Participant and Setting

Harrison was a White 4-year-old male diagnosed with level 1 ASD who communicated in approximately 3–6-word phrases. Harrison was 3 years old at the start of the study but turned 4 during phase 2; he participated in all three phases. At the time of the study, Harrison had received applied behavior analytic services for approximately 6 months, for 4 hours a week. At the onset of the study, Harrison scored 114 points on the Verbal Behavior Milestones Assessment and Placement Program (VB-MAPP; Sundberg, 2008), scoring primarily in the level 2 (18–30 months) range. Harrison scored 69 on the intraverbal subtest (Sundberg, 2008), which examines responding to various questions in sequence of difficulty, from fill-in-the-blank responses to answering multiple-part questions. Further, a language sample completed around the start of the study reported Harrison’s mean length of utterance at 4.25–4.77 words. The Preschool Language Scale-5 (Zimmerman et al., 2011), a norm-referenced assessment measuring speaker and listener skills for children between birth and 7 years 11 months, was completed 5 months before the start of the study. Harrison’s scores on the Preschool Language Scale-5 auditory comprehension, expressive communication, and combined standard scores were 89, 74, and 80, respectively (M = 100, SD = 15). According to the assessment scores, Harrison’s auditory comprehension score was within normal limits, while his expressive communication score was below average, indicating the need for intervention for speaker behavior.

All sessions occurred in a small therapy room measuring 2.1 m x 1.8 m at a midwestern university-based autism clinic. We conducted two sessions per week for 23 weeks (about 5.5 months) for all three phases for approximately 60–120 min each.

Materials and Stimuli

Each room contained a table, chairs, preferred items selected by Harrison each day (e.g., cars, blocks), instructional materials (e.g., Webber® cards), and session materials (e.g., clipboards, data sheets, timers, and pens). Stimulus cards were selected from the Webber® cards set and were 8.25 cm x 10.79 cm pictures printed on a white background. For phase 1, profession pictures were selected (e.g., a doctor) and divided into three sets (antecedent, consequence, and PT only). For phase 2, house item pictures were selected (e.g., broom) across two sets (antecedent and consequence). For phase 3, everyday items were selected (e.g., umbrella) and distributed across two sets (antecedent and consequence). Each set consisted of five targets related to the specific category.

Response Measurement

The dependent variables for all phases were the percentage of correct independent responses for ST, PT, and intraverbals, and the number of trials-to-criterion across the antecedent, consequence, and PT-only conditions. A correct independent response was defined as Harrison emitting the vocal response corresponding to the SD within 5 s. Trials-to-criterion were defined as the number of trials for Harrison to meet the mastery criteria for each phase. During phase 1, sessions consisted of 5-trial blocks with mastery criteria for the PT at three consecutive blocks at 80% or more (4/5) independent correct responses. Mastery criteria for the ST were one block during probes at 80% or more (4/5) independent correct responses. During phases 2 and 3, sessions consisted of 15-trial blocks with mastery criteria for the PT and ST at three consecutive blocks at 80% or more (12/15) independent correct responses. Mastery for the intraverbal responses was one block at 80% or more. Intraverbal responses were considered the terminal goal given that responding would occur in the absence of nonverbal stimuli (pictures). Thus, in phase 2, we used a nonparallel stopping criterion for the ST, meaning we ceased teaching in the antecedent condition before observing the ST mastery criterion, as mastery of the optimal response (i.e., intraverbals) was observed.

Interobserver Agreement (IOA) and Procedural Fidelity

A second, independent observer collected data, and trial-by-trial IOA was calculated across conditions for each phase. Trial-by-trial IOA was calculated by dividing the number of trials with agreements by the total number of trials multiplied by 100 for each condition. An agreement was defined as both observers recording the same response for the same trial, scored as a 1, and disagreements were scored as a 0.

IOA for phase 1 was collected for approximately 46.6% of trial blocks across all conditions. IOA was high for the antecedent (M = 94.6%, range: 86.7–100%), consequence (M = 91.7%, range: 86.7–100%), and the PT only (M = 100%, range: 93.3–100%) conditions. IOA was collected for approximately 44.4% of trial blocks for all conditions for phase 2. IOA was high for the antecedent condition for both ST probes (M = 98.2%, range: 97–100%) and intervention (M = 100%), and for the consequence condition for the ST probes (M = 100%), and intervention (M = 96.6%, range: 90–100%). IOA was collected for 100% of trial blocks across all conditions for phase 3. In the antecedent condition, IOA was high for probes (M = 98.5%, range: 93.3–100%) and intervention (M = 100%); and were high for the consequence condition probes (M = 100%) and intervention (M = 99%, range: 93.33–100%).

