Teacher praise is self-explanatory and may appear to be even a bit trite. After all, don't we all know that teacher praise is important? The answer is that yes, it is important but many people don't seem to realize this clearly when they student teach.

I regularly see student teachers who believe that they are "praising" by saying "Good" until it is mechanical-sounding; who praise while turning away from the student or not even looking at him/her. Here are some guidelines for avoiding that trap:_

(1) Be advised that presenting the lesson is only a small part of teaching. You must do much more than carry out the lesson plan. Many student teachers treat the lesson plan as a script that must be followed as if they are in a play. Treat it instead as a guideline for doing improvisational street theater. This will keep you on track but make you much more sensitive to your audience (the kids). _

(2) Praising requires practice. Good praise is sincere, is associated with a smile and perhaps a pat on the back or other sign of acceptance. Try to individualize praise. Students cannot all be praised equally well in the same way. If a child is interested in skateboards and sports, use that in the context of your praise (e.g., high fives, physical displays of pleasure with the child's performance, and references to sports figures). Many children like praise written out so they can look at it over and over (smilely faces, stickers, notes, etc.) Use variety in your praise and individualize it.

(3) Ensure that praise comes from significant others. Arrange for peers to praise each others' work. Many music teachers have group lessons where students listen to each other's work and then remark on what the performer did well. This is easily built into classes. Let the principal, parents, and visible people from the community praise the children. Also, train the children to praise themselves.

Clearly praise involves manipulating consequences. Therefore it concentrates on the response==>consequence relationship.

Being a consequence, it can be a motivational variable. (Note that the issue of motivation is largely a question of reinforcers and history of learning-remember we acquire many of our motivators e.g., praise, games, interaction, etc.)

Applications are typically appropriate in virtually any setting. Praise is a common form of reinforcement in the normal, natural environments of children. Therefore it is an especially good reinforcer because it means that children will be sensitive to the contingencies in their future environments.

When praise doesn't work, typically one or more of these suggestions will help:

(1) Be sure the praise is lively, not mechanically delivered, or a generic "Good" or "Way to go". Try to individualize the praise and accentuate it with emphasis, enthusiasm, and a meaningful context e.g., "Your reading is wonderful! That's real progress".

(2) Pair the praise with hugs, pats on the back, etc. that are age-appropriate and meaningful to the child. For example, secondary students sometimes prefer quiet praise on a one-to-one basis because teacher attention and compliance with rules may lower their status with peers. Children, on the other hand, may enjoy public recognition.

(3) Have the principal, counselor, and parents give praise. Sometimes the praise from these significant others is more powerful than when it comes from you.

(4) Have a time during the day when peers recognize and praise each other's work. Have them find something good that the student did each day. This not only increases the efficacy of the praise given, it has students practice praising each other.

(5) Praise immediately.

(6) State the behavior being reinforced.

(7) Use a variety of secondary, age-appropriate reinforcers.

(8) Use reinforcers that will be used in other settings. This will help generalization and maintenance.

(9) Eliminate competing contingencies. For example, if your reinforcement contingency is for on-task behavior, separate the child from others that may distract until the on-task behavior is so stable that it can withstand such competition.

(10) Reinforce every instance of the response.

Defined: Teachers send home daily or possibly weekly or bi-weekly reports on a child's progress.

Examples: Teachers send home a variety of reports from simple one line remarks to elaborate scorings of points earned in each class and comments regarding student demeanor. As always, use the least amount of programming that will do the job. If one line reports are sufficient, use them. If more elaborate systems are desirable, use them. This is not just for your own convenience. You want to use as little programming as possible so that the student does not become dependent upon such interventions. It is for the student's sake that you minimize programming.

Behavior analysis: Consequences are dealt with here but more can be done (see typical applications). These consequences can be of any type (reinforcement, punishment, extinction, etc.), but typically positive reinforcement is intended.

(1) The child can score their own reports thus learning to discriminate good and bad performances. This also decreases the amount of time between the behavior and the consequences for it. So the student learns to identify good performances, the consequences are administered more quickly (thus increasing the pace of learning), and students have a clear picture of what is required.

(2) The students can compare their reports with those of the teacher to refine notions of what is an is not an appropriate performance and why.

(3) Parents can reward (praise or provide privileges) based on these reports. Therefore home-based reinforcers (some of which you cannot provide in school, like slumber parties, staying up late on Friday night, etc.) can be brought to bear on in-class performance.

The main problems encountered are these:

(1) The reinforcers are too weak. Be sure to constantly assess your students' reinforcers. They change often and in unpredictable ways.

(2) The parents are not consequating the behavior. Be sure that the parents (or other parties charged with providing reinforcers) are fulfilling their requirements.

The Premack Principle states that A high probability response can be used to reinforce a lower probability response". This is much simpler than it sounds, it means that desired (rewarding) activities, such as free time play or access to playground equipment, can serve as a positive reinforcer for lower probability responses e.g., doing math or reading assignments.

This deals with response-consequence relationships and identifies a class of secondary reinforcers that are themselves behaviors i.e, the Premack Principle states that responses can reinforce responses. This means that powerful reinforcers can be found in what the child prefers to do._

Typically, Premack Principle rewards are simply one of several that children can earn as rewards. No special status is given to them.

Premack Principle rewards are often used within programs like Daily Report Cards where parents determine whether a child is free to obtain reinforcers like late night TV, having friends over, and staying out later in the evening on school nights. These reinforcers are all behaviors used as reinforcers and therefore are within the domain of the Premack Principle.

Similarly, token economies often have activity rewards e.g., 5 tokens purchases 10 minutes of free play with a friend (play here being the activity reward).

If the activity rewards are not rewarding, then several problems may be present:

(1) The children have free access to many other rewards, such as peer attention for inappropriate behavior. The activity rewards in your program may not be powerful enough to overcome these alternative and undesirable sources of rewards.

To deal with this, see if it is possible to use these more powerful rewards. For example, if someone loves to tell jokes, let them (and others) have the opportunity to do stand-up comedy. If talking to friends competes with completion of math lessons, use opportunities for peer conversation as an activity reward for correctly completing math assignments.

(2) The activities used as rewards are not very rewarding. Again, sample the students' reinforcement preferences via reinforcement surveys, conversation, and observation. Very often they will simply tell you what is rewarding.

Just as individuals can be reinforced and punished for responding, so can groups be reinforced and punished for their collective performance. Therefore the group as a aggregate must meet some criterion or criteria for members to be reinforced or punished.

Numerous group contingencies can be arranged:

(1) One person can earn a reinforcer for the entire group.

(2) Each person must meet all criteria for everyone to receive a reinforcer.

(3) Subgroups can earn reinforcers for all group members.

Just about any variation is possible here.

Group contingencies are an extension of the definitions of reinforcement of punishment. Individuals' behaviors are being modified but as part of a larger contingency that sets group-not individual-requirements for reinforcement.

Group contingencies can be used with virtually any other technique we will cover. Often teachers prefer to use group contingencies because they are more convenient-you have one criterion for reinforcement instead of one for each child.

Another good reason for using group contingencies is that they are closer to what occurs in regular classrooms and the real world. Almost certainly you will use group contingencies when teaching.

(1) The problem is one student. One student's behavior may consistently result in the loss of group reinforcers. This individual, especially if he/she is ED, may not be remorseful or responsive to the requests of other students to comply with rules.

The best tactic is to remove this individual from the group contingency and institute a program just for them. Note that this program should have as a goal the person's reentry into the group contingencies.

(2) The "unfair" argument. Students who individually meet criteria may state that they should not pay the price for a few persons who don't meet criteria.

The best tactic is to have group rewards that only groups can receive e.g., field trips, time in the gym, extra time a recess, early dismissal for lunch. Indicate tot hem that they can only receive such rewards if you are there to supervise them and since you can't be in two places at once, all or none must receive the reward.

Yoked contingencies are a kind of group contingency. Here the performance of one person determines reinforcement or punishment for another person. None of the consequences come to bear on the person who earns the rewards of punishments. For example, Jake can earn free time for Harry and Harry can earn free time for

Jake. If they both earn the free time for each other, they can have it together.

At least two contingencies are required for a yoked contingency. Person A's performance determine consequences for person B and vice versa.

(Jake) Response== =====>Consequence (Earned by Harry)_

\\//_

//_

(Harry)Response==//\\=====>Consequence (Earned by Jake)_

Suppose that Jake and Harry like to talk during class and thereby disrupt everyone else. Mrs. Jones, their kindly second grade teacher, makes two observations:

(a) Talking is a reinforcer for them (a'la the Premack Principle).

