By Matt Harrington, BCBA · Behaviorist Book Club · April 2026 · 12 min read

Multiple Alternative Prompts for Tact Training: Clinical Evidence and Practice

Overview & Clinical Significance

Tacting — labeling objects, actions, and events in the environment — is a foundational verbal operant with far-reaching effects on language development, social communication, and academic readiness. For individuals with autism spectrum disorder, tact deficits are among the most clinically significant barriers to communication, as tacts enable individuals to comment on and share their experience of the world with others. The development of effective, efficient tact training procedures is therefore a high-priority area within verbal behavior programming.

Dr. Leaf's work comparing multiple alternative prompting (MAP) to standard vocal verbal prompting for tact training addresses a specific and practically important question: does providing the learner with multiple choice alternatives during tact training produce better outcomes than single vocal prompts, without increasing error rates or teaching time? The results, derived from an adapted alternating-treatments design across three individuals with ASD, indicate that both procedures were effective relative to a no-intervention control, that MAP did not increase errors or extend teaching duration, and — critically — that MAP produced superior maintenance of tacted responses over time.

For BCBAs designing verbal behavior programs, this research has direct implications for procedure selection in tact training protocols. Maintenance — the persistence of acquired skills after the training phase ends — is a frequently underaddressed outcome in behavior analytic research. Skills that are acquired during intensive training but fail to maintain in naturalistic conditions have limited functional value. A training procedure that produces equivalent or superior acquisition with better maintenance represents a clinically meaningful advance.

This presentation is relevant to practitioners at all levels who design or supervise tact programs, particularly those working with learners in the early stages of verbal behavior development or those who are transitioning from receptive to expressive language programming.

Background & Context

The tact is one of the six primary verbal operants described by Skinner in Verbal Behavior (1957). A tact is controlled by a non-verbal antecedent stimulus — an object, action, property, or event — and is maintained by generalized conditioned reinforcement (social approval). In contrast to the mand, which is controlled by motivating operations, or the intraverbal, which is controlled by verbal antecedents, the tact is primarily a stimulus-discrimination behavior that requires the speaker to attend to and label features of the physical environment.

Vocal verbal prompts are among the most common prompting strategies in tact training. The instructor presents a stimulus, provides a vocal model of the correct label, and reinforces correct responding or imitation. Over time, the prompt is faded using procedures such as time delay or most-to-least prompting, with the goal of establishing stimulus control by the antecedent object or event rather than the instructor's vocal model.

Multiple alternative prompting provides additional response options alongside or in place of the single vocal model. By presenting multiple possible labels and requiring the learner to select the correct one, MAP introduces a discrimination component into the prompting procedure itself. This is functionally analogous to a modified errorless learning approach in which the learner must actively engage with the stimulus set rather than passively imitating a single model.

The adapted alternating-treatments design used in Leaf's study is the gold standard for comparing two or more procedures for the same learner, controlling for confounds associated with multiple treatment comparisons. By rapidly alternating between MAP and vocal verbal prompt conditions within sessions and across participants, the design provides convincing within-subject evidence that any observed differences in acquisition or maintenance are attributable to the procedure rather than to participant characteristics, history effects, or time in treatment.

Maintenance data — collected after the training phase without continued prompting or reinforcement schedules — is the most clinically relevant indicator of durable learning. Leaf's finding that MAP produced better maintenance suggests that the discrimination and active selection components embedded in MAP may enhance the stimulus control acquired during training in ways that passive vocal imitation does not.

Clinical Implications

The finding that MAP produced better maintenance without increasing errors or teaching time has several practical implications for BCBA program design. First, it suggests that the additional response alternatives in MAP do not function as distractors that slow acquisition — a common concern when considering multiple-choice formats for learners with limited discrimination repertoires. Instead, they appear to produce equivalent or faster acquisition with more durable outcomes.

For practitioners designing tact programs, this research supports considering MAP as a viable alternative to single vocal prompts, particularly for learners who show inconsistent maintenance of tacts taught under standard prompting procedures. The practical implementation of MAP involves presenting two or three alternative labels alongside the target stimulus and reinforcing selection of the correct one, with the alternatives selected to be meaningfully distinct from the target to support discrimination without excessive difficulty.

The prompt fading considerations for MAP differ somewhat from those for single vocal prompts. Because MAP involves a selection response rather than pure imitation, the fading procedure should systematically reduce the response-differentiation support provided by the alternatives — for example, by increasing the phonological or semantic similarity between the target and foils, or by increasing the delay before alternatives are presented, moving the learner toward independent labeling of the stimulus.

