Topic 7: Differential reinforcement Flashcards
Differential reinforcement
-Operant training procedure in which some behaviors are systematically reinforced and others are not
-Reinforcement + Extinction
-Through practice extinction can be punishment instead
4 different types of differential reinforcement
-Low rates of responding (DRL)
-High rates of responding (DRH)
-Other behavior (DRO)
-Alternative behavior (DRA)
DRL
-Reinforcer is delivered contingent upon a low rate of responding during certain period of time
-Behavior is reinforced only if it occurs no more than a set maximum
-Results in less of the behavior but does not completely stop behavior
-A behavior needs to first be learned before reinforcement can occur
Examples of DRL
-Reinforce rat pressing lever only if it has been at least 10 seconds since last lever press
-Reset clock with each response (reinforced or not)
-On road trip, child will get treat in end if they don’t ask “Are we there yet” more than once an hour
-If they exceed the limit, no reinforcement
3 different types of DRL
1) Full-session DRL
2) Spaced-responding DRL
3) Interval DRL
1) Full-session DRL
-Reinforcement delivered if fewer than a specific number of responses occur in a period of time
-Rate within session does not matter
-eg) kid can ask all three questions right away, or they can space them out
2) Spaced-Responding DRL
-Reinforcement delivered if a set amount of time passes between responses
-Wait at least “X” times between responses
-Child only raises their hand after 15 minutes passes
-Low rates come from spacing out responses, therefore slowing down occurrances
-Timing more controlled
3) Interval DRL (least discusses)
-Session divided into time intervals and reinforcement delivered if behavior does not occur more than once per interval
-eg) Child raises hand only once per 15-minute interval throughout classes
-Mix between full-session and spaced-responding
Steps for DRL
1) Identify target behavior
2) Evaluate if DRL is best procedure
3) Determine acceptable rate of the behavior
4) Full session or spaced responding DRL?
5) Once plan is set, inform client
6) Implement treatment and provide feedback
DRl use
-Useful for reducing, but not eliminating, rates of undesired behaviors
-Longer intervals and lower thresholds produce lower rates
-Can start with less strict criteria and change over time
-Concerns with superstitious behavior
-Reinforcement delivered but not contingent on the behavior happening, but on the behavior NOT happening
DRH
-Reinforcer is delivered contingent upon a high rate of responding during a certain period of time
-Eg) The behavior is reinforced only if it occurs AT LEAST a set minimum number of times in a set time period
-Results in more of the behavior, increases desired behavior
Examples of DRH
-Reinforce rat pressing lever only if it has been pressed at least ten times within 10 seconds
-Keep track of number of responses within set time period and deliver reinforcement once criteria are met
-Child gets ice-cream at the end of the day if they answer at least two questions
DRO
-AKA differential reinforcement of ZERO responding
-Reinforcement is delivered contingent upon the absence of the behavior during certain period of time
-Decreased responding to ZERO
-Useful if extinction is not possible
-If we can’t remove reinforcer, we give a more appealing reinforcer for stopping the behavior
Examples of DRO
-Reinforce a rat if it does not press the lever in the box ten seconds
-Reset clock if response made
-On road trip, child will get ice-cream at the end if they don’t ask “Are we there yet” at all
-If they ask, then no response will be made
“Other behavior”
-Extinction of undesired behavior
-Only works if we are able to withhold reinforcer
-Not reinforcing an “Other behavior”
-Reinforces absence of behavior
-However, will engage in other behaviors naturally if avoiding undesired behavior
Short VS long DRO interval
-Use short DRO intervals for frequent problem behavior
-Use long DRO intervals for infrequent problem behavior
-Determined based on baseline
Steps for DRO
1) Identify target behavior
2) Evaluate if DRO is best procedure (DRL?, Extinction?)
3) Determine effective reinforcer
4) Choose initial DRO interval
5) Inform client, this helps eliminate chance of superstitious behavior to occur
6) Implement treatment and provide feedback
2 types of DRO
1) Whole interval DRO
2) Momentary DRO
1) Whole interval DRO
-Problem must be absent for the whole interval for reinforcement
eg) No talking for entirety of class period
2) Momentary DRO
-Problem must be absent at the end of the interval for reinforcement
eg) At specific times you wander through the class to ensure everyone is quiet
DRA
-Reinforcer is delivered contingent upon the occurrence of the desirable behavior
-Occurrence of undesired behavior is put on extinction
-Reinforcement of wanted behavior and extinction of unwanted behavior
Steps for DRA
1) Identify target behaviors to increase and decrease
2) Evaluate if DRA is best procedure
3) Determine effective reinforcer
4) Determine method of extinction or devaluation
5) Inform client
6) Implement treatment and provide feedback
Is DRA appropriate?
-We need to be sure that we have a desirable behavior to increase to use DRA
-That desirable behavior needs to happen at least sometimes for it to be reinforced
-Can the undesired behavior to extinguished?
-Can we withhold the reinforcer used and if not, can we devalue/overpower it?
Desirable behaviors
-Desirable behaviors should require less effort than the undesired behavior
-Especially important if we can’t use extinction on undesired behavior
eg) Healthy food easily accessible whereas junkfood is hard to reach
Effective reinforcers
-You could use the same reinforcer maintaining the problem behavior for the desired behavior
-This can reduce the occurrence of the undesired behavior
Preference assesments
1) Single stimulus assessment
2) Paired stimulus assessment
3) Multiple stimulus assessment
1) Single stimulus assessment
-Potential reinforces presented individually multiple times in random orderings
-On each exposure, record if stimulus was approached or not
-Percentage of approaches are calculated
2) Paired stimulus assessment
-Potential reinforcers are presented in pairs
-Each stimulus is presented with every other stimulus multiple times
eg) A+B, B+C, C+A
-Percentage of approaches are calculated
3) Multiple stimulus assessment
-Potential reinforcement presented in full array
-Items removed as chosen
-Process repeated with varied orderings
-Items chosen first most likely to be reinforcing
Prompting desired behaviors
-Prompt the desirable behavior before undesirable behavior occurs, not after
-Prompt when important EO’s and S^D’s are present
-Draw attention to presence of antecedents that should make the behavior more likely to occur
-EO’S make behavior more appealing
-S^D’s tell subject, if behavior occurs, it will be rewarded
extinction
-if the behavior can be withheld by change agent, use extinction as normal
-If behavior cannot be withheld, get creative
Maintenance and Generalization
-Switch from continuous reinforcement to intermittent reinforcement for maintenance
-Intermittent schedules more resistant to extinction
-Generalization procedures to help ensure the desirable behavior occurs in all relevant situations and with all relevant agents
DRA when problem behavior is negatively reinforced?
-Negative reinforcement= taking away something bad to increase behavior
-Differential negative reinforcement of alternative behavior (DNRA)
-When reinforcement is the removal of an aversive stimulus (escape behavior)
-Use the escape behavior as a reinforcer
Self Injurious behavior (SIB)
-Behavior in which subject causes harm to themselves
-Attention delivered every 15 seconds without SIB (DRO)
-Breaks from academic tasks delivered every 20 minutes for the absence of problem behaviors
2 variations of DRA
1) DRI
2) DRC
1) DRI
-Differential reinforcement of Incompatible behavior
-Behavior that is incompatible with the unwanted behavior is reinforced
-Increasing rate of desired behavior decreases rate of undesired behavior because the two cannot occur simultaneously
eg) Nail biting, sit on hands
2) DRC
-Differential reinforcement of communication
-Communication response is reinforced to replace the problem behavior
-Communication response delivers the reinforcer more rapidly than the problem behavior
-eg) Child yells for parents attention and is ignored until they speak politely