(b) Punishment has too many side effects to use in this context.

So Mrs. Jones takes Jake and Harry aside after class and cuts a deal. If Jake does not talk out for 5 minutes, then he earns one minute of free time for Harry to talk to someone at the end of class. Likewise, if Harry does not talk out for 5 minutes, then Jake earns a minute of free time to talk to someone. So if Jake and Harry each refrain from talking out for 25 minutes, they can have the last 5 minutes of class free to talk with each other.

Mrs. Jones gets a quiet class, Jake and Harry get to talk, and overall we have a kinder, gentler educational system. We also have a more efficient system that Jake and Harry will like more than if Mrs. Jones had followed Mrs. Hatchett's suggestion to lower the boom by writing out the dictionary, staying after school, and having the parents in for a "conference".

Yoked contingencies and other techniques: As with other techniques, yoked contingencies are easily combined with group contingencies, response cost, and most of the other techniques discussed in this handbook.

Rewards aren't earned. If the students are not earning reinforcers for each other, the program will quickly fold and you should change rewards used. If changed rewards are also ineffective, change to individualized programming for each student.

Technique Description

Timer programs involve using a timer to manage schedules of reinforcement.

Several types of schedules addressed here are these:

(a) Variable time (also called momentary time-sampling)

(b) Differential reinforcement of low rates.

(c) Differential reinforcement of high rates.

Each will be dealt with in turn:

(a) Variable time (VT). This schedule involves setting a timer randomly and when it sounds doing the following:

(1) Checking to see if the child is engaging in a targeted behavior.

(2) Consequating behavior accordingly.

Note that the child's behavior is not monitored during the intervals between timer soundings.

Note that one can deliver a positive reinforcer if a response is occurring or if a response is not occurring. Therefore reinforcement can be used to increase the probability of a response or decrease the probability of a response. Likewise, punishment can be delivered if an undesirable response is occurring.

(b) Differential reinforcement of low rates (DRL). DRL schedules are used to reduce rates of responding. They involve the following procedure:

(1) Determining how often the child engages in a given behavior.

(2) Setting an interval slightly shorter than the usual intervals between the responding.

(3) Delivering a reinforcer if no response occurs during the interval.

If a targeted behavior does occur during the interval, the timer is reset and the interval starts over. After behaviors consistently do not occur during the interval, the interval length can be slowly increased until response rates fall to desired levels. Such a deceleration is what gives this schedule the name differential reinforcement of low rates.

(c) Differential reinforcement of high rates (DRH). DRH is like DRL except that two responses must occur within the set interval for reinforcement to occur. It is run just like DRL except that the interval size is decreased so that response rates increase.

The function of DRL and DRH schedules is to change ongoing response rates-not to establish behavior. This rate modification occurs because naturally occurring responding is differentially reinforced depending upon whether it meets the criteria for DRL or DRH reinforcement.

This raises an interesting point-overall response rate changes can be produced by reinforcing specific interresponse intervals. So again, the key function of DRL and DRH is to modify response rates.

Typical application(s)

Anytime that you want to change response rate, timer programs are appropriate. Remember, this is not a technique for establishing a response, it is for changing rates of responses that are already in the person's repertoire.

(1) The rate doesn't change fast enough. Be sure to start each timer program by reinforcing response rates that the person can already do. In all behavioral programming, you must start where the person is, nothing else is possible.

Also check whether the amount of behavior change is too large. Increments of behavioral change should be so slow that the individual doesn't notice the differences at each level.

(2) The timer disrupts class. Use a digital watch with a timer that is quieter.

(3) The response rate has changed, but academic performance is no better. The criteria for an adequate performance should be added to the timer program. For example, is the child is on-task for 10 minutes and receive credit for this, then bonus points should be added for doing the task correctly. In short, add another contingency (reward) for task completion.

The student, teacher, and any other relevant parties develop a contract that specifies at least the following:

(a) Target behaviors

(b) Situations where the target behaviors are to occur

(c) Rewards/punishments (any consequences) for the behaviors.

(d) How data will be kept

The contract takes the form of a formal agreement in that it is written, discussed, and signed by all relevant parties. Its role is to clarify a program's contingencies.

Often contracts also specify contingencies that bear on authority figures such as teachers. For example, if the teacher does not deliver the agreed upon reinforcer (such as free time or other consequence) the child has an appeal process (e.g., to the principal) and then receives double reinforcers or some other compensation.

Contracts have several advantages:

(1) It clarifies contingencies for students and teachers.

(2) It systematizes the teacher's consequation of behavior thus organizing the teacher's program.

(3) It builds trust between students and teachers. The students can be sure that the contract will be carried out.

The contract formalizes contingencies therefore it acts as rules in rule-governed behavior. The contract is an antecedent the response is the behaviors specified (for student and authorities) and the consequence(s) are only those cited in the contract.

(1) Classroom management. Students and teacher contract for reinforcers. These can be short- or long-term programs.

(2) Vocational education projects. On-the-job training programs typically require students to (a) be on time for the job, (b) complete school assignments before going to work, (c) calling the employer when they are absent from school, etc. The teacher's target behaviors may include (a) providing transportation to and from the job, (b) finding a job, (c) supervising the on-the-job performance, etc.

(3) Specially arranged programs. Sometimes a great motivator is to tell LD students how they can be removed from LD. This requires citing the behavioral criteria for returning full-time to the classroom. A contract can specify this and provide targets for both the teacher and student.

(1) Target behaviors are not met. Several variables should be checked:

(a) Are the reinforcers effective? Does the student state that they are what he/she wants to earn?

(b) Are the behavioral criteria within the student's reach? Does the student have prerequisite behaviors for completing the tasks? Does the teacher need to provide prosthetic programming or equipment to support such behavior?

(2) The student doesn't trust the authority figures. Many special education students do not trust the school's authority figures. The contract, therefore may not be believed. The solution is to set some very short-term goals that are easy for the student to meet and then be sure to liberally reinforce the student for meeting them. This correspondence between the authority's statements and actions enhances trust.

Schedules of reinforcement are rules for reinforcing responses. There are two types of schedules of reinforcement:

(a) Continuous reinforcement: Each response is reinforced.

(b) Intermittent reinforcement: Not every response is reinforced. There are many types of intermittent schedules of reinforcement. Only some of them are commonly used in education so we will limit our discussion here to them:

(a) Differential reinforcement of incompatible behavior (DRI): This schedule of reinforcement involves reinforcing a response that is appropriate and incompatible with an inappropriate behavior you wish to eliminate. Therefore it is schedule of reinforcement that is used to (a) decrease the rate of an unwanted behavior and simultaneously (b) increase the rate of a desirable behavior.

For example, reinforcing on-task behavior such as reading assigned material will decrease the rate of off-task responding such as talking out. Therefore we can eliminate talk-outs by reinforcing reading or some other on-task behavior.

(b) Differential reinforcement of decreased rates of behavior (DRD): Sometimes reinforcing incompatible behavior is not enough. Under these circumstances, we can use the DRI and the DRD together. In the DRD, we reinforce a child for emitting fewer responses within a specified period of time.

For example, if Jake emits 20 talk-outs per hour on average, we would deliver a reinforcer to him when he drops his rate to 18 or 17. We would not expect all responses to be eliminated immediately, therefore a DRD allows us to move the response rate down slowly until it achieves a level we want.

Those of you have had my lectures on single subject designs will recognize the DRD as a kind of changing criterion design, where changes in criteria for reinforcement as sequentially required and appropriate levels of responding are reinforced.

(c) Differential reinforcement of high rates: This schedule of reinforcement requires reinforcement when a set number of behaviors occurs within a given time span. It is really quite straightforward: The teacher sets a number of responses, for example, 5, to occur with, for example, 10 minutes. If the set number occurs, then reinforcement is delivered. For example, Jake may do his math problems accurately but slowly. So we (a)

take a baseline to determine how quickly he currently does is math, (b) set the first response rate within this current limit, (c) slowly increase the number of responses for reinforcement until a normal rate is achieved. Therefore this schedule of reinforcement is the like the DRD but the rate increases instead of decreases.

(d) Differential reinforcement of other behavior (DRO): A DRO is like a DRI except the response needn't be incompatible-it need only be different. For example, suppose that we set a timer at random intervals that have a mean of 4 minutes. Each time the timer rings, we check to see whether a targeted student or students are engaging in some inappropriate behavior. If they are not, then a reinforcer is delivered. The behaviors rewarded needn't necessarily be incompatible with inappropriate responses, they just need to be different i.e., "other" behavior hence the name "DRO"._

Maximizing effectiveness with these schedules of reinforcement:_

(1) Make sure that the reinforcer is strong. Changing behavior often requires a more powerful reinforcer than maintaining it.

