Cognitive Psych - Exam 4

Semantic memory

organized, general knowledge about the world; important for chunking, inferences, building expertise
- used for building categories and concepts

What can semantic memory refer to

general knowledge (facts), lexical knowledge (word relations), and conceptual knowledge (abstract concepts)

Category

set of objects or ideas that have common underlying factors
ex. grape --> fruit, wrench --> tool

Concept

mental representation of a category; essential for understanding (integrate new information you learn into concept)
ex. memory --> sensory, working, long term

Feature

specific attribute of a dimension

Similarity based approach

judge the similarity between the target object and some standard memory

Classical view

items belong to certain categories if they contain certain features
- must contain certain features
- when classifying a new object, compare it to list of features

Limitations of classical view

typicality effect

Typicality effect

some items are seen as better members of a category than others, all members of a category are not created equal - graded structure, classical view says it's either in a category or not

Typicality ratings (Rosch and Merus 1975)

basic method
- given pictures of basic level categories
- rate most (1) to least (7) typical
results
- most: chair, least: mirror --> furniture
- most: car, least: wheelchair --> vehicle
- most: orange, least: blueberry --> fruit
- most: gun, least: whip --> weapon
good for well defined categories
- limitations: categories have fuzzy boundaries (not all categories have "necessary" conditions)

Prototype approach

based on abstractions and prototypes

Abstraction

based on previous experience

Prototype

idealized "average" example that includes typical features
- represents best member of that category
- formed via abstraction
- not concerned with necessary features

Prototype formation (Posner et al. 1967)

method
- present array of dots
- prototype never presented
test
- recognition test with old and new arrays
- include prototype
results
- high mis-identification of prototype
conclusion
- extraction of prototype based on presented examples
- typicality

Typicality

members can be more or less like the prototype

Family resemblance

objects in a category resemble one another in some way
- no single feature is shared by all members of a category
- but they all share at least one feature
- the most prototypical items tend to be the ones that share the most features

Characteristics of prototypes

have privileged status, judged faster, semantic priming effect
- semantic priming facilitates responses to prototypes more so than non-prototypes

Superordinate level

broad (ex. animal)

Basic level

more specific (ex. dog and cat)

Subordinate level

more specific (ex. lab, pug, and siamese)

Structure of categories

superordinate level, basic level, and subordinate level

Basic level has special status

generally most useful (ex. hammer vs. tool), most often used to name objects (ex. dog vs. animal), and most likely to produce semantic priming effect

Influence of society on typicality ratings (Novick 2003)

measured the length of time since the 9/11 attacks
task
- rank order 10 vehicles in terms of typicality most (1) to least (10)
groups of students
- prior to 9/11
- 5 hours after, 6 days, 1 month, 4.5 months
results for "airplane"
- less typical before, more typical close after, less typical as more time persisted
conclusion: society and culture can impact typicality of exemplars

Prototype approach limitations

unclear how they're formed, what's in a prototype, not everything we categorize is compared to some "idealized" mean, seems like we store some specific information

Exemplar approach

no single prototype, many examples; membership based on similarities to exemplars; we compare new items against exemplars (rather than one "idealized" prototype, does not require abstraction); exemplar chose can depend on context
- can explain typicality
- cognitive economy
- no abstraction required (abstraction requires loss of details, if we store exemplar, don't lose details)

Combination approach

for some categories we store exemplars (categories with fewer members (ex. tropical fruit)) and for other categories we store prototypes (categories with many members (ex. fruit))

Systematic memory organization

concepts represented as nodes
- individual units in network
- each node represents a category/concept
interconnections among related nodes
pattern or spreading activation

Evidence for network models (Collins and Quilian 1969)

method
- sentence verification task
- correct or not
- manipulated how far participants had to mentally travel to make decision
results
- the further people had to travel in the semantic network, the longer it took (people were slower to respond to questions about properties compared to category membership)

Parallel distributed processing (PDP) approach

network of neuron-like units
- distribution: activation occurs in many parts of the brain
- parallel: activation occurs simultaneously
- address how various categories interact with each other (ex. chicken belongs to food, animal)
- can take context into account (hearing a person talk about a chicken in a restaurant activates different parts of the network compared to a farmer talking about a chicken)
- important features: connection weights (modified with each new experience), a unit can influence neighbor units via inhibition (-) or excitation (+)

Evidence for network models

typicality effects (faster to confirm because nodes for typical items are closer) and priming effects

Priming effects in network models (Meyer and Schvaneveldt 1971)

lexical decision task
- participants see 2 strings of letters and respond yes if both are words or no if they're not
- reaction time: faster for related words
why: related words are closer together than modes for unrelated words

More complex priming (Bargh et al 1996)

method
- participants completed a word scramble task
conditions
- polite words, rude, neutral
- when finished, participant finds researcher who is having a conversation
DV
- % of participants who interrupt the researcher
results
- polite: 20%
- neutral: 40%
- rude: 60%
why: taks activated rude concepts, which primed participants to be rude

Schema

generalized, well integrated knowledge about a situation, event, or person

Script

one type of schema, sequence of events, high familiarity

Schemas and scripts can affect what we remember

memory for schema inconsistent items (vivid or surprising things), reconstruction error (schema consistent)

Schema inconsistent influence on memory (Brewer and Treyens 1981)

method
- left alone (35s) in office
- ask to remember what's in the office
results
- remembered items with office schema
- forgot inconsistent (ex. picnic basket)
- reconstruction errors (books when not there)

Memory integration

our background knowledge encourages us to take new information in a schema consistent fashion

Stereotyping

over simplified image or idea of a particular category of people

Implicit association task

without awareness, pair related words faster than unrelated

Congruent mapping

people are faster and more accurate at classifying stimuli

Incongruent mapping

people are slower and less accurate at classifying stimuli

Memory accuracy

when asked about how memory works, 63% agreed it was like a video camera; 37% thought one confident eyewitness was enough evidence to convict

Constructivist approach

we construct memories by integrating information to form a coherent event

Brown v. Board (1954) - history of psych and law

first time psychological research was cited in a legal case
- Kenneth Clark and Mamie Phipps Clark presented their "doll test" studies where black children preferred white dolls over black dolls

Advisors

expert testimony, trial consultant, create amicus curiae briefs to inform the court about recent findings

Evaluator

evaluate the effectiveness of programs or public policies, clinical diagnoses

Reformers

gather evidence to change policies

Daubert v. Merrell Dow Pharmaceuticals (1993) - expert testimony

established criteria for judges to decide on the admissibility of expert testimony
- science to be presented is falsifiable
- must have undergone peer review
- must have a known rate of error
- should be generally accepted by the scientific community
these criteria are vague and judges often lack the scientific background to make educated decisions

Judicial system goal and problem

goal: identify and prosecute the guilty, protect the innocent
problem: innocent are prosecuted and guilty set free
- why: eyewitness

How important eyewitness is (Loftus 1974)

presented a hypothetical court case
three groups
- no eyewitness (18% conviction)
- eyewitness (72%)
- discredited eyewitness (68%)

Eyewitness ID before DNA (Wells et al. 1998)

first 40 cases in the U.S. in which DNA was used to exonerate a previously convicted person, each case represents an innocent person who was convicted and served time in prison
- 90% involved eyewitness ID (these were wrongful convictions)

Confidence and accuracy (Leippe and Eisentadt 2007)

low correlation between confidence in memory and accuracy of eyewitness (.3-.5%)
- jurors are more likely to believe a confident eyewitness

Variables affecting accuracy

1. estimator variables: not under the control of the justice system (ex. characteristics of the witness, event, media coverage)
2. system variables: under the control of the justice system (ex. line up composition, line up instructions)

Source of memory distortion

encoding (quality of the viewing conditions, emotional stress), storage (introduction of new information, misleading/suggestive questioning), and retrieval (retention interval - time between event and testimony, line up composition/format)

Yerkes-Dodson curve

moderate arousal improves memory, low (ex. invisible gorilla study) and high (ex. high stress) arousal impairs memory, can disrupt quality of encoding

Presence of a weapon (Loftus, Loftus, Messo 1987)

method
- participants were instructed that they were going to view a series of slides that their eye movement would measure
- shown 18 studies of a "taco time" restaurant
groups
- control: man buys something with check and gets change
- weapon: robbery with gun
results
- eye fixation: more on gun than check
- fixation duration: small effect, gun more often
- weapon focus effect: people tend to focus more on the weapon relative to a mundane stimulus

Presence of a weapon 2 (Loftus, Loftus, Messo 1987)

same as first, but given MC test 15 minutes later and given a 12 person lineup
results
- memory test: control remembered more
- identifications: weapon group worse memory and ID
why
- unexpected/unusual object
- highly stressful event
- pulls attention from other aspects of the event (including the perpetrator)

Influence of stress on ID accuracy (Morgan et al. 2004)

method
- 50 military personnel undergoing POW survival training, 1 day leaving 15 person line up and ID
conditions
- high stress: physical confrontation
- low stress: tricked but not physically confronted
results
- low stress: more correct choice
- high stress: more incorrect choice

Arousal matters

type of arousal
- experienced emotion (ex. joy (can eliminate own race bias), fear, neutral)

Social contagion effect

others' memories can "infect" ours
-

Social contagion effect (Principe, Kanaya, Ceci, and Singh 2006)

method
- preschoolers (3-5)
- see failed magic trick pulling bunny out of hat
interviews
- suggestive: "what did the rabbit eat"
- neutral
groups
- overheard: overheard adults saying magician failed because rabbit was loose in the school
- classmate: classmates of the preschoolers in the overheard grouo
- control: no additional exposure
- witness: saw it but didn't hear anything
results
- did you see the bunny?
- suggestive: control < overheard < classmate < witness
- neutral: control < overheard < classmate < witness
more false memories when influenced by others' memories
- true of neutral and suggestive questioning

Misinformation effect

distorts memory for original event, form of retroactive interference

Post event encoding (Loftus, Miller, and Burns 1978)

method
- included red Datsun at stop sign and the car later hitting a pedestrian
groups
- consistent info: stop sign
- inconsistent info: yield sign
- asked to select slide previously seen
results
- consistent: more likely to select correct slide (stop sign)
causes for misinformation effect
- occurs between event and recall (retention interval)
- incorporate into remembering of event

Suggestive questioning (Loftus and Palmer 1974)

method
- watch 7 videos of traffic accident
- complete questionnaire focusing on speed of vehicles
- "how fast were the cars going when they ___ each other" (smashed/collied/bumped/hit/contacted)
results
- less aggressive word = slower speed (more reported broken glass for "smashed")

Personal memory (Wade et al. 2002)

method
- shown 4 pictures from childhood events
- 3 were real and 1 fake (photoshopped into hot air balloon ride)
- interviewed participants 3 times about events
results
- interview 1: more memory for true event, some endorsement of false event
- interview 3: similar amount of memory for true and false events, slightly increased memory for false event over time

Lineup identification (goal and outcomes)

goal: identify the perpetrator
four outcomes
1. correct identification (hit)
2. incorrect identification (false alarm)
3. identification failure (miss)
4. correct rejection if perpetrator not there

Age and unconscious transference (Perfect and Harris 2003)

method
1. each person committed a crime, given details, told to remember as much as possible
2. mug shot photos: did you see this person
- 1 week later: memory test for perpetrator
results
- young: no sig. difference
- old: strong own age bias
why
- expertise
results for perpetrator absent lineup
- correction rejections: young > old
- transference errors: old > young
- other foil errors: old > young
aging effect: young made more correct rejections
transference errors: more by older adults, own age effect for young adults
- why: face is familiar because you've seen it before

Confirmatory feedback (Wells et al.)

rate confidence
- yes, good job vs. no feedback
increased eyewitness confidence
alters memory for original event
- rate original view as better
- indicate they could make out more facial details

Characteristics of a line up

line up format can be sequential or simultaneous; line up composition (suspect present or absent)

Sequential line up

see one picture (person) at a time; make decision on each

Simultaneous line up

see all pictures (people) at a once; probably a relative process (deduction), more likely to make IDs but also increased mistaken and correct IDs

Line up (wells 1984)

view mock crime, 1/2 viewed suspect present, half viewed suspect absent
- results
present: correct 54%
absent: filler ID 68%

Problem

occurs when there's an obstacle between a present state and goal state, and it's not immediately obvious how to get around the obstacle

Problem solving

the process necessary to move from an initial state to a goal state

Key questions from Gestalt approach

how do people represent problems in their mind and how do people organize and structure their mental representations?

Problem representation

key point: understanding what the problem is; to understand a problem, people must pay attention to relevant information and ignore irrelevant information
ex. elevators in NYC
- complains elevators are too slow --> added mirrors outside elevators
ex. baggage claim
- complaints about waiting for bags --> create baggage claim
real problem: getting bored waiting

Problem restructuring

how do people organize and structure their mental representation of a problem

Insight

a sudden comprehension, realization, mental representation of a problem to a mental representation that wasn't initially obvious

Insight (Metcalf and Wiebe 1987)

method
- give insight problems (not stepwise)
- give algebra (non-insight problems - stepwise, see getting closer)
- make warmth judgments every 15s (1 not close to 7 very close)
results
- insight: sudden sharp increase in warmth judgments
- non-insight: steady increase in warmth judgments

Information processing approach (Newell and Simon 1972)

logic theorist computer program; computer program designed to stimulate human problem solving
- problem solving can be broken down into a series of steps

Tower of Hanoi (Newell and Simon 1972)

added intermediate state: any state a process is in but before goal state
operators: actions that take problem from one state to another (usually governed by rules)
subgoals: small goals that help get you to go the goal state

Mental set

a preconceived notion about how to approach a problem, often determined by a person's experience of what has worked in the past

Water jug problem (Luchins 1942)

mental set group: solved all following same formula followed by easier formula
no mental set: only solved easier problems
results
- mental set: 23% used simple formula at end
- no: 100% used easier problems

Candle stick problem

included fixation and functional fixedness
- when some people were given an empty box, they were 2x as likely to solve the candlestick problem

Fixation

when people focus on a specific characteristic of the problem that keeps them from finding a solution

Functional fixedness

occurs when the ideas a person has about an object's function inhibits their ability to use the object for a different function

Mental fixation

inability to move past a particular thought or cue

Mental fixation (Smith, Ward, Shumacher 1993)

method
- draw a never before seen alien life form
groups
- no examples
- examples
- examples (told not to conform)
results percent including parts
- no example: least amount of included parts
- example: medium amount
- not conform: largest amount

Algorithm strategy

step-by-step procedure for solving a problem
ex. LSSTNEUIAMYOUL
- try to identify first 2 letters
- reject unlikely letter combinations (ex. LS and LT)

Exhaustive search strategy

trying out all possible combinations until you find a solution
- will always get you a solution but may not always be efficient

Analogy strategy

using a solution from a similar problem to solve a current problem; includes underlying and surface structure
- can only be used when the problems have the same underlying structure
- people often focus on surface structure, which makes us bad at analogical problem solving

Underlying structure

the heart of the problem

Surface structure

the physical features of the problem

Analogical problem solving from Duncker (1945) radiation problem (Gick and Holyoak 1980)

same radiation problem from Duncker (1945), except some participants read a story about a general wanting to attack a fortress before computing the radiation problem
- without story: 10% solved radiation problem
- with story: 30% solved radiation problem

Analogical problem solving

includes the target problem and the source problem; 3 steps
1. notice the analogous relationship
2. map the correspondences
3. applying to generate a parallel solution

Target problem

the problem you're trying to solve

Source problem

another problem that shares similarities with the target problem and that illustrates a way to solve the target problem

System 1 (heuristic model)

automatic, non-conscious; thinking quickly and without much deliberation; efficient but imperfect

System 2 (analytic mode)

effortful, conscious; thought that's relatively slow, deliberate, and controlled; more cognitively demanding

Inductive reasoning

moving from specific pieces of data toward a general conclusion or rule
ex. my boss hates when I'm late to work, I'm stressed when I'm late to work
- general conclusion: being late is not good