Go back to Higher Education Articles & Chapters

 

Behavioral Systems Analysis and Higher Education
Richard W. Malott1

Behavior Analysis Program
Department of Psychology
Western Michigan University

Download Word version of this article

Abstract


I suggest that we behavior-analysis college professors practice our preaching, that we apply behavior analysis, organizational behavior management, and behavioral systems analysis to our university instruction. I suggest that we college professors apply to what we do most (teaching) the approaches and philosophy we know works everywhere else—behavior analysis and all it implies.

Critique


1950 A.D.
Fred R. Malott, MD: The practice of medicine wouldn’t be so bad, if it weren’t for the damn patients.

1965 A.D.
Prototypical Manager: The practice of management wouldn’t be so bad, if it weren’t for the damn workers. Labor and management are usually at war. (Based on remarks by participants in organizational behavior management workshops I conducted at the University of Michigan for business managers from around the country.)

1970 A.D.
OBM Safety Consultant: The reason workers injure themselves is so they can get insurance benefits.
Behaviorman: OBM’ers, please don’t blame the victims.

1975 A.D.
Western Michigan University faculty organizing to form our first labor union.

Irate WMU Professor: The people in the administration are basically bad human beings.

Come on, man, they’re just faculty members, like you and me, only they happened to get promoted to administration.

Irate WMU Professor: No, it takes a certain type of nasty personality to become a university administrator.

The faculty and administration are usually at war.
Then as a single organism, the faculty grabbed picket signs and marched around our administration building singing “We Shall Overcome.”

Implications: The plight of these highly educated, white, privileged university professors was comparable to that of the oppressed, segregated, discriminated-against, black Americans who marched in protest on Birmingham, Alabama in 1963.

Implications: Our protest was as noble as the protest of those black Americans. Implying that our faculty union organizer was a reincarnation of Martin Luther King Jr.

This was one of my most embarrassing moments. I’d have given a lot not to have found myself among those singing faculty protesters. We professors are capable of such sanctimonious BS.

But I digress. The point is: The faculty is usually at war with the administration AND the students. The faculty feels that teaching wouldn’t be so bad if we poor faculty members weren’t caught between a lazy, incompetent, psychopathic administration and a a lazy, incompetent, psychopathic student body, neither of whom appreciates how wonderful we faculty members really are and how hard we work.
Awww, poor babies.

Still not convinced, huh? Need some more of those hardcore scientific data? Here they come.

1980 A.D.
Prototypical Faculty Member: Teaching wouldn’t be so bad, if it weren’t for those damn students. Students today are not as serious as when we were students. And they are not as well prepared as when we were students. Why don’t we have good students here, like the ones at
Harvard University?

Behaviorman: Please don’t blame the victims.
Faculty and students are usually at war.

1985 A.D.
Famous Behavior Analyst: Many students fail to study enough for my courses. That’s because other things, like their social lives, have a higher priority. Those students have decided to pursue their social lives, rather than their academic career.

Behaviorman: Come on, man, that’s just cheap cognitive rationalization. The reason they don’t study is that you haven’t made the effort to arrange effective performance-management contingencies to support their studying.
Behavior Woman: Behavior analysts, please don’t blame the victim.

2000 A.D.
New WMU Psych Faculty Member: The grad students at WMU are a bunch of whiners. When I assign them a few books to read, they complain. Not like when I was in grad school.

Charles Darwin: Like there’s been a major genetic drift in the five years since you got your PhD.
Behaviorman: New PhDs, please don’t blame the victims.
The point: Even behavior analysts fail to appreciate B. F. Skinner’s dictum: The subject is always right.

If the students aren’t pressing the lever the way we think they should, it’s our fault, not the students. There’s no such thing as a dumb or lazy rat. And there’s no such thing as a dumb or lazy student. There are only dumb and lazy professors.

2002 A.D.
WMU Faculty Union Newsletter. Article by a distinguished behavior analyst: Today’s students aren’t serious about their education. All they care about is their new surround-sound entertainment centers, in their lavishly furnished, upscale apartments, and their new sports cars. To support their decadent lifestyles, they work at outside jobs 20 to 40 hours per week. The result is they have neither the time nor the energy to study, and they fall asleep when they do manage to make it to class.
W. H. Auden, British Poet, 1907-1973: A professor is one who talks in someone else’s sleep.

I’ve done a lot of academic and career counseling with WMU undergrads, and I often find very serious students. If they work fulltime, it’s because they must, in order to go to school. If it weren’t for table waiting and bartending jobs, half our students couldn’t afford to go to school. These working students also often take full-time course loads and still do great work at school and get excellent grades. And they only rarely fall asleep in class, in spite of the chronic state of sleep deprivation most serious students suffer.