(2) Combine these schedules with other procedures.

(3) Make sure that other environments will support the behavior i.e., that other teachers, parents, etc. will also reinforce as you are reinforcing.

(4) Thin the schedule slowly to ensure that the behavior does not disintegrate.

Theoretical analysis

Schedules of reinforcement are obviously related to reinforcement, but the situation is much more complex than just that. Schedules of reinforcement are basic learning phenomena that span all species down to one-celled organisms. Hence it is a very basic phenomena and a power behavior change tactic if used correctly.

The DRI, DRD, DRO, and DRD schedules of reinforcement are used in classroom management and academic areas. They are primarily concerned with changing the rate of one response by changing the rate of other responses-hence they can decrease the rates of inappropriate behavior without punishment! Clearly this is a great advantage to teachers and students.

The key variables that go awry when using such schedules are these:

(1) The reinforcer is weak.

(2) The schedule is not used consistently.

(3) The behaviors changes required for reinforcement are too large.

The solutions to such problems are obvious.

Less obvious is that students can often apply these schedules to themselves as part of a self-management program. For example, the student can evaluate their own behavior and engage in self-reinforcement procedures. They can set their own reinforcers and can choose reinforcers. In short, they can participate in virtually every aspect of the program.

This technique involves having the teacher reinforce good behavior in full view of other targeted students. In essence, the good behavior of other students constitutes a modeling trial and the probability of imitation by others increases when you publicly reinforce good behavior.

Modeling is a behavioral repertoire acquired like any other. If the student observes reinforcement of another, the probability goes up that the student will model the response.

In any situation where on-task behavior is reinforced, the teacher should state what is being reinforced, liberally reward the response, and ensure that others clearly identify what was rewarded and why.

This has applications in academics and classroom management. Any response that can be modeled is a candidate for this tactic.

If this tactic fails:

(1) Check to determine whether incompatible responses are being reinforced by peers. A teacher's vicarious reinforcement may be weak compared to peers' acceptance for deviance. Therefore your rewarding must be more rewarding than the competition available from other students.

(2) Be sure that the person being rewarded is a high status model. Students who are unpopular are not readily modeled.

(3) Do not use implicit punishment. For example, I have seen teachers state "Jake is really working, the rest of you could learn by his example." This has punishment built into it and may even decrease the likelihood that Jake will continue to behave well because it makes him different from his peers.

Positive practice overcorrection: Engaging in a corrected version of the inappropriate behavior.

Restitutional overcorrection: Setting things right.

Overcorrection is most easily seen as the opposite of the Premack Principle. A low probability behavior is engaged in following a higher probability behavior (the opposite relationship defines the Premack Principle). Two kinds of overcorrection are possible:

(1) Positive practice overcorrection: When an inappropriate behavior is engaged in, an appropriate form of behavior similar to the inappropriate response in engaged in ad nauseum.

This technique is typically used with such responses as stereotypic responding, for example hand flapping and head twirling. A girl who twirled her head round and round was required to move it forward, backward, right, and then left on cue for 20 minutes. The inappropriate behavior was the twirling; the positive practice overcorrection was moving her head in a direction on cue for 20 minutes (this was prompted). The result was that head twirling decreased.

Do not confuse this procedure with overlearning. Overlearning involves doing a response well beyond the point when it is fully established and stabilized. Overlearning is not a procedure that consequates an inappropriate response.

(2) Restitutional overcorrection: This involves having the individual more than remediate for an inappropriate behavior. For example, if a child hits another, they must do much more than simply apologize. They must make additional restitution such as serving the child lunch, carrying the child's books, and repeatedly making verbal admonitions that they are sorry for the infraction.

A child that vandalizes a bathroom would not only replace the broken materials but clean the all the school's bathrooms for two weeks. Stealing money would result in not only returning the money but also paying "interest".

To maximize over correction:

(1) Apply consistently

(2) Apply immediately

(3) Arrange environment to block escape.

(4) Combine with other, positive techniques for changing the behavior.

Theoretical analysis

The best analysis is that overcorrection is the opposite of the Premack Principle. Low probability behaviors serve as punishers that are required contingent upon inappropriate behavior. The notion of restitution has a lot of appeal for many people who want more than an empty "I'm sorry" and a meaningless punishment (like going to the office or staying after school). The courts have also pick up this tactic and are applying it juvenile and adult court.

(1) Positive practice overcorrection: This procedure is typically used with the severely mentally retarded or to deal with stereotypic responding. For example, suppose a child flapped his arms and stared into space. The child may be prompted to hold his arms at his side, and then cross them-for 20 minutes each time stereotypic responding occurred.

In short, some acceptable behavior is prompted and this behavior has something in common with the behavior punished.

(2) Restitutional overcorrection: Here the individual engages in behavior that more than remediates. For example, breaking a window requires paying for it, helping install it, and washing all windows in the classroom for two weeks.

Hitting others requires excessive apologizing, carrying the offended student's books, obtaining his/her lunch, etc.

(1) Restitution requires money the child does not have. The child should then earn the money needed by doing tasks around the school (perhaps petty cash can be used). The parents should definitely be notified and asked for the necessary funds.

(2) Positive practice is time consuming. It is time consuming in the long run but it is a technique that involves no corporal punishment and can be effective if used consistently.

If prompting the positive practice overcorrection is not possible, set a timer and have the child do the overcorrection while you supervise from a distance.

Technique Description

Essentially, this tactic has teachers pay no attention to inappropriate behaviors.

Ignoring is one of those techniques that was inappropriately elevated to the level of principle by educators. Their dictum was this: "If the child's behavior is ignored, it will decrease". This is a very dangerous misinterpretation of behavioral theory and technique. I will explain why next.

The key principle underlying ignoring is extinction. The assumption is that attending to the child reinforces the misbehavior. This may be true, but here is what often happens:

(a) The behavior is ignored and an extinction burst occurs. The increased level of inappropriate behavior results in the teacher breaking down and attending to the child thus reinforcing an increased rate! This then sets up even more inappropriate behavior. (Some of you may recognize this as thinning a reinforcement schedule-it has the effect of strengthening the response even more, which is the exact opposite of the intended

outcome.)

(b) Another common occurrence is that the extinction burst results in the other children attending to the child. Then the other children's attention becomes the reinforcer and the teacher now has the double problem of controlling the deviant child and the entire class' attending. Finally, extinction bursts sometimes come the attention of principals and other teachers who respond thus again increasing the response rate to an even higher level.

Ignoring is often attempted in classrooms, it fails, and then the teacher resorts to negative tactics to decrease the now even higher rates of inappropriate responding.

Ignoring and other techniques: Ignoring should never be used in isolation. Classrooms are just too uncontrolled to be assured that the deviant behavior will not be reinforced by another person. Therefore be sure to reinforce a behavior incompatible with the deviant response (technically this is a DRO schedule of reinforcement). This will decrease the rate of inappropriate behavior while ensuring that another, more desirable response,

replaces it.

When ignoring is used without other techniques, it typically does fail, so the solution to this problem is obvious: Use ignoring with techniques for rewarding incompatible responding.

This is the removal, usually temporary, of a positive reinforcer. Think of response cost as a "fine".

Examples: A child can lose access to a privileged area for a short time; lose access to a favored toy; be restricted from engaging in a behavior (such as talking, playing with a certain child, etc.); and be required to pay 10 cents each time they swear.

Clearly this is a type of punishment where a positive reinforcer is removed contingent upon a response.

Response cost if often used within token economies. Typically several tokens are removed for breaking rules.

As with all punishment, it is important that it be used concurrently with rewards for incompatible responses.

When response cost doesn't work, typically one of these problems is occurring:

(a) A reward is not being lost. The thing or activity lost must be rewarding if its loss is to serve as a punishment. The solution is to assess students' rewards often (reinforcement menus, observation, conversations, etc.)_

(b) So many reinforcers are already available in the classroom or home that loss of one or a few has little effect. Under these circumstances, the child must lose more than one reward, as occurs in timeout procedures (see timeout entry).

Timeout's full name is "timeout from positive reinforcement", therefore it is the contingent removal of the chance to obtain reinforcers in a given environment. "Timeout" is a general term used to describe two procedures:

(1) Seclusionary timeout: This requires removing a person from a reinforcing environment. For example, if a child is talking out, off task with friends, etc. that child is clearly in a very reinforcing environment. Exclusion contingent upon inappropriate behavior can result in reduced response rates.

This is the most familiar form of timeout where a "timeout" area" or "timeout room" is used. Please note that schools and institutions often have very strict regulations governing timeout's use. Be aware of them if you plan to use timeout.

(2) Exclusionary timeout: This involves terminating the child's chances of receiving reinforcement within a given environment. It does not involve removing the child from the environment._

To maximize effectiveness:

(1) Remove all reinforcers during timeout.

(2) Make time-in as reinforcing as possible.

(3) Do not punish with reminders of past inappropriate behavior. Let by-gones be by-gones and reinforce the student as much as possible.

(4) Use the shortest timeout duration that will do the job.

(5) Combine timeout with positive procedures.

(6) Apply timeout consistently.

(7) Calmly state the reason why the child is going to timeout prior to placing them in timeout.

(8) Do not release from timeout until inappropriate behavior (such as yelling, swearing, etc.) stops for at least 4 minutes. This will avoid reinforcing inappropriate behavior by releasing the child from timeout when they are being inappropriate.

Timeout's description indicates that it is a punishment technique (access to an environment's positive reinforcers is restricted). This occurs by either isolating the child (seclusionary timeout) or excluding the child from reinforcement (exclusionary timeout).

(1) Isolation (seclusionary timeout) typically involves placing the child in a secluded area where positive reinforcers are not available.

Timeout "boxes", timeout "rooms", etc. have been abused by individuals who do not understand the consequences of using punitive techniques. Therefore timeout has had a bad press because of a few untrained persons. This had led to developing due process procedures for using timeout. Check with your school to see if they have any.

(2) The less used but better procedure is the exclusionary timeout procedure that incorporates a "timeout ribbon". Each child is given a ribbon to wear while in class. If they misbehave, the ribbon is taken from them for a short period (usually 15 minutes or so) and during this time the student can earn no positive reinforcers but is still required to complete all tasks required of others. In effect, then, the child is working without being paid.

(1) Child continues to misbehave Timeout muse be from positive reinforcement. If the classroom is punitive (and many are) then timeout releases the student from a negative setting thus negatively reinforcing inappropriate behavior (i.e., here the "timeout" is "timeout from punishment").

Therefore be sure that positive control exists in the classroom. Indeed, if the classroom were primarily punitive, no additional punitive procedures (like timeout) should be used-concentrate on instituting positive classroom management programs.

Again, unless the classroom is reinforcing in the first place, timeout will not result in reduced response rates because the "timeout" is not "timeout from positive reinforcement". It is instead "timeout from boredom (or worse)" and that would be negative reinforcement, which would make the problematic behavior even more likely to occur.

So if you decide to use timeout, make sure that it is timeout from positive reinforcement or the technique may backfire. Also be advised that any punishment procedure should evidence results within a week or it will not work.

(2) Time limits for timeout. I have seen timeout programs run for months with no evidence that it had any affect on behavior. I even saw one where the response rate was increasing. This is unethical and perhaps illegal. It certainly reflects incompetence. I have seen timeout used with children who self-stimulate. When the child flaps their hands or masturbates, he was put into a timeout room. Guess what he did there! Of course, he flapped his arms and masturbated at will and the geniuses running this program couldn't understand why they were having a problem!

One more time: Putting a child in a timeout room is not timeout unless the room has fewer reinforcers available than the environment the child left.

This is punishment procedure that is to be used as a last resort. Place a video camera in the back of the room and tell the students that if they break class rules more than once, the video of their class disruptions will be taken to their home and played. Then if rules are violated, do exactly what you say you will do-take it to their homes, take a VCR if necessary, and play the disruptive behavior over and over. I have almost never had to use this technique more than once.

Here are some tactics for making this technique more effective:

(1) Replay ad nauseum the misbehavior. Each replay should have a CBS sports-type reaction from you with questions like "Did you see that?" and rhetorical questions like "Can you see why I need your (the parents) help in running a more effective classroom?"

(2) Make sure that the parents are not blamed-you want them on your side so that they will help run a program.

(3) The parents may not be helpful, therefore other persons who serve as parents (aunts, relatives, caretakers, etc.) may be the appropriate audience.

Clearly this program is an escape-avoidance contingency for the parents and the student. The antecedent is the tape and your leading questions, the response is agreement with your leading questions about he need to do something, and the consequences involve parents helping you to avoid having to see more tapes of this type. Obviously the child student also has escape-avoidance contingencies where the antecedent is the tape and questions, the response is agreement that something must be done, and the consequence is the termination of the tape.

Clearly this is necessary where classroom management problems exist at an extreme level. Do not use as a first technique unless baselines exist indicating the need for fast, extreme action.

It is also a good technique when parents are uncooperative or support the child's lies about how you mistreat him/her in class.

This technique is not appropriate if the parents or other caregivers are already cooperative. This is a means of punishing inappropriate in-class behavior and unfounded parent support of their child's inappropriate behavior.

(1) Parents already cooperative. This technique is not appropriate for use when parents support your programs.

(2) Parents refuse to cooperate. Tell the parents the truth: If the problems persist, you will recommend that the student be placed in a more restrictive setting e.g., semi-institutional center such as a day program, or the need for family counseling as prerequisite for continuing in your program. Then contact the school social worker and carry through on these suggestions._

As with vicarious reinforcement, modeling requires the use of high status models and reinforcement of the modeled response. Typically teachers serve as the model and the students are expected to emulate.

Modeling is a repertoire that is acquired. People are reinforced for doing what others do. In short, matching the models behavioral sample resulted in the same kind of reinforcement.

S_d_===>R=======>Consequence

The model's=>Modeled==>Same or similar behavior=>response consequence

Applications are typically academic. The teachers models how to carry numbers, use a metacognitive technique, draw an object, etc. Then the students are asked emulate the behavior. The uses of modeling are so common and wide-ranging that they may appear a given. But good modeling procedures are seldom found as the trouble shooting section indicates.

(1) The model should be high status. Do not use unpopular students or individuals that are not liked by the student. Advertising uses models very effectively-sports stars sell everything from tennis shoes to cars. They are high status persons depicted in situations where they are being reinforced by a product. Clearly it works and it can work in the classroom too if used properly.

(2) Modeling requires that the student be able to do what is modeled. In short, do not use modeling to establish new behaviors unless you are sure that all necessary prerequisite responses are in the child's repertoire.

(3) Modeled behavior is often weak. Make sure that you reinforce the response often following modeling.

(4) Use shaping/fading procedures with modeling e.g.:

(1) Model parts of a response and have the students do each part of the total modeled response-don't model the entire response and expect students to do it in toto.

(2) Model only part of a response and ask students to complete it on their own (fading the modeled prompt).

(3) Have students describe the response they are modeling. These additional cues will help maintain the response as you fade the modeled cues.

In the early days of radio, one some stations simply read the time. The radio announcers who took this job quickly found other positions because of the extreme boredom associated with reading the time every 15 seconds for an eight hour shift.

This technique is similarly a punitive procedure and therefore is to be used as a last resort: Have the students read the time into a tape recorder every 15 seconds for 10-20 minutes. When they have completed this task, check it for accuracy, if they missed an interval or were more than 5 seconds late with one interval, have them redo it. Then have the student complete missed work.

This procedure has the advantages of not being directly associated with academic tasks, does not allow the student to simply avoid school tasks-it delays these tasks.

The consequence of reading the time every 15 seconds is clearly punishment.

This is a punitive technique that should be used as a last resort to deal with classroom disruptions.

As with any punitive technique, be sure to use positive approaches concurrently. This will reduce the chances of inappropriate escape-avoidance behavior and help countercondition the unwanted behavior.

The misbehavior does not decrease in frequency.

(a) The student is clearly not being punished-try increasing the total time spent reading the time.

(b) It is possible that the avoidance of the tasks in class is so reinforcing that even counting the time is more palatable than doing the lessons. Be sure that meaningful positive reinforcers are in the child's classroom situation.

(c) The child may not have prerequisite behaviors needed to do the lessons therefore he/she is not on-task. Be sure that a baseline indicates that all prerequisites are established.

First, a very thorough task analysis is done. The goal is to create so many steps that the student will make no mistakes in the acquisition of the response.

Second, prompts and modeling are used to guide the child's behavior during the early stages of response acquisition.

Third, prompts and modeling are faded slowly-only after the response is well established.

The prompts and modeling are antecedent control that occasion a response that is reinforced:

Antecedent=====>Response========>Consequence

(Prompts) Step in the Positive

(Modeling) Task Analysis Reinforcer

(Later)

Antecedent==>Response========>Consequence

(Prompts Step in the Positive

& Task Analysis Reinforcer

Modeling

are Faded)

Note the fading-no response is really established if teacher-mediated prompts are still used.

The applications are so numerous that they can only be categorized here: Reading, writing, social skills, math, i.e., any area you wish to work with can benefit from errorless discrimination.

Example 1: Reading

The word to be taught is "House". The teacher may do some or all of the following:

(1) Do a baseline to be sure that each letter of the word "house" is recognized.

(2) Model reading the word house.

(3) Prompt the students to read the word "House". During the student reading, the teacher will state the word "house" along with the students so that no one makes an error.

(4) The teacher will then again ask the students to read the word "house" and the teacher will fade her verbal prompt by saying the word "house" in a softer voice.

(5) Place the words "House" and "garden" on the board ("garden is a word the children can already read accurately). Again the reading of "house" will be prompted.

(6) Additional words will be paced on the board and the students will make appropriate discriminations between them.

(7) The word "house" will be read in a sentence where it is the only new word.

(8) Other objectives will be covered that relate to the word "house" such as definitions, use in context, and derivations of "house" e.g., "Housing", "Housed", "House broken", and "House about them Cubbies?" [just kidding]

(1) Mistakes still occur. The keys to making errorless discrimination work are the task analysis and the baseline. You must be sure that the students have all necessary prerequisites and that the task analysis has broken the task into the smallest reasonable steps.

(2) The child is unmotivated. This is a question of reinforcers. No technique will work if the reinforcers are weak.

Note: Shaping, fading, and chaining typically are used together so I will treat them together here.

Prompts. Prompts are S_d_s that increase the likelihood that a response will occur. They are typically faded.

Many kinds of prompts can be used including these:

(1) Within stimulus prompts: These are prompts that change some important aspect of an antecedent. For example, suppose that Jake does not attend to the "+" and "-" in math problems. He then adds when he should subtract and vice versa. The teacher then changes the color of the "+" and "-" to red. Jake attends, does the problems correctly, and the teacher fades the colors to purple, navy blue, and then black just like the rest of the problem.

Note that an aspect of the stimulus was changed-nothing was added, the S_d_ was simply modified to accentuate one of its properties.

(2) Extra-stimulus prompts: Extra-stimulus prompts are additions to the S_d_. For example, instead of changing the color of the "+" and "-" in the last example, Jake's teacher could have added an arrow so that the problems looked like this:

23

==>+ 44

This arrow could then have been faded. Typically, within stimulus prompts are better, but the difference is not large and you may find some students who work better with extra-stimulus prompts. Therefore experiment and see which work for a given student.

Shaping involves these steps to (a) establish a new behavior, (b) change the rate of an existing behavior, or (c) modify the form of an existing behavior.

(a) Doing a task analysis. This requires writing out the steps involved in doing a the targeted behavior.

(b) Reinforcing successive approximations. The "successive approximations" are the steps in your task analysis. You reinforce each step toward the final form of response that you wish to build.

For example, writing the letter "Y" involves several steps:

(1) Making the "\".

(2) Making the "/".

(c) Making the "|"

Of course each of these must be located in the correct place relative to each other. One could shape such behavior by using a dot-to-dot tactic such as this:

(1) (2)

* *

* *

* *

*

* <=(3)

*

The dots serve as guides for making the component parts and the numbers indicate the order that the parts should be completed. After the child can do this, the dots would be faded in steps perhaps like this:

Step 1: (1) (2)

* *

* *

* *

*

* <=(3)

Step 2: (1) (2)

* *

* *

* *

Step 3:

* *

*

Step 4: *

The actual steps to use should be determined by how well the child is learning the response. You will find that this is best done by close observation and your judgment based on you background with the child.

Fading: Fading was slipped into the discussion above so it needs little development here. "Fading" refers to the fading of prompts. In the example above, the prompts are the dots and numbers and the fading was the systematic removal of these prompts across steps.

Therefore "fading" refers to the removal of prompts. Note that these prompts can be any of the types discussed below.

Chaining: Chaining is like shaping but instead of building a new behavior, you link existing behaviors. For example, a 16 year old has all the responses needed to drive a car. They can push pedals, turn wheels, move levers, etc. but they cannot drive a car until these responses are chained together, usually in driver's education._

So the process for chaining behaviors is like shaping:

(a) Do a task analysis.

(b) Reinforce successive approximations that are built from behaviors already in the person's repertoire.

Many of the behaviors you will establish involve chaining. Seldom will you establish a behavior that has virtually no relationship to any other behavior in a person's repertoire. Only when children are very small or if you are dealing with handicapped persons, especially the mentally retarded, will you do a great deal of shaping.

To maximize shaping:

(1) Start with behaviors in the student's repertoire.

(2) Start with behaviors the most closely approximate the final form of the targeted behavior.

(3) Remain at a given step only long enough to incorporate it into the student's repertoire. then move on.

(4) If a step cannot be reached, reanalyze the response's task analysis-do a more molecular analysis.

(5) Provide ample opportunity to practice new behavior.

Prompts The key here is S_d_ control and fading. Prompts function to make responses more likely to occur and then are faded so that the child's responding is controlled by stimuli like those that will be encountered in future environments.

Shaping/Chaining Clearly the key principle operating here is reinforcement. Since these new behaviors are so weak, you should use very powerful reinforcers on a CRF schedule. This will quickly establish the responses and then later you should thin the schedules of reinforcement as much as possible. You always want the maximum the child can give for the least amount of reinforcement. Be sure, though, that you are not using aversive

means to stretch the child's skills. Just thin the schedules slowly and the behavior will stabilize without need for aversive motivation such as direct or implied threats.

Also operating here is discriminative stimulus control. The prompts are S_d_s that temporarily guide behavior until they can be faded.

Most of what you do as a teacher will involve shaping, chaining, and prompts therefore there is no "typical" application". Lets hit some common areas, though:

Math: Steps for completing problems are the components of a task analysis (this term is covered in this handbook). These are taught in a step-by-step fashion with teacher prompts being verbal and written. As the child's accuracy increases, the prompts are faded, the child prompts him/herself and thus becomes his/her own teacher.

Reading. Modeling is used to shape reading, chaining is used when students read known words in new sentences, prompts are common when children have decoding problems, and rate of reading is increased thus shaping reading rate.

These examples are, of course, merely a few of many ways that shaping, chaining, and prompting is used by teachers. Most of your day will involve prompting, fading prompts, shaping and chaining new behaviors._

(1) Progress stops. The steps of the shaping or chaining program should be inspected. Are they too large?

Is the child bored because they are too small?

Is there adequate reinforcement for correct responding? Perhaps reinforcers should be used twice or three times throughout the response chain if the chain is new or difficult. This would improve motivation and provide the mediators needed to complete the task.

(2) Retention is poor. This typically occurs when the new behavior is trained in isolation. Check the curriculum-is the behavior required to do other tasks or is it just a single, isolated response? Is the response a knowledge level task on Bloom's Taxonomy? If it is, it probably wasn't merged into a larger repertoire of related behaviors. Behaviors are like the walls of a house, they gain their strength by being connected.

Technique Description.

Many different kind of prompts are possible:

(1) Physical prompts. Here the teacher physically prompts the response to occur. This can be holding the child's hand while making a letter to improve handwriting, moving the child's body in a gymnastics exercise, or places the child's finger on the word they are reading.

In short, the teacher makes the response occur.

(2) Verbal prompts. These are typically rules or other verbal cue that make a response more likely to occur. These and physical prompts are often used in errorless discrimination.

(3) Within stimulus prompts. These are modification of an antecedent. For example, the child reads "horse" when presented with the word "house". The letters "ou" in house could be capitalized thus "hOUse" and the student required to read all the letters in the word before reading the word itself. The change of letters' properties is a "within-stimulus prompt".

Such prompts are, of course, faded as quickly as possible.

(4) Extra-stimulus prompts. These are additions to an anatecedent . For example, in the "house"-vs-"horse" example given above, the teacher could have used an arrow to point to the "ou" in "house".This arrow is an additional stimulus-not a modification of a stimulus; therefore it is an "extra-stimulus" prompt.

Note that the data suggest that LD students learn more quickly with within stimulus prompts-so use them whenever possible.

Prompts are antecedents that increase the likelihood that a response will occur. They are therefore sometimes called "antecedent stimulus control" because they rely primarily on the antecedents, not the consequences, to control behavior.

As you probably guessed already, no prompts will be of any use in a program where the reinforcers are weak.

As with errorless discrimination, prompts are used throughout education and are so often used that they are often forgotten about as a technique.

Clearly all academic areas, social skills training, and physical activities involve prompts at some point.

