Red Flags and Green Lights: How to Evaluate a Supervision Program

Why Meeting Every Supervision Requirement Still Is Not Enough

Most supervision talk starts with hours and forms, and those things do matter. Anyone who has sat through an audit knows how fast the paperwork becomes the point. Feick opens with a harder claim: a program can meet every rule and still fail the trainee. Hours are the floor of the work, not proof that anyone learned to think.

She calls the risk the compliance trap. The hours are done, the forms are signed, and the meetings are all logged. The competency checklist may even be finished. Then comes the question Feick asks: then what? A trainee can log every required fieldwork hour (work hours done under a supervisor). They can still freeze on a real clinical call. They can define a term cleanly and still use it poorly with a client.

Her test is blunt and useful. Picture a trainee who meets the final hour tomorrow and passes the exam on Sunday. Could they practice well on Monday, with real decisions and no one to check them? If the answer is no, the paperwork did its job and the program did not.

Attendees named their own signs in the chat, and they were good ones. Can the trainee generalize (use a skill in new settings) when the scenario is novel? Can they find answers to new questions without being handed them? Can they judge their own limits and ask for help at the right time? Can they speak the language fluently and still coach an RBT (a front line therapy worker) in real time?

They help establish accountability, they demonstrate that we did actually do the things.

From the talk — Amber Feick

Green Lights One and Two: Purpose and a Plan Built for One Person

The first green light is purpose behind every supervision activity. Say you hand a trainee a treatment plan to review. What skill does that task build? Are you teaching clinical reasoning (thinking through a clinical decision), or the ability to defend a choice? If you cannot explain why a task sits in the plan, change it or drop it.

The second green light is supervision shaped around one person. Trainees do not arrive at the same level, and transfer hours hide a lot. A new trainee may bring 500 logged hours from another organization. Those hours tell you little about whether the work was active or meaningful. A trainee who is strong with data and shaky with parents needs a different plan. They may need a different BCBA (a certified behavior analyst) for part of it.

Needs also move over time. A trainee who looks solid with one client can struggle with the next profile. Skills grow, so the plan has to grow with them. A shared curriculum can set a floor, but it should not freeze the plan.

So every supervision activity should have a reason for occurring.

From the talk — Amber Feick

Green Lights Three to Five: Active Practice, Usable Feedback, Fading Support

Green light three is active supervision, not supervision made of conversation. The trainee needs to practice the skill, get feedback, and show what they learned. Treat it the way you would treat any staff training you run. You model the skill, watch them do it, give feedback, and then run it again.

Feick uses a preference assessment (finding what a client likes) as her example. Explaining one is fine, and a good trainee can explain it back to you. But if they never run one while you watch, the learning stays flat. The skill lives in the doing and in the feedback that follows.

Green light four is feedback that actually changes behavior. Make it specific, timely, behavioral, and something the trainee can act on. Telling someone that was good gives them nothing to repeat. Name what they did, say why it worked, and say what it means next time. Good feedback should carry to the next client and the next RBT.

Green light five is growth, with support that fades on purpose. Early on you model heavily and hold the structure tight. Over time you step back and let the trainee carry more of the decision. The sign off question keeps you honest here. Would you let this person write a plan or supervise an RBT alone tomorrow? If not, the gap is yours to close before the hours run out.

we want to be modeling we want to be giving that feedback we want them to be practicing again and again

From the talk — Amber Feick

Red Flags: Supervision That Lives in Meetings, Forms, and One Fixed Plan

The first red flag is supervision that happens only in meetings. Meetings have real value, and some feedback should not land in front of the RBT who is also there. Feick saves that kind of message for a private conversation. The problem shows up when the meeting is the entire program.

Ask three questions about any standing supervision meeting. What skill are we targeting today, and how is performance being measured? How will we know the feedback we gave actually worked? If all three answers are a conversation, you are grading talk instead of practice.

The second red flag is a program that runs on paperwork. Hours get tracked, forms get closed, and admin work fills the time. Very little is left for the thing the trainee came for. Feick is clear that records matter, but they are not the purpose of the work.

The third red flag is one program for everyone. A shared bank of assignments gives real consistency, and that is worth having. It goes wrong when the bank replaces judgment about this trainee right now. Assigning a task the trainee already mastered only proves that you assigned it.

