Evidence-Based Supervision Practices Most BCBAs Skip

Sakai (2024) surveyed 317 BCBA (Board Certified Behavior Analyst) supervisors, scored them on 46 recommended supervision practices, and tied those practices to fieldwork pass rates, and only 7 of the 46 showed a weak but real link to whether trainees passed the exam. Most of the high-frequency practices were the easy ones, the things you can say out loud in the moment. The ones that got skipped were the ones that take 20 minutes at your desk on a Sunday. This page walks through what the survey found, why it lines up with how a lot of supervision actually runs, and a 30-minute weekly template that hits the seven practices without adding hours to your week.

The 46-practice survey nobody talks about (Sakai 2024)#

The survey took the BACB (Behavior Analyst Certification Board) ethics code section 5, the supervision section, and pulled it apart into 46 recommended supervisor behaviors. Things like communicating supervision conditions up front, using behavioral skills training (BST, the model-rehearse-feedback teaching loop) to teach skills, giving immediate praise, doing written evaluations, checking credentialing requirements, reviewing literature with the trainee, and so on. Then they asked 317 BCBA supervisors which of those 46 they actually do, and how often. Then they cross-checked those answers against whether the supervisor's trainees passed the certification exam.

The most frequent were communicating supervision conditions, so clear contracts, expectations, termination clauses, designing effective training, use of behavioral skills training to actually teach things, and then providing feedback. Immediate praise, correction was high. Written evaluation still occurred, but was relatively speaking less common. In that mid-range, delegation didn't come up too much. From the talk — Matt Harrington

That gives you the shape of the field. Contracts at the top. BST in the middle of the day. Praise and correction constantly. Written evaluations less often. Delegation almost never. None of that is a moral failure on supervisors. It is the natural pattern of what gets done when you are busy and what gets pushed to "next week."

What 317 BCBAs actually do in supervision (ranked)#

Here is the rough ranking the survey describes, written as a working list a supervisor can sanity-check themselves against.

1. Communicating supervision conditions. Most supervisors do this at the start. Clear contracts, hour expectations, what counts as restricted versus unrestricted activity, and a written termination clause.
2. Designing effective training. Most use some version of BST, even if they do not call it that. Model the skill, watch the trainee try, give feedback, repeat.
3. Providing immediate feedback. Praise when something goes well. A short correction in the moment when it does not. This is the most frequent supervision behavior in the dataset.
4. Written evaluations. Still happens, just less often than the verbal stuff. Often quarterly. Sometimes only at the end of fieldwork.
5. Delegation. Rarely shows up. Most supervisors do everything themselves rather than building a structured chain where senior trainees teach junior ones.
6. After-the-fact systems work. Setting up the next week's supervision schedule, picking the literature to review, planning the written evaluation. The work that happens at a desk, not in a session.

The pattern is clean. The first three are things you do with your mouth while a session is happening. The last three are things you do with a calendar open and the door closed. The first three get done. The last three do not.

Why in-the-moment feedback gets done and everything else gets skipped#

If you have ever finished a Friday wondering why the written evaluation never happened, this is the answer. The supervision behaviors that survived the week were the ones that fit inside a moment that was already on the schedule. The ones that did not survive were the ones that needed their own moment.

The talk frames this as a structural problem, not a motivation problem. Nobody decided not to write the evaluation. The evaluation just never got a slot.

This is also why two supervisors who feel like they are doing the same job can have very different fieldwork outcomes. The one who blocks 30 minutes a week for the desk work and the one who does not are doing different jobs, even if their in-session behavior looks identical.

The fix is not "try harder." The fix is to give the desk-work practices a recurring slot the same way you gave the session a recurring slot. If it does not have a calendar block, it does not exist.

The 7 supervision behaviors weakly linked to higher pass rates#

This is the part of the survey that should change a supervisor's calendar. Out of 46 practices, the survey found a small group with a weak but statistically real link to higher exam pass rates. None of the links were strong enough to call a magic bullet. All of them point in the same direction.

There were seven supervision behaviors that weakly, but did statistically significant, linked to higher pass rates. So what I'm saying is essentially there were some weak correlations, right? Nothing concrete, but checking credentialing requirements, pursuing new training, reviewing literature, including ethics, and frequent evaluations of client performance, all somewhat related to better pass rates. From the talk — Matt Harrington

The seven cluster into three buckets, which makes them easier to design a weekly habit around.

Bucket one is staying current. Checking credentialing requirements. Pursuing your own new training. Reading the literature with your trainee, including ethics. These are the practices that say "I am still learning, here is what I read this month, here is what we are going to try because of it."