(1) The response does not maintain. The key to making prompts work is how quickly they are faded. The fading should occur over only when the response is clearly stabilized._

Also check the child's prerequisites. Are all prerequisites established? Was the task analysis complete? Be sure that the curriculum included a balance of taxonomic objectives.

(2) The child depends on the teacher's prompts. Part of any good program is teaching the child to be independent. When teacher prompts lead to a stabilized response, then have the child prompt him/herself as the teacher did. This is the first step toward self-management.

To maximize prompts, do the following:

(1) Make sure all goals are achievable.

(2) Increase each goals slowly, only after the previous step has been achieved.

(3) Teach rule-use whenever possible.

(4) Use stimuli familiar to the student (avoid vague vocabulary, obscure references. Prompts should not be difficult to interpret.

(5) Fade prompts as quickly as possible.

Task analysis is citing the component parts of a response. For example, completing the following addition problem may involve:

23

+12

(1) Determining what kind of problem it is (addition).

(2) Finding the first two numbers to add (3 and 2)

(3) Determining whether carrying is necessary (no).

(4) Finding the second two numbers to add (1 and 2)

(5) Checking the answer with a calculator._

The steps 1-5 are one possible task analysis of such an addition problem._

Social skills can also be task analyzed. For example, dealing with an authority figure who has accused one of violating a rule could be handled thus:_

(1) Determining what one is accused of (e.g., stealing)._

(2) Maintaining a calm demeanor (eye contact with the authority figure, a calm tone, smiles to counter the aversive nature of the situation, a clear and concise statement of why one could not have stolen the item).

(3) Stating to the authority figure that one can understand how he/she could have been mistaken.

(4) Shaking hands with the authority figure and parting in good terms.

A task analysis is the elements of a shaping or chaining program i.e., the components of what will be a larger, more complex behavior or series of behaviors.

Therefore task analyses are "curriculum" in the best sense of that word.

Typical application(s)

Task analysis is used in at least the following:

(a) Chaining programs

(b) Shaping programs

(c) Curriculum development, especially when writing IEPs.

(d) When developing procedures or strategies for students to use in "metacognitive" procedures.

(1) The student fails to make subsequent steps. Several problems may be at fault here:

(a) Prerequisites are not established. The student lacks responses that are not contained in the task analysis.

(b) The reinforcers are weak. Shaping and chaining responses cited in task analyses requires good reinforcers just like ant other program.

(c) The steps in the program are too large. This is easily resolved by simply doing a more molecular analysis of the task.

This when student use rules, principles, tactics, strategies, or other verbal S_d_s as prompts to guide them through a task, their behavior is "rule governed". Therefore "rule" here has a very broad meaning.

For example, suppose that a student uses "i before e except after e" to determine whether the correct spelling is "weigh" or "weigh". The rule (a verbal antecedent) is part of the stimulus complex that controls correct responding.

Such rule use is the defining characteristic of "metacognitive" procedures (see this entry). Here rules are used to either (a) choose the correct tactic for solving a problem (e.g., determining that addition, rather than subtraction, must be done to solve a problem) or (b) using a tactic to actually solve the problem (e.g., using an addition tactic to add the problem).

Rule-governed behavior describes a contingency where the antecedent is at least partially verbal e.g., a rule, procedure, tactic, mnemonic code, etc. Therefore the key stimulus control variable is verbal.

Many times a behavior will be established as a rule-governed response but then become so efficient that the rule is no longer necessary. For example, the student may, after several usages of the "i before e" rule, simply spell "weight" without having to use the rule. A this point the response is no longer rule-governed.

Typical application(s)_

(1) Metacognition (see this entry and the above)

(2) Social skills. Skillstreaming is rule-governed behavior as is any program that has a student follow a proscribed procedure for solving a problem.

(1) Rules for social skills are not used in the natural setting-only in counseling or the social skills class. Generalization training is necessary. See the skillstreaming entry and pay special attention to "announced" and "unannounced" barbs.

(2) Rules for academics are not used outside of special class. Again, generalization training is necessary. The rule use must be reinforced outside of class-nothing else will work. Therefore work with regular education teachers and others to have them prompt and reinforce rule use.

Must people seem to believe that concepts are somehow not mysical. A closer look, though, reveals that concepts are very such real things, or more precisely, properties of things. In fact, we can redefine "concepts" as "properties".

For example, the property "round" is a part of numerous objects from balls to planets. Being round is clearly a property of a physical thing. Therefore the concept "round" is antecedents control exerted by the property "round".

More complex concepts are simply more subtle types of antecedents control. For example, "love" is different for all of us, but it is nonetheless, a property of things for all of us. The easiest way to see this is to take all the possible uses of the word "love" and determine how one uses it. I can say that "I love my Corvette". This means that I an reinforced by the presence of my Corvette, the acceleration, the comfort, the great stereo, etc.

This accumulation of antecedents constitutes why I say "I love my Corvette". Note that it is not simply one of these properties (speed, power, comfort, etc.), it is all of them.

"I love my kids" is clearly different. Here the stimulus control is their behavior, their affection, their learning, their play, etc. that positively reinforce me. These properties of my children's behavior constitute the reason why I say "I love my kids".

In short, the properties of stimuli that control the use of "conceptual" words like "liberty", "freedom", "democracy", etc. are the meaning of these words-not some mentalistic, ephemeral,

inaccessible "thing", but rather specific physical properties of things. Understand the antecedent(s) controlling the use of "conceptual" words, and you will understand what that word "means" to the person.

Total antecedents aren't necessary to control behavior-properties of things can also control behavior. When such properties are shared by more than one thing (e.g., freedom involves properties from voting, speaking out, town meetings, etc.), then the property defines a "concept".

All teachers have to train concepts. To understand how to train them most effectively, you must understand what they are-they are not nebulous, mentalistic things, they are properties of physical things. As you may recall from my early lectures, "Reduce abstractions to the concrete level". Concepts are such abstractions that are only understood when reduced to the physical events that control "conceptual" words such as freedom, love, liberty, kindness, fellowship, etc.

(1) The key stimulus property does not gain control. This problem is typically linked to (a) not presenting enough examples with the property or (b) not presenting enough examples of stimuli without the property, and (c) not adequately comparing events with and without the property. In short, more examples are needed.

Also be sure to check the reinforcer's strength._

Technique Description_

Discrimination training involves teaching the student to respond only when a given stimulus is present.

This involves using positive and negative examples defined thus:

(a) Positive examples have the key stimulus present.

(b) Negative examples do not have the stimulus present.

For example, a student may read "horse" when presented with "house" or "horse". Clearly the key stimuli "ou" is not controlling his reading. Discrimination training involves reinforcing "house" when the child is presented with positive examples such as "house" in differing type fonts and sizes or different reading contexts.

Conversely, reading negative examples such as "Horse", "Hurry", and "Harold" in similar contexts are not read as horse or read correctly.

An example from math would be having the student circle all the problems that are addition (the "+" is the controlling antecedents) and crossing out all the subtraction problems ("-" is antecedent). Then the student would describe how to solve the circled problems (a "metacognitive" procedure-see that entry)-vs-how to solve the

crossed out problems (again, a "metacognitive" tactic)._

An obvious discrimination occurs when a substitute teacher arrives-often classroom management deteriorates until controls are reestablished. The antecedent for good behavior (the regular teacher) is absent to other so, typically, classroom demeanor deteriorates. Of course some teachers have established classroom

management tactics that do not depend on just their presence (which is typically and antecedent for punishment if misbehaviors occur) and instead have training self-control procedures, which is what you want.

These examples indicate that discrimination training is a tactic for learning to respond one way when a given stimulus is present and respond another way when that stimulus is not present. It is used constantly in all academic, social, and self-control situations. It is so basic a process that it could easily be missed.

Discrimination training may seem obvious and simple but too little of it occurs in schools and basal series. Too little discrimination training occurs in taxonomically balanced curricula and too little occurs in maintenance training.

Two contingencies are involved in discrimination training:

S+ (the positive stimulus)==> Response appropriate =>Consequence to S+

S- (the negative stimulus)==> Either a response ==>Consequence appropriate to the S-

or_

Simply not doing the

response for S+

Note that the consequences can be positive or negative. When the consequence is punishment, the probability that a given response will occur in the presence of an antecedent decreases. For example, children do not do certain misbehaviors in the presence of their teacher (the antecedent for punishment of such behavior) but will do such behaviors with their peers (an antecedent for reinforcement of such behaviors).

Discrimination training occurs in all academic, social, and self-control areas. It is so basic that it must be dealt with no matter what you are doing.

