What Good Supervision Looks Like: BCBA Supervision Requirements
Imagine learning to drive with an instructor who rode along for one out of every twenty trips. Those rides were always scheduled in advance, always on a Sunday morning, always on the same empty residential streets.
You'd pass. You'd have the hours. Your logbook would be spotless.
And the first time it rained on a Tuesday at five o'clock, you'd be alone in a car with a skill nobody had ever actually watched you use.
That's an uncomfortably fair description of how a lot of behavior-analytic supervision works. Not because supervisors are lazy. Most of the ones we meet are drowning. It's that the field has quietly agreed to measure supervision the way you'd measure attendance. Hours logged, contacts made, boxes checked. All of it verifiable. None of it evidence that anyone can actually do anything.
Think of the requirements as a floor. Floors are useful. They're also the thing you stand on, not the thing you aim for.
Everyone Knows the Numbers. Almost Nobody Reads Past Them.
Let's start with what the rules actually say, because a surprising number of agencies are running on hallway rumor.
For RBTs, the RBT Handbook requires ongoing supervision for a minimum of 5% of the hours spent delivering behavior-analytic services in each calendar month. That supervision has to include at least two face-to-face, real-time contacts, and the supervisor has to observe the RBT working with a client in at least one of them. One of the two monthly contacts can be a small group, as long as the group is genuinely interactive and holds no more than ten RBTs. Observation can happen over a web camera. Passive video monitoring, with no real-time interaction or feedback, doesn't count as supervision at all.
The documentation requirements are heavier than most people expect. Dates and times services were provided, supervision dates and duration, format, observation dates, supervisor names, and the supervisor's relationship to the client, all retained for seven years. The Board isn't subtle about the stakes, either. An RBT found substantially noncompliant with supervision requirements is subject to immediate termination of certification or of recertification eligibility.
Who can supervise whom is just as specific. RBTs are supervised by BCBAs or BCaBAs. BCaBAs are supervised by BCBAs. Fieldwork trainees are supervised by BCBAs only. And before any of it starts, the supervisor has to complete the 8-hour supervision training built on the Board's Supervisor Training Curriculum Outline 2.0.
That outline is worth reading sometime. Its objectives include describing ineffective supervision strategies, determining your own capacity before you accept a supervisee, and applying behavioral skills training with real rehearsal and feedback. The Board isn't describing a meeting. It's describing a behavior-change program where the learner happens to be a clinician.
What Changes in January 2027
Anyone supervising fieldwork should have the 2027 BCBA requirements on their radar now, because they reshape the structure of supervision more than the volume of it.
Starting January 1, 2027, candidates complete either 2,000 hours of Supervised Fieldwork with 5% supervision, or 1,500 hours of Concentrated Supervised Fieldwork with 7.5% supervision. Observation of the trainee with a client is spelled out directly: 60 cumulative minutes per supervisory period for standard fieldwork, 90 for concentrated. At least half of all supervised hours have to be individual rather than group. At least 60% of total fieldwork hours have to be spent in unrestricted activities. A supervisory period runs a minimum of 20 hours and a maximum of 160.
Notice what the Board chose to get specific about. Not just how much supervision happens, but how much of it involves watching someone work, and how much of it happens one-on-one instead of in a group of eight. Those are quality parameters written as compliance requirements.
One myth worth killing while we're on the topic. The 2026 RBT updates did not change supervision requirements. They changed the training curriculum, the competency assessment, and recertification cycles. Supervision is still 5%, still two real-time contacts, still one observation with a client. Anyone telling you otherwise skimmed.
Compliance Measures Attendance, Not Ability
Now the uncomfortable part.
You can meet every requirement above, perfectly and verifiably, and still hand the field a newly certified analyst who can't run a competent preference assessment, write a defensible treatment plan, or tell a technician that something needs to change without either apologizing for it or steamrolling them.
Five percent isn't a quality standard. It's an attendance policy. It tells you a supervisor was present for one hour out of twenty. It tells you nothing about what happened in that hour, whether anybody practiced anything, whether any performance got measured, or whether the supervisee walked out able to do something they couldn't do before.
