From Hours to Competence: Assessing and Developing Supervisee Skills

Why Hour Counts Do Not Prove Competence

Supervision is easy to count in hours. The log shows when a trainee (someone earning hours toward certification) showed up. It does not show what that person can do alone. Lauren Donovan built this whole talk around that gap. Time is the input. Competence is what we owe the field.

Exposure is not mastery. A trainee can watch you write a plan one time and still not be able to write one next month. Fieldwork (supervised hours before board certification) is supposed to build an independent clinician, not just a person who is eligible to sit for the exam. Most new BCBAs (certified behavior analysts) start practicing on their own fast. The support they get after that varies a lot from place to place.

So the stakes sit with the supervisor. The people you train will serve real learners next year. Many of them will supervise other people a few years after that. Donovan said supervision quality shows up in the programming those learners receive. Weak fieldwork passes down. Strong fieldwork passes down too.

Completing an activity does not mean it can be performed.

From the talk — Lauren Donovan

The Barriers Supervisors Name First: Time and Billable Hours

Donovan polled the room on their biggest barriers to supervising trainees. Time came out biggest in the word cloud, and she said that is always the answer. Scheduling, follow through, consistency, and motivation came next. Several people named unrestricted hours (fieldwork tasks outside direct service) as the hardest part of the job. Others named a missing curriculum or a system that is too complicated to keep.

The money math drives a lot of it. Unrestricted work is often not billable. Prep time, planning time, and the supervision meeting itself usually are not billable either. Trainees still have targets to hit, and so do you.

Funding rules add a second wall. Some work has to carry the certified person's signature before it goes to a funder, so you cannot hand that task to a trainee at all. Setting narrows things too. A trainee who is in one home with one learner sees a thin slice of the field. Donovan said she never ran a functional analysis (a test of behavior's cause) during her own fieldwork and had to learn it after she was certified. She pointed to Lydden et al. 2020 for options. The top two there were nonprofit work and salaried trainee or apprentice roles.

None of that's billable and you have to get billables every month, every weekend.

From the talk — Lauren Donovan

Set Expectations Before You Accept a Trainee

Donovan front loads the work. Name both roles at the start. Name how goals get set and how progress gets checked. Name the clinical opportunities you can offer and the ones you cannot. Put all of it in the supervision agreement (a written contract with your trainee). Be honest that some fieldwork will land outside paid time, because for many trainees it will.

She is direct about limits. Trainees ask for more hours, and supervisors have far less time than the demand. Decide your cap before the request arrives. If you are new to this, take one person and build your system around that one person. Ask them for feedback on how it is going. Then grow the number. Saying no to a request is part of doing this well.

Scope matters just as much. You do not have to teach the whole test content outline (the exam's topic list), and you are not teaching to the exam at all. You are building clinical skill on real cases. If a skill sits outside your scope, say that out loud. Some trainees seek a second supervisor on purpose to widen what they learn.

Teach what you know and teach what you know well.

From the talk — Lauren Donovan

Start With a Baseline, Not a Blank Page

Two trainees rarely start in the same place. One may be brand new to the field. One may be transferring in with a stack of hours already logged. So Donovan opens with a baseline (where their skills start today). She reviews whatever records came from the prior supervisor, has the trainee rate their own skills, then runs her own assessment.

Direct observation is the part people skip. A self rating and a live look often disagree, and that gap is the useful part. It tells you where teaching should start. It also tells you which claims need a second look before you sign off on anything.

She keeps the test content outline nearby as a guide rather than a syllabus. Items get crossed off as they are covered in real work. Nobody covers everything in a month. The list keeps your coverage honest across a full year of supervision.

One might be brand new. Another might have a thousand hours.

From the talk — Lauren Donovan

Track Knowledge, Application, and Performance Separately

This is the heart of her system. Donovan splits her tracker into three columns. Knowledge is what a trainee can define and explain. Application is what they do with a scenario or a graph in your meeting. Performance is what they do with a real case in the room.

She split them for a reason. Some people can define any term and explain it beautifully to a parent, then freeze in the moment with a learner. Others run great sessions but cannot tell you why any of it worked. A BCBA has to do both. Tracking only one column hides the hole in the other one.

The knowledge questions are simple. Can they define it? Can they explain it to a teacher or a parent? Can they describe the ethical issues inside it? Donovan folds ethics into every standard instead of teaching it as a separate unit. Application happens during your meeting. Talk through scenarios, pull up a real graph together, and ask what the data show and what they would change next. She notes that AI can draft practice scenarios, as long as you check them and keep student information out.

Performance means show me. Run a staff training. Run a parent training. Make a program change on time. Write a treatment plan, behavior guidelines, or an FBA (a study of why behavior happens). Take permanent product data (data from work left behind). Every one of those gets rated on the same four point scale, and the file lives somewhere the trainee can see it so you set the next goal together.

I use that same rating scale for every thing so it's just a four-point Likert scale, not demonstrated. Emerging, proficient, and and independent.

From the talk — Lauren Donovan

Choose What to Teach From Coursework, Caseload, and Goals

Donovan does not follow a packaged curriculum. She builds the plan around each trainee and calls it a puzzle. Four pieces fit together: the test content outline, the courses they are taking right now, what the job and the organization actually need, and their caseload, goals, and interests.

Coursework is the easiest hook. Look at the syllabus with them. If the class is on measurement this month, plan the next assignments around measurement. The schoolwork itself does not count as fieldwork hours. The applied version of that same topic does, and the bridge between the two is where learning sticks.

She revisits the plan as things change. New course, new competencies, new caseload, new responsibilities. Spend the last five minutes of each meeting on next steps and ask what they are studying right now. She credited Garza et al. for part of this approach and added her own pieces on top.

Unrestricted hours should be built. On skills for independent practice. Not just to accumulate hours.

From the talk — Lauren Donovan

Turn One Stuck Program Into Weeks of Unrestricted Hours

Supervisors keep saying unrestricted hours are hard to find. Donovan's answer is to mine the clinical work already sitting on your desk. One program that is not moving can carry weeks of meaningful tasks.

The room proved it live. For a learner who is stuck on labeling objects, attendees listed a fidelity check (is staff running it as written), interobserver agreement data (two people score the same thing), error analysis (looking at patterns in mistakes), a preference assessment (finding what a learner likes), probes with new materials, a prompting review, and a check on prerequisite skills (skills needed before this one). For rising aggression they listed ABC data sheets (antecedent, behavior, consequence notes), another fidelity check, an intervention review, and a full FBA.

Big tasks get broken into steps. An FBA can fill several weeks on its own. One week on writing definitions. The next on building the ABC sheet. Then graphing the data and learning to read it. Donovan also assigns a time estimate with each task, because a job that should take an hour should not take ten. She noted the 2027 requirements are stricter, and the work has to involve a real learner rather than made up data.

The last move is teaching trainees to bring you their own proposals. They still need your approval before they run anything. But a trainee who asks to run a fidelity check across staff has saved you the assignment and is already motivated to do it. That is a skill you teach on purpose, one generalization at a time.

One clinical situation can create so many meaningful unrestricted opportunities.

From the talk — Lauren Donovan

Use BST Without Missing Your Deadline

BST (tell, show, practice, then feedback) is the standard here. The board names it as a core part of effective supervision, and we all know the research. Donovan's honest read is that most of us drop it the second a deadline shows up.

Her fix is to run both tracks at once. You write the report and meet the deadline, and your version is the one that gets submitted. The trainee writes on the same data as a teaching task. Then you compare the two documents and walk through how you got to yours. If a full plan is too much, give them one section instead of the whole thing.

She also warns against handing out tasks before you have taught them. Prepare the trainee. Model the work. Give the background knowledge first. Let them watch you make the clinical decision out loud and narrate the steps. Train one thing at a time if that is all the week allows.

Then the whole thing loops. Assess, set goals, teach with BST, observe, give timely feedback, and reassess. Donovan pointed out that this is the same logic we already use when we program for learners. Her argument is that strong systems cut the administrative drag, which leaves more of your attention for the supervision itself.

Someone's working through their first treatment plan. You still have a deadline so I I suggest you do both, right?

From the talk — Lauren Donovan