(1) The response appropriate to S+ or S- occurs when it shouldn't. Several problems can cause this:_

(a) The reinforcers are not powerful enough._

(b) The differences between the stimuli need to be accentuated (see "within stimulus prompts" and "extra-stimulus prompts" in the "prompts" entry).

(c) Sometimes you merely have to point out what you are trying to do: Be sure to simply explain to the child what you want them to learn i.e, state outright the kind of discrimination you want them to make: "Jake, you will circle all the addition problems-the ones with a `+'-and cross out all the subtraction problems-the ones with the `-'."

(2) The discrimination doesn't maintain or generalize. Generalization, like any response, has to be trained. Be sure to at least do the following:

(a) Reinforce it in new situations.

(b) Have other reinforce it in new situations.

(c) Do enough examples distributed over time to ensure that the discrimination is well established.

(d) Train self-control procedures (such as rules) that the student can use to make the proper discrimination when it is required e.g., "i before e...."

Bridging Term: Beyond the classroom

Generalization and maintenance training can be accomplished in a number of ways:

(1) Train the response to a high level of efficiency.

(2) Review the response often and within a taxonomically balanced set of objectives.

(3) Reinforce the response in other settings-this is a key tactic.

(4) Use a wide range of positive and negative examples ("exemplars" is the technical name).

(5) Train the response until tightly monitored contingencies are no longer necessary.

(6) Train so that reinforcers occurring in natural settings will strengthen the response. For example, reinforcers the regular education teacher, peers, and parents use.

(7) Minimize errors during training (see "errorless discrimination" entry)

(8) Thin reinforcement to an intermittent schedule.

(9) Train the student to use rules to guide responding in novel situations.

Generalization is a natural process-you are born able to do it. The tactics listed in 1-8 above will help that natural

behavioral process operate more efficiently.

Typical application(s)

Generalization training should probably occur for all responses. If the response is valuable only in the classroom, be sure to clearly think through its justification. The job of a teacher is to prepare people who can function outside of schools and tightly controlled setting such as classrooms.

If the response does not generalize or maintain-review the suggestions 1-8 above. Be sure that all have been used.

Also be sure that the reinforcers used are powerful enough to affect behavior. Most teachers do not effectively use or even identify the courses of reinforcement at their disposal and often do not individualize the reinforcers for maximal effect.

Self-control procedures are packaged in a number of ways including skillstreaming, metacognitive strategies, aggression replacement training (see entries on these). But in all cases, the self-control training is essentially the same:

(a) A rule or strategy is taught.

(b) The student learns to discriminate when that rule is to be used.

(c) The student uses the rule and evaluates his/her use of it.

(d) The student self-reinforces (or, less frequently self-punishes).

Antecedent stimulus control is important here, therefore discrimination training is also important (see that entry). Note that steps b-d above each require that a discrimination be established. In b, the discrimination is where the rule should be used; in c, the evaluation of the performance; in d, the evaluation in c is used to determine whether reinforcement is to be provided.

Please see these entries for examples:

(a) Rule-governed behavior

(b) Aggression replacement training

(c) Skillstreaming

(d) Metacognition

Note that self-management is often used in classroom management to train students to control their own behavior, set reasonable goals, and generalize appropriate responding across settings.

(1) Discrimination. The discrimination training needed to identify when an appropriate response has occurred is essential. Many times the student cannot determine when a correct response has or has not occurred. They also do not accurately judge how they have affected others, therefore they cannot assess the social damage their inappropriate behavior has had.

(2) Self-reinforcement is a problem. Students typically cannot reinforce themselves very effectively. Their self-praiseis a weak reinforcer that must be strengthened by pairing it with activity reinforcers, praise from you and other significant people.

(3) Self-control does not generalize. See the entries under generalization and be sure to have done the following:

(a) Use rules to help mediate control (see rule-governed behavior entry).

(b) Use confederates to reinforce the behavior in multiple settings.

Students earn the opportunity to be in a drawing for a large reinforcer. For example, each homework assignment with an "A" or "B" grade earns one entry form that is put into a barrel. Then after two weeks, a student pulls out a name of a person who wins a large reinforcer (tickets to a movie, tokens for use at video arcades, two hours of freetime shooting baskets with a friend of their choice, etc.)

Such large magnitude reinforcers motivate a lot of behavior and only one reinforcer needs to be arranged._

Clearly positive reinforcement is the key here. The large reinforcer motivates and simplifies programming. The key is to ensure that the drawing is truly random so that charges of being unfair cannot arise.

As with state lotteries, you may do well to have several different lotteries with differing probabilities of being a

winner. The cost of each should, of course, correspond to the potential gain.

(1) Group contingencies. Groups of individuals can be reinforced with daily behavior reports from regular educators. These reports then serve as lottery tickets if they reflect some criterion level performance.

(2) Yoked contingencies. Individuals can earn lottery tickets for each other. Then if one wins, they both share in the prize.

Note that typical uses involve reports from individuals working with your students. Therefore the lottery is a way that you can motivate good behavior in settings other than your own.

(1) No one cares. Check the reinforcer. It clearly is not motivating.

(2) Too many losers. Have several winners or have winners share the prize with others e.g., 50 minutes of free time on a Friday could be spent with someone else in one-on-one basketball during the last hour of the day. Do more drawings.

Token economies are a large class of programs that have this feature in common: A medium of exchange (such as a token or points) is used to purchase reinforcers.

Tokens are awarded following targeted responses. Students then save the number of tokens needed to purchase a reward e.g., freetime, extra recess, popcorn party, lunch with the teacher at McDonalds, or even home-based rewards such as a friend staying over, late night TV with a friend, etc.

Personal note: Most classroom management systems are one form or another of token system. Typically teachers have a set of class rules that are posted in the room. Following the rules results in earning points (tokens) that are exchanged for privileges (Premack Principle reinforcement) or other primary and secondary reinforcers. Response cost takes the form of fines for inappropriate behavior, thus children lose points.

This type of system is usually lauded by the educational establishment, so we should be wary right off the bat. The facts are these:

(1) Token economies were developed to control the behavior of back-ward schizophrenics at an institution in Illinois. This group of people had never received treatment and a powerful and radically different program had to be initiated.

(2) Token systems can reduce the likelihood that a child using such a system in special education will engage in such behaviors in any other setting.

(3) Token economies may set the child back developmentally. Here is what I mean: As we move toward the independence required by adulthood, our reinforcers change. We are under many primary reinforcers as an infant but by age 1, we have many secondary reinforcers (attention, praise, interaction with others) that serve as the kinds of reinforcers we will receive throughout our lives. To use tokens with a child who can function without them is to send the child back to an earlier stage of behavioral development, to a dependence on more material reinforcers instead of the kind of secondary reinforcers that will be used in the child's culture. This handicaps the child who must eventually function in that culture.

The problem is that once again education has taken a technique, misused it, and harmed children. Here are some

important guiding principles you should apply when using any technique:

(1) Create a continuum of environments that will lead the child to independence, not dependence. This means teaching the child to function on their own, without the support of a teacher.

(2) Use techniques that have reinforcers like those in the child's future environments. Don't hamstring a child with a token system when they already function without one.

(3) Be sure that you have complete knowledge of the child's reinforcers. Observe the child, what do they like to do? Have the child tell you what they like and use reinforcement surveys, paper and pencil measures of what a child says they like.

(4) Shun the simplicity of powerful programs like token systems that simply make life easy. Anyone can find "the big reinforcer" and motivate a lot of behavior. Real teachers know that their task is not to look good at any cost, but rather to prepare children for the future environments they will face. Therefore build classroom environments that approximate those future settings and training children to function within them.

The token is obviously a secondary reinforcer because it is paired with other reinforcers (the ones purchased). The fact that the token is paired with such a wide range of reinforcers makes it very powerful. This analysis also applies to money. Our culture is on a token economy that generates a full range of appropriate, productive behavior as well as thieves.

Token systems have been used in classroom management systems, on wards with severe behavioral problems, to manage corporate executives, teachers, teacher aides, etc. In short, token systems have been applied to most of the problems faced by educators and educational psychologists.

When token systems go awry, they typically do so for these reasons:

(1) Too many tokens are required for a reward. This results in no one ever receiving a reward therefore frustration sets in.

The solution is obvious: Make sure that sufficient tokens can be earned, especially when the program starts. Then slowly require increased numbers of tokens for purchases.

(2) Record keeping becomes too cumbersome. Make sure that the tokens are easy to keep track of. For example, instead of using poker chips, have students score their points on a chart at the end of each class-and be sure that it is done at the end of each class! Failure to regularly record data means that you lose track and the students can claim they earned more tokens than they actually did.

