Prepare by translating terminology into decision rules: the observable trigger, the nearest competing term, and what excludes the competitor. Verify eligibility and administrative details in the BACB's current BCBA Handbook, and review the BACB's BCBA page and its 2027 requirements resources for upcoming changes before finalizing your plan. Administrative logistics belong to the BACB; your study time belongs to applied decision-making.
Decision rules for near-neighbor terms: DRA versus DRI as the template
Convert each term into a rule with three parts: the observable features that trigger it, the nearest competing term, and the specific feature that rules the competitor out.
Case-based application is where near-neighbor terms create genuine difficulty: a single case description can plausibly match two terms, and only a discriminating feature separates them. Preparation that stops at flashcard definitions leaves you without that feature at decision time. After studying each term, write a two-sentence rule naming what you would see in a scenario that makes the term apply, and what would have to change for the neighboring term to apply instead.
For example, DRA reinforces an alternative behavior while withholding reinforcement for the problem behavior; DRI requires the alternative to be physically incompatible with the problem behavior. The trigger is identical—reinforce a replacement. The discriminating feature is incompatibility. When a scenario notes that the learner cannot clap and hit at the same time, DRI fits; when the alternative is simply a different response, such as requesting a break, DRA fits. Practicing that discrimination, term by term, is the core work.
Descriptive assessment versus functional analysis: keeping evidence tiers straight
Descriptive assessment records antecedents and consequences as they occur; a functional analysis actively arranges and tests contingencies. Name the evidence tier before making any function claim.
A functional behavior assessment is an umbrella process that may include record review, interviews, direct observation, and systematic manipulations. Descriptive methods capture antecedent-behavior-consequence sequences in the natural setting and generate hypotheses about correlated events. A functional analysis goes further: the analyst alternates conditions such as attention, escape, tangible, and a control, then compares responding across them. Because the analyst controls the contingencies, a functional analysis supports stronger claims about behavioral function than descriptive data do.
Apply this as a tier-matching rule. If a teacher's log shows problem behavior typically follows task demands, say the data suggest an escape hypothesis worth testing; do not write that the behavior is escape-maintained as an established fact. Conversely, if conditions were actually manipulated and produced differentiated responding, cite the manipulation itself. Mislabeling the tier leads to treatment rationales that outrun the data, and in written reports, to causal claims your own methods cannot support.
Matching measurement dimensions and IOA formulas to the behavior
Select the dimension that answers the question—rate for cross-session comparison, duration for length, latency for onset, interresponse time for pacing—and match the agreement formula to the recording method.
Frequency counts alone mislead when observation periods differ in length, which is why rate (count divided by time) supports comparisons across sessions. Duration fits behaviors that vary mainly in how long they last, such as tantrum episodes; latency measures the time from instruction to response onset; interresponse time describes pacing between responses. In practice examples, state the dimension and the reason together: converted to rate because session lengths varied from ten to twenty minutes.
Interobserver agreement has its own matching rule. Total-count agreement suits event recording and permanent products; interval-based agreement suits partial-interval, whole-interval, and momentary time sampling; trial-by-trial agreement fits discrete-trial data. Watch the recording system's known bias: partial-interval recording tends to overestimate high-rate behavior, while whole-interval tends to underestimate it. A useful drill is to reconstruct two observers' tallies from a short practice sample and compute agreement the same way the data were collected.
Single-case design selection: reading the scenario's constraints
Choose the design from the scenario's constraints rather than preference: withdrawal only when behavior is reversible and safe to reverse; multiple baseline when it is not; alternating treatments to compare interventions.
Treat each design's constraint as the decision trigger you scan for in the scenario. Language about a skill that will not be lost, safety concerns about withholding treatment, or instructional settings where withdrawal is impractical points toward a multiple baseline. A request to compare two treatments quickly points toward alternating treatments, provided the conditions are discriminable. Questions about whether behavior tracks a gradually raised standard point to a changing criterion design.
