lecture 9- problem-solving & reasoning

Question 1

what are the two states in problem solving?

Answer 1

1. current position (where are)
2. goal (where want to be)

route between them not clear and takes multiple steps

Question 2

what are the properties of problem solving?

Answer 2

• purposeful (goal-directed, have to be aware of goal, not driven by habit)
• controlled process (not fully automatic)
• BECAUSE we lack appropriate knowledge to generate immediate solution
• e.g making tea in new kitchen- route not clear as don’t know where everything is so multiple subgoals

Question 3

what did Goel & Grafman (2000) find about the role of frontal lobe in problem solving?

Answer 3

• architect with frontal lobe lesion
• asked to design office space
• layered, incomplete drawings
• some experience/knowledge from past so can put something down, but not clear
• struggled to progress from problem structuring to problem solving
• linguistic abilities and memory fine

Question 4

what are the categories of problems?

Answer 4

• well-defined: current position, possible moves, goal well-specified (chess- know possible moves for each piece, goal to take king)
• ill-defined: current position, possible moves, goal not well-specified (being happy- multiple ways, what is happy)
• knowledge-rich: only solvable via relevant knowledge (chess)
• knowledge-lean: can be solved without needing prior knowledge, all information contained in presentation of problem (lab experiements)

Question 5

what is an example of a knowledge-lean problem?

Answer 5

• monty hall
• 3 doors- goat behind 2 of them, car behind the other
• pick a door, reveals goat- do you keep door you selected or switch to remaining door?

Question 6

what is insight?

Answer 6

• sudden, discrete change in mind that contributes to solution
• subjective experience- ‘ah ha!’ moment where goes form not solved -> solved

Question 7

what did Metcalfe & Wiebe (1987) find when trying characterise how people got insight?

Answer 7

• p’s rated ‘warmth’ during problem solving task (warmth = proximity to solution_
• for problems including insight, sudden increase in ‘warmth’ (rapid, sudden progress)
• for problems solved without insight, gradual increase in ‘warmth’ (slow accumulation to goal)

Question 8

what did Jung-Beeman (2004) find about the neural associate to insight?

Answer 8

• gave 3 words: fence, card, master
• gave word that can precede: post
• indicates ‘insight’ for particular trials
• increased activity in superior temporal gyrus
• correlates, doesn’t mean causates

Question 9

is insight a separate process, or does it just differ in phenomenology (i.e subjective experience?)

Answer 9

• insight is ‘real’ in sense that people experience it (can measure- there is a neural correlate)
• but may not be separate cognitive process (higher order processes may gradually arrive at solution, but only become aware once threshold reached)
• gradual accumulation concept may differ in sense depending on nature of problem (well-defined, ill-defined etc.)

Question 10

what study did Ellis et al. (2011) do on gradual accumulation towards insight?

Answer 10

• anagram task: KAFMS, asked to solve anagram to find word with one letter not being in answer (MASK-F not in)
• record eye movement
• p’s reported sudden solution
• but gradual decrease in fixations on non-relevant letter before report (look at F less and less)
• gradual accumulation of knowledge towards solution despite subjective ‘a ha’ experience

Question 11

what is incubation?

Answer 11

• stop directly thinking about the problem for a period of time
• most have experienced the feeling of solving a problem during this problem
• Sio & Ormerod: incubation led to small but consistent improvements in problem solving
• Wagner: sleep is a version of incubation, sleep increased insight

Question 12

how can we facilitate insight/ increase probability of solving a problem?

Answer 12

hints, incubation and sleep

Question 13

what did Newell & Simon find about strategies for problem-solving?

Answer 13

• limited STM capacity (cannot hold too many pieces of info in mind at the same time- steps)
• complex info processing happens serially
• forced with tradeoff between accuracy, computational complexity and time
• humans rely on heuristics- computationally cheap rules of thumb that produces reasonably accurate answers

Question 14

what is means-end analysis?

Answer 14

• form subgoal in between to minimise distance between current location and goal, then move towards subgoal
• BUT this requires info about location of final goal
• what if better strategy to move away from final goal
• as can set wrong subgoal that moves further away from end goal before moving back towards

Question 15

what is hill-climbing?

Answer 15

• change present state to a state closer to end goal
• simpler than means-end as no explicit ‘subgoal’
• analogous to climbing hill by always moving to next highest point
• can get caught in local maxima when can go around which is less effort but get to global maximum easier

Question 16

what strategy also helps that’s not heuristics (means-end, hill-climbing etc.)? ehat did Koppenol-Gonzalez et al. (2010) do?

Answer 16

• planning before try to reach goal (useful when goal involves sequence of behaviours)
• tower of london task- certain arrangment in limited no. moves
• p’s who spent longer deliberately planning moves before acting performed better with fewer errors

Question 17

how does progress monitoring help problem solving? MacGregor et al. (2001)

Answer 17

• performance worse when p’s thought progress was being made
• when realised progress was slow, most likely to switch strategies
• but how is progress measured or defined?

Question 18

how is expertise important in problem-solving?

Answer 18

• lab tasks often knowledge-lean (abstracted away from real-world knowledge)
• expertise important for solving real-world problems that are knowledge-rich
• expertise definition- high level of knowledge and performance in a given domain acquired through a long period of systematic study/practice

Question 19

does expertise aid fast or slow problem-solving? use chess examples

Answer 19

• BOTH
• fast: experts eye movements fixate to relevant/critical pieces quicker in first few seconds (Charness et al. 2001)
• experts still good at blitz chess- high correlation between blitz and normal (Burns 2004)
• slow: as time to make move decreases, skill differences less predictive of outcome- longer means expertise has mor eeffect (Van Harreveld 2007)
• longer time to make move leads to better performance (Moxley et al. 2012)

Question 20

how does expertise look in medical decisions?

Answer 20

• breast biopsies: reduction in number of fixations, fewer fixations on non-diagnostic regions (Krupinski et al. 2013)
• mammograms of possible breast cancer: less than 1 sec to fixate on cancer for experts, correlation between speed of fixation on cancer and overall detection performance (Kundel et al. 2007)

Question 21

when does experience become bad?

Answer 21

• functional fixedness- the inflexible focus on the usual function of an object- can’t see novel use of object to solve problem
• mental set- tend to use familiar strategy that was successful in the past, even when it’s not appropriate (Bilalic et al. 2008- chess experts don’t always find quickest route to victory, often find longer solution based on familiar strategy)