(3) Response cost contingencies create red ink. Here is what happens: Response cost involves fines e.g., 10 tokens are taken away for hitting, for example. Eventually, the child is fined into the hold-they owe more tokens than they have! This creates trouble because the student has nothing to lose if they misbehave-the teacher already has it all!

The solution is to make sure that response cost is not the centerpiece of the token system. Be sure that students can earn "bonus points" for especially good work so that they can avoid losing so many tokens that no purchases re ever possible. Finally, be sure that criteria for earning tokens are within each student's reach-this may require having different contingencies for each student or for groups of students-that's O.K.

Metacognition teaches students to follow rules, mnemonic codes, and other assists (like "i before e except after c"). Therefore it stresses two responses:

(1) Identifying which technique should be used.

(2) Using the technique correctly.

I shy away from the term "metacognitive" because it does not describe what the student does. Instead, I prefer the term "Rule-governed behavior" because the student is following rules to solve problems-typically rules the student is stating to themselves.

Theoretical analysis

(1) Metacognition is rule-governed behavior. The S_d_ for metacognitive responding includes a verbal component.

(2) The self-control literature fits under the "metacogition" umbrella.

(3) The key to establishing metacognitive repertoires is rewards you give for correctly identifying which

strategy or tactic should be used and then for using it.

Once these rules can be identified and used, you must train the children to do the self-monitoring and self-reinforcement required for metacognitive techniques to succeed.

Metacognitive procedures are used in all academic areas, social skills training, and self-management tactics. Therefore it has academic and clinical applications.

The "SQ3R" procedure is a common example. It stands for "survey", "question", "read", "write", "review". It tells the student to survey a reading assignment before starting. This includes reading headings and attempting to determine what kind of material is being covered.

Question requires writing a question that will be answered while reading. This question is, of course, based on the major headings just surveyed.

Read indicates that one should read to answer the question.

Write has students write summaries of key theses and answers to the questions.

Review is reviewing notes and the chapter.

Many such codes are used at all levels. Doctors in training create codes to memorize names of bones and families of medications-you may have been given a prescription that resulted from the use of a mnemonic code. Students can use similar codes that describe procedures for answering questions.

(1) Code is not easily learned by the student. Have the student come up with the code. The code must relate to something he/she already knows, and preferable likes.

(2) Too many codes are required if material is to be covered. Codes are typically used when memorization is unwieldily. If too many codes are required, it is probably because too many items have to be memorized and this suggests that too many knowledge level objects are being required. Distribute the taxonomic levels. If this is not done, the material will be quickly forgotten, anyway.

Peer tutoring and peer monitoring collectively are most of or more than what is currently called "cooperative learning". In essence, peer tutoring and peer monitoring are organized cooperative learning i.e., a teacher is appointed and trained as such whereas in cooperative learning such roles just tend to emerge.

A peer tutor therefore:_

(a) Has training as a tutor-they are not simply someone who has demonstrated that they can do the task. As you well know, teaching and doing are two different skills.

(b) Is reinforced for changes in another person's behavior-just as you are.

Probably the best way to train peer tutoring is to do the following:

(a) Be sure that the tutor can do the task.

(b) Give the tutor a format for teaching the skill e.g., a series of steps to follow written on a 3" X 5" card.

(c) Monitor the tutor's work and shape the tutor's teaching.

(d) Reinforce the tutor.

The tutor:

(a) Is a model, therefore he/she serves as an antecedent.

(b) Reinforces, i.e, the tutor delivers consequences.

(c) Evaluates the tutee's work, therefore the tutor needs an evaluation system.

Typical application(s)

(1) Review work where serving as a tutor is a reinforcer. The tutor, therefore earns the privilege of tutoring either peers or younger children.

Do not use a tutor to establish new responding-that is best left to you because working with weak behavior is typically far more difficult that maintaining and reviewing well-established responses.

(1) Tutee does not acquire targeted behaviors. Tutors sometimes present more hints than they should. Be sure that the antecedents to be used in training are clearly understood by the tutor.

(2) Tutee motivation falls. Be sure that:

(a) Being tutored by a peer is not punishing (embarrassing)

(b) The tutor is using effective reinforcers.

(c) The tutee and tutor sometimes switch roles so that students do not begin to label tutees as "dumbies" or otherwise inferior.

(3) Tutor and tutee misbehave. Yoke the tutor and tutee so that changes in tutee behavior have reinforcing consequences not only for the tutee but also for the tutor. Therefore it is in both their interests for learning to occur.

Skillstreaming and aggression replacement training are Dr. Arnold Goldstein's techniques for developing social skills in children, adolescents, and adults. The key difference between these approaches is the types of behaviors worked with-Skillstreaming deals with social skills, aggression replacement training deals with controlling aggression.

The essentials of these techniques are these:_

(1) Create group. Groups of about 5-8 students with social skills problems are created._

(2) Assessment. The group leader determines what the social skills problems are from referrals, discussions with other teachers, and typical types of problems that persons that age actually have.

(3) Vignettes. The problems found in (b) are written into vignettes that the students read. Note that these vignettes are based on real problems-but do not identify who specifically is having the problem.

(4) Brainstorming. The group makes suggestions for handling the problems cited in the vignettes. For example, the vignette include a student who being tempted by peers to do a B & E (breaking and entering/Burglary). The student wants to keep these friends but does not want to get into any more trouble with the law (he has a history). How will the student handle this?

Alternatively, how does a girl handle a boy who wants to have sex? She wants to keep him as a boy friend, but not have sex. Similary types of problems involving drugs, others' aggression, gangs, etc. are created.

(5) In-session modeling. The students act out the vignettes during the session. They try a number of alternatives for solving the problem and are encouraged to speak freely about the efficacy of each.

(6) Generalization training. The in-class training can usually establish appropriate social behaviors, but these

responses do not generalize. Therefore Goldstein has developed "barbs".

Two kinds of barbs are used:

(a) Announced barbs. Announced barbs are situations like those in vignettes practiced in the session except that (a) they occur outside the session and (b) the person creating the difficult situation announces to the student that a "barb" will take place. Therefore a vignette-like problem is posed by someone other than the in-class trainer (principal, friend of the student, a teacher, etc.) and the problem is labeled as only a "practice run" by calling it a "barb".

(b) Unannounced barbs. Unannounced barbs are like announced barbs only they are not announced as such._

(7) Hassel logs. Hassel logs are records of interactions that cause the student trouble. They are set up thus:_

Situation What you did Consequences_

________________________________________________________________

________________________________________________________________

Etc._

Obviously this is a contingency analysis of some interaction. It identifies problematic situations, responses, and the consequences. Hassel logs then become the foundation for new vignettes and in-class sessions._

Clearly several basic principles are operating here:

(a) Modeling constitutes stimulus control and should be maximized (see that entry).

(b) Discrimination training is needed to cite key elements of contingencies and to assess whether appropriated responding occurred (for self-reinforcement).

This technique is a package that is used to train social skills in children, adolescents, and youth.

(1) Generalization does not occur. The number of and realistic aspects of barbs needs to be considered. Most children and adolescents can readily tell when a barb is very unlikely (they are out of town, the person creating the problem is someone they don't know, etc.) Therefore the key to generalization is self-reinforcement and lots of practice. Unless doing the new social skill is really a reinforcer of someone else will eventually reinforce the response liberally, the probability that the response will occur is greatly damaged. Therefore like always, maximize the reinforcers' efficacy.

(2) Student don't participate in role plays. Sometimes students won't participate in role plays because they are shy, required to be in the social skills program and they don't want to be there, etc. The best tactic is not to force

participation-just let then observer all the other students being reinforced, participating and enjoying it, and obtaining some benefits. Also make sure that the vignettes you are working on deal with problems central to the nonparticipant's life-thus making the skillstreaming more relevant.

Personalized systems of instruction (PSI) have these characteristics:_

(1) Individualized pacing: Each student completes modules at his/her own pace. Testing out of modules is possible and there's no reason that material tested out of couldn't be replaced with other material of the student's choosing.

(2) Complete task analysis: The curriculum is clearly delineated and includes all levels of Bloom's taxonomy. The curriculum thus created serves as the curriculum-based assessment tool for assessing student progress.

(3) Typically each class is run this way:

(a) A pretest is given. Passing the pretest allows the student to be excused from the class._

(b) Students do projects, discuss readings, etc.

(c) A posttest is given

Students failing the posttest must do remedial work and pass the posttest. Such remediation is often done by teaching assistants, programmed instruction, or computer-assisted instruc