RBT Turnover: What the Research Shows and What ABA Employers Can Change

The most-cited number in ABA staffing conversations — a national RBT turnover rate — is not a number our research could verify.

Novack and Dixon's systematic review, published in 2019 in the Review Journal of Autism and Developmental Disorders, identified 33 articles on behavior technician retention, turnover, burnout and job satisfaction, and only 5 studies met its inclusion criteria.

Across those five studies the sample was 351 behavior technicians, 89% to 100% female, drawn from California, Ireland and the United Kingdom among other places.

The review's bottom line on the headline question: estimated turnover rates for behavior technicians had not been reported, and none of the five studies reported one.

The figures you will see quoted have specific sources, and the sources limit them:

• 48%. A 2018 Behavioral Health Center of Excellence (BHCOE) figure for direct care staff — not RBTs specifically — cited in Slowiak and DeLongchamp's 2022 peer-reviewed article. The underlying report's sample and method were not described in what we could read, so treat it as context about the era, not a benchmark for your clinic.
• 70.7% to 77%. These come from 1994 and 2002 studies of community direct-care staff serving people with developmental disabilities — not certified RBTs in modern ABA organizations.
• 38% likely to resign. Kazemi, Shapiro and Kavner's 2015 survey measured intention: 38% of the behavior technicians sampled said they were highly or somewhat likely to resign. Intention to leave is not the same measurement as actual turnover.

Burnout research fills in some of the picture: in Slowiak and DeLongchamp's survey of 826 ABA practitioners, 72% reported medium to high levels of burnout — a convenience sample of ABA practitioners generally, not RBTs only.

Before you repeat any turnover figure internally, trace it.

Each of the figures above describes a different workforce, a different era of the field, or something other than measured turnover.

The one turnover number you can act on is the one you compute from your own roster — the last section covers how.

The strongest current dataset is the BACB's own: in February and March 2025 it surveyed 30,018 people whose RBT certification expired in 2024 and who held no other BACB credential; 1,386 responded.

That is a 4.6% response rate, so read every number below as what respondents reported, not a census.

Compensation and workload topped their own list: 57% of respondents named inadequate pay, 38% named high work demands with clients and 34% named unpredictable pay.

Respondents could select multiple reasons, so the percentages do not sum to 100.

Organizational reasons included: concerning issues in the workplace (42%), limited opportunities for professional growth (41%), too many or too few weekly hours (41%), an unpredictable schedule (38%) and travel or commuting between sites (26%).

Workplace-culture reasons included: poor treatment by the company (44%), lack of supervisor support (41%) and isolation in the workplace (27%).

One result belongs in every internal debate: 58% of respondents chose "I like the work, but life took me in a different direction."

Life circumstances are not all within an employer's control — the workplace items on the same list are.

Asked what would bring them back, respondents overwhelmingly named better pay, consistent and predictable scheduling, and more professional support.

The BACB's own conclusion was that improving pay, scheduling predictability and supervisor support may enhance RBT retention and re-engagement — a short list, and mostly one an employer controls.

The interview research points the same way.

Nastasi and colleagues interviewed 11 Florida RBTs (published online in December 2024) and identified four themes behind burnout and turnover: trouble establishing and keeping competency, difficult working conditions, a transient or little-known career, and dissatisfaction with pay and benefits.

With 11 participants these are themes, not rates — but they are the vocabulary the technicians interviewed used.

These numbers measure technicians, not licensed clinicians: counselors, social workers and psychologists are a different population, and our clinician retention guide covers that workforce separately, from different surveys.

Supervision shows up twice in the evidence: as the certification requirement that structures the working relationship, and as a factor the studies link to whether technicians stay, though not in every study.

Start with the rule, because it is a floor, not a retention program.

RBTs must receive supervision for at least 5% of the hours they spend providing behavior-analytic services each calendar month, including at least two face-to-face, real-time contacts per month, with the supervisor observing the RBT providing services in at least one of them.

Those are BACB requirements for keeping the credential.

The BACB lists what supervision is meant to include: developing performance expectations, observation with behavioral skills training and feedback, and ongoing evaluation of the effects of supervision.

The operational mechanics — scheduling, records, what an RBT cannot do when supervision lapses — are covered in our RBT supervision guide.

The research suggests what surrounds that minimum matters.

Across the studies in the 2019 review, perceived supervisor support was associated with lower burnout, higher job satisfaction and lower intention to leave — though not significant in every study.

In the BACB exit survey, lack of supervisor support was named by 41% of respondents.

The Ethics Code adds a supervisor duty that reads like a retention tool: standard 4.08 requires behavior analysts to collect and document ongoing performance-monitoring data on supervisees, give timely feedback and, when performance problems arise, develop and evaluate an improvement plan with clearly identified procedures.

That is an ethics duty of the supervising behavior analyst, not an employment-law requirement — but a technician who only hears about problems at termination has not been supervised in this sense.

Supervisor turnover and technician turnover move together.

Cymbal, Litvak, Wilder and Burns, analyzing BHCOE accreditation data in a study published online in 2021, found that technician and supervisor turnover rates correlated with each other.

There is a structural reason: an RBT may practice only while an RBT Supervisor or RBT Requirements Coordinator is on record with the BACB, so when your supervising BCBA or BCaBA resigns, the technicians under them cannot continue until you get a new supervisor on record.

Treat a supervisor resignation as a staffing event with that deadline attached: plan coverage before you accept one.

ABA schedules are built around client sessions, school days and authorized hours, and they move — clients cancel, school schedules shift, authorizations end.

The research splits that volatility into two parts, how many hours a technician works and how predictable those hours are, and the two do not behave the same.

In Kazemi's 2015 survey, technicians who reported being under-scheduled were more likely to intend to leave, while the number of hours worked per week was not correlated with intention to leave.

The BACB exit survey names both: 41% of respondents named too many or too few weekly hours as a reason for leaving, and 38% named an unpredictable schedule.

Pay shows the same perception-versus-reality split.

In the exit survey, inadequate pay was the top of the compensation and workload reasons at 57%, and 34% named unpredictable pay.

But in Kazemi's data, satisfaction with pay predicted intention to leave while technicians' actual hourly wages did not.

Neither finding tells you what to pay — our research has no sourced RBT pay benchmark, and BLS has no occupation series that isolates RBT pay.

Both point at the same split: the rate on the paycheck and the predictability of the paycheck are separate levers.

We found no published benchmarks for the operational fixes, so treat these as decisions to make deliberately, not standards to match:

• How far in advance schedules are published, and how technicians are told about changes.
• What happens to a technician's hours and pay when a client cancels — and whether cancellations concentrate on particular technicians.
• Whether technicians have a core of predictable weekly hours, and what that commitment costs against the hours you can actually schedule.
• How much travel a schedule carries — 26% of exit-survey respondents named travel or commute between sites — and whether technician territories can be drawn to shrink it.

When a resignation lands, the backfill is its own craft: recruiting RBTs covers sourcing, screening and what to show in the ad, and the behavioral health hiring hub indexes every guide in this series.

The oldest ABA-specific finding on this page is about onboarding.

Kazemi's team found satisfaction with training was linked to lower intention to leave, and satisfaction with training corresponded with 30 or more hours of initial training.

The credential's training sits above that mark: since January 1, 2026, new RBT applicants must meet revised eligibility requirements, including a 40-hour training that meets the updated training requirements.

We found no study measuring whether the updated training changes retention.

What the evidence supports is narrower: training substantial enough that technicians report satisfaction with it, followed by the feedback-based coaching the BACB's supervision activities describe.

Recertification creates a natural check-in.

Under the 2026 rules, RBT recertification moves to a two-year cycle with a new requirement to earn 12 professional development units (PDUs) in each cycle; RBTs already certified recertify as usual in 2026 before the new cycle applies to them.

Whether your practice funds PDUs and protects time for them is now a visible signal of how investable the job feels.

Growth is the quiet driver: 41% of exit-survey respondents named limited opportunities for professional growth.

We found no research measuring whether a career ladder reduces technician turnover, so no percentage belongs here.

But the credential system has built rungs you can use.

The noncertified RBT Supervisor role has been eliminated — RBT Supervisors must hold BCaBA or BCBA certification — and before supervising, RBT Supervisors and RBT Requirements Coordinators must complete training based on the BACB Supervisor Training Curriculum Outline (2.0) and upload it to their BACB account.

In practice that makes supervision a genuine rung for a technician who goes on to earn BCaBA or BCBA certification, with the training, caseload and compensation conversation that come with it.

Smaller rungs count too: lead-technician duties, mentoring new hires, owning part of the scheduling or materials work.

If you want the rungs in detail — supervisor eligibility, fieldwork hours, the contracts involved — our BCBA fieldwork guide covers the employer side, and the behavior technician career path covers the credential rungs from the technician's side.

Since the research here turned up no verified national rate, the number worth managing is yours — and consistency matters more than the formula.

Pick a period (a quarter, or a trailing twelve months), count separations, divide by average technician headcount over the period, and keep the definition fixed so quarters compare.

Decide once whether resignations and terminations count together or separately, and whether you measure RBTs alone or blended with BCaBAs and BCBAs — a blended number hides this population.

Then cut it into pieces you can act on:

• By supervisor. Technician and supervisor turnover moved together in the BHCOE accreditation data, and a supervisor departure freezes the technicians they oversaw — supervisor-level rates show where the risk sits.
• By tenure cohort. Separations in the first months point at onboarding and training; separations after a year point at growth, pay or schedule.
• By site and schedule pattern. The exit survey's organizational items — hours, unpredictable schedules, travel between sites — are the kind of items that can differ site to site inside one company.

Code the reason behind every separation using the same categories the BACB used — pay, workload with clients, hours, schedule predictability, supervisor support, growth, treatment by the company, commute — and your counts become comparable to the survey's proportions, not anecdote.

A satisfaction survey is not a substitute: in the same BHCOE-data study, staff satisfaction was not a significant predictor of turnover; the same study tied technician turnover to lower wages and declining caregiver satisfaction.

Exit conversations, and their stay-interview counterparts, are where the reasons surface — keep the questions on work conditions rather than health.