Participants

Nine boys (P1–P9), aged 3.2–6.2 years (M = 4.8 years), diagnosed with autism spectrum disorder (ASD), language delays, or intellectual disability participated in the present experiment. Seven participants with ASD (P1, P2, P3, P5, P6, P7, and P8) received 18–25 h of early intensive behavioral intervention per week. The two remaining participants with language delays (P4) and intellectual disability (P9) received individual educational sessions approximately 10 h weekly––consisting of language training involving increasing their vocabulary and language comprehension (i.e., listener responses). All participants were monolingual (Norwegian). However, four of them were exposed to a second language at home (i.e., P4 Iranian, P5 Somali, P6 Syrian, and P7 Somali).

Before the experiment was initiated, the caregivers provided informed consent and the study was approved by the Norwegian Agency for Shared Services in Education and Research. Inclusion criteria for the present study included prerequisites for Inc-BiN: (1) spontaneous echoing of words, (2) visual-visual MTS responses, (3) at least 20 different tacts, (4) 20 manded tacts, and (5) 20 listener responses––to everyday objects, actions, body parts, animals, functions, peers, and relatives (Catania, 2013; Horne & Lowe, 1996; Miguel, 2016; Sivaraman & Barnes-Holmes, 2023; Yoon et al., 2023). The prerequisites were assessed using the Assessment of Basic Language and Learning Skills-Revised (ABLLS-R; Partington, 2006; Appendix A). The results of the ABLLS-R show that the participants’ scores ranged from 85 to 100% mastery in the domain cooperation and reinforcement effectiveness and ranged from 49 to 100% mastery on visual performance, indicating that all participants matched identical visual stimuli. In the domain of receptive language (i.e., listener responses) and labeling (i.e., tacts), the participants obtained scores ranging from 63 to 98% and from 44 to 97% mastery, respectively, which means that they were able to tact and respond as a listener far above the inclusion criterion for the present experiment. Further, the participants’ scores on vocal imitation (i.e., echoic) ranged from 52 to 100% mastery, indicating that all participants echoed words spontaneously. Finally, all participants were able to emit short sentences and scores ranged from 27 to 82% mastery, and they were able to request preferred items and avoid non-preferred stimuli (i.e., mands) within the range of 53% and 96% mastery. Detailed scores on ABLLS-R and characteristics of each participant are provided in Appendix A.

Setting and Materials

The experiment was conducted in the participants’ teaching rooms in their daycare centers. The teaching rooms differed across participants as they attended different daycare centers. However, P5, P7 and P9 attended the same daycare center but were taught in different teaching rooms. Similarly, P6 and P8 attended the same daycare center and shared the same teaching room. All teaching rooms were approximately 3 × 2 m and were equipped with a shelf with books, bins with materials for each participant’s teaching program, preferred toys, and bins for toys used during breaks. The experimenter (the first author) also brought a binder with materials used during the experiment, a medium-sized suitcase, and a bag with preferred toys, including a tablet with games which served as positive consequences during instructional conditions (see preference assessment below). The teaching rooms included an adult-size table with one chair at each side of the table. The participants used adjustable chairs (e.g., Stokke trip-trap chairs), which were adjusted for a good working position for each child at the table. A third chair was placed in a corner for the observer (a staff member employed at the daycare centers). Only the participant, the experimenter, and a staff member at the daycare center were present in the teaching room. In addition, we used a video camera to record trial blocks for reliability and treatment integrity assessments.

The primary materials were laminated photos (7 × 5 cm) of unfamiliar stimuli for each participant. The photos were individualized for each participant. All stimuli were selected from categories of everyday objects (e.g., vegetables, sport-related items, and flags). The sequence of novel stimulus sets was randomly assigned across Inc-BiN probe blocks and MOI blocks (c.f. Cariveau et al., 2021, 2022) by blindly drawing envelopes with the images of the relevant sets in an effort to reduce experimenter bias (e.g., Ledford et al., 2019).

Each set consisted of five novel stimuli. Up to eight sets were identified for each participant (see Pre-experimental Procedures), for a total of 40 novel stimuli. All stimuli identified within each set for each participant are displayed in Table 1. Sets 1–3 were baseline Inc-BiN probe sets and were used in baseline and post-MOI, when applicable. Sets 4–6 were MOI sets, Set 7 was designated for generative Inc-BiN probes, and Set 8 was used to conduct follow-up Inc-BiN probes to evaluate the maintenance of Inc-BiN skills.

Table 1 Overview of stimuli used in the experiment

During data collection, we used prepared pencil-and-paper scoring sheets for each condition. A video camera was used in some sessions, when an observerer from the daycare center was not able during the sessions. The observers from the daycare centers were previously trianed to collect trial-by-trial data. After supervision by the first author, the observers from the daycare centers assessed relability and procedural fidelity.

Dependent Measures

The primary dependent measure was the number of correct responses during all types of Inc-BiN probes, which included the total number of correct listener responses, tacts, and manded tacts. During each probe block we assessed responding to one stimulus set of five different pictures, which were presented four times each. All probe blocks consisted of one 20-trial block of the naming experience, followed by a 20-trial block of listener probes, a 20-trial block of tact probes, and a 20-trial block of manded tact probes. One block of baseline Inc-BiN probes––across three stimulus sets––consisted of 180 trials in which no programmed consequences followed. That is, each probe set included 60 trials (i.e., 20 tacts, 20 manded tacts, and 20 listener responses; each of the five stimuli in a set was probed four times each for tacts, manded tacts, and listener responses) in the absence of programmed differential consequences. The Inc-BiN probes are described in greater detail below.

A secondary dependent measure was the number of correct responses during MOI, in which four response types were rotated within each 20-trial block. Two different listener responses (selection responses during MTS and point to responses) and two types of speaker responses (tacts and manded tacts) were rotated from trial to trial. MOI trials are described in greater detail below.

Data Collection

Frequency data were collected across all experimental conditions by hand, using prepared data-collection sheets. All trial blocks consisted of 20 trials and were recorded on a trial-by-trial basis. To be scored as correct, the response had to be emitted within 6 s of the onset of the discriminative stimulus. If no response occurred within 6 s after the instruction was presented or an incorrect response was emitted, the trial was scored as incorrect. During the naming experience (matching-to-sample tasks) and MOI blocks, a correct trial required that the participant responded correctly without any prompts.

Experimental Design

The design was selected to enable a thorough analysis of the extent to which repeated MTS-naming experiences followed by Inc-BiN probes alone strengthened Inc-BiN skills, or alternatively, whether MOI was a critical independent variable. In the present experiment, we used a multiple probe design across three stimulus sets for each participant to answer the research questions. The design was set up to randomly assign nine participants to three different conditions––two brief-baseline conditions, Condition 1 and Condition 2, and one extended-baseline condition, Condition 3. Participants assigned to Condition 1 were designated P1, P2, and P3, while those assigned to Condition 2 were designated P4, P5, and P6, and those assigned to Condition 3 were P7, P8, and P9 (see Fig. 1).

Fig. 1

Overview of the experimental sequence. Note. P stands for participant and the numbers indicate an individual participant. P1–P3 were assigned to Condition 1, P4–P6 were assigned to Condition 2, and P7–P9 were assigned to Condition 3. * Indicate exceptions: P3 in Condition 1 received five baseline Inc-BiN probes and in Condition 2 P4 received three baseline Inc-BiN probes

Condition 1, a brief-baseline condition, consisted of probing speaker responses (tacts and manded tacts) before listener responses. Condition 2, a brief-baseline condition, consisted of probing listener before speaker responses. In Condition 3, an extended-baseline condition, listener responses were probed first.

After completion of baseline Inc-BiN probes, MOI was introduced for the participants who were randomly assigned to the brief-baseline conditions. The two brief-baseline conditions (Conditions 1 and 2) consisted of 2–3 baseline Inc-BiN probes, until a stable baseline was achieved before MOI was initiated. Baseline was considered stable if scores did not differ more than four responses within each set across baseline Inc-BiN probes. When responding met stability during baseline Inc-BiN probes for participants assigned to Conditions 1 and 2, MOI was initiated. If the stability criterion was not met, the participant would be exposed to additional baseline Inc-BiN probes, up to a maximum of five baseline Inc-BiN probes (as for P3), across Sets 1–3.

For the participants who were assigned to the extended-baseline condition (Condition 3), exposure to baseline Inc-BiN probes continued up to a maximum of five probes. For ethical reasons, the baseline was not extended without programmed consequences beyond the five baseline Inc-BiN probes. In Condition 3, if Inc-BiN skills did not meet the emergence criterion within the programmed number of baseline Inc-BiN probes, MOI was introduced (as for P9). Regardless of the length of the baseline, a maximum of three blocks of MOI (Sets 4–6) were trained.

Pre-experimental ProceduresPreference Assessment

During the matching-to-sample tasks in the naming experience and MOI, preferred toys, edibles, or tablet games were used. Preferred stimuli were assessed using a multiple stimulus assessment without replacement procedure (DeLeon & Iwata, 1996). The preferred stimuli were delivered contingently on correct responses along with typical social consequences, such as enthusiastic praise, high fives, tickles, and smiles.

Identification of Novel Stimuli

Tacts, manded tacts, and listener responses were assessed, in that order, for each pictorial stimulus. During identification of novel stimuli, MTS responses were not tested (see inclusion criteria above). Tacts were assessed by holding up the unknown stimulus in front of the participant and scored as novel if no responses or incorrect responses occurred within 6 s after a stimulus was presented. Manded tacts were tested in a similar manner, except that the vocal instruction “What is this?” was presented approximately 2 s before presenting the test stimulus. Listener trials were conducted by presenting five pictorial stimuli on the table in front of the participant and the participant was asked to point to one of them (e.g., “Point to the opal”). The response was considered correct if the participants pointed to the picture that corresponded with the instruction (e.g., the picture of the opal) within 6 s after the instruction was presented.

Each operant was tested three times with each of the five stimuli within a set. Thus, testing a stimulus set for novelty consisted of a total of 15 trials across the three operants. The stimulus was considered as novel if no correct tacts and manded tacts were emitted within 6 s following stimulus presentation, and a maximum of one correct listener response in the presence of the same stimulus. With five stimuli available for point-to responses, random listener responses could be emitted correctly by chance. Throughout the listener trials, all responses to a stimulus included in the probe sets had to be below chance (20%). If a correct listener response occurred more than once to a stimulus in a set during the identification of novel stimuli, it was used during MOI blocks. If more than one listener response or a correct tact response occurred to a stimulus, that particular stimulus was replaced by another novel stimulus.

Novel stimuli were identified under extinction meaning that no differential reinforcement was delivered contingent on correct responses. During identification of novel stimuli, as well as during all Inc-BiN probes, maintenance tasks were interspersed after every third to fifth trial to avoid fatigue and reduced responding to trials. The maintenance tasks were targets that the participants had previously mastered, and preferred toys along with praise were used as consequences. Examples of maintenance targets included pointing-to and tacting relatives, animals, body parts, tacting functions of everyday commonplace objects, and answering personal knowledge questions (e.g., “How old are you?”).

Experimental Procedures

An overview of the experimental conditions is displayed in Fig. 1.

Baseline Inc-BiN Probes

The purpose of baseline Inc-BiN probes (Sets 1–3 stimuli) was to evaluate the effect of repeated probing and MOI. The Inc-BiN probes consisted of two subphases: (1) Naming experience, and (2) Inc-BiN probes––testing the emergence of untaught listener responses, tacts, and manded tacts (each subphase is described in detail below). In the present experiment, all the Inc-BiN probes were procedurally identical in that the naming experience was implemented first and then the Inc-BiN probes were conducted. This sequence of presenting the naming experience followed by Inc-BiN probes occurred for all probe types (i.e., baseline Inc-BiN probes, post-MOI Inc-BiN probes, generative Inc-BiN probes, and follow-up Inc-BiN probes).

Naming Experience 

During simultaneous (auditory + visual)-visual MTS, the experimenter placed the five comparison pictorial stimuli on the table in front of the participants. Then, the experimenter presented the participant with the sample picture card facing the participant, and presented the tact (e.g., “Okapi”). Next, the experimenter handed the sample picture card to the participant. The correct response was defined as placing the sample picture card on the positive comparison stimulus (i.e., a selection response). During the naming experience, the participants were exposed to each of the five novel stimuli and corresponding tacts four times within each set, which constituted one 20-trial block for each baseline set. Thus, the naming experience across Sets 1–3 consisted of at least 60 trials per set which were interspersed with play breaks and Inc-BiN probes (as described below). Correct selection responses during MTS trials were differentially reinforced with praise and items identified through the preference assessment.

If an incorrect selection response occurred during the naming experience (MTS), a correction procedure was implemented by re-presenting the trial and then prompting the correct selection response by pointing to the corresponding comparison. During the following trials, the prompt was faded using a 0–6 s progressive delay prompting procedure until independent responding occurred. If an incorrect response occurred, a 0-s prompt delay was presented on the next trial. Then the same trial was repeated with a 6-s time-delay. This sequence was repeated until a correct response occurred without a prompt within 6 s. During the naming experience, the mastery criterion was 18 of 20 (90%) correct MTS responses for two consecutive 20-trial blocks, or 20 of 20 (100%) accuracy for one 20-trial block. Between the naming experience and the Inc-BiN probes, the participants were given a play break for 10–15 min.

Inc-BiN Probes 

During Inc-BiN probes, the tacts, manded tacts, and listener responses were probed separately under extinction in 20 trial-blocks. Tacts and manded tacts were probed by presenting a novel stimulus on the table in front of the participant. In the case of manded tact trials, the instruction “What is this?” was presented 2 s after a picture was presented. A correct tact or manded tact was scored if the correct tact occurred within 6 s after the discriminative stimulus was presented. Listener trials were conducted by placing five pictorial stimuli on the table in front of the participant, who was then asked to point to one of them (e.g., “Point to the opal”). A correct response was scored if the participant pointed to the picture that corresponded with the instruction given by the experimenter.

The emergence of listener responses, tacts, and manded tacts were scored the same way as during identification of novel stimuli. The emergence criterion for the acquisition of Inc-BiN was a minimum of 14 out of 20 correct responses (70%; c.f. Morgan et al., 2021) for listener responses, tacts, and manded tacts.

Test Sequences

P1–P3 (Condition 1 participants) were exposed to tact probes first, then manded tacts and, finally, listener responses. The remaining six participants, P4–P9 (Conditions 2 and 3), were exposed to the test sequence probing listener responses before speaker responses (i.e., testing point-to responses, then tacts, and finally manded tacts).

Generative Inc-BiN Probe

The purpose of a generative Inc-BiN probe was to measure the extent to which Inc-BiN skills occurred using a novel stimulus set. All participants were exposed to a generative probe. The generative Inc-BiN probe was administered either after post-MOI Inc-BiN probes (as for P1–P6 and P9) or after the fifth baseline Inc-BiN probe (P7 and P8). The probing procedure was procedurally identical to the baseline Inc-BiN probes.

Follow-Up Inc-BiN Probe

For eight of the nine participants, a follow-up probe with another set of novel stimuli was conducted one month after the generative Inc-BiN probe to assess for maintenance of responding. For P2, the experiment ended before a follow-up probe was conducted because he left the daycare center to enter primary school. Follow-up Inc-BiN probes were procedurally similar to the baseline Inc-BiN probes.

Mixed-Operant Instruction (MOI)

During the 20-trial MOI blocks, the operants were trained until mastery in the following order: echoics and selection responses during MTS trials, tacts, manded tacts, and listener responses. Rather than training the four operants separately, the operants were intermixed across training trials within a trial block. Additionally, the stimuli within a set were rotated. For example, in the presence of the first stimulus in the set (e.g., an image of a sapphire), a selection response (i.e., listener response) and an echoic were trained during auditory/visual-visual MTS. Next, a tact was taught in the presence of the second stimulus (e.g., an image of a grenade), a manded tact in the presence of the third stimulus (e.g., an image of a jade), and finally, a listener response in the presence of the fourth stimulus (e.g., an image of an opal). On the fifth stimulus in the set (e.g., emerald), the sequence was reset, beginning with an MTS task with training echoic and selection response. In this way, the four operants were rotated across the five stimuli.

During MOI, the five stimuli within each training set was presented four times per operant, which constituted 20 trials per trial block. Thus, each operant was trained in four trials in each trial block. The mastery criterion was 18 of 20 (90%) accurate responses for two consecutive trial blocks, or 20 out of 20 (100%) correct responses in one trial block.

If Inc-BiN did not emerge during the first block of post-MOI Inc-BiN probes following the first MOI-set, a second set of MOI was introduced. Further, if Inc-BiN skills did not emerge during the second block of post-MOI Inc-BiN probes, following two sets of MOI, training with the third MOI-set was initiated.

Correction procedure 

First, echoics and selection responses were trained during (auditory + visual)-visual MTS tasks. In addition to a correct selection response, the participant was required to echo the researcher’s tact of the sample stimulus (e.g., saying “sapphire”) within 6 s of the vocal antecedent. If an incorrect selection response occurred during MTS, the prompt-fading procedure used during the naming experience was implemented. However, if the participant correctly selected the comparison stimulus without an echoic, we scored the trial as incorrect, and the researcher prompted the correct echoic in the next trial by saying the name of the relevant stimulus (e.g., said “sapphire”) and by withholding the visual sample stimulus until the participant emitted the correct echoic (e.g., saying “sapphire”). A correct echoic resulted in the presentation of the visual sample stimulus. Initially, the researcher used a 0-s prompt delay procedure. Next, the vocal echoic prompt and withholding of the sample stimulus were faded according to a progressive delay prompting procedure. The prompts were faded by increasing the delay from 0–6 s. However, if an echoic response did not occur after the researcher had withheld the stimulus for 6 s, the prompting sequence was re-implemented starting with a 0-s prompt delay. On the next trial, the experimenter modeled the correct response (i.e., the experimenter said the name of the sample stimulus and placed it on top of the identical comparison) and then said, “Your turn.” If the participant failed to emit the correct response within 6 s, the prompted trial was re-presented with an instruction (e.g., “Repeat after me, sapphire”) and the sample stimulus was withheld. A correct echoic (e.g., saying “sapphire”) within 6 s resulted in the presentation of the sample stimulus. Echoic prompts were faded by first fading the instruction (e.g., “Repeat after me…”) to merely uttering the initial syllables of the relevant words (e.g., “sa…”), followed by only the first sound (e.g., “s…”) and finally, all vocal prompts were faded.

Second, during tact trials, the experimenter held up the training stimulus in front of the participants. The correct response was to tact the stimulus within 6 s (e.g., saying “grenade”) of the presentation of the relevant stimulus. Third, during a manded tact trial (e.g., saying “jade”), a vocal antecedent was added (“What is this?”). During both types of tact trials, the correct response was saying the name of the stimulus presented. Novel tacts and manded tacts were initiated by providing an echoic prompt immediately following the presentation of the stimulus (i.e., a 0-s delay prompt). The echoic prompting procedure used to establish tacts was procedurally identical to the procedures described above to teach echoics during (auditory + visual)-visual MTS trials.

Finally, the pointing selection responses were taught during auditory-visual MTS. Stimuli were presented and instructions were provided in the same way as during the identification of novel stimuli, pre-experimental procedures, and during listener probes. Pointing or position prompts were used to teach listener responding. If an incorrect pointing response was emitted, the correct response was prompted on the following trials using a progressive prompt delay procedure. That is, correct pointing responses to the training stimuli were immediately prompted (0-s delay). Then, the researcher repeated the same trial by progressively expanding the delay between the antecedent and the prompt until the correct response occurred within 6 s. Across trained operants (selection response with echoic, tact, manded tact, and pointing responses), if three consecutive prompted trials did not produce a correct response, the researcher moved on to the next trial.

Post-MOI Inc-BiN Probes

After the mastery of an MOI-set, post-MOI Inc-BiN probes were conducted. These probes were procedurally identical to the baseline Inc-BiN probes, described above. Only participants who had received MOI were exposed to post-MOI Inc-BiN probes. These included all participants in Conditions 1 and 2. In Condition 3, P9 also received MOI, because five baseline Inc-BiN probes were not sufficient for speaker and listener responses to meet the mastery criterion. The same stimulus sets used during baseline Inc-BiN probes were used during post-MOI Inc-BiN probes.

Interobserver Agreement

Trial-by-trial interobserver agreement (IOA) was scored from videotapes and in vivo by an independent observer during 55.9% (range 22–100%) of trial blocks selected randomly. In the present study, each procedural phase was assessed for IOA (i.e., identification of novel stimuli, the naming experience, Inc-BiN probes, and MOI trial blocks). All Inc-BiN probes (i.e., baseline Inc-BiN probes, post-MOI Inc-BiN probes, generative Inc-BiN probes, and follow-up Inc-BiN probes) and all MOI blocks were calculated together. The IOA was calculated as the number of trials with agreement divided by the total number of trials scored, multiplied by 100. The IOA criterion was at least 90% agreement. If the IOA scores fell below 90%, training and calibration between the two observers would be implemented. During identification of novel stimuli and the naming experience, the mean IOA was 100% agreement. During Inc-BiN probes the mean IOA was 99.7% agreement (range 98–100%). Finally, across MOI trial blocks IOA was 98.2% (range 95–100%). Across participants and conditions, the total agreement was 99.5% (range 96.6–100%).

Procedural Fidelity

Procedural fidelity (PF) was collected for at least 30% of all trial blocks across all conditions and participants and was calculated by dividing the number of correctly implemented training components by the total number of training components and multiplying by 100. During identification of the novel stimuli phase, during tact trials, we assessed whether the stimuli were presented without any vocal supplements and that there was only one correct listener response emitted per stimulus. Additionally, during this phase, we also assessed whether the experimenter delivered reinforcement or implemented the prompting procedure. The PF during identification of novel stimuli was 100%. During the naming experience, we measured the extent to which the experimenter required only selection responses during MTS tasks and PF was 100%. During Inc-BiN probes we assessed whether the experimenter arranged an extinction condition with interspersed maintenance trials, and whether they implemented the correction procedure or provided consequences. Additionally, during Inc-BiN probes, we checked whether the experimenter used the programmed probe sequence. PF across Inc-BiN probes was 100%. During MOI, we assessed PF on intermixing of trial types within the trial blocks, that the correction procedure was used accurately for each operant, and that the relevant stimulus was presented during tact trials. For all seven participants who received MOI, PF across trial blocks was 100%. Finally, correct continuation criterion was examined for MTS (i.e., naming experience) and MOI trial blocks and that the criterion was held constant throughout the experiment.

Although the standard procedure during Inc-BiN probes included the interspersal of maintenance tasks on every third to fifth trial, the procedure was adjusted for P8 and P9 to include the interspersal of maintenance tasks on every second trial to increase motivation and responding to the experimental tasks. In addition, because of consistent incorrect responses to two specific stimuli (e.g., pencil case and ruler) by P4, P6, and P9, those two stimuli were trained in isolatation in one or two 20-trial blocks before being incorporated back into the MOI-trial blocks.