Then match your interpretation to the design's logic. A reversal supports a claim about treatment control when behavior tracks the phases on and off. A multiple baseline relies on staggered onsets: change should appear in each tier only after introduction there, so predicted stability during extended baselines is doing the inferential work. Interpretation rules matter as much as selection, because a design chosen for the right reason can still be read incorrectly in the write-up.
| Design | Choose when the scenario says | Avoid when the scenario says | Interpretation anchor |
|---|---|---|---|
| Reversal / withdrawal | Behavior is reversible and can be safely withheld temporarily | Skill persists once learned, or withdrawal poses risk | Behavior tracks treatment on and off phases |
| Multiple baseline | Withdrawal is impractical; several behaviors, settings, or participants are available | Only one behavior or tier exists; baselines cannot be staggered | Change appears in each tier only after introduction there |
| Alternating treatments | Two or more interventions need rapid comparison | Conditions are not discriminable or carryover obscures effects | Consistent response differences between compared conditions |
| Changing criterion | The question is whether behavior tracks a gradually raised standard | Criterion steps are too large or baseline responding is unstable | Level shifts after each criterion change |
Worked scenario: an undifferentiated functional analysis graph
When test conditions are elevated but similar to each other, report an undifferentiated pattern and gather more evidence rather than assigning the single highest condition as the function.
Scenario: a functional analysis for a school-age learner shows low responding in the play condition, moderate elevation in the attention condition, and nearly identical elevation in the escape condition, with tangible near control levels. The tempting call is attention-maintained, so the team should build an attention-based intervention, because attention scored highest. The mistake is treating a rank ordering as a differentiated outcome when two test conditions are functionally similar to each other and both exceed control.
The stronger decision is to describe the pattern accurately—an undifferentiated elevation across attention and escape relative to control—and state what follows: descriptive data collection, a modified or extended analysis, or a brief contingency test to separate the hypotheses. This matters because treatment selection is function-based; choosing attention-based reinforcement when escape may actually drive responding risks reinforcing escape-maintained behavior, and the written rationale would claim more than the graph shows.
- Report which test conditions exceeded control, and by how much, before naming any single function.
- Note whether attention and escape were similar to each other, not merely higher than play.
- State the next evidence step: descriptive ABC data, a modified or extended analysis, or a brief contingency test.
- Write the summary so the strength of the function claim matches the tier of evidence behind it.
Worked scenario: a supervision lapse disguised as initiative
A supervisee improvising a novel response to escalation is a supervision and treatment-integrity issue: the obligation is preparation, training, and documented oversight, not retrospective praise.
Scenario: a BCBA writes a plan that includes a crisis protocol. An RBT, facing escalation the steps do not cover, improvises a response on the spot; afterward, the BCBA compliments the initiative and files the session note. The plausible mistake is treating improvisation as admirable flexibility. The analyst's relevant obligations run earlier in time: behavior-change plans should account for predictable challenges, and supervisees should be trained and assessed for competency on the procedures they are expected to implement.
The better decision is to review the plan for missing contingencies, train and assess the RBT on an updated protocol, increase temporary observation of the procedure, and document both the change and the supervision provided. This matters on two fronts—client safety, because untrained responses to escalation carry risk, and professional accountability, because the supervising analyst owns the plan and the oversight. Principles in the Ethics Code for Behavior Analysts, such as supervisory responsibility and supporting effective treatment, frame exactly this kind of decision.
A self-check rubric and an adaptable preparation sequence
Build a term-to-decision-rule sheet, drill graph and vignette decisions, then run mixed sets; score yourself with a rubric that rewards stated alternatives, not just the right term.
Exercise: select ten practice items you have already answered. For each, write the decision rule you used in two sentences, name the closest competing term or design, and state the feature that excluded it. Score against this rubric, one point each: the rule cites a trigger feature from the scenario; a named alternative is identified; the exclusion reason is specific; an implication for treatment, design, or ethics is noted. Expected observation: early passes feel slow and rules run long, but after two or three passes the competing terms come to mind automatically and your written exclusions shorten.
An adaptable sequence: weeks one and two, build the decision-rule sheet across your content domains and drill measurement and agreement examples; weeks three and four, graph interpretation and design selection using the table above; weeks five and six, ethics vignettes organized by the Ethics Code's named principles; final weeks, mixed timed sets, updating the decision rule for every miss. Self-check scores are learning milestones for pacing yourself, not predictions about any passing standard.
- Rubric checkpoint 1: every rule cites a trigger feature visible in the scenario itself.
- Rubric checkpoint 2: a named competing term, method, or design is identified.
- Rubric checkpoint 3: the exclusion reason is specific, not simply that the other term did not fit.
- Rubric checkpoint 4: an implication for treatment, design, or ethics is stated.
- Readiness check: on mixed timed sets, you can state the competing candidate and the exclusion rule aloud before choosing an answer.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.