Generalization programming is a critical complement to tact acquisition data. Tacts acquired under specific training stimuli must generalize to novel exemplars of the same category — a learner who tacts "dog" with one picture must do so with dogs of different breeds, sizes, and contexts. MAP may support broader generalization by exposing learners to multiple related labels during training, strengthening the categorical relations underlying tact behavior. However, this hypothesis requires direct assessment through generalization probes with untrained exemplars.

Ethical Considerations

The ethics of tact training procedure selection center on the obligation to use evidence-based practices that maximize client outcomes while minimizing the cost of treatment in terms of time, effort, and potential for errorful responding. Code 2.01 requires practitioners to use evidence-based practices, and Code 6.01 requires recommending the least intrusive, most effective procedures. Leaf's research contributes directly to this evidence base by providing comparative data on MAP versus standard prompting.

A related ethical consideration is the documentation of maintenance outcomes in clinical programs. Many behavior analytic programs focus their data collection on acquisition — the percentage correct during teaching sessions — without systematic maintenance probing after skill mastery is declared. This practice can produce an inflated picture of treatment effectiveness and may lead to early termination of programs for skills that have not actually achieved durable stimulus control. Building scheduled maintenance probes into program design is both a quality assurance measure and an ethical practice aligned with the obligation to provide effective treatment.

For learners in the early stages of verbal behavior development, the selection of tact training procedures has compounding effects — skills built on a foundation of strong stimulus control generalize more readily, require fewer re-teaching episodes, and produce faster overall verbal behavior expansion. Choosing procedures based on evidence of their maintenance properties, as Leaf's research supports, is therefore not a minor procedural detail but a decision with long-term clinical significance.

Research participation considerations apply to the three participants in Leaf's study. Alternating-treatments designs expose participants to all conditions including less effective ones, which raises the question of whether continued access to the better-performing condition should be prioritized over the scientific value of the comparison. This is addressed in research ethics through the use of best-treatment follow-up phases and careful monitoring for participant welfare throughout the comparison phase.

Assessment & Decision-Making

Before selecting a tact training procedure, BCBAs should assess the learner's current tact repertoire, imitation skills, and discrimination capacity. Learners with robust vocal imitation and strong receptive discrimination are good candidates for standard vocal prompt procedures. Learners who show inconsistent imitation, high error rates under standard prompting, or poor maintenance of previously acquired tacts may be better served by MAP or other errorless formats that reduce the response demand while supporting active discrimination.

An adapted alternating-treatments design at the clinical level — comparing MAP and standard prompting for a small set of novel tact targets for a specific learner — can generate individualized data about which procedure is more effective for that client without requiring population-level research. This kind of clinical single-case comparison, while less rigorous than published research, provides functional information that can guide program design decisions.

Maintenance probing should be built into the tact program from the beginning. A reasonable protocol involves probing mastered tact targets at 1-week, 4-week, and 8-week intervals after mastery criteria are met, without prompting or supplemental reinforcement during the probe sessions. Targets that fail maintenance probes should re-enter the acquisition phase with modified procedures, which may include MAP if the initial acquisition procedure was standard vocal prompting.

Error analysis during tact training provides additional decision-making information. Consistent error patterns — such as perseverative responding with a high-frequency foil, or failure to attend to the stimulus in favor of selecting the same alternative repeatedly — indicate that the alternative set or the prompt fading procedure requires modification. MAP implementations where foils are too similar or too familiar may produce these error patterns and should be adjusted.

What This Means for Your Practice

If you supervise tact programs, maintenance data is your most important outcome measure — more so than acquisition-phase accuracy. A learner who achieves 90% accuracy during training but cannot tact the same items three weeks later has not learned the behavior in any functionally meaningful sense. Building maintenance probes into every verbal behavior program, reviewing that data regularly, and having a clear re-teaching protocol for maintenance failures are baseline standards of quality tact programming.

For practitioners who have relied exclusively on vocal verbal prompts for tact training, Leaf's research is an invitation to experiment with MAP in your own programs. The implementation is not technically demanding — it requires preparing alternative labels in advance and a clear decision rule for foil selection. The primary clinical question is whether MAP produces better maintenance for specific learners on your caseload, which you can evaluate with your own single-case data.

For supervisors, this CEU also raises broader questions about how tact programs are designed and monitored in your agency. Are prompt fading procedures specified and monitored for fidelity? Are maintenance probes scheduled and reviewed, or does mastery declaration effectively end ongoing data collection? Are there explicit decision criteria for what triggers a re-teaching episode versus a generalization probe? These programmatic questions are as important as the specific prompting procedure chosen.

Voicing these questions in team meetings and clinical supervision is part of building a culture of evidence-based practice. Leaf's work exemplifies the kind of within-subject comparative research that is most directly applicable to clinical decision-making — it generates answers about specific procedures for specific populations that can be applied to individual learners with manageable additional assessment effort.