The Board actually says all of this, in the place people are least likely to look. Ethics Code standard 4.06 requires supervision to be evidence based, focused on positive reinforcement, and individualized. Standard Standard 4.08 requires ongoing, documented, evidence-based data collection and performance monitoring of supervisees, with timely praise and feedback, plus a written improvement plan when performance problems come up. Standard 4.10 requires supervisors to continually evaluate their own supervisory practices, using feedback from others and from client and supervisee outcomes, and to document those self-evaluations.
We measure client progress obsessively. We measure supervisee performance almost never.
Read those three together and the picture that emerges looks nothing like a monthly meeting. It looks like a behavior plan, with a pinpointed target, a measurement system, a reinforcement schedule, and a data-based revision process. For the supervisee.
We measure client progress obsessively. We measure supervisee performance almost never.
An Experienced Supervisor's Number Is Four Times the Minimum Requirements
We asked a consultant on our team, someone who supervises across multiple agencies and has watched this go right and wrong for years, what number she actually uses.
She refused to give one at first, which is the correct instinct. Then she said something more useful. With a large technician team on a complex case, she'd need to be supervising closer to 20% of their hours, not 5%, and she'd need to concentrate her time there.
That's an experienced practitioner's working estimate landing at four times the regulatory minimum.
She isn't arguing the Board got it wrong. The minimum exists to be a minimum. It's one number that has to cover a seasoned technician on a stable case and a six-week hire on a complicated one, which is exactly why it can't double as a target. A floor built to work for everyone is, by design, not enough for the hard cases.
Her factors were specific. Caseload complexity came first. Then whether a technician works across more than one client, whether there are senior technicians she knows and trusts, whether team meetings count as paid administrative time, and whether the supervisor is also carrying a billable clinical load. Her rough tells for a caseload getting away from someone: cases running five or more technicians, or three or more clients with high-intensity, complex needs.
On fieldwork trainees she was blunter, and she did give a number. No more than two at a time, knowing how much time it takes on her end to do it properly.
The Ethics Code Won't Give You a Number, and That Isn't a Loophole
Standard 4.03, Supervisory Volume, says behavior analysts take on only the number of supervisees or trainees that allows them to provide effective supervision and training. No ratio, no formula, no cap.
At eleven o'clock at night, while in the midst of creating a staffing model, it might seem like avoiding the question. However, that's not the case. Publishing a specific ratio would quickly be manipulated, as any established number turns into a target to meet rather than a decision to evaluate. Agencies nationwide would adhere strictly to that number, regardless of whether the circumstances justified it.
What most people miss is that 4.03 doesn't stop there. It names factors you're expected to weigh on an ongoing basis, including your current client demands, your existing supervisee or trainee caseload, and your time and logistical resources. Then it goes further. When you determine you've reached your threshold volume for providing effective supervision, the standard requires you to document that self-assessment and communicate the result to your employer or other relevant parties.
Read that twice if you supervise inside an organization. Deciding you're at capacity isn't a private conclusion you reach and quietly resent. The Code expects you to write it down and hand it to somebody. Most supervisors we meet have never done that, which is usually the real story behind an unmanageable caseload. Nobody decided it. It accumulated.
The Warning Signs Are Behavioral, and They Show Up Early
Over-capacity announces itself well before it reaches client outcomes, assuming somebody's watching for it.
There are two indicators our consultant mentioned that are worth noting. The first is formal reviews beginning to miss their deadlines. The second is more acute: indirect complaints from technicians you have a good relationship with, regarding someone else not receiving what they need. People seldom describe their own supervision as lacking; instead, they comment on a colleague's situation.
Add the ones we observe across client organizations. Supervision sessions are rescheduled more frequently than they occur. Feedback becomes increasingly vague, transitioning from "your prompt delay was inconsistent on the second trial block" to "things looked good." Notes are compiled in batches on the last day of the month. And the most subtle sign, a supervisee who no longer asks questions. This is rarely a sign of mastery. More often, it indicates someone who has realized that asking questions isn't worth the effort.
None of this requires a new data system. It requires someone whose job it is to look.
Almost Nobody Teaches a Supervisee How to Give Feedback
This is the failure we'd put at the top of the list, and it's the one we hear discussed least.
A trainee spends two years learning to conduct assessments, design interventions, graph data, and defend treatment decisions. Personnel Supervision and Management is a real domain on the BCBA exam, roughly 11% of it. So trainees learn about supervision. They can define behavioral skills training. They can list the components of effective feedback.
After certification, they are assigned a team of technicians and are required to provide immediate corrective feedback to an adult during a session, in front of a client, without harming the relationship or the session.
Many of them have never performed this task under observation.
As our consultant put it, people are often not good judges of how they come across when they communicate, and it becomes a real problem when they get thrown into giving feedback to technicians. That matches what we see. One new BCBA softens every correction into a suggestion so gentle the technician doesn't register that a change was requested. Another is so anxious about being firm that she lands as cold without meaning to. Neither of them is a bad clinician. Both were assessed on knowing about feedback and never assessed on giving it.
The solution is straightforward and aligns perfectly with what we already know. Behavioral skills training is as effective for supervisees just as it is for technicians. It involves instruction, modeling, rehearsal, and feedback. This means someone must observe a trainee providing feedback and then offer feedback on their performance. It may feel awkward initially, but it becomes a skill like any other with practice.
Improvisation Has Predictable Gaps
Ask a supervising BCBA what their fieldwork curriculum is and you'll usually get a slightly defensive answer, because for most of them the honest one is "whatever came up this week."
Our consultant noted this straightforwardly, and we suspect it's a common occurrence. None of the companies she has worked with have invested in a supervision curriculum. Each BCBA follows their own approach.
Improvisation isn't automatically bad, but it can lead to predictable gaps, turning this into a systemic issue rather than an individual one. Behavior Assessment is never overlooked because it frequently arises and is engaging to teach. However, Experimental Design is often skipped since it constitutes a smaller portion of the exam, it rarely surfaces on its own in a Thursday afternoon meeting, and it's genuinely harder to present in an engaging way. The trainee finds that hole two years later, at a testing center, when it's expensive.
A structured curriculum doesn't necessarily enhance a supervisor's skills, but it ensures comprehensive coverage as part of the system, rather than relying on the supervisor's memory during a difficult week. This distinguishes a program from a habit.
It's also why we built ours: a sequence covering measurement systems, experimental design, assessment, verbal behavior, ethics, and supervision itself. It's available as the Supervision Mastery Vault (18 lessons), Supervision Essentials (12 lessons), or Fieldwork Foundations (6 lessons) for supervisors who'd rather start smaller. It's content you deliver to trainees, not training for supervisors. It doesn't satisfy the 8-hour requirement, and we'd rather say that plainly than let anyone assume otherwise.
Medical Necessity Is a Clinical Skill, Not a Billing Detail
One more gap, because this one costs real money and is often overlooked intentionally.
New analysts routinely request service hours based on when the client is available rather than what's clinically indicated. It feels accommodating. It's scheduling dressed up as clinical judgment. And it works fine right until circumstances change, the client genuinely needs more intensive services, and the funder looks at a sudden request for increased hours and asks the obvious question. Why now?
"The family has more availability" doesn't hold up in that discussion. By then the documentation trail shows the original recommendation was built around a calendar, and the request for more hours appears driven by desire rather than need.
Supervisees need to learn medical necessity as a clinical construct, tied to assessment findings and defensible in writing, well before they're the one signing the request.
What Good BCBA Supervision Is Not
It isn't niceness. A supervisor who's beloved and never corrects anything is producing clinicians who will get corrected later by a funder, a family, or a licensing board, none of whom will be as kind about it.
It isn't harshness either. Rigor and severity get confused constantly. Standard 4.06 specifically calls for a focus on positive reinforcement, which isn't a suggestion to be pleasant. It's a statement about what actually changes behavior.
It isn't paperwork. Documentation is a record of supervision. It isn't supervision, any more than a receipt is a meal.
And it isn't hours. Two thousand hours of unobserved practice produces two thousand hours of whatever habits formed in the first month.
An Important Caution
None of this is a Prisma Dimensions Group (PDG) standard, and the 20% figure above is one experienced practitioner's judgment about complex cases. It isn't a benchmark, it isn't a requirement, and it shouldn't go into a policy manual without hard thinking about your own context.
