GHABA Proficiency II common mistakes
GHABA Proficiency II · Understand the traps and how to avoid them.
Defining a behavior in terms of an internal state or a broad trait (e.g., "was aggressive," "seemed anxious," "lacked motivation") instead of writing an observable, measurable, dead-man's-test-passing operational definition with clear onset and offset criteria.
Candidates carry everyday mentalistic language from their training or workplace notes into exam items, and under time pressure they pick the option that sounds most clinically descriptive rather than the one that specifies topography, boundaries, and measurement.
What to do instead
Before answering any definition or measurement item, silently apply three tests: Can two observers independently agree on occurrence? Does the definition pass the dead-man's test (a dead person could not do it)? Are onset and offset specified? Eliminate any option containing mental-state verbs (wants, feels, knows, is motivated) or unobservable constructs. Practice rewriting ten of your own workplace behavior descriptions into observable terms.
Confusing the three-term contingency with the four-term contingency, or mislabeling the motivating operation as a discriminative stimulus (or vice versa) in scenario items, and then selecting an intervention that manipulates the wrong term.
The MO and the SD both occur before behavior and both increase the likelihood of responding, so candidates pattern-match on temporal position rather than function: the SD signals availability of reinforcement, while the MO alters the value of the consequence and the momentary frequency of the behavior.
What to do instead
For every antecedent-consequence item, write out Sd, MO, behavior, and consequence on scratch paper before reading the options. Ask two questions: Does this stimulus change the value of the outcome (MO), or does it signal that the outcome is currently available (SD)? Then check which contingency term the proposed intervention actually targets.
Selecting a reinforcer or treatment based on staff preference, ease of implementation, or what worked for a previous client, rather than conducting a stimulus preference assessment and then validating the presumed reinforcer with a reinforcer assessment.
Preference assessments and reinforcer assessments are frequently taught together, so candidates conflate them; they also assume a high-preference item functions as reinforcement without empirical verification, and exam items deliberately reward the distinction.
What to do instead
Memorize the decision rule: preference assessment ranks stimuli, reinforcer assessment demonstrates that the stimulus increases behavior. When an item asks what to do before using an item as a reinforcer, the answer involves a reinforcer assessment, not just a preference assessment. Also be able to name which preference assessment format (single stimulus, paired choice, multiple stimulus without replacement, free operant) fits a given learner profile.
Reporting a functional analysis result as conclusive when the data show undifferentiated or inconclusive patterns, or skipping the functional analysis entirely and asserting a function from a single indirect source such as a teacher interview or rating scale.
Candidates want a clean answer and feel pressure to produce a function statement; they also over-rely on indirect measures because those are faster, and they underweight the requirement that function be demonstrated through experimental control or converging evidence.
What to do instead
When an item describes undifferentiated FA data, the correct response is usually to refine conditions (e.g., modify contingencies, extend sessions, conduct a brief or trial-based analysis, or evaluate an idiosyncratic variable) rather than to declare a function. Treat indirect and descriptive data as hypothesis-generating only. Always ask: what is the evidence of experimental control here?
Confusing conditional and trial-based functional analysis formats, or misidentifying which condition (attention, escape, tangible, alone/automatic, play/control) is being described when the item reports consequences delivered contingent on behavior.
The conditions share a common structure and differ only in the consequence delivered and the antecedent arrangement; candidates skim for keywords like "demand" or "attention" without checking who delivers the consequence and when.
What to do instead
Build a comparison table from memory: for each condition, list the antecedent, the consequence contingent on the target behavior, and the control comparison. Then, for any scenario, identify the consequence actually delivered contingent on behavior, not the setting or the person present. Rehearse the logic of the alone condition and why it tests automatic reinforcement.
Writing intervention goals that are not socially valid or not anchored to the assessed function, such as targeting a topography reduction without addressing the maintaining contingency, or selecting a punishment-based procedure when a function-based reinforcement procedure is indicated.
Candidates default to suppression of behavior because that is what referral sources ask for, and they lose sight of the requirement that intervention selection follow from the functional analysis and consider least-restrictive, reinforcement-first alternatives.
What to do instead
For each intervention item, state the hypothesized function first, then ask whether the proposed procedure matches that function (e.g., FCT or extinction for attention-maintained behavior; demand fading, choice, or NCR for escape-maintained). Eliminate options that reduce behavior without a plausible function-based mechanism. Confirm the plan includes a replacement behavior, a reinforcement schedule, and a plan for generalization and maintenance.
