Lecture 23 - Experimental analysis of behaviour Flashcards
experimental analysis of behaviour
Systematic study of relation between behaviour and its consequences.
the skinner box
The Skinner Box
Want to study precisely the context the behaviour was in and the behaviour itself
Only measures what you want as you restrict the environment as much as you can
Sparse environment to look at relationship of the behaviour (in relation to what is in the box)
Recording in the Skinner Box = drum and paper going around it, as drum rolls under the paper it drags the pen along it, response causes the pen to go up
Schedule of reinforcement is …
A schedule of reinforcement is a specific pattern of presenting reinforcers over time.
Continuous reinforcement
Continuous reinforcement (CRF)
Every instance of a response is reinforced.
Useful for learning new behaviours and influencing ongoing patterns of behaviour quickly.
e.g. start off with CRF during behaviour modification
Partial (or intermittent) reinforcement
Partial (or intermittent) reinforcement
A designated response is reinforced only some of the time.
Useful for maintaining behaviours.
Motivation maintained
Same amount of reinforcer for more behaviour
Not every response is guaranteed to get a reinforcer
Four of the most simple schedules:
Ratio - Depends on numbers of responses (fixed and variable). - animal given reward after a certain number of times
Interval - Depends largely on the passage of time (fixed and variable). - time between rewards is set
Four simple schedules of partial reinforcement
Four of the most simple schedules:
Ratio - Depends on numbers of responses (fixed and variable). - animal given reward after a certain number of times
Interval - Depends largely on the passage of time (fixed and variable). - time between rewards is set
Fixed ratio schedule
The reinforcer is given after a fixed number of non-reinforced responses.
Lab: A rat receives food for every tenth response (FR 10). (Every tenth response there is a reward)
Cumulative record: Post- reinforcer pause, “burst” of responses until next reinforcer.
Whack out set usually to get reward and then pause after
Down check - reward given
Life?
e.g. a loyalty card
Variable ratio schedule
The reinforcer is given after a variable number of non- reinforced responses. The number of non-reinforced responses varies around a predetermined average.
Has unpredictability around it
Lab: A rat is reinforced for every 10th response on average, but the exact number required varies across trials.
After each reinforcer there is no way of predicting when the next one is going to be given
Cumulative record: High steady rate of response. Occasional pauses.
Life?
Slot machines, a lot of gambling is based on variable-ration schedule
Games - free play with occasional reinforcers of winning big for example (1 in ten players win a deluxe character etc), encourages people to spend money as they play
Fixed interval schedule
The reinforcer is given for the first response after a fixed period of time has elapsed.
Reward for 1st response after certain time
Lab: Rat is reinforced for the first lever press after 2 minutes has elapsed since the last reinforcer
Wait a set amount of time even if they are pressing the lever correctly
Cumulative record: Pause after rft, steadily increasing response rate as interval elapses.
Reward, wait, hit lever with increasing intensity over time
Looks a little like a fixed ratio, it shows you some discrimination of the passage of time by the organism, then you get an increase in the rate of responding over time
Accelerated response rate in between the reinforcers
Scalloped
Life?
Less common
Waiting for the bus, look more frequently closer to the due time
Pasta boiling - know how long it takes, keep looking more and more closer to the time
Variable interval schedule
The reinforcer is given for the first response after a variable time interval has elapsed. The interval lengths vary around a predetermined average.
Lab: Rat is reinforced for the first lever press after 1’ on average, but the interval length varies from trial to trial.
Cumulative record: High steady rate of response, although not quite as high as comparable VR schedule. Occasional pauses.
Different time levels, unpredictability encourages high steady rates of activity
VR only matters about number so can go faster whereas VI is paced
Life?
Fishing - wait certain amount of time, no way to speed it up
Applied issue of schedules
Engineering compensation systems: Effects of commissioned versus wage payment. (Gaetani et al., 1986, J. Org. Beh. Man).
Looking at the effects of different compensation systems
Setting: 2 machinists - high performance auto machining shop.
Baseline, basic wages($5-$7 per hour).
$5 for a junior, $7 for experiences
Feedback phase.
Workers were told how much work they got done each day based on invoicing
Return to baseline
Take feedback out
Commissioned compensation with daily feedback and quality control. If worked at baseline rate, would still get equivalent of baseline “wage”.
Changed wage system, wage dependent on invoicing
More work/invoicing = more money each week otherwise the same amount of work that they were doing still gave them the same baseline wage
However, must avoid competition between workers, especially if there is a limited amount of work (waiters and trappers)
Phase 1 = baseline (steady)
Phase 2 = feedback
Phase 3 = returning to baseline
Phase 4 = commissioned salary (produces alarming changes in work output)
Needs specific conditions to work = plenty of work and no competition
Waiters - end up not being very professional to customers due to competitive nature of wage system
Trappers - pay for each trap, initially lots of pests and over time this goes down to the point where people might actually start poaching. Initially lots of money but as the number decreases there is less money and more competition
Extinction
Reinforcers are no longer delivered contingent upon a response, and the strength of the response decreases.
e.g. job saying that you can keep working there however they will no longer pay you, individuals are less likely to go back and work there
Partial-reinforcement extinction effect
Partial-reinforcement extinction effect
Partial-reinforcement extinction effect:
Partial reinforcement schedules (FR, VR, FI,VI) provide greater resistance to extinction.
Behaviours maintained on high rates of reward have higher resistance than low rates of reward (usually variable behaviours)
Example - graph pictured
Initially both groups are fast, remove food for both groups, group that has food there 100% of the time drops speed quickly, variable group takes longer to distinguish behaviours - persistance is seen in groups with intermittent reinforcers
Drag resistance experiment
Particial reinforcement groups maintains its speed for much longer
This resistance to extinction is useful from an applied perspective.
If you are trying to teach a behaviour for example to someone with down syndrome, if you are teaching social interaction behaviours etc, if you can reward these more intermittently then they are more likely to persist in certain situations and over time rather than rewarding behaviours every single time
Why this partial reinforcement extinction effect? – Shouldn’t more reinforcement increase strength of response?
1. A subject trained with partial reinforcement has learned that reinforcement follows non- reinforcement.
2. Learned to persist in the face of frustration produced by absence of reinforcement. (Persist in the face of an emotional response)
3. Discriminates that situation is different. ( discrimination of the change in the word)
Applied issue extinction
The side-effects of extinction.
Extinction “bursts” of responses.
Especially if a high rate of reinforcement
Get initial increase in responding
Extinction induced aggression.
Animal gets aggressive
Increase in response topography.
One paw to two paws to full body for example
Increase in intensity
Typical problems using extinction.
Inability to control all sources of reinforcement for the behaviour.
Need to control sources of reinforcement for good extinction
Failure to provide alternative (appropriate) behaviours leading to the same reinforcer.
As well as trying to extinguish other behaviour
Spontaneous recovery.
Like classical conditioning
Temporary if you continue to extinguish