But sometimes I fall asleep in my own class.

335 B.C.
Aristotle: The new generation is not nearly as good as the earlier generations.
What a dumb ass.

The older and younger generations are usually at war.
The older generation has been dissing the younger generation for at least 2,400 years. If, for the last 2,400 years, each generation had gotten worse than the previous, by now, we’d all be sitting on tree limbs, picking at cooties, not writing our scholarly papers, nor presenting those papers at wonderful behavior-analysis conferences.

Behavior Woman: Older gen, please stop blaming the victims.
Now, just a little more of those hard-core scientific data:

The Faculty’s Favorite Comment
The Faculty’s Favorite Comment: Half the students failed my midterm.

The Faculty’s Favorite Inference: That shows what a bunch of scumbags our students are.

The Faculty’s Second Favorite Inference: That shows what high standards I have.

Cheap Thrills: Self-righteous indignation is one of our biggest reinforcers. Professors get it off by being indignant about half their students failing the midterm.

Cheap Thrills: I get it off by being indignant about professors failing half their students and then blaming those students.

Behavior Woman: Professors, please don’t blame the student victims. It’s really we professors who have failed.

Recommendations:


At last, time to shift from critique to recommendations:
Remember Skinner’s dictum: The subject is always right.

The students are always right. The students you have to work with are the students you have to work with. So design educational systems to accommodate your students’ entering repertoires and values.
As a teacher, your semester’s goals should be that your students learn as much behavior-analysis as is humanely possible and that they
love behavior analysis.

Let me rephrase that: Your semester’s goals are that your students ACQUIRE AS MANY BEHAVIOR ANALYSIS SKILLS as is humanely possible and that they love it.

The less we talk about learning and the less we talk about education and the more we talk, and think, and teach in terms of skills training, the more successful we will be in helping our students acquire a functional and lasting repertoire. We should think of ourselves as running a trade school, even though that trade may involve very complex, subtle, higher-order skills. But we should not think it easy to train complex, subtle, higher-order skills. We will have limited quantitative and qualitative success in such training.

These issues lead to the following three models.

The Three Models

The Skills-training Model of Education
The philosophy behind the Skills-training Model is well expressed in a message from a Chinese fortune cookie: Education that does not lead to action is wasted.

The Skills-training Model of Education has four components:
o concept and principles training
o strategy training
o knowledge teaching
o appreciation enhancing
Traditional university education stresses knowledge learning and ignores skills training and appreciation.
As a teacher, your semester’s goals should be that your students become eager, skilled behavior analysts. But that’s a big task for your students, which leads to another model.

The Performance-Management Model of Task Accomplishment
The 5 steps to accomplishing any big task:
o Divide the big task into many small tasks.
o Specify exactly what each small task is and how to accomplish it.
o Have frequent deadlines, for the tasks.
o Monitor the task accomplishment.
o Have small but significant outcomes for each monitored task.
Incidentally, this model applies not only to tasks in education but also to motor-skills training (e.g., sports), OBM, self-management, behavioral medicine, and clinical interventions.

Professor Tradition: Unnecessary coddling. Just tell the little brats what they’re supposed to do and toss them out of school, if don’t do it. I certainly manage to teach my behavior analysis classes and prepare my scholarly papers without procrastination; so I have no sympathy for these procrastinating students.

Behaviorman: Ninety percent of the professional behavior analysts don’t get their Association for Behavior Analysis presentation proposal in, until 5 days before the deadline. Seventy percent don’t get their papers finished until they’re at the ABA conference. And 60% should have spent 20 more hours preparing their papers. But, no doubt, you’re the exception, Professor Tradition.

Behavior Woman: Show me a man who says he doesn’t procrastinate and I’ll show you a liar, or at least a person who has impressively poor self-awareness.

Then why does procrastination rein?

The Three-Contingency Model of Performance Management
The Three Contingency Model of Performance Management explains why we procrastinate and how we can prevent procrastination (Malott, 1989; Malott, & Garcia, 1991). We procrastinate, because the outcomes of our behavior do not reinforce that behavior, because those outcomes are too small or too improbable. The delay of the outcomes are irrelevant.

The Ineffective Natural Contingency
Suppose a student has an assignment due one month from now (e.g., an exam or a term paper). Will he start working on the assignment right now? To answer this, we need to consider the contingency operating on his starting to work on the assignment. This contingency is shown as the first diagram in Figure 1. It is an avoidance contingency: The student’s working on the assignment will avoid his getting a few minutes closer to class without some preparation done. But when the student is a month away from the due date, getting just a few minutes closer to that distant time with no preparation (or no new preparation) is not very aversive, probably not aversive enough to support his starting to work (after all, he can always start working on the assignment a few minutes from now, or a few minutes after that, or a few minutes after that, etc.). So the student procrastinates. Thus this natural contingency is ineffective; the difference between the before and after conditions are too small to reinforce the avoidance response.

