An Evaluation of a Game-Based Treatment Package on Intraverbals in Young Children with Autism

Abstract

The purpose of the current study was to evaluate a game-based treatment package on the acquisition of intraverbals in young children with autism. The treatment package was composed of using a listener response training game (i.e., bingo), providing verbal praise that contained the label for the listener response, and modeling a pretend play action related to the answer. During posttreatment probes, participants vocally answered wh- questions without any supplementary stimuli present and maintained responses during follow-up probes.

Discrete-trial training (DTT) is often the treatment of choice within early intensive behavioral intervention (EIBI) for children with autism spectrum disorder (ASD). Numerous studies have demonstrated that DTT is an effective procedure for teaching language skills to individuals with ASD (Sautter & LeBlanc, 2006). Practitioners and researchers have identified the limitations of DTT when teaching young children with ASD. One of these limitations includes the use of artificial or functionally unrelated reinforcers rather than natural reinforcers. Also, the highly structured arrangement can evoke problematic behaviors (e.g., stereotypy, noncompliance, eloping) that interfere with instruction (Carbone, Morgenstern, Zecchin-Tirri, & Kolberg, 2010). For many children with ASD, it is likely that the DTT arrangement resembles situations in which escape and avoidant behaviors have been reinforced in the past.

In response to some of the limitations of DTT, researchers developed naturalistic behavioral treatments, including incidental teaching (McGee, Morrier, & Daly, 1999) and game-based instruction (Foxx, McMorrow, & Schloss, 1983). These naturalistic behavioral interventions embed discrete trials within games or routines, use functionally related reinforcers, and have been shown to increase generalization (Delprato, 2001). Foxx et al. (1983) used a behavioral treatment package that incorporated a modified board game to teach individuals with developmental disabilities to engage in several social skills. There are a number of practical advantages to using a game format to teach individuals with ASD. First, several skills can be targeted while playing games, including mands, turn taking, joint attention, motor imitation, and social skills. Additionally, embedding a variety of learning opportunities and teaching in a game format may increase the social validity of applied behavior analysis (ABA) procedures among stakeholders (e.g., general education teachers, family members). Last, using commercially available broad games reduces the time needed to develop teaching materials for practitioners.

Another problem practitioners often encounter when implementing an echoic-to-intraverbal training procedure is that children with ASD do not echo the verbal model. Behavioral researchers have developed and evaluated emergence procedures. Emergence procedures include directly training a response as one verbal operant (e.g., tact) and subsequently assessing if the individual can make a similar response as a different verbal operant (e.g., intraverbal). If the individual can do so, the second response is described as an “emergent” response because it was not directly trained. Several studies have evaluated the extent to which intraverbals emerge after tact or listener training. In these procedures, participants are trained to select a nonverbal stimulus (e.g., picture card of an apple) when presented with an auditory discriminative stimulus (S^D; e.g., “What do you eat?”). After participants have acquired the listener response, intraverbal probes are completed in which the researcher presents the auditory stimulus without picture cards to assess if the participant can answer the question.

Smith et al. (2016) examined the extent to which intraverbals emerged following listener training in children with ASD. Specifically, the target intraverbal was to answer six wh- questions, such as answering “tomato” when asked, “What do you eat that is red?” Listener training consisted of teaching participants to select a nonverbal stimulus (e.g., picture card) when presented with the instruction. After participants had demonstrated mastery of the selection response, probes were conducted in which no teaching materials (e.g., picture cards) were present. Four of the five participants demonstrated emergent intraverbals, whereas one participant required additional intervention that included tact training during the listener training.

It seems tenable that teaching young children with ASD a listener response while playing a game would lead to emergent intraverbals. Therefore, the current study evaluated a game-based treatment package on the emergence of intraverbals in young children with ASD. The treatment package was developed in response to both participants engaging in high rates of problematic behaviors and their failure to make adequate progress during direct intraverbal training. The treatment package included three components: (a) teaching participants to play a commercially available wh- bingo game using graduate guidance, (b) providing verbal praise that included the auditory S^D and target response, and (c) modeling a pretend play action related to the target response. The verbal praise and pretend play action were designed to be naturally and functionally related reinforcers.

Method

Setting, Participants, and Materials

Sessions took place at two EIBI centers housed in a general education preschool. Eight children with ASD were enrolled in each center and received 30 hr of ABA therapy each week. The daily schedule was similar to an early childhood classroom and included large-group and one-on-one instruction, snacks and lunch, and recess. Each child had an individualized program that included multiple language, social, play, cognitive, imitation, and functional routine targets. Each participant spent approximately two hours each day in a general education classroom with one-on-one support. Experimental sessions were conducted one on one within the therapy room at various times throughout the day; one to two sessions were implemented daily and lasted approximately five minutes. Experimental sessions were carried out by ABA master’s students who had worked at the EIBI centers for over two years and had passed procedural integrity checks on various evidence-based practices, including DTT, naturalistic behavioral interventions (NBI), and various prompting procedures, including graduated guidance (which was used in the current study). Each researcher was trained in the current procedures by the first author, who is a Board Certified Behavior Analyst at the Doctoral level (BCBA-D) and has over 10 years of experience implementing DTT and NBI. The first author implemented the first few intervention sessions with each participant and recorded procedural integrity for each research team member to ensure a high fidelity of treatment.

Two children, Jed and Esau, participated in the study and were both 4 years old. Participants were included in this study because each had demonstrated high rates of problematic behavior and had stopped making progress during direct intraverbal training that used an echoic and textual prompt in a DTT arrangement. Assessment results prior to the start of the study showed that both participants could tact over 25 items, had generalized echoic and imitative repertoires, and could answer a few intraverbal questions that had been taught.

The Ask and Answer “Wh” Bingo by Super Duper Publications (Greenville, South Carolina) