7/20
. Aversive control (ch.6)
. Why punishment has many side effect?
. reinforce <-> punishment (flip side)
. decrease of behavior
. Primary Aversive stimuli(US)
. stimuli that as soon as they are encountered, are
escaped from, and later avoided
. Unlearned
ex) pain, bright light, extreme cold, heat, or loud noise
. Conditioned aversive stimuli(CS)
. Once neutral stimuli, that through pairing with
a primary aversive stimuli, come to be avoided
(US+CS)
. ex) verbal reprimand, police lights, etc
Some people don't like vegitables
. Primary SR+ - We can change Primary SR+ into conditioned aversive stimuli, we can flip
. Secondary SR+ - some people like new music.
. functional definition
. explain the difference between the peoples
. we have to test, before define
. A review of punishment
. positive punishment
. the addition of an aversive stimulus(behavior) as a consequence,
which causes a reduction in the behavior preceding it.
. negative punishment
. the removal of a reinforcing stimulus(behavior) as a consequence,
which causes a reduction in the behavior preceding it.
ex) Jail(Judicial system)
If it decrease the behavior, it's a punishment.
But some people never stop commiting crime.
. Effective Punishment
. Similar to reinforcement, punishment can be more effective by
1. introducting punishment abruptly
ex) 3-strikes rule is not useful,
mideast - cut the hand,
singapore
2. increasing the intensity of punishment
ex) jail time
3. decreasing the time between the behavior and punishment (latency)
ex) immediatly put people into jail,
sometime it takes more than 1.5 years.
4. punishment continuously (not intermittently)
. other factors to consider
1. the motivation for the response
. deprivation/satiation
ex) poor people steal money to survive. no alternative
2. response alternative
. Is this the only way to get reinforcers?
. Choice
ex) if homeless people go to jail, system give them food and drink
in jail (and shelter)
volent gangs - they have many friends in jail. (Mexican jail)
talk more about crimes.
. Negative reinforcement
. contingent removal of an aversive stimulus(behavior)
results in the increase of that behavior.
. escape : already in the presence of an aversive stimulus
. avoidance : engage in the behavior so that no contact with the
aversive stimulus occurs.
. escape -> learn -> avoidance
ex) headset
. positive reinforcement : enjoy music
. negative reinforcement : get rid of external noise
. functional analysis를 하고, 둘 중 1개만 선택하여 test
. Research on avoidance
. Sidman avoidance task (1960s)
. shocks are periodically given unless a specified behavior occurs,
which postpones the shocks for a specified amount of time
. S-S interval : time between shocks(constant)
. R-S interval : time between behavior and next shock
(very important variable)
ex) S-S : 20 sec
R-S : 5 sec (R-S < S-S) avoidance learns very poor, during learning it increase shocks(give more shocks)
25 sec (R-S > S-S)
. Problem with aversive control
. ethics (really sticky issue, moral, cultural difference)
. behavioral persistence
. interaction with classical conditioning
. learned helplessness
. aggression
. social disruption
. behavioral persistence
. remember the example about the contingency trap
. respondent-operant interactions
. now, remember the little albert experiment
. establish fear of a white rat in a child that wat not
afraid of rats before using classical conditioning.
. why not habituation (sensitivity)
. little albert experiment is too young. (2 years old)
. learned helplessness
. an organism is subjected to aversive stimulation which it cannot escape or avoid
. later, the organism is given the same aversive stimulation, but with a chance to escape
. however, there is no escape behavior.
. similar result with humans
. subject took longer to finish problems, etc...
. has been considered a basic model for clinical depression.
(hopeless, give-up, sleep more)
. ways to overcome learned helplessness
. put subject in contact with escape contingencies before procedure
. physically force them to maintain contact with contingencies.
. Aggression
. Providing aversive stimuli to two subjects in a confined space elicits aggression
. the more frequent the shocks, the more aggression
. the more intense the shocks, the more aggression
. the closer together, the more aggression
. if the subjects could escape, aggression would cease
. one person punishing another person may emit operant aggression.
. the person being punished want to escape punishment,
so they aggress toward the person punishing them.
. this aggression is maintained by negative reinforcement
. aggression is cyclical.
. Social disruption
. Is the punished behavior the only behavior affected?
. The context in which punishment occurs may
become a conditioned aversive stimulus
. Likewise, the punisher, may become a conditioned aversive stimulus.
. Both of these conditions lead to avoidance
ex) drop-out the social, not go to school,
don't want to meet parent, not go home.
7/25
. schedules of reinforcement
. most operant behavior is not reinforced after every occurrence of the operant
. intermittent schedules
. continuous reinforcement(CRF)
. Are the results of providing intermittent reinforcement predictable?
. what happens if every 10th operant ends in reinforcement?
. continuous reinforcement
. when every occurrence of the operant is reinforced
. abbreviated CRF
. exhibits response stereotypy.
ex) door open
. extinction of CRF operant leads to operant variability
ex) door not opening to the classroom
. extinction burst
. operant diversify
. History of schedules
. first "studied" by B.F.Skinner
. ran out of food
. went the cheap way : reinforce every 10th operant
. found stereotypic patterns of behavior
. programmatic study carried out by Ferster & skinner(1957)
. studied different ways in which food (reinforcement)
. results & interpretation
. results of manipulating the contingency had observable and replicable results
. Internal validity - rat, pigeon - high internal validity - cage, age, predictable
. External validity - not high - how generalizable?
(다른 animal과 human이 same result)
. single subject research
. no control group
. each animal, each cage, every level of variable
. interpretation was descriptive
. the behavior observed as a function of the environment,
rather than a potential intervening variable
(i.e. thoughts, feelings)
. variables in schedules of reinforcement
. dependent variable(s)
. rate of responding
. pattern of responding
. independent variable
. when reinforcement is scheduled
. every nth response
. first response after x minutes
. graphing patterns of behavior
. cululative recorder(1940~60)
. displays a real-time account of the total
(cumulative) number of operants emitted by the subject over the session
. the rate of response corresponds to the slope of the cumulative record
. both dependent variables can be measured.
. natural contingencies
. (operant) behavior we emit during our daily lives are on same schedule of reinforcement
. can our behavior be explained by the schedules of reinforcement?
. ratio & interval schedules
. ratio schedules of reinforcement
. after the nth number of operants, reinforcement in given
. the faster the organism responds, the more reinforcement is available
. interval schedules of reinforcement
. after x amount of time, the first operant produces reinforcement
. increasing response rate does not significantly affect reinforcement rate
. Fixed ratio schedule
. reinforcement only occurs after a certain number of operants
. abbreviated FR (with the number of operants required after FR)
. for example, FR 10 requires 10 instances of the operant to produce one reinforcer.
. CRF can be written FR 1, since every (one) operant ends in reinforcement.
ex) learning foreign language
write same character same times.
. FR attern of responding
. "Break-and-run" pattern (ex - FR 10)
. Why the "break" in the FR schedule?
. expirically shown to be a function of the ratio requirement
. the larger the ratio (FR 100), the longer the post-reinforcement pause(PRP)
. also called pre-ratio pause, since research ahs shown that
the upcoming schedule exerts more control over the pause than the previous schedule
. Fixed interval schedules
. reinforcement can only occur for an operant after a certain amount of time elapses.
. abbreviated FI(with the time of the interval and time parameter after the FI)
. for example, an FI 25 sec schedule would schedule reinforcement for the first operant that occurs after 25 seconds have passed since the last reinforcer.
. unless operants are occurring at a very high rate FI 1 sec = CRF
. FI pattern of respoinding
. scalloped pattern of responding
. animals and infants up to about 4~5 years of age show the stereotypic scalloped pattern of response on FI schedules, but adults usually don't. why?
. adults can time intervals, if they are short enought.
we have symbolic machenism. ex - clock, watch.
. Thus, only one response is required to obtain reinforcement, as soon as the interval times.
. variable schedules of reinforcement - ex) gamble
. both interval and ratio schedules can have reinforcement arranged for an average interval or ratio requirement
. on any single trial, the interval or ratio requirement may be more or less than the average
. adding all of the trials together and dividing by the number of trials will give the average
. real-world example
. Fixed ratio
. all 10days work -> paycheck
. fixed interval
. studying for quiz -> grade
. variable ratio
. insert coin -> get money (slot machine)
. variable interval
. throw a rod -> catching fish
. behaviral momentum
. behavior pesisting in the presence of a particular stimulus despite disruptive factors
. the higher the response rates, the less disruption during chaning reinforcement conditions
. response rate is analogous to mass in physics
. the more mass a moving object contains the harder it is to stop
. PRE revisited
. variables effecting PRE
. # of reinforcements(behavioral momentum)
. discrimination between reinforcement and extinction
. problem with PRE
. not enough variables controlled (internal validity)
. theories on response rate differentiation
. why do ratio schedules generally produce higher rates of responding than interval schedules?
. what happens between (individual) responses?
. what happens over an entire trial or session of responding?
. no complete answer
(no one has a complete answer to these questions.)
. molecular theory
. the time between responses - IRT(interresponse time)
is what is being reinforced. (moment by moment)
. IRTs have been shown to be an operant.
. Molar theory
. the overall response rate is a function of the overall reinforcement rate it produces.
. reinforcing non-smoking behavior
. not smoking (abstinence) is an operant that can be reinforced?
. does the schedule of reinforcement for abstinence matter?
. 3 groups were studied on different schedules of monetary reinforcement for not smoking.
(Does the schedule of reinforcement matter?)
. Independent variable : schedule of reinforcement
. control group : non-contigent money
. fixed group : same amount of money for each passed CO test
. prograssive ratio : increasing amounts of money for each consecutive passed CO test. (amount was reset for a failed CO test)
. Result
. in fixed and progressive groups : ~ 80% of CO tests were passed
. in control group : ~ 40% passed CO test
. Once abstinence occurred (3 straight CO tests)
only 22% of progressive group started smoking again
(60% and 82% for fixed and control groups respectively)
. about 50% abstinence throughout study for progressive
(only 30% and 5% for fixed and control groups.)