Professor Tradition: Even so, my good man, I still have no sympathy for the procrastinating proletariat. I always get my lectures prepared in time for class.

Behavior Woman: And you always start preparing in sufficient time that you end up with the high-quality lecture you had hoped for? Maybe, but probably not.

If this ineffective, natural avoidance contingency is so pervasive, how do students manage to get anything at all done? Why don’t they procrastinate on everything, to the point of no return? Because, eventually, the student gets close enough to the due date that being just a few minutes closer without some preparation will be fairly aversive. And now, the decrement in aversiveness from the before to the after condition is large enough to support the avoidance behavior of starting to work on the assignment. Then the natural avoidance contingency becomes effective. But often the before condition does not become aversive enough soon enough, and the student does not have enough time left to prepare adequately before the due date. The student’s preschool behavioral history determines how close to the deadline he or she must get for the proximity to be sufficiently aversive to support the avoidance behavior of working on the assignment. (Malott, in press)

Professor Tradition: Stop right there. We professors can’t do anything about our student’s preschool behavioral history.

Behaviorman: No, but we can use performance management to get rid of most of the student’s procrastination, and some of our own too.

Let us now apply the Three Contingency Model of Performance Management to the prevention of procrastination.

The Effective Performance-Management Contingency

Recall the Performance-Management Model of Task Accomplishment.
The 5 steps to accomplishing any big task:
o Divide the big task into many small tasks. Suppose the big task is reading and studying a textbook for a final exam. The professor might divide the textbook into single-chapter assignments.
o Specify exactly what each small task is and how to accomplish it. The professor might provide a set of detailed study objectives for each chapter.
o Have frequent deadlines, for the tasks. The professor might give a quiz over a new chapter at each class meeting.
o Monitor the task accomplishment. Of course, the professor would then administer and evaluate those quizzes.
o Have small but significant outcomes for each monitored task. A few points that count toward the final class grade could be contingent on each quiz score.

This performance-management effort results in the second contingency in Figure 1, the effective performance-management contingency, an avoidance contingency: By starting to work on the assignment, the student will avoid the loss of the opportunity to earn the course points.
At first glance, this contingency looks like an analog to a reinforcement contingency, where starting to work on the assignment produces the delayed reinforcer of course points or at least the delayed reinforcer of the opportunity to earn the course points. But the imposition of the deadline drastically changes the contingency. With the deadline, if the student does not start to work soon enough to finish the assignment (e.g., three hours before class), he will lose the opportunity to earn 100% of the course points, an analog to avoidance of the loss of a reinforcer.

Thus, with the deadline, our savior, aversive control, comes to our rescue; for, without the deadline and the resultant aversive control, the student would be in forever-procrastination land. Without the deadline, the resulting analog to a reinforcement contingency would be too tolerant of procrastination. It is deadline-induced analog avoidance contingencies that keep our world in its orderly orbit.

By analog contingencies, I mean indirect-acting contingencies where the outcome is too delayed to reinforce or punish the behavior. Research suggests that outcomes delayed much more than sixty seconds after the response are too delayed to reinforce or punish that response. Yet, some sorts of contingencies with delayed outcomes are effective in controlling our behavior but, if and only if, we know the rules describing those contingencies. I call such effective, delayed-outcome contingencies, analog contingencies. For example, suppose the student knows that, if he starts studying three hours before class, he will avoid the loss of the opportunity to get a good grade. The outcome is too delayed to reinforce that avoidance response, yet this rule-governed analog to an avoidance contingency will probably control the student’s behavior.

It is not that we couldn’t, in theory, control assignment preparation with direct-acting reinforcement contingencies; but, in fact, we usually do not: If the student were deprived of food to 80% of his free-feeding weight, and, if every time he read a paragraph or rehearsed a definition, we gave him a bite of food, then we could readily get a high rate of assignment preparation, with simple reinforcement contingencies. But that ain’t going to happen. And with instrumental reinforcers like course points, rather than hedonic reinforcers like food, and with the delayed delivery of those instrumental reinforcers or the delayed opportunity to earn them, we have analogs to reinforcement contingencies; and such analog reinforcement contingencies result in procrastination and poor student performance.

