Class9 Instrumental Conditioning

8/6/2019 Class9 Instrumental Conditioning

http://slidepdf.com/reader/full/class9-instrumental-conditioning 1/24

LEARNING

THEORIES

Behaviorist Cognitive

Classical

Conditioning

Instrumental

Conditioning



Behaviorist/Associative Learning «

Learn to develop automatic Response to a Stimulus (stimulus-response learning)

This type of learning occurs through forming associations between a new

stimulus (CS) and an old stimulus (US)



I nstrumental (Operant) Conditioning

Whether you perform a behavior again depends on what happened when youperformed that behavior in the past

Skinner ¶s bird experiment:

Suppose there are two buttons -- bird can choose button A or button B (so two different behaviors)

Choice of A followed by food (reinforcer/reward)

Choice of B followed by shock (punishment)



I nstrumental (Operant) Conditioning

Over time, Button A will be chosen automatically!

Button B will be avoided automatically!



Learning has taken place «

Learning results in automatic response to a stimulus (avoiding or choosing

button)

Again: stimulus-response learning



Learning through I nstrumental

Conditioning Reward for behavior increases chances

of behavior

Punishment for behavior decreases chances of behavior



I nstrumental Conditioning also

involves CS-US association Classical Conditioning:

Cowboy --- Liked

Marlboro --- Liked (US, CS, CR??)

I nstrumental conditioning: Food ------- Like

Button ------ Hitting button

US, CS, CR??



H ow can Marketers use IC?

I nstrumental conditioning is used to influence behaviors

What behavior do you want consumers to do ??

H ow? Rewards ?

Punishments ?



M arketing Example of IC Buy Brand ---- Get Discount ---- Buy

Brand again

What is US, C S, C R?



M arketing Example of IC Discount --------- Like

Brand -------------- Buy Again



If Success ful «.

I nstr um enta l C onditioning C an produ ce a u to m atic bu ying and

repu rchase o f yo u r brand (ju st l ike bird

wi ll a u to m atica ll y hit bu tton A and not B)

Au to m atic repu rchase (witho u t

consideration o f other a l ternatives) =what type o f decision m aking?



I ss u es in I nstr um enta l Learning

Reinforcement Schedules: Continuous schedule: reward every time

behavior is performed

partial schedule: reward sometimes

Example: giving discount/coupons etc.after purchase Continuous reinforcement = discount every

time

Partial schedule = discount onlyoccasionally.



C ontin u o u s vs . Partia l Schedul e Which produces faster learning? (I.e.,

which one will make the consumerslearn the ³brand ± discount´ association

faster?)

Which produces more permanentlearning (I.e., which one will lead to

more long-term positive effects onconsumers¶ desired behavior ± buyingyour brand?) Airlines??



I ss u es in I nstr um enta l C onditioning «

Reinforcement vs. Punishment

Instrumental Conditioning occurs

when behavior is followed byreinforcement or punishment



R ein f orce m ent «

Reinforcement = rewarding a behavior toincrease future possibility of that behavior

Two Types of Reinforcement«



Tw o types o f R ein f orce m ent «

Positive reinforcement (behavior followed bygood consequence; e.g., good after salesservice) Here, ³Reward´ = Getting something positive

because of behavior

Negative reinforcement Negative State First ± then Behavior ± then Removal

ofN

egative Stage e.g., taking medicine for headache

Here, ³Reward´ = removal of negative state becauseof behavior



P u nish m ent «

Punishment = Punishing a behavior todecrease future possibility of that behavior

behavior followed by negative consequence: e.g, bad

after sales service

Note: Punishment is different from NegativeReinforcement (even though the words soundsimilar!)

Negative Reinforcement increases chances ofbehavior

Punishment decreases chances of behavior



R ein f orce m ent and P u nish m ent

Punishment can be used bymarketer

example???



I ss u es in I nstr um enta l C onditioning «

Shaping Shaping = Reinforcing a behavior

at every step

So: break down an overall behaviorinto its component parts, andreward each component, so that

the overall behavior is learnt Bird example?



S haping « M arketing Exa mpl e

Suppose you own a car dealership(showroom). You want to make aconsumer buy a car from your

showroom Can you think of how you can use

³shaping´ to do this? (hint: start

from the consumer entering yourshowroom)



Other I ss u es ...

Discriminative Stimu lu s: Stimu lu s that occ u rs in the presence of a reinforcer bu t not in its absence Example: bu y Reebok---satisfaction (reward)

Learn to associate the brand name ³Reebok´ with the reinforcer/reward (satisfaction)

Brand name = discriminative stimu lu s

Helps to discriminate (differentiate) between different stimu li Reebok gives satisfaction, bu t Nike gives

dissatisfaction

I learn to discriminate based on brand name



S ti mulu s Discri m ination andS ti mulu s Genera l ization

S ti mulu s Discri m ination Using a discri m inative sti mulu s to differentiate

between different sti mul i (Nike vs . Reebok)

Can a l so u se discri m inative sti mulu s to treat different ite m s as si m i l ar (S ti mulu s Genera l ization)

Reebok shoes give m e satisfaction S o: I try Reebok c l othing a l so, ju st based on

brand na m e

Here: brand na m e is being u sed to treat different ite m s as si m i l ar



Classical vs. Instrumental

� Similarities

± both behavioral learning;

± number of trials increases learning

� Differences

± passive learning vs. active learning

� Instrumental conditioning: you have to perform an actionbefore any learning can occur (behavior ± consequence ± learning)

� Classical conditioning: your action is not needed. Learningwill occur if the US and CS are presented together to youby some external agent

± Example (watching advertising)



F or Next Cl ass F or next c l ass:

Extinction and F orgetting: pg . 114-118

Habit: pg . 122-133

Class9 Instrumental Conditioning

Documents