Procedural fidelity was assessed by a second independent observer during 55.5% of sessions in phase 2 and 45.4% of sessions in phase 3. During each intervention session, the observer recorded whether the experimenter (a) had materials prepared, (b) ensured client was orienting to the clinician, (c) used the correct color-correlated stimuli based on the condition in effect, (d) used various instructions for the PT, (e) used various instructions for the ST, (f) implemented the correct ST locus in the instruction based on the specific condition, (g) delivered verbal praise and specific reinforcement for correct responses, and (h) delivered the error correction procedure for incorrect responses. During the baseline and probe sessions, steps a-e were scored, but the experimenter was also scored as to the absence of feedback provided following Harrison’s responses. Correct implementations were scored as 1, and incorrect implementations were scored as 0. Procedural fidelity was calculated per session as a percentage by dividing the sum of correct implementations by the total number of events multiplied by 100. For phase 2, procedural fidelity for the antecedent condition was a mean of 91% (range: 69–100%), and a mean of 100% for the consequence condition. For phase 3, the mean procedural fidelity for the antecedent condition was 100% and a mean of 93% (range: 86–100%) for the consequence condition.

Experimental Design

For all phases, an adapted alternating treatments design (Sindelar et al., 1985) was used to compare the effects of antecedent ST, consequence ST, and PT only (phase 1 only) on the levels of correct ST and intraverbal responses. All intervention conditions were randomly alternated with no more than two consecutive presentations of the same condition. For each condition, targets were equally distributed across sets based on Cariveau et al. (2020) guidelines using a logical analysis to control the number of syllables and words for all targets (Table 1).

Table 1 Logical analysis for professions for phase 1 – primary and secondary targetsProcedure

At the beginning of the study, a list of items or people (PT) and related functions (ST) across all three categories was probed to identify unknown targets. Targets that were incorrect for both the PT and ST were selected for phases 1 and 2 and distributed across conditions using a logical analysis. For phase 3, Harrison scored correctly on two targets when testing, and these targets were counterbalanced across both conditions.

Before each session and periodically during breaks, the experimenter asked Harrison what he wanted to play with during work time. Harrison then selected two to four toys that were used as reinforcers and delivered during teaching trials on a fixed-ratio 5 schedule (i.e., each target was presented one time). As such, reinforcers varied daily depending on Harrison’s responses.

Baseline

Figure 1 (top panel) depicts the sequence of steps during baseline. The experimenter presented the PT picture and asked: (a) “Who is this?” or (b) “What do you see?” The experimenter provided no prompts or consequences regardless of the response. After a brief 5-s pause, the experimenter re-presented the picture and asked: (a) “What does it/they do?”, (b) “What does he/she do?”, or (c) “What do you use it for?” for the ST. All targets were interspersed, and the experimenter intermittently provided praise for working (e.g., “Great job sitting in your chair!”) and interspersed maintenance trials to provide Harrison with opportunities to contact reinforcement. No praise or feedback was provided to Harrison for responding to the PT and ST during baseline trials. Once stability in baseline responding occurred for each set, we introduced the teaching conditions.

Fig. 1The alternative text for this image may have been generated using AI.

Flow chart depicting the three intervention procedures. Note: + correct response, - error, P prompt; Primary target (PT) depicted with the image card and finger pointing to it; Secondary target (ST) depicted in brown

General Tact Teaching (PT)

The experimenter presented the PT as described in baseline. If Harrison responded correctly within 5 s, the experimenter provided behavior-specific praise (e.g., “Good job! This is a plumber”). If Harrison answered incorrectly or did not respond within 5 s, the experimenter implemented error correction by pointing to the picture and modeling the correct vocal response (e.g., “This is a plumber”). The experimenter then removed the picture for 2 s and repeated the trial until a correct response occurred or up to three times. Trials were presented in pseudorandom order.

Antecedent Condition

Figure 1 (middle panel) depicts the sequence of steps during the antecedent condition. For every trial, the experimenter first presented a picture and delivered the ST (e.g., “They fix pipes”). No consequences were delivered for echoing or responding to the ST. After delivering the ST, the experimenter delivered one of the instructions for the PT and conducted tact teaching as described above.