The Direct-Acting Contingency
A theoretical question remains: How is it that knowing the rule describing the effective performance-management contingency causes that contingency to control the behavior, even though it is only a delayed analog to an avoidance contingency, rather than a direct-acting avoidance contingency where the outcome would be immediate enough to reinforce the response? I think the control of such a performance-management contingency is indirect. The statement of the rule describing the contingency creates an aversive condition--proximity to the deadline without starting to work on the assignment. This is a conditional aversive stimulus; being close to the deadline is maximally aversive, only if the student has not started working on the assignment, in other words, conditional on the student’s not having started working. The student can escape this conditional aversive stimulus by starting to work on the assignment; and the resultant, immediate reduction in the aversiveness reinforces that escape response (see Fig. 2). That does not mean the student is home free; that does not mean being close to the deadline is not still somewhat aversive, even though the student is now working on the assignment; it is just a little less aversive than if he were not working, enough less aversive to reinforce the escape response of starting to work.

It is a fact that this contingency exists. It is an inference it is in an effective escape contingency. In other words, it is an inference that the before condition is an aversive condition for the student and that there is a sufficient difference in the aversiveness between the before and after conditions to reinforce the escape response of starting to work on the assignment. But, I infer that if the indirect-acting, rule-governed, analog-to-avoidance performance-management contingency is effective, it must be because of this escape contingency.

I think it is important to drill down into the conditional aversive before condition that “motivates” the escape behavior (working on the assignment). But the terminology is awkward. I find it handier to simply use the common-sense term fear to label that aversive before condition (see the third contingency diagram in Fig. 1). However, anyone uncomfortable with the fear terminology but otherwise comfortable with the three-contingency model is free to use the conditional-aversive stimulus terminology, which is more specific.

As I mentioned earlier, how close students get to the deadline before starting to work is a function of their early behavioral histories. We can understand that better, in terms of the inferred, direct-acting escape contingency of Fig. 1. Students differ greatly in terms of how close to the deadline they get before they start working on their assignments. I suggest that this means they differ greatly in terms of how close to the deadline they get, before fear kicks in, before the conditional stimulus of proximity to the deadline and not working on the assignment becomes sufficiently aversive. And I suggest that this difference among students is a function of differences in their early behavioral histories, differences in parental role modeling and direct parental intervention.

For example, if Mama gets hysterical when she has not started working on the task, even though she is a long way from the deadline, this analog pairing of aversive hysteria and a distant deadline will cause such conditions to be aversive for the child with his own distant deadlines. But more to the point, suppose Mama also gets hysterical and starts shouting at the child or merely gets subtly critical of the child or starts making dire predictions of failure, when the child has not started working even though the deadline is still distant. That pairing of aversiveness with not working even in the presence of a distant deadline will result in a child, a college student, an adult who will begin to fear failure much earlier than many of his compatriots, and who will thus begin working on assignments or other projects much earlier than his compatriots, and thus a person who will be much more successful as a student and as a professional than will his compatriots. But that success has a price—a life much fuller of fear of failure (again, for more details see Malott, in press).

Thus, we have seen how the Three Contingency Model of Performance Management explains not only why we procrastinate but also how we can prevent procrastination, when necessary. We prevent procrastination by adding performance management contingencies; and such performance management is necessary for those of us who have not been raised by appropriately hysterical parents (Malott, in press).

An Example: BATS


Most of the details of this approach to higher education have been previously described (Heward, 1994; Jackson, & Malott, 1994; Malott 1993, October; Malott, Vunovich, Boettcher, & Groeger, (1995). Therefore, I will briefly summarize some of those details and then add a little enrichment.

Over the 40 years, my students and I have evolved four different, “major” instructional systems, sometimes they run concurrently, but usually I get bored with one, after five or ten years and feel guilty because I have not been doing enough writing and publishing. Then I hand off the instructional system or close it down, so I can create my own personal, publish-or-perish, academic monastery. But then I feel guilty again, this time because I am not training enough students; so I get an idea for a great new instructional system that will clearly save the world. But the new system turns out to be only a minor variation of the previous ones, which is fine because I think the previous ones were good enough.

My current and probably final system is BATS, the Behavior Analysis Training System, which started around 1990, with about three MA students and one PhD students. We started BATS when I realized I had already broken my vow never to take on any more grad students and instead just teach a few non-demanding undergrad courses and concentrate on my writing. And now, as of 2003, BATS consists of 3 PhD students, 20 MA students, and 25 BA thesis students.

The main goal of BATS is to produce and professionally place as many well trained behavior analysts as possible, at the BA, MA, and PhD levels. We divide this main goal into two sub-goals.