Bucket two is feedback structure. Written evaluations on a real cadence, not just at the end. Frequent supervision meetings that are scheduled, not opportunistic. This is the desk-work pattern from the section above.

Bucket three is showing the trainee how a behavior analyst thinks about their own ongoing growth. The trainee is watching. If the supervisor never reads anything new, the trainee learns that the credential is the finish line. If the supervisor does read, the trainee learns that the credential is the start.

A BCaBA (Board Certified Assistant Behavior Analyst) supervising RBTs (Registered Behavior Technicians, the entry-level direct-care credential) can use the same three buckets. The exam in question changes. The structure does not.

The one supervision practice that hurt pass rates (and why)#

This is the surprising one. One supervision behavior in the dataset showed a negative correlation with pass rates. Spending more supervision time on more frequent evaluation of the client in front of you.

That sounds wrong at first. More client review should mean more learning, right? Read the finding carefully and it makes sense. If every supervision meeting becomes a status update on this one kid, the trainee never gets exposure to the broader skill set the exam tests. They get good at the case. They do not get good at the credential.

The fix is not "stop talking about clients." Clinical review still matters. The fix is to put a cap on it. Maybe 10 of your 30 weekly minutes on this client. The other 20 on a practice item, a literature piece, a written eval, or a credentialing check. The trainee is preparing for an exam that covers concepts. The supervision time has to cover concepts too.

If the trainee comes in every week wanting to talk about the same client, that is a signal, not a request to fulfill. They are anxious about that case. The clinical solve for the case is probably outside supervision. Inside supervision, you protect the time for the practices that are tied to passing.

A 30-minute weekly supervision template that hits the seven practices#

Here is a template a supervisor can drop into a recurring meeting on Monday. It is built around 30 minutes because that is what most caseloads actually allow per trainee per week, and it intentionally protects time for the desk-work practices that the survey says move pass rates.

Minutes 0 to 5. Open with a check-in on the trainee, not the caseload. How is the week. Anything blocking. This is not a clinical question. This is a relationship one.

Minutes 5 to 15. Clinical review of one case. Cap it at 10 minutes. Use a written agenda the trainee fills out beforehand so you do not spend the time deciding what to discuss. Capture decisions in the supervision log on the spot.

Minutes 15 to 22. Literature or ethics review. One article, one section of the BACB ethics code, or one chapter of a current text. Trainee summarizes. Supervisor adds the part the trainee missed. This is the practice the survey ties most clearly to better outcomes and the one most likely to get cut. Treat it as the 7-minute item that does not move.

Minutes 22 to 27. Written feedback. Not a full quarterly evaluation. A short written note in the supervision log of one thing the trainee did well this week and one thing to work on next week. Written, not just said. The trainee leaves with something on paper.

Minutes 27 to 30. Plan next week. One credentialing or training item the trainee will do on their own. One reading. One specific clinical skill to practice. Put it in writing.

That structure hits five of the seven practices every single week. Written evaluation. Literature review. Ethics review. Credentialing check. Frequent scheduled evaluation. The other two, pursuing new supervisor training and modeling ongoing growth, the supervisor does on their own time and brings into the meeting as content.

Frequently asked questions#

Which BCBA supervision practices are required by the BACB ethics code section 5?

Section 5 of the BACB ethics code requires supervisors to be qualified for the area they are supervising, to provide a written supervision contract that lays out scope, conditions, and termination terms, to use evidence-based instructional methods, to deliver ongoing performance feedback, and to keep documentation of the supervision provided. The Sakai survey discussed above is built on the 46 specific practices that map back to those broader requirements. The code is the floor. The survey is what the field actually does on top of that floor.

Does a fieldwork supervisor need to be certified for a minimum number of years?

A BCBA can serve as a fieldwork supervisor once they complete the BACB-required 8-hour supervision training and meet ongoing supervision continuing education requirements. There is no built-in waiting period of years, but supervising trainees in your first year of certification is a real workload question, not just a credential one. If you do supervise early, lean harder on the structured template above, because the desk-work practices are exactly the ones a new supervisor is most likely to skip.

How much of BCBA supervision should be group versus individual?

The BACB allows a mix, and both have a place. Individual supervision is where the written evaluation, the case-specific feedback, and the trainee-supervisor relationship live. Group supervision is a good fit for literature review, ethics scenarios, and BST role-plays where peer modeling helps. A common split is one 30-minute individual meeting per week plus one 60-minute group meeting every two weeks. Whatever the split, the seven practices tied to pass rates still need a home in someone's calendar.

Watch the full hour for the rest of the survey#

The CEU walks through the full set of 46 practices, the demographic patterns the survey found, and the parts of section 5 the dataset is built on. Watch the talk for the rest of the numbers and the BACB code mapping.