Copyright © 2007-2018 Russ Dewey
Applied Behavior Analysis
In the preceding section of this chapter you were introduced to basic concepts of operant conditioning. A psychologist who applies this knowledge outside the laboratory is called an applied behavior analyst.
The applied behavior analyst has two main tools: (1) systematic arrangement of consequences (reinforcement and punishment) and (2) careful analysis and arrangement of antecedents (S+ and S-).
The arrangement of consequences for behaviors is called contingency management. A contingency is a dependency between events, such as delivery of food when a behavior is performed.
Contingency management is used whenever creatures are motivated by incentives (such as getting paid for a job) or penalties (such as paying a fine for doing something wrong).
One professor started a graduate class on applied behavior analysis by generating a huge list of problems. The professor told the students, "Think of any problem a person can have, and we will design a behavioral therapy for it. Let's get a big list on the board."
At first the class responded slowly. Someone suggested "marriage problems" and the professor said, "Put it in terms of behaviors." "OK," said the student, "let's say the problem is not enough time spent together."
The professor wrote that on the board. Another student suggested the problem of eliminating "writer's block," defined behaviorally as increasing writing output.
Another student suggested the problem of eliminating unwanted involuntary movements (tics). Other students mentioned eliminating nailbiting and quitting cigarettes.
How did a professor start a class on applied behavior analysis, and what was the point?
As the list grew, the students realized this process could take quite a while. The list of possible human problems is neverending.
Most problems can be defined in behavioral terms. In other words, most problems can be described in terms of some observable, measurable activity (behavior) that people wish to make more frequent or less frequent.
Behavior therapy consists of applying conditioning techniques to make the desired changes. That means altering (managing) contingencies.
After the list filled the board, the professor gave the class its assignment. Each student had to select a problem and, by the end of the term, design a behavioral treatment for it.
The professor was making a point, not just an assignment. Behavioral approaches can be applied to virtually any conceivable problem.
The first step in applied behavior analysis is to analyze the problem in a behavioral way. That means stating the problem in terms of behaviors, controlling stimuli, reinforcers, punishers, or observational learning: the concepts we have covered in this chapter.
Antecedent and consequent stimuli must be identified. After this analysis, one can make an educated guess about which intervention strategy or "treatment" might be best.
Lindsley's Simplified Precision Model
Green and Morrow (1974) offered a convenient, four-step guide to successful behavior change. Developed by Ogden Lindsley, it is called Lindsley's Simplified Precision Model.
1. Pinpoint the target behavior to be modified.
2. Record the rate of that behavior.
3. Change consequences of the behavior.
4. If the first try does not succeed, try and try again with revised procedures.
What is Lindsley's Simplified Precision Model?
The first step in Lindsley's list was to pinpoint the behavior to be modified. This is often the most crucial step. If you cannot specify a behavior, how can you modify it?
Behavior modifiers and therapists sometimes go to heroic lengths to identify specific, modifiable behaviors. For example, a team of behavior therapists at a speech clinic came up with 49 different, specific verbal problems. The checklist included the following:
1. Overtalk. A person speaks considerably more than is appropriate.
2. Undertalk. A person doesn't speak enough.
3. Fast talk. A person talks too fast.
4. Slow talk. A person speaks too slowly.
5. Loud talk. A person speaks too loudly.
6. Quiet talk. A person speaks too softly.
7. Singsong speech. A person talks as if singing or chanting.
8. Monotone speech. A person speaks with unvarying tone.
9. Rapid latency. A person speaks too quickly after someone else.
10. Slow latency. A person responds only very slowly.
11. Affective talk. A person talks with great emotion, crying, whining, screaming, or speaking with a shaky voice.
12. Unaffective talk. A person speaks in a flat, unemotional voice even when emotion is appropriate.
13. Obtrusions. A person too often "butts in" to conversation.
...and the list goes on, up to #49, which is "illogical talk."
(Adapted from Thomas, Walter & O'Flaherty, 1974, pp. 248-251.)
What "heroic" efforts are exemplified by the list of speech problems?
The problem is specified in terms of something that can be measured or detected, a behavior. Then a therapist can attempt to change the antecedents or consequences of the behavior, to alter its frequency.
After specifying a behavior to be changed, the next step is to take some measurements to see how frequently it is occurring before any attempt at intervention. Baselining is keeping a careful record of how often a behavior occurs, without trying to alter it.
The purpose of baselining is to establish a point of reference, so one can determine later if interventions have any effect. A genuine behavior change (as opposed to a random variation in the frequency of a behavior) should stand out sharply from the baseline rate of behavior.
What is baselining? What is its purpose? How long should baselining continue, as a rule?
As a general rule, baselining should continue until there is a definite pattern. If the frequency of the behavior varies a lot, baseline observations should continue for a long time. If the behavior is produced at a steady rate, the baseline observation period can be short.
While taking baseline measurements, an applied behavior analyst should pay careful attention to antecedent stimuli: those that come before the behavior. As we saw earlier, discriminative stimuli (both S+s and S-s) can control behavior, turning it on or off.
One weakness of Lindsley's Simplified Precision Model is that it does not mention antecedents. It only mentioned changing consequences of a behavior.
In what important respect was Lindsley's model incomplete?
Often the relevance of antecedents will be obvious, once they are noticed. For example, if a child's cries every day when dropped off at a nursery school, a behavior analyst might target that event.
With a little ingenuity, it might be possible to arrange a reinforcing event to occur when the child is dropped off. That might reduce the child's distress.
Self-Monitoring as a Method of Behavior Change
Sometimes behavior will change during a baseline observation period, due to the measurement itself. Measurement of one's own behavior is called self-
For example, many people wish to lose weight, but few actually keep a detailed record of calorie intake. People who keep a record of every calorie consumed often find that they lose weight as a result, with no other intervention.
What is self-monitoring? What sorts of problems respond well to self-monitoring?
Self-monitoring works especially well with impulsive habits, like snacking, cigarette smoking, or TV-watching. These are all things a person may start to do without thinking.
Self-monitoring also draws attention to the consequences of behavior. "If I eat this, I am over my limit," or "If I start watching TV I won't get my homework done."
Self-monitoring, as a behavior change procedure, lacks any specially arranged reinforcement or punishment. However, it forces attention to natural reinforcements and punishments, and often that is enough to produce behavior change.
Green, J. & Morrow, W. (1974). Precision social work. In E. Thomas (Ed.) Behavior modification procedure: A sourcebook. Chicago, IL: Aldine Publishing Company.
Thomas, E. J., Walter, C. L., & O'Flaherty, K. (1974). A verbal problem checklist for use in assessing family verbal behavior. Behavior Therapy, 5, 235-246.
Write to Dr. Dewey at firstname.lastname@example.org.