One sub-goal is to train undergrads in the principles of behavior, the principles of organizational behavior management (OBM) and behavior systems analysis, the general applications of those principles, and the specific skills of discrete-trial training with preschool autistic children. We accomplish this with three sequential undergrad courses that teach the equivalent of about 400 students per year (in fact, many of the students take all three courses, so the number of different students is about 200). Therefore, we have designed BATS to provide those three undergrad courses and related services to 200 undergrad students each year.

The other sub-goal is to train at a more advanced level the BA, MA, and PhD members of BATS in the principles of behavior, the principles of OBM and behavior systems analysis, the general applications of those principles, the specific skills of discrete-trial training with preschool autistic children as well as the skills of training and management of other populations with behavior problems, and specific OBM and behavior-systems-analysis skills. To a large part, we accomplish this as a result of the BATS participation of these advanced undergrad and grad students. They teach the undergrad courses, supervise and provide adjunct services to those courses, and run BATS and its various components.

We are effective in achieving these two sub-goals to the extent that we accomplish another sub-goal—that BATS be an exemplary system, an exemplary organization, illustrating the best of OBM and behavioral systems analysis, a place where we practice our preaching. When we started BATS, we achieved perhaps 15% of that goal—15% reality, 85% hype. Over time, we have reversed those percentages; now we are about 85% of being an exemplary system (OK, maybe only 80% exemplary).

For years, I taught OBM and behavior systems analysis in regular courses and had about as much success as if I had been teaching roller-skating by the lecture method. The only way I have been able to teach OBM and behavior systems analysis so that it melts into the soul of the student is to intensely immerse them in OBM/behavior-systems-analysis systems like BATS. After a year or two in BATS, many, maybe most, students acquire an OBM/systems worldview and OBM/systems skills at an impressive level.


OBM
By OBM (organizational behavior management), I mean setting up procedures that effectively manage the behavior of people in organizations (Malott, Malott, & Shimamune, 1993; Malott, Shimamune, & Malott, 1993). And the people in or served by BATS are the students in our undergrad courses, the MA students teaching those courses, the advanced MA and PhD students supervising those courses, the undergrad and grad students managing subsystems that serve the undergrad and grad students, and the CEO--me. And for BATS, OBM entails using the performance-management model of task accomplishment along with the three-contingency model of performance management.

And all of this is based on a strong commitment to the avoidance of victim blaming. So, if a group within our system is failing to do the tasks they are supposed to do, we do not blame them. Instead, we blame ourselves and implement performance-management systems and performance-management contingencies to support the desired performance, not as easy nor as emotionally gratifying as victim blaming, but much more effective.

For example, we have already discussed performance-management contingencies to support studying for quizzes, but we also have found it necessary to implement contingencies for class attendance because, for a few students, the quiz contingencies are not enough, even though you must be in class to take the quiz. So students who have more than three absences from our 30 class meetings automatically have their final grade reduced by half a letter grade. This prevents all but about 2% of our students from unintentionally drifting into a lower grade than they wanted because they had unintentionally slept-in a few too many times.
At the MA teaching-assistant level, we have found that the TAs do not read the TA procedure manual as reliably as they need to, because the natural contingencies are insufficient. So rather than throw a conniption fit about what a bunch of unconscientious TAs this new generation of grad students are, we implemented review quizzes over the TA manuals. On the other hand, the natural contingencies usually support the TAs reviewing the days assignment they will be teaching, so we have not needed to add a performance-management contingency for that.

In hierarchical organizations such as BATS, performance often falls apart at the top, because the CEO reports to no one (even when there is a board of directors, it would seem). And BATS is no different. So I now vow, that starting with the new semester, I, as CEO of BATS, will tighten up the performance-management contingencies on my own BATS performance.

Behavioral Systems Analysis
In much the same spirit as avoiding victim blaming, in BATS, we also try to avoid complaining about things not happening the way they should. If things aren’t right, implement a subsystem to make them right. This approach has resulted in many subsystems within BATS; here are a fews:

Subsystems

GRE Course
For example, one of our goals is to place as many students into grad school as possible. But a major obstacle is that too many good students, even those getting the top scores in our courses, were not getting into the grad school of their choice because of low GRE scores, in spite of well-earned, high GPAs. So, we implemented a voluntary GRE prep course, mainly a performance-management program that ensured our students would do the hard work of spending the necessary 100 to 150 hours studying standard GRE prep books that would raise their GRE score an average of 100 points (Groeger, In press; Miller, Goodyear-Orwat, & Malott, 1996; Vunovich, 1996).
Behavioral Academic Career Counseling (BACC)

Another problem with placing our BA graduates in grad school and jobs is that they had no idea of what opportunities were there and how achieve those opportunities. So I started giving a semesterly lecture on grad schools, jobs, and how to get there from here. At the end of each lecture I would invite the students to set up a BACC appointment with me for personalized counseling. Maybe one or two in 50 would. So I instructed the TAs to explicitly set up appointments BACC for me with the top students in their seminar sections. They came, they appreciated, and they got into grad school.