We'd also push back on reading this as a piece about bad supervisors. Most supervisors we meet are carrying a billable caseload, a supervision load, and an administrative load that were each designed as if the other two didn't exist. When supervision degrades, it's usually the third priority losing out to two things with clearer deadlines. That's an organizational design failure, not a character failure, and it doesn't get fixed by asking individuals to try harder.
If your agency wants better supervision, the question isn't whether your BCBAs care. It's whether anyone has protected the time, funded it, and made it something a person is actually evaluated on.
The Cheapest Fix Costs Nothing but Attention
If you do one thing after reading this, do this one. Write down what a supervisee has to be able to demonstrate, not what you intend to cover.
Our consultant described walking into organizations where a newly certified analyst plainly didn't know what they were doing, then finding, when she asked, that the supervisor had gone over the skills. There was just no formal expectation of what had to be shown. Covered and demonstrated are different verbs, and only one of them is measurable.
A one-page list of demonstrable competencies costs nothing and takes an afternoon. It turns supervision from a topic list into a criterion, and once it exists, it's obvious who's met it and who hasn't.
Add one unannounced observation a month and you have most of the benefit of a formal quality system for the price of paying attention.
The Bottom Line
The requirements exist to catch the worst outcomes, not to produce the best ones. Five percent, two contacts, one observation. That's what the Board can define and enforce across thousands of very different organizations, and it's a real achievement. It's also nowhere near sufficient.
Good supervision looks like a behavior plan aimed at a clinician. Pinpointed skills, real measurement, deliberate practice with feedback, and a supervisor honest enough to check whether any of it is working. Sometimes that costs 5% of the hours. Sometimes, on a hard case with a new team, it costs four times that.
The floor is where you start. It was never meant to be where you build.
A practical next step Pick one supervision group, one month, and one competency. Define what competent performance looks like, observe for it directly, and review whether the supervisee improved. If that small test exposes gaps, it has already done its job.
Want the Supervision Audit Checklist?
We put together a one-page checklist for auditing your own supervision system: what to look for, what to measure, and what to write down.
That link opens an email with the subject line already filled in. Send it and I'll reply within one business day with the checklist attached. No form, no signup, no newsletter, and nobody here will follow up trying to sell you anything.
If you'd rather not email a stranger, BACB's Fieldwork Checklist for BCBA and BCaBA Supervisors is free and thorough, and it'll tell you in about ten minutes whether your current system has holes in it. The link is in the sources below.
And if the curriculum gap in this article sounded familiar, our supervision lessons live in the curriculum shop.
Sources
Behavior Analyst Certification Board. Registered Behavior Technician Handbook. https://www.bacb.com/rbt-handbook/
Behavior Analyst Certification Board. RBT 2026 Requirements. https://www.bacb.com/wp-content/uploads/2025/07/RBT-2026-Requirements_250723-a.pdf
Behavior Analyst Certification Board. 2027 BCBA Requirements. https://www.bacb.com/bcba-2027-requirements
Behavior Analyst Certification Board. Supervision, Assessment, Training, and Oversight. https://www.bacb.com/supervision-and-training/
Behavior Analyst Certification Board. Supervisor Training Curriculum Outline (2.0). https://www.bacb.com/wp-content/Supervision-Training-Curriculum/
Behavior Analyst Certification Board. Ethics Code for Behavior Analysts. https://www.bacb.com/wp-content/uploads/2022/01/Ethics-Code-for-Behavior-Analysts-240830-a.pdf
Behavior Analyst Certification Board. BCBA Test Content Outline (6th ed.). https://www.bacb.com/wp-content/uploads/2022/01/BCBA-6th-Edition-Test-Content-Outline-240903-a.pdf
Behavior Analyst Certification Board. Fieldwork Checklist for BCBA and BCaBA Supervisors. https://www.bacb.com/wp-content/fieldwork-checklist-supervisors
Behavior Analyst Certification Board. 7 FAQs for RBTs Pursuing BCBA or BCaBA Certification. https://www.bacb.com/7-faqs-for-rbts-pursuing-bcba-or-bcaba-certification/
Comments