Kenji Part 1 Flashcards
(26 cards)
Background to problem solving
Two states: where you are and where you want to be. Route between 2 not clear, can take multiple steps. Doesn’t have to be physical. Purposeful and goal directed,a controlled process, not habitual. Lack knowledge to gen immediate solution
Frontal lobes
Main location for problem solving. Goel and grafman 2000: architect w frontal lesion asked to design a lab, struggled to progress from problem structuring (id what the issue is) to problem solving but poor final design
Types of problems- not separate categories, a scale
Well defined: current position, goal and possible moves all well specified (chess) or ill defined where not e.g. I want to be happier. Well defined easier to research but less realistic. Knowledge rich: can only solve via relevant expereince like chess or knowledge lean: can be solved without
Monty hall problem also the 2 string problem
3 doors 2 w goats, 1 w Ferrari , mentally pick one in your head. Experimenter reveals a diff one to be goat, do you keep or switch to the other last option? 2 strings hanging, need to hold both at once but too far apart, have hammer on table- use hammer as pendulum
Insight part 1
Moment of realisation w reorganisation of elements of mental reps of a situation to yield a non obvious interpretation. Metcalfe & weibe 87: ps rated how warm they were to goal, for problems w insight, have sudden increase in warmth, for those without, saw a slow gradual increase. Jung beeman 2004: remote associates test, has insight, have 3 words and have to come up with word that can immediately follow all 3. When ease in superior temporal gyrus (but correlation/don’t know if experience or phenomenology)
Insight part 2
Ellis 2011: anagram task w one irrelevant letter, eye tracking. Ps reported insight but saw gradual decrease in fixation on irrelevant letter before decisionnot sharp. Suggests unconscious gradual accumulation of knowledge. Insight real in the way ppl expereince it but not separate, higher order may gradually arrive at solution but aware when threshold reached- but for all problems?
Things that facilitate insight
Hints, incubation- stop thinking about it and come back. Sio & ormerod 2009: meta anal: incubation led to small but consistent improvements in problem solving, particularly for creative solutions. Sleep a version of incubation. Wagner 2004: sleep increased insight. Number sequences where have to work out rule. Those who did training then sleep then test had sig more insight than those who no sleep. Don’t know spec to insight or problem solving, don’t know about the brain processes
Strategies
Newell and simon 72: problem solving from info processing perspective like computer : limited short term memory capacity, complex info processing, serial, can’t hold all steps at once. Trade off between accuracy and computational complexity, also time. Rely on heuristics , less processing
Heuristics
Means end analysis:id diff between state and goal, try to reduce by setting sub goals. Require info about location of final goal and what if better strategy. Hill climbing: choose the next step that brings you closer to goal but no long term plans. May get stuck at local maximum (point higher than surrounding but not highest possible/global max) think at best solution when not
Other strategies
Planning: useful when need sequence of steps. Koppenol gonzalez 2010: Tower of London task where ps who spent longer planning moves performed better and made less errors. Progress monitoring: track progress toward goal and switch if progress slow: macgregor 2001: performed worse when thought progress being made, when thought slow, more likely to switch strategies. Progress , fast and slow not defined
Expertise background
Definition: high level of knowledge and performance in a given domain acquired through a long period of systematic study/practice
Lab tasks often knowledge lean, abstract but problems in real world often knowledge rich
Expertise: chess
De groot 65: able to remember position so n cheese board but not better memory in general. Charness 2001: eye moves in first few seconds to more relevant prices in experts. Burns 2004: high correlation between normall and blitz chess/shorter version, this and char example of fast processes. Slow processes: van harreveld 2007: as time to make move decreases, skill diffs less predictive of outcome, moxley 2012: longer time to make move leads to better performance
Medical expertise
More serious,have false negatives in diagnosis by x ray- 30%. Krupinksi 2013: training pathologists breast biopsies, reduction in number of fixation per image and fewer fixation on non diagnostic regions. Kundel 2007: mammograms, mean search time 27s, time to fixate on cancer 1.13. Correlation between speed of fixation on cancer and detection performance
Expereince gone bad
Functional fixedness: inflexible focus on the usual function of an object, cant see new sue for it
Mental set: tend to use familiar strategy that was successful in the past even if not appropriate
Bilalic 2008: chess experts dont wlays fine quickest route to victory- often found longer solution based on familiar strategy
Wason selection task
4 cards on table- r,g,2,u, each has a letter on one side and number on the other, need to decide if r is always on the side of 2 but turning over few cards
Inductive reasoning
Drawing out general conclusions from particular examples/extrapolation E.g. alll swans I’ve seen are white so all swans are right. How science works as you gen hyp based on limited data. Data should be popperian-Karl popper says need falsification of hyp in contrast to confirmation. Confirmation can’t support hyp as can’t claim all swans are white but falsification can prove wrong
Examples of inductive- wason 246
Have to guess the rule that generated the numbers, ps give 3 numbers and experimenter says if follows rule or not -testing hyp. Most never guess it’s just increasing in magnititide due to confirmation bias, don’t try to falify by guessing something they don’t think
Weaknesses of inductive
Not real world as get immediate feedback and in world not fully informative. Rule is general so confirmation not appropriate. Other research says confirmation bias not present when someone else creates hypothesis
Heuristics in science
Science don’t always use hyp testing as unusual heuristics- pay more attention to unexpected result. What if testing-know some ideas would never make sense. Fuselgang 2004: looked at 417 results in bio, 223 found not what they expected so should say falsified by 88 blamed methods, but if other studies replicated, 61% change theory
Deductive reasoning
Also called conditional reasoning. Use symbols to stand for sentence and apply logical operators like if, and to reach concs e.g. p then q. P is the antecedent/premis and q is consequent/conc. consequent is binary- affirmed or not based on antecedent. Modus ponens- if p then q, p therefore q. Modus tollens- not q therefore not p- both correct but ppl refute mt more
Deductive reasoning part 2
Ven diagram with p inside q. Types of denial are affirming the consequent e.g can’t say q therefore p, if dog can’t say it must be pug. Or denial of antecedent- not p therefore not q isn’t true E.g. not a pug so not a dog. People do this in real word and use probabilities but logic is binary. If circles same size, more likely to accept in word
Deductive reasoning and false concs
Doesn’t matter if statements are actually true, the logic still follows e.g Sarah in Rio not Brazil so in Rio therefore not Brazil, need real world knowledge. De neys 2005: if brake pressed, car slows but gave counters like no gas or disabled like car broken. If ps use deductive, these should be ignored but they rejected/affirmed consequent. Markovits 2013: people accept that finger bleeds so must be cut more than window broken so must be rock and less possibilities p- circles smaller
Inductive reasoning
Going from antecedents to consequents logically. Wason more indicative as gen hyp and test it. 5-10% correct, maybe as cards mentioned in rule-matching bias or people may not understand deductive rules
Informal reasoning background
People used to just say deductive/inductive based on what ppl should do but now informal is what’s seen from behaviour. Based on knowledge like more likely to say drug safe as no evidence not than ghosts exists bc no one proved haven’t. Informal is probability like 80% chance of being correct. Straw man fallacy- picking apart someone else’s argument makes yours seem more likely.