And, as is my tendency, I faded out of most of the BACC appointments. Now the TAs for the undergrad seminars do the appointments, and to encourage the students to encourage the students make appointments, we give them 10 optional activity points which they can substitute for participation in some required course activity. Now over half the students have BACC appointments and they evaluate the appointments as being quite valuable.

Yes, they should not need an individualized appointment, because not only do I give the lecture but all that material is in the last chapter of one of our texts, Elementary Principles of Behavior (Malott, Malott, and Trojan, 19??); but most of them do need the one-on-one counseling session. And they should need a special, individualized invitation for the appointment, but most of them do. And they should not need the optional-activity points for attending, but most of them do. We try not to get to hung up on should; instead, we try to do what it takes to achieve our goals, providing high-quality training in behavior analysis to as many students as possible and making sure they end up in positions where they can make good use of that training, whether it is in grad programs or behavior-analytic jobs. That exemplifies our goal-directed systems approach.

The Behavioral Research Supervisory System (BRSS)
The biggest problem grad students face is completing their theses and dissertations in a timely manner or completing them at all. By the time students finish up their PhD degree, they should have sufficient time-management skills that they would not need have trouble completing writing projects, not even projects as big as a dissertation; but most of them do not have those time-management skills. In fact, most of their faculty advisors do not have those time-management skills. In fact, I do not have those time-management skills. So, instead of blaming the victims, we implemented another subsystem, BRSS, to solve the problem using the previously described performance-management model of task accomplishment and the three-contingency model of performance-management (Garcia & Malott, 1988; Garcia, Malott, & Brethower, D; 1988). The research and writing of all BATS students are very effectively supported by BRSS; this includes undergraduate honors theses, MA projects, MA theses, and PhD dissertations. Almost all students produce quality products in a timely manner, a rarity without BRSS or the equivalent.

Welcome Wagon
Many subsystems, components, and procedures in BATS are a result of grad student initiatives. They identify a problem, propose a solution, and implement that solution, often though I think there is no problem or their solution will not work. But, if I think they will do no harm, I allow them to give it a try. And often I am wrong, there was a problem or an unrealized need and their solutions do work. Now many of the institutionalized subsystems, components, and procedures of BATS are a result of such student initiatives. This semester, Melinda Sota, a new MA student from St. Cloud University, identified a problem and proposed to develop a solution. The problem is the lost, freshman like feeling new grad students from other schools have when they enter BATS (and probably any other system in any other grad school). So Melinda is developing the BATS Welcome Wagon, a set of information that will be useful to students considering enter our program and useful to them once they are here. In addition, she will be setting up some sort of big-brother, big-sister program for the advanced grad students to adopt and mentor the incoming grad students.

Super-A
Many of our subsystems are designed not so much to fix problems but rather to enrich our current educational/training system. For example, I noticed that, in our principles course, a reasonable percentage of elite students were performing much better on their daily quizzes than was needed to achieve our 92% criterion for an A. I thought we should recognize this accomplishment and provide an incentive for anyone else wishing to stretch out a bit. So we set up the Super-A program. Students who meet the requirements for a Super-A can sign up next semester to receive one credit of A for their accomplishment. The requirements are that they earn an A in the regular course and earn 500 optional-activity points (about 50 hours worth of work) for other activities and accomplishments that semester, including 40 points for each cumulative quiz percentage point above 92%. About 25% of our students have joined the expanding Super-A elite.

A Letter to Mom and Dad
Another enrichment component is the letter to mom and dad. About the only thing Mom and Dad get from the university is bills and a certain amount of perplexity about just what it is their child is studying, especially if there child is studying behavior analysis. So, our TAs get the names and addresses of the top performing two or three students in each seminar and write letters of congratulations to those parents which both the TAs and I sign. Those top two or three students are excellent, serious, hard-working students who deserve that little bit of recognition and whatever praise will now bounce off Mom and Dad. But those super-students did not invent themselves; they are a result of excellent child rearing and nurturing by excellent, serious, hard-working parents; so those parents also deserve a little recognition and the opportunity for some bragging rights on their kids. In addition, Mom and Dad may now be a little more supportive of the super-kid who is thinking about going to grad school to study behavior analysis; maybe being a lawyer or a doctor is not the most important thing after all.

