of wording on choices that are logically the same: a treatment described as having a 90% survival rate is preferred to one with a 10% death rate. Hindsight bias is the feeling, after an outcome is known, that you "knew it all along", which makes past events look more predictable than they were. Overconfidence leads people to rate their judgements as more accurate than they are; answers given with high certainty are often wrong more often than the certainty implies. Belief perseverance is clinging to a belief even after the evidence for it has been discredited. Deliberately considering how you might be wrong reduces several of these errors.
Worked example: naming the bias
Identify the bias in each case. 1. After a stock-market crash, a commentator insists the warning signs were obvious, though he had predicted growth: hindsight bias. 2. Asked whether English has more words beginning with K or with K as the third letter, most people say beginning, because those come to mind more easily, though third-letter K words are more common: availability. 3. A quiet, tidy man who loves books is judged more likely to be a librarian than a farmer, ignoring that farmers vastly outnumber male librarians: representativeness and neglect of the base rate. 4. A seller lists a used car at $12,000 and the buyer's counter-offer ends up close to that figure: anchoring.
Common mistakes to avoid
First, assuming heuristics are always bad: they are fast and often accurate, and fail in predictable situations. Second, confusing algorithms and heuristics: only an algorithm guarantees a solution. Third, mixing up mental set (sticking with a past method) and functional fixedness (seeing only an object's usual use). Fourth, mistaking availability (ease of recall) for representativeness (resemblance to a stereotype). Fifth, thinking intelligence protects you: these biases also affect experts and highly able people. Sixth, testing only cases that would confirm your idea, as in the 2-4-6 task.
Key takeaways
Problem solving uses trial and error, algorithms (step-by-step, guaranteed, sometimes slow) and heuristics (fast shortcuts such as working backwards, sub-goals and means-end analysis). Kahneman's System 1 is fast and intuitive and System 2 slow and deliberate. Obstacles include mental set (Luchins's water jars), functional fixedness (Duncker's candle) and confirmation bias (Wason's 2-4-6 task). Tversky and Kahneman showed how availability, representativeness, anchoring and framing distort judgement; hindsight bias, overconfidence and belief perseverance add further errors. Shortcuts are often useful, but considering the opposite and testing your ideas against disconfirming evidence improves decisions.
Flashcards
Flashcard
A plan of action for finding a solution
Test your understanding
Question 1 of 5
Which statement correctly contrasts an algorithm with a heuristic?
Question 2 of 5
In Duncker's candle problem, many people fail to use the box that held the drawing pins as a shelf. This is an example of:
Question 3 of 5
In Wason's 2-4-6 task, why did most people fail to find the rule "any three rising numbers"?
Question 4 of 5
A car is listed at $15,000. Buyers who see that price offer more than buyers shown the same car listed at $11,000. Which bias is this?
Question 5 of 5
A doctor's patients prefer a surgery described as having a 90% survival rate to the same surgery described as having a 10% death rate. This shows:
