DRAGENwiki · Article
Seven dimensions of ABA
Customize your experience
Sources checked. Clinical review pending.
In 1968, three researchers named seven qualities that good applied behavior analysis should have: applied, behavioral, analytic, technological, conceptually systematic, effective, and showing generality. Each one is a question that a family, a teacher or a BCBA can ask about any plan. Ten years later, one of the three argued that the people served should judge whether the work matters to them. Twenty years later, all three checked the list again.
The seven dimensions in 60 seconds
- Seven words, one test. Baer, Wolf and Risley (1968) said an applied study must be applied, behavioral and analytic. It should also be technological, conceptually systematic and effective, and it should show some generality. [1]
- Each word is a question. Does this matter to the person? Is it something they do that people can count? Can we show the plan caused the change? Could someone else run it? Does it fit known principles? Is the change big enough? Does it last and spread? [1]
- The people who live with the behavior help judge "enough." How much change is needed, the 1968 paper said, is "not a scientific question, but a practical one." [1]
- Social validity asks whether people accept the goals, the methods and the results. Montrose Wolf made the case for it in 1978, in a paper whose title ends "how applied behavior analysis is finding its heart." [3]
- The list was checked again. In 1987 the same authors called it "still prescriptive, and to an increasing extent, descriptive." [2]
- The words guide a plan. They do not replace the person. What the person wants comes first, and a BCBA reviews every clinical decision.
Jump to: Where the seven came from · The seven, one at a time · Social validity · Twenty years later · At home and school · A made-up example · Myths · Words to know · Sources
Where the seven came from
In 1968, the Journal of Applied Behavior Analysis (JABA) published its first issue. Montrose Wolf later recalled asking his colleague Don Baer what the new journal was for. Baer's answer, which Wolf put in the journal, was that it was for applications of the analysis of behavior "to problems of social importance." [3]
That first issue carried a short paper by Donald Baer, Montrose Wolf and Todd Risley. It explained how to judge a study that calls itself applied behavior analysis. Such a study, they wrote, must be applied, behavioral and analytic. It should also be technological, conceptually systematic and effective, and it should show some generality. [1] They called these qualities dimensions, and their 1987 follow-up kept the word in its title. [1][2]
The paper set applied work beside basic laboratory research, the kind described on the experimental analysis of behavior page. A basic study may choose a behavior because it helps answer a question about theory. An applied study chooses a behavior because it matters to the person and to society. [1] The methods can be the same. What makes a study applied, the authors wrote, is society's interest in the problem, not the research procedures. [1]
Ten years later, in 1978, Wolf argued that the field should also measure what the people it serves think of its goals, its methods and its results. [3] In 1987, the three authors wrote "Some still-current dimensions of applied behavior analysis." They went back over each word, said what had changed and what had not, and named the problems still open. [2] Those three papers are the backbone of this page.
The seven, one at a time
Each dimension below has three parts: what it meant in 1968, what the authors added in 1987, and a question you can ask today.
Applied
1968. A study is applied when the behavior is chosen because it matters to the person and to society, not because it matters to a theory. [1] The authors' test was a plain question: "how immediately important is this behavior or these stimuli to this subject?" [1]
1987. The authors looked again at what "applied" means. In their view, a problem becomes an applied problem when someone displays it and explains it, and other people respond. To understand that fully, they wrote, the field would need analyses of how problems are presented, how decision-makers are influenced, how support is recruited and how a crisis can set the stage for change. [2]
Ask today: Why does this goal matter to this person's life right now, and who said so?
Behavioral
1968. Applied behavior analysis studies what people do. The authors put it simply: the question is "not what he can say, but what he can do." [1] Behavior has to be measured, and in applied settings people often do the measuring. [1] So the authors asked readers to check "whose behavior?" really changed. If people are recording, a change on the graph could be a change only in how they observe and record, not in the learner. That is why they called measuring observer reliability "a prime criterion" of whether a study was truly behavioral. [1]
1987. The authors wrote that valid measurement means measuring the behavior that led someone to bring the problem forward in the first place. [2] They also noted that everyday labels such as attention, readiness or anxiety often point to "some behavioral reality not yet analyzed as such." [2] The job is to find the behavior behind the label. DRAGEN adds one check: the person whose behavior it is should also have a say in what counts as a problem.
Ask today: What exactly will people see and count, and would two observers count it the same way? See measuring behavior and operational definition.