Behavioral-systems Approaches


Goal-Directed Systems Design
Goal-directed systems design is the specification of the goals a system is to accomplish and the design of that system so all its components lead to the accomplishment of those goals (Malott, & Garcia, 1987). Simple and obvious, but rare. Most systems have no clear goals and are not designed but, rather, evolve as a result of historical tradition and momentary expedience, certainly most systems in higher-education, from the level of the course, to the department, to the college.

But each component of each course and each supporting subsystem is designed to help us achieve our goal of saving the world with behavior analysis by training as many behavior analysts as well as we can and helping them find positions where they can save the world using their behavior-analysis and behavior-systems-analysis skills.

Continuous Quality Improvement
In implementing a goal-directed systems design, it is important to evaluate that system and all of its components, down to the most minute, to determine if all those subsystems and components are contributing to saving the world with behavior analysis. Often, we find that they are not functioning quite as planned. So after the negative evaluation, we must recycle, we must redesign the component or subsystem, re-implement, re-evaluate, etc. until it is working as it should or at least as well as we can get it to work (we are not able to solve all of our problems).

This continuous quality improvement is a fascinating process because it points to so many things that what should work but do not until they are revised. The evolution of our Behavioral Academic Career Counseling subsystem is an excellent example of this, as we discussed earlier.
The greatest amount of our continuous quality improvement efforts go into our instructional materials--the texts (Bosch, 2001; Malott, & Suarez-Trojan, in press; Suarez, 2001), homework, computer-based instruction, study objectives, and quizzes. If 20 or 30% of our students miss a quiz question, something is wrong. At times, it is even hard for us not to blame the victim, as reducing the error rate can be frustrating. But our performance-management contingencies are sufficiently effective that we can reasonably assume that our students are studying and doing the best they can; so the errors are not their fault; they are ours. Then we examine, with an eye toward revision, the quiz question itself, the relevant study objective, and the relevant homework and text material. We make the changes that seem appropriate. We implement the revised component the next semester. We revaluate. And we recycle until we have fixed the problem or run out of solutions.

Occasionally, we conclude that it is too much sugar for a cent. For example, we might have one high-error-rate quiz question based on one study objective and one paragraph of text. But after recycling several times, it looks like we would need to expand the paragraph to an entire chapter, because the concept is much more difficult than we had anticipated. In such cases we may conclude that the concept is not sufficiently crucial as to justify that much course time; so we either eliminate it or put it in an optional advanced enrichment section, at the end of the relevant chapter.

All professors revise their instructional material a little bit from semester to semester; but my observation is that most simply blame the students for poor quiz performance and then proceed to update their lectures, and they do that more on the basis of the latest developments in the subject matter of their field than in order to improve the effectiveness of their instruction. I have known very few professors who do the sort of molecular-continuous-quality improvement described here. Therefore, I think few professors realize how difficult are the concepts they are teaching and how difficult it is for to communicate to sincere, conscientious, hard-working students. Sometimes I find myself saying, how can those undergrads be so dumb as not to understand this concept; and my grad-student TAs say, we don’t understand it either. Oh, yes, don’t blame the victim.

One final comment about our continuous quality improvement: We evaluate almost every detail of a very large instructional system. That means that our evaluations and fixes are rough and ready, at best. They never achieve anything like the rigor appropriate for a research journal such as the Journal of Applied Behavior Analysis. Instead we note problems suggested by confusions in the seminars and homework, student questions and complaints, quiz performances, and student social-validity evaluations on anonymous questionnaires. Then we try to fix the problem based on our behavior-analytic-educated guesses. And we try again the next semester. In other words, this is real-world R & D (research and development), not scholarly research. But it is real-world research and development, not scholarly research that most behavior analysts do, even those with PhDs; therefore, working in BATS provides our students with very appropriate job skills (Malott, 1992 a, 1992 b).

Conclusions


What I advocate (and therefore, what I do; or is it the other way around?) is that faculty not be teachers of courses but, instead, be instructional systems designers and managers. Our job is not to present subject matter but to help students acquire complex, subtle professional repertoires and values. Such an approach makes little, if any use of lectures; instead it involves designing and implementing learning opportunities, in the form of readings, homework, seminars, quizzes, and practica.

Such an approach requires much more work than traditional teaching; it is very labor intensive. But the solution is to give a large number of grad students and even undergrad students an opportunity to help in the analysis, design, implementation, evaluation, and recycling of the training system. In so doing, the undergrad students in the courses get much better training, and the undergrad and grad students helping with the system get even better training.

And though the system is labor intensive, it can be more cost effective than traditional approaches to teaching. For example, BATS has very few paid assistants. Instead, most of the work is done by students who are receiving academic credit, for which they pay.

And though the system is labor intensive, in the need for graduate TAs to run the individual seminars, the system is not labor intensive in terms of faculty time. While the system takes much faculty time, doubling the number of students the system serves only slightly increases the amount of needed faculty time, not close to doubling the amount of needed time.

Finally, a training system such as BATS is a populist meritocracy. There is room for almost everyone to participate and contribute and benefit, and those students with exceptional skills have the opportunity to really stretch out, to really develop, hone, and demonstrate those skills. I am impressed with the number of undergrad and grad students who, when given a chance in the right context, turn out to be very proactive, responsible, dedicated, creative, personable behavior analysts/ behavior systems analysts/ managers/ trainers/ therapists/ people. I have great love and respect for our students.


References


Bosch, S. (2001) Making of a textbook on behavior analysis and autism : a behavior analytic approach. Doctoral dissertation, Western Michigan University.

Garcia, M. E., & Malott, R. W. (1988) Una solucion al fenomeno "todo menos tesis y disertacion." Revista Intercontinental de Psicologia y Educacion, 1, 205-216.

Garcia, M. E., Malott, R. W., & Brethower, D. (1988). A system of thesis and dissertation supervision: Helping graduate students succeed. Teaching of Psychology, 15, 186-191.

Groeger, C. (In Press). Preparation for the GRE. Journal of Behavioral Education.

Heward, W. L. (1994) Three "low-tech" strategies for increasing the frequency of active student responses during group instruction. In R. Gardner, D. M. Sainato, J. O. Cooper, T. E. Heron, W. L. Heward, J. Eshleman, T. A. Grossi (Eds.) Behavior Analysis in Education: Focus on Measurably Superior Instruction. (pp. 283-320) Pacific Grove: Brooks/Cole.

Jackson, M. & Malott, R. W. (1994). Helping high-risk black college students. In R. Gardner (ed.) Behavior analysis in education: Focus on measurably superior instruction. (pp. . 349-363) Pacific Grove, CA: Brooks/Cole.

Malott, R. W. (1984) In search of human perfectibility. In W.L. Heward, T.E. Heron, D.S. Hill, & J. Trap-porter (Eds.) Focus on Behavior Analysis in Education (pp. 218-245). Columbus: Charles E. Merrill.

Malott, R. W. (1989) The achievement of evasive goals: Control by rules describing contingencies that are not direct-acting. S. C. Hayes (Ed), Rule-governed behavior: Cognition, contingencies, and instructional control (pp. 269-322). New York: Plenum.

Malott, R. W. (1992 a). Should we train applied behavior analysts to be researchers? Journal of Applied Behavior Analysis. 25, 83-88.

Malott, R. W. (1992 b). Follow-up commentary on training behavior analysts. Journal of Applied Behavior Analysis. 25, 513-515.

Malott, R. W. (1993, October). The three-contingency model applied to performance management in higher education. Educational Technology, 33, 21-28.

Malott, R. W., (in press). What OBM needs is more Jewish mothers. Journal of Organizational Behavior Management.

Malott, R. W. & Garcia, M. E. (1987). A goal directed model approach for the design of human performance systems. Journal of Organizational Behavior Management, 9, 125-159.

Malott, R. W. & Garcia, M. E. (1991). The role of private events in rule-governed behavior. L. J. Hayes & P. Chase (Eds.), Dialogues on verbal behavior (pp. 237-254). Reno, NV: Context Press.

Malott, R. W., Malott, M. E., & Shimamune, S. (1993). Comments on rule-governed behavior. Journal of Organizational Behavior Management, 12, 91-101.

Malott, R. W., Shimamune, S., & Malott, M. E. (1993). Rule-governed behavior and organizational behavior management: An analysis of interventions. Journal of Organizational Behavior Management, 12, 103-116.

Malott, R. W., Vunovich, P. L. Boettcher, W. & Groeger, C. (1995). Saving the world by teaching behavior analysis: A behavioral-systems approach. The Behavior Analyst, 18, 341-356.

Malott, R. W. & Suarez-Trojan, E. W. (in press) Elementary principles of behavior (fifth edition). Upper Saddle River, NJ: Prentice Hall.

Miller, J. M., Goodyear-Orwat, A., & Malott, R. W. (1996). The effects of intensive, extensive, structured study on GRE scores. Journal of Behavioral Education 6(4): 369-379.

Suarez, E. T. (2001) A behavioral systems analysis of textbook quality improvement. Doctoral dissertation, Western Michigan University.

Vunovich, P. L (1996). Fluency training on quantitative skills tested by the graduate record examination. Doctoral dissertation, Western Michigan University.