Topic 4 Learning Theories - Content Flashcards
What is Classical Conditioning?
Learning by association:
- pair a new stimulus with an existing stimulus response link -> learn to associate the two stimuli & respond in a similar manner to both.
What’s the unconditioned stimulus, unconditioned response, neutral stimulus, conditioned stimulus & conditioned response?
Unconditioned Stimulus = Stimulus producing a natural unlearned response
Unconditioned Response = Response that occurs naturally without any form of learning (reflex action)
Neutral Stimulus = Env stimulus that does not produce a response itself (no association)
Conditioned Stimulus = Stimulus associated with an unconditioned stimulus so that it now produces the same response
Conditioned Response = Behaviour that is shown in response to a learned stimulus
———– EXAMPLE:
Food for your cat -> Unconditioned stimulus
Salivation from the cat/Response to the food -> Unconditioned Response
(No learning required as salivation is an automatic response)
Sound of opening can of cat food -> Neutral stimulus
(Produces no effect as yet & no innate reflex response to this sound)
(Neutral stimulus must be paired with UCS to evoke a response -> forming an association between both)
Neutral Stimulus becomes conditioned stimulus
Unconditioned becomes conditioned response
Thus, cat salivates wen presented wit food -> sound of opening can paired with food multiple times -> trigger conditioned response
What is Stimulus Generalisation?
Conditioned stimulus produces same behaviour to similar after response has been conditioned
- stimulus triggering a reaction doesn’t have to be the exact one involved in the process of learning -> more similar it is, more likely it will produce a conditioned stimulus
E.g. cat running to any tin being opened (not just cat food)
What is Stimulus Discrimination?
E.g. Runs to can at a certain time of day/Cat only responds to a tin of food but not a glass jar
What’s Extinction (in terms of classical conditioning)?
- Removal of a behaviour
- If CS constantly presented without UCS -> gradually learn to disassociate the two stimuli’s -> e.g. not salivate on hearing a can opening
- BUT association may not be entirely lost
What’s Spontaneous Recovery?
- Accelerate form of learning association
- After extinction, if old CS paired with UCS -> association quickly learned
How does Classical Conditioning link to the evolutionary approach?
E.g.
- Ancestors avoiding specific berries that lead to consequences & also avoid slightly different berries that looked similar to the previous ones to avoid risk
- Also if they took risk of eating those berries & had no negative consequences -> able to make similar distinction in future & provide hunter -> provide hunters/gatherers with another valuable food -> survival
Explain the experiment of Pavlov (1927) salivation with dogs
Found when a dog encounters stimulus of food -> saliva starts to pour
- Observing dogs drooling without the proper stimulus (food) by reacting to the lab coats of his assistants -> assistant presented the food when wearing lab coat -> dogs responded of the lab coat (without food) by drooling
What are the strengths for Pavlov (1927) experiment?
- Controlled experiments -> env of dogs controlled except variables being tested
- Found any stimuli would produce conditioned responses -> so controlled all stimuli other than the stimulis being studied
- Objective, Scientifically credible results
Pavlov repeated many classical conditioning experiments -> reliable -> continually found that conditioned stimuli would produce conditioned responses using buzzer, etc.
What are the weaknesses for Pavlov (1927) experiment?
- Unable to measure brain activity in any direct way -> assumed what was happening in cerebral cortex
- Lack validity -> highly controlled -> not natural -> dogs in a chamber & no other stimuli were present -> not real life behaviour
- Ideas about evolution led Pavlov to generalise dogs findings to humans -> both are different -> humans have a different higher order processing
What’s Operant Conditioning?
Learning by Consequence
What is the study of Thorndike that produces the law and effect?
Thorndike (1911) -> ‘instrumental learning’
- Puzzle learning
- Kitten had to solve puzzle to escape & receive reward
- More trials = faster
Law of effect -> Behaviour followed by nice consequence will be replicated, behaviour followed by bad consequence will be withdrawn
What’s the Skinner study?
Placed rats in a ‘Skinner box’ containing a lever, light & food dispenser
- If rat pressed lever -> light came on & food rolled down the dispenser -> positive reinforcement -> encourages rat to press lever more
Another Variation: - Negative reinforcement -> electric shock turned off when rat presses lever -> encourages rat to press lever more
ABC Model:
Antecedent -> skinner box presents stimulus that triggers behaviour
Behaviour -> response made that can be observed as an outcome of the antecedent
Consequence -> Reward/Punishment after behaviour
Stimulus -> response link is only learned if positive consequence, weakened if negative
What are the types of reinforcement & what do they mean?
Positive Reinforcement -> Giving smth good (reward)
Negative Reinforcement -> Removing smth bad (to make conditions better)
Positive Punishment -> Giving smth bad
Negative Punishment -> Removing smth good
What are the Schedules of Reinforcement?
Continuous Reinforcement -> desired behaviour reinforced every time it occurs
Partial Reinforcement -> desires behaviour only reinforced some of the time
Behaviour learned through partial takes longer to learn but is more resistant to extinction
Partial Reinforcement can be broken into 4 schedules, what are they?
Fixed Interval -> rewarding of a behaviour after a preset amount of time has passed
Variable Interval -> rewarding after a set amount of time has passed
Learning takes longer when fixed BUT response rate is higher towards the end
With fixed, there’s a scalloping effect -> response rate drop dramatically
Fixed ratio -> Behaviour reinforced after a preset number of responses
Variable ratio -> Behaviour reinforced after a number of correct responses
Skinner said variable is good for maintaining
What’s Behaviour Modification?
- Extinguish undesirable behaviour & replace with desirable behaviour & reinforce it
Shaping: Skinner developed ‘Successive Approximations’
- related, general, desirable behaviours rewarded
- rewards become more selective
- step-by-step, gradual process
Used as therapy -> ADHD & OCD
Token Economy:
- Encourage desirable through reward
- Reduce undesirable through punishment
- Tokens are secondary reinforcers that can be exchanged for primary reinforcers
- token only given for desirable behaviour
Used in schools/prison
What’s Social Learning Theory?
Learning by Observation
- Behaviour of a model is observed then imitated -> more likely to copy model if observer identify with model (age, sex, status)
More likely to copy if consequences are rewarding rather than resulting in a punishment -> vicarious reinforcement
What are the Stages of Social Learning?
- The behaviour must be modelled which means it must be carried out by a role model like a parent, friend or celebrity.
- The observer must identify with the role model – normally because they are similar in appearance, gender, interests, etc
- The behaviour must be observed
Attention: you must be attending to the behaviour
Retention: you must retain it in your memory
Reproduction: you must be capable of carrying out the behaviour
Motivation: you must have a reason to carry out the behaviour (eg a reward) - Behaviour is imitated
What’s the aim & procedure for Bandura (1961) original Bobo doll experiments?
Aim: investigate if social behaviors (i.e., aggression) can be acquired by observation and imitation.
Procedure: Lab exp
- 36 boys & 36 girls, aged: 3-6 - matched children in terms of aggression, grouped them -> matched pairs design
Stage 1: Modelling -> 24 children watched a female/male model behaving aggressively towards a toy called ‘Bobo doll’ -> some cases used hammer, throwing it in the air
Another 24 children exposed to a non aggressive model -> played in a quiet manner -> ignoring bobo doll
Final 24 children -> control group -> not exposed to doll
Stage 2: Aggression Arousal -> each child taken to a room with attractive toys -> as soon as child plays with toys, experimenter told child these were experimenter’s best toys
Stage 3: Delayed Imitation
- Room contained some aggressive toys & some non-aggressive toys
- child in room for 20 min & observed through a one way mirror
What are the results for Bandura (1961) original Bobo doll experiments?
- 10% modelled aggression
- 13% moddeled non agression
Children who observed the aggressive model made far more imitative aggressive responses than those who were in the non-aggressive or control groups. - The girls in the aggressive model condition also showed more physical aggressive responses if the model was male, but more verbal aggressive responses if the model was female.
- Boys were more likely to imitate same-sex models than girls. The evidence for girls imitating same-sex models is not strong.
- Boys imitated more physically aggressive acts than girls. There was little difference in the verbal aggression between boys and girls.
What are the strengths & weaknesses for Bandura (1961) original Bobo doll experiments?
STRENGTHS:
Controlled/Standardised procedure - lab exp - variables controlled (gender, time) -> replication -> reliability
Cause & effect can be established -> Model did have an effect on child’s subsequent behaviour because all variables other than independent variable are controlled
Application to therapy & real life situations -> useful in showing effects of role models on aggressive behaviour
WEAKNESSES:
Lab exp -> lack ecological validity -> situation of child & adult: limited social interaction (child no chance to influence adult), stranger (not family, usually take place with family)
Unethical -> encouraging aggressive behaviour in children & may lead to long term consequences in children
Generalisability: American children -> cannot be generalised to other cultures -> culturally biased - conducted in one nursery setting -> may not apply to children who don’t regularly attend nursery
What’s Bandura 1963 experiment?
- To find out if children would become more aggressive if exposed to an aggressive role model in film or in a less-realistic cartoon compared to watching a live model
- 96 children, 48 boys and 48 girls, aged 3-5, recruited from Stanford University Nursery School (an opportunity sample).
- Live model
- Filmed model (of real person) -> adult female dressed as a cat
- Cartoon model
- Control -> No model
-> Children put into model room -> there were toys -> exp conditions then saw a model enter the room with a Bobo doll -> acting aggressively (through the diff conditions, cartoon, filmed)
-> Then children placed in arousal room -> where there were some appealing toys (told after a few min not to play with toys -> making them frustrated)
-> Finally - brought into Observation room -> contained mixture of aggressive & non-aggressive toys
results: - Children will imitate filmed/cartoon aggression in the same way as live aggressive role models (experimental groups)
- Control group half amount of aggression
- Watching violence is not cathartic -> encourages more violence
strengths & weaknesses for Bandura (1963) exp?
STRENGTHS:
- > standardised procedure -> same rooms for pp, same model for every pp in that specific condition
- > Inter-rater reliability -> filmed the study -> can be watched by anyone
Application:
Can be applied to real life in terms of bringing up children. It shows that if a child’s role models are not aggressive, they are also less likely to be aggressive. Cartoon (Tom & Jerry) influencing violence on to children
-> Catharsis argument -> violent sports on tv defend violence as it allows boys to “vent” their aggression -> Bandura disagrees -> catharsis causes a build up of more aggression
Generalisability:
- 96 children -> large sample
WEAKNESSES:
Ethics -> distress to pps, influence violence behaviour (chance that they will be more violent in the future) -> normalising unhelpful behaviours, no valid consent gain & couldn’t withdraw
Reliability: no guarantee all children saw all the actions
Generalisability:
- children can’t be generalised to adults -> children more quickly to pick up new behaviours due to them still learning