Consequence Condition

Figure 1 (bottom panel) depicts the sequence of steps during the consequence condition. The trial procedures were the same as in the antecedent condition, except the ST was delivered after verbal praise (e.g., “Good job! It is a plumber! They fix pipes”). Consistent with past research (e.g., Delmolino et al., 2013), ST were delivered only for independent correct PT responses. In other words, if Harrison erred on or required a prompt for the PT, the ST was not provided at that time. Instead, error correction or prompting was implemented as described above. Once Harrison responded correctly to the PT (within the three repetitions), the experimenter delivered the ST following the praise (e.g., “You’re right! It is a plumber. They fix pipes”). If no independent, correct attempts were emitted within three trials, the experimenter moved on to the next trial.

Target Probes

Once Harrison reached mastery criteria for a PT set, the experimenter probed the corresponding ST. ST probes were conducted as described during baseline. If ST mastery was not initially demonstrated, the experimenter conducted three more trial blocks of each teaching conditionFootnote 1, repeated the ST probes, and continued this sequence until mastery of the ST was obtainedFootnote 2.

Intraverbal Probes

Starting in phase 2, once mastery of the ST occurred, we conducted intraverbal probes based on the ST and PT (i.e., in the absence of the picture cards). The experimenter randomized the trials within blocks and asked questions such as, “What do you use a [PT] for?”, or “What do you use to [ST]?” Correct responses included the ST for PT intraverbals (e.g., “What do you use a [sink] for?” “to wash hands”) or the PT for ST intraverbals (e.g., “What do you use to [wash hands]?” “sink”). No reinforcement or prompts were programmed. In phase 3 only, intraverbal probes were conducted during baseline to compare performance to post-teaching.

Best Teaching Condition

The type of trial configuration in which Harrison reached ST mastery first was used to teach the ST for the control set in phase 1. In other words, the intervention condition that Harrison met ST mastery criteria before the other intervention condition was considered the “best intervention” and replicated to teach the control condition target set.

Maintenance Probes

All phases incorporated maintenance probes at the end of teaching. In phase 1, maintenance probes were conducted 1 month after mastery of PT and ST were observed. In phase 2, we completed maintenance probes 1 week and then 1 month after mastery of antecedent and consequence intraverbals were observed. Finally, in phase 3, we completed maintenance 1 week and 1 month after mastery of either the consequence ST or the antecedent intraverbal was observed. The trial arrangement and procedures for maintenance probes were identical to the baseline condition.

Phase Specifications

Some minor procedural differences were present across phases as illustrated in Table 2 and described below.

Table 2 Phase differences across the three phases Phase 1

During phase 1, we incorporated the PT-only condition to serve as a control for the conditions that included ST. Sessions were also distinct from the other phases based on the number of trials in a block (5) and the mastery criteria for the ST (one block at 80% or more [4/5] independent correct responses).

Phase 2

During phase 2, we incorporated intraverbal probes after mastery of the PT and ST were observed to assess for the transfer-of-stimulus control from PT and ST tacts to intraverbals. Sessions consisted of 15-trial blocks with mastery criteria set at three consecutive blocks at 80% or more (12/15) independent correct responses for both PT and ST. Mastery for the intraverbal response was one block at 80% or higher.

Phase 3

During phase 3, we added intraverbal probes during baseline so that we could compare responding prior to intervention and post-intervention. We also added color-correlated stimuli to each ST condition (i.e., blue and yellow cards) to help enhance Harrison’s discrimination across conditions. Sessions and mastery criteria were the same as in phase 2 for PT, ST, and intraverbal probes.

Concurrent-Chains Procedure

At the end of phase 3, we added a concurrent-chains procedure to assess Harrison’s preference for the delivery of the ST. Targets from both conditions were quasi-randomly mixed and counterbalanced in three, 10-trial blocks to control for potential confounds related to stimulus/item biases and decrease the probability of order effects. The trial started by presenting the yellow and blue cards and asking, “Do you want to do the picture with the yellow or the blue square?” Harrison could point to or vocalize his response. Harrison then experienced a typical tact-teaching trial with the ST delivery in the antecedent or consequence portion of the trial based on his color selection. For example, if Harrison selected yellow, the experimenter would conduct one trial placing the ST in the antecedent portion of the trial; if Harrison selected blue, the experimenter would deliver the ST in the consequence portion of the trial. We then compared the differential selection of one condition over the other, suggesting a possible preference for that ST configuration (Auten et al., 2024).