Meetings can be very valuable.

From the talk — Amber Feick

Red Flags: Avoided Feedback, Supervisor Dependence, and No Visible Progress

Red flag four is avoided feedback, and the reasons for it always sound kind. You do not want to discourage someone new, or you are still building rapport. The skill gap does not pause while you wait for a better moment. The trainee keeps rehearsing the wrong version of the task. Feick practices hard conversations on purpose, and she names the discomfort out loud.

Red flag five is a trainee who runs on your answers. If every clinical question gets an instant answer, you have taught one skill: ask the supervisor. Flip it with questions of your own. What have you considered, and what information would help you decide? How will we know if the change you suggest is working? You can still correct a weak answer after you hear their thinking.

Red flag six is low psychological safety (feeling safe to ask questions). Questions, mistakes, and I do not know all have to be survivable. Feick says she will throw herself under the bus first and reach out for help. Passing the exam does not hand anyone every answer, and trainees should hear that.

Red flag seven is a program that cannot show progress. If your only evidence is the hour count, you are missing most of the picture. You should be able to name skills built, progress made, and areas still open. One side effect is worth planning for in advance. Strong trainees often finish with more hours than the minimum requires. Feick needed 1,500 hours and signed off closer to 2,300 of her own.

it's creating an environment where questions mistakes feedback are all okay.

From the talk — Amber Feick

The Supervision Cycle: Before, During, After, and In Between

Supervision works as a cycle, not as a standing appointment. Before the session, check what the trainee needs and write down your plan. What feedback did you give last time, and are you there to see if it stuck? Walk in knowing what you are looking for.

During the session, watch real performance and give feedback in the moment. Afterward, decide what has to change and what needs more practice. Do not simply hope the next visit looks better. Schedule the practice the same way you schedule the observation.

The days in between matter just as much. Say you supervise on Monday and are back again on Friday. What are you watching during those days in between? Do you pass them in another session and see independent work? Do you catch them coaching staff well without being asked? The real test is whether the skill holds with no supervisor in the room.

It's not just an isolated event.

From the talk — Amber Feick

The Supervision Stoplight Audit: Ten Domains to Score

Feick's tool is a stoplight audit across ten parts of a program. You score each part red, yellow, or green, and you score it honestly. Goals are red when unclear, yellow when generic, and green when tailored and measurable. Observation is red at bare minimum, yellow when it happens but randomly, green when planned and frequent.

Feedback is red when vague or rare, yellow when somewhat specific, green when frequent and usable. Skill development is red when it lives in discussion and definitions alone. Some real practice moves it to yellow, and scored checks of real work make it green. Fit is red at one size for all and yellow with small tweaks. It turns green when the plan follows the trainee's history and current needs.

Independence is red when the trainee leans on your answers for every call. Some emerging independence is yellow, and building it on purpose is green. Professional growth runs from limited, to occasional, to intentional and ongoing. Safety is red when the reply is figure it out, and green when it is steady. Records are red at compliance only and green when they show meaningful progress.

The tenth domain is the program itself. Do you evaluate whether your program works at all? No review at all is red, and stories about past trainees are yellow. Real data on how your trainees perform is the green light. Score each part on its own, because a program can be green on forms and red on observation.

those things are re-evaluated, they're rescheduled, and they still are frequent.

From the talk — Amber Feick

Reading Your Score: Three Questions and One Change in 30 Days

Read the whole card at once after you score it. Mostly red means the program needs real work, not small adjustments. Mostly yellow means you are compliant with clear room to improve. Mostly green means the base is built on quality and not only on rules.

Three questions sit underneath the entire audit. Are we meeting the requirements, which is the compliance question? Are we building competence, which is the quality question? Could this trainee work well without us, which is the independence question?

Independence does not mean never needing another person. Ten years in, Feick still reaches out when a case pushes past her limits. The goal is a professional who knows those limits and seeks support early.

Then pick one thing and start it this week. Choose a single green light behavior and work it for the next 30 days. Reassess at the end and see what actually changed in the program. The aim is not a perfect program by Friday. The aim is the habit of checking your own work and fixing it. Weak supervision copies itself into the next generation of supervisors. Feick would rather plant a garden the next person keeps growing.

a strong supervision program needs to address all three.

From the talk — Amber Feick