Analytic
1968. A study is analytic when it gives "a believable demonstration of the events that can be responsible" for the change. [1] The paper described two common ways to do that:
- Reversal. Measure the behavior until it is steady. Apply the change. Then stop or alter it, and see whether the behavior depends on it. The authors noted that in real settings, the number of reversals may be limited. [1]
- Multiple baseline. Measure several behaviors over time, then apply the change to one behavior at a time. The authors said this is especially useful when a behavior seems irreversible or reversing it would be undesirable. [1]
Both designs show the same effect more than once. In the authors' words, "replication is the essence of believability." [1]
1987. Twenty years later, the authors said "analytic" and "conceptual" had grown together. The field now counted as analytic when it showed convincingly how to change a behavior, and when its methods made "systematic, conceptual sense." [2] They also warned against rigid rules. They told of a study whose design clearly showed that a training program was very effective. Many people in a convention audience argued instead that it was not a "proper" multiple baseline design. The authors suggested that "convincing designs should be more important than 'proper' designs." [2]
Ask today: What would show that the plan made the difference, and not time, chance or something else in the person's life? See reading a behavior graph.
Technological
1968. A procedure is technological when every part of it is "completely identified and described." [1] A label such as "play therapy" or "social reinforcement" is not enough. For social reinforcement, a write-up has to say what was given, under what contingency and on what schedule. [1] The rule of thumb is "whether a typically trained reader could replicate that procedure well enough to produce the same results." [1] A good description also says what to do when the person responds in some other way. [1]
1987. For the most part, the authors judged, published procedures were now described well enough to give a reader a fair chance of replicating them with the same results. [2] They saw a gap, though. Data on whether procedures were carried out as written were rarely shared outside the team that ran them. [2] They also named a tension. Keeping to the studied procedure is safer, because that version is known to work. Letting people adapt a program matters too, because people who cannot fit a program to their situation may come to dislike it, and "if a program is not used, it cannot be effective." [2] Each view, they wrote, is likely right in some contexts and wrong in others. Their open question: "What is the range of variation of a program's procedures that still allows sufficient effectiveness?" If that range is large, they wrote, flexible use can be encouraged. If it is narrow, fidelity will be required. [2]
Ask today: Could a new aide or a grandparent run this plan from the written page? Which parts may be adjusted, and which must stay the same?
Conceptually systematic
1968. Procedures should be tied to the basic principles they come from. The authors urged published work to "strive for relevance to principle." [1] Their example: describing exactly how colors are changed, step by step, to move a child from telling colors apart to telling shapes apart is good. Also naming the principles involved, such as fading and errorless discrimination, is better, because it shows the reader how similar procedures can be built from basic principles. [1] That is what makes the field "a discipline rather than a collection of tricks." [1] A pile of tricks, they said, is hard to expand, hard to learn and hard to teach. [1]
1987. As noted under Analytic, the authors said the line between analytic and conceptual work had become hard to keep. A good method should make conceptual sense as well as produce a change. [2]
Ask today: Which principle explains why this plan should work? For example, does it rest on reinforcement, and does it fit the four-term contingency?
Effective
1968. Applied work has to make a difference big enough to matter. If an application "does not produce large enough effects for practical value, then application has failed." [1] Basic research can value a small but reliable effect when it shows that a variable matters for theory. In applied work, what counts is practical importance. [1] How much change is enough? The authors called that "not a scientific question, but a practical one." Its answer is likely to come from the people who must deal with the behavior every day. [1]
1987. "The hallmark of any applied discipline ought to be effectiveness," the authors wrote. [2] They also admitted a possible blind spot. The field may have taught social skills without checking whether the person's social life improved, and safety skills without checking whether the person was actually safer afterward. [2]
Ask today: Is the change big enough to make daily life better for the person, and who decided how much is enough?
Generality
1968. A change has generality if it lasts over time, shows up in many settings, or spreads to related behaviors. [1] The authors were clear that this does not happen by itself: "generalization should be programmed, rather than expected or lamented." [1]
1987. Generality was still crucial, they wrote, though now for making the field as effective as possible rather than for its survival. [2] It should come "by suitable programming, rather than by good luck." [2] They were also honest about the state of things: "The problem is far from solved." [2]
Ask today: What is the plan for this skill to last, and to work at home, at school and with new people?
Social validity: finding ABA's heart
The 1968 paper already said that the people who deal with a behavior help decide how much change is enough. [1] Ten years later, Wolf took the idea further. Critics had asked whether the field's changes were really important. Social importance, Wolf argued, is a value judgment, and only society can make it. [3] So he proposed asking consumers, the people who take part in, live with or pay for a program, about three things: [3]
- The goals. "Are the specific behavioral goals really what society wants?" [3]
- The procedures. "Do the ends justify the means?" That is, do participants, caregivers and other consumers find the methods acceptable? [3]
- The effects. "Are consumers satisfied with the results?" Wolf added: all the results, including any no one predicted. [3]
He called these judgments of social validity. [3]
Wolf gave an example from the research team's own work. In the Achievement Place research project, many communities said they wanted teaching-parents with "warmth," people who knew how to relate to young people. At first, the team disagreed and argued that skillful use of the point system mattered more. Their later research tended to support the communities' commonsense view. [3]
Wolf wanted systems that let consumers tell researchers how a program relates to their values and to what they find reinforcing. [3] He did not see this as leaving science behind: "This is not a rejection of our heritage." [3] Asking people for their views, he explained, is not a return to explaining behavior by inner causes. It is a practical way to assess complex reinforcers. [3]
In 1987, the three authors described social validity as how much everyone served by a program likes its goals, targets, effects, procedures and staff. Its purpose is to predict and avoid rejection. [2] They wrote that "social validity is not sufficient for effectiveness but is necessary to effectiveness." [2] Being liked does not make a program work, but they saw it as a requirement for working. They also warned that the measures of the day were very basic and could find social validity where it did not really exist. They warned that hoping to find social validity is far riskier than "searching for any sources of social invalidity." [2]
At DRAGEN, the person receiving support is a consumer too. Wolf's own list names participants. [3] A learner's "no," a complaint or a preference is social validity information. See the right to say no.
Twenty years later
The 1987 paper looked back on two decades. The authors judged that their list was "still prescriptive, and to an increasing extent, descriptive." [2] It still told the field what to aim for, and more of the field now did it.
They also described what was still hard: making sense of who defines a problem, sharing data on whether plans were followed, balancing fidelity with flexibility, programming generality, and measuring social validity well. [2] They pointed to the Teaching-Family model, as old as the journal itself, as an apparently durable way to spread a program. It created 12 regional training centers to support 215 replications of the original Achievement Place program, and it wrote plain-English training manuals rather than journal articles. [2]
They added that the field should learn to tell apart problems that one short intervention can fix from problems that need support over a lifetime, much as medicine treats appendicitis differently from diabetes. [2] And they ended with confidence: "It seems clear that we can do what remains to be done." [2]
What it means at home and school
You do not need to be a researcher to use the seven. Each one becomes a question you can ask at a team meeting:
- Applied: Why does this goal matter for my child's life now? Who picked it?
- Behavioral: What exactly will people see and count? Do two people count it the same way?
- Analytic: How will we know the plan made the change, and not something else?
- Technological: Is the plan written so that a new staff member, a teacher or I could follow it?
- Conceptually systematic: Which principle explains why it should work?
- Effective: Is the change big enough to make daily life better? Who decides how much is enough?
- Generality: What is the plan for the skill to last, and to work at home, at school and with new people?
And one more, for social validity: do we, and does my child, accept the goals, the methods and the results?
In DRAGEN's practice, this looks like:
- Choose goals with the person and the family, and write down why each goal matters to them now.
- Describe behavior in words that two people could check, and count it the same way each time.
- Read a learner's graph against their own earlier data, not against other children.
- Keep the plan on one page that a new aide or a grandparent could follow, and say which parts may be adjusted.
- Plan for generalization from the first day: other places, other people, later dates.
- Ask the person and the family, more than once, whether they accept the goals, the methods and the results.
- Treat "no," a complaint, a question or a request to stop as information about what the person needs. It is not "noncompliance," and it is never a target to reduce. See the right to say no.
- Keep immediate safety separate. If someone could get hurt, deal with safety first, then look at the pattern.
- A BCBA reviews every clinical decision. This page is general information, not medical or clinical advice.
A made-up example
This example is invented. It does not describe a real person.
Learner E is 8. At lunch in a loud cafeteria, she often covers her ears and leaves the table. Her team uses the seven dimensions as a checklist.
- Applied. E's family and teacher say lunch is the hardest part of her day. E says it is "too loud." Together with E, they agree on a goal: she can ask for a break or for headphones, and lunch becomes a time she can manage.
- Behavioral. The team writes down what asking for a break looks like: saying "break," handing over a break card, or signing it. Each lunch, an adult counts how often E asks and how often she leaves without asking. On some days, two adults count at the same time to check that they agree.
- Analytic. The team starts teaching in the cafeteria first, while still counting in music class, which is also loud. A few weeks later, they start in music class. If asking goes up in each place only after teaching starts there, that is good evidence that the teaching helped. They never stop honoring her requests just to test the plan.
- Technological. The plan fits on one page. When E asks, the adult says "Sure," and she gets two minutes in the hallway with an adult, or her headphones. If she asks again, the answer is still yes. If she leaves without asking, the adult goes with her calmly, keeps her safe and offers the card later, with no penalty. Any aide can follow it.
- Conceptually systematic. Asking works because it gets her out of the noise. That is negative reinforcement, and the plan is a form of functional communication training.
- Effective. After a month, E asks on most days and leaves without asking less often. E says lunch is "okay now." Her teacher says she stays with her friends for more of the period. The family and E agree the change is big enough to matter.
- Generality. The team checks again at a school assembly, with a new aide, and a month later.
When E says no to wearing headphones, the team writes that down as information and offers a quieter table instead. Her "no" is not a behavior to reduce. A BCBA reviews each decision.
Common myths
- Myth: The seven dimensions are a kind of therapy. Fact: They are standards for judging applied work. The 1968 paper offered them for judging studies, and DRAGEN uses them to judge plans too. They describe what good work looks like, not one method. [1]
- Myth: If behavior changed, the plan worked. Fact: A change alone does not show its cause. The analytic dimension asks for a believable demonstration that the plan was responsible. [1]
- Myth: Skills spread to new places on their own. Fact: Generalization has to be planned. The 1968 authors said so directly, and twenty years later they still called the problem unsolved. [1][2]
- Myth: "Effective" means statistically significant. Fact: In the 1968 paper, effective means a change big enough for practical value, judged with the people who deal with the behavior. [1]
- Myth: If the family likes the program, it must be working. Fact: Being liked is necessary but not enough. A program also has to produce real change. [2]
- Myth: The list was fixed in 1968. Fact: The same authors revisited it in 1987, kept it, and added questions about who defines a problem, fidelity, flexibility and social validity. [2]
Words to know
- Dimension: one of the seven qualities used to judge applied behavior analysis. [1]
- Applied: chosen because it matters to the person and to society. [1]
- Behavioral: about what a person does, measured directly and reliably. [1]
- Analytic: with a believable demonstration that the procedure caused the change. [1]
- Technological: described completely enough for someone else to repeat it. [1]
- Conceptually systematic: tied to the basic principles of behavior. [1]
- Effective: big enough to matter in daily life. [1]
- Generality: lasting over time, showing up in other settings, or spreading to related behaviors. [1]
- Reversal: applying a procedure, then stopping or changing it, to see whether behavior depends on it. [1]
- Multiple baseline: measuring several behaviors and starting a procedure on one at a time. [1]
- Fidelity: carrying out a plan the way it was written. [2]
- Social validity: whether the people served accept a program's goals, methods and results. [3]
Need a PDF?
Create it here, for free: the DRAGEN app. Use it at your next session.
Or print one we already made, also free:
- Goal-writing template (PDF)
- Generalization plan (PDF)
- BST fidelity checklist (PDF)
- Behavior plan at a glance (PDF)
- Every free PDF
More from DRAGEN
- Lesson: How to write a goal a stranger could measure
- Lesson: How to make a skill work everywhere: generalization
- Lesson: How to read a behavior graph in sixty seconds
- Lesson: How to check your data with interobserver agreement (IOA)
- Lesson: How to teach anyone a skill in four steps
- Lesson: How to teach a child to ask instead: functional communication training (FCT)
- Free app for parents: DRAGENparent
- BCBA care: DRAGENtherapy
- Books: DRAGENbooks
- Talk with Chase: Contact
- Home: dragen.cloud
Related DRAGENwiki pages: applied behavior analysis · experimental analysis of behavior · behaviorism · radical behaviorism · reading a behavior graph · measuring behavior · operational definition · reinforcement · four-term contingency · the right to say no · writing behavior goals · how progress is measured · how to choose an ABA provider · history of ABA
What remains unknown
- How far a program's procedures can vary and still work well enough was an open question in 1987. The answer decides whether flexible use can be encouraged or fidelity is required. [2]
- Making generalization happen by design was still unsolved in 1987. [2]
- Ways of measuring social validity were basic in 1987, and Wolf expected later studies to show the shortcomings of some measures. [2][3]
- When the people served disagree, for example a learner and the adults around them, these papers do not settle whose view should win. DRAGEN puts the person's own voice first and handles immediate safety separately.
- This page summarizes the main lines. It does not settle these debates.
- Last checked on November 12, 2026.
Sources
- Baer, D. M., Wolf, M. M., & Risley, T. R. (1968). Some current dimensions of applied behavior analysis. Journal of Applied Behavior Analysis, 1(1), 91–97. https://pmc.ncbi.nlm.nih.gov/articles/PMC1310980/
- Baer, D. M., Wolf, M. M., & Risley, T. R. (1987). Some still-current dimensions of applied behavior analysis. Journal of Applied Behavior Analysis, 20(4), 313–327. https://doi.org/10.1901/jaba.1987.20-313 (full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC1286071/)
- Wolf, M. M. (1978). Social validity: The case for subjective measurement or how applied behavior analysis is finding its heart. Journal of Applied Behavior Analysis, 11(2), 203–214. https://doi.org/10.1901/jaba.1978.11-203 (full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC1311293/)