Designing or evaluating a behavior intervention plan that omits generalization and maintenance planning, or that specifies a consequence strategy without specifying the antecedent manipulations, teaching procedures, and data-based decision rules.
Exam items often present plans that look complete because they contain a reinforcement schedule; candidates accept them without checking the full plan architecture, and they underweight the requirement for proactive strategies and for objective decision criteria.
What to do instead
Use a checklist for every plan item: operational definition, baseline data, function statement, antecedent strategies, replacement behavior teaching, consequence strategies, reinforcement schedule, crisis/safety plan where relevant, generalization plan, maintenance plan, data collection method, and decision rules with timelines. If any element is missing, the option describing the plan is likely incomplete.
Miscomputing or misinterpreting interobserver agreement and treatment integrity data, such as using the wrong IOA formula for the recording method, or treating low treatment integrity as irrelevant to interpreting outcomes.
Candidates memorize formulas without the conditions of use (total count vs. interval-by-interval vs. trial-by-trial) and forget that IOA and procedural integrity are separate indices answering separate questions; they also confuse IOA with accuracy.
What to do instead
Match formula to method: exact count/total for frequency and duration, interval-by-interval for partial and whole interval, trial-by-trial for discrete trial data. Remember IOA measures agreement between observers, not correctness against a standard. When an item reports low treatment integrity, the correct interpretation is that the independent variable was not implemented as intended, so conclusions about the intervention's effect are limited.
Choosing a single-subject design that cannot answer the question posed, such as using an AB design to claim experimental control, or failing to identify the design from a described graph (reversal, multiple baseline across behaviors/settings/participants, alternating treatments, changing criterion).
Candidates focus on the visual pattern of the data rather than the logic of the design, and they forget that demonstration of experimental control requires replication of the effect within the design.
What to do instead
For each design question, ask: where is the replication, and what is the comparison? An AB design has no replication and cannot demonstrate control. Learn to identify designs from structural features (number of tiers and staggered introduction, presence of a return to baseline, rapid alternation of conditions, stepwise criteria). Also be ready to interpret level, trend, and variability and to state whether a change is clinically and statistically meaningful.
Overlooking informed consent, assent, confidentiality, and mandated-reporting obligations embedded in scenario items, or applying the wrong regulatory framework (e.g., citing the wrong body's guidelines, or asserting that a supervisor's approval substitutes for consent).
Candidates read ethics items as clinical items and answer the therapeutically appealing option; they also blur the distinction between consent (legally authorized person) and assent (the client's agreement), and between organizational policy and enforceable regulation.
What to do instead
For every ethics scenario, identify in order: who is the client, who can legally consent, whether assent is required, what documentation exists, and which obligation is triggered (report suspected abuse, protect confidentiality, avoid dual relationships, obtain consent before recording). Choose the option that satisfies the most binding obligation first, and remember that supervision does not transfer accountability for consent.
Treating supervision as informal advice-giving: failing to specify supervision goals, observation of the supervisee's actual performance, documentation of supervision contacts, or a remediation plan when competencies are not met.
Candidates with clinical experience but limited supervisory training equate supervision with case consultation, and they underestimate the documentation and competency-verification requirements that exam items probe.
What to do instead
Memorize the supervision cycle: contract and goal-setting, direct observation of the supervisee's performance, feedback, documentation of contacts and content, evaluation against stated competencies, and a written remediation plan if benchmarks are not met. For scenario items, prefer the option that includes direct observation and documented feedback over the option that relies on self-report.
Failing to establish role clarity and communication structures in interdisciplinary scenarios, such as accepting recommendations from another discipline without evaluating their compatibility with the behavior plan, or declining to share data that would inform a team decision.
Candidates either defer entirely to authority (medical, educational) or defend their own plan territoriality; both errors stem from not having a clear model of the behavior analyst's scope and the team's decision-making process.
What to do instead
In collaboration items, identify the behavior analyst's scope (behavior assessment, intervention design, data-based decision making) and the complementary scopes of other professionals, then choose the option that shares objective data, seeks input, and documents the rationale for the final plan. Prefer options describing a coordinated, data-driven team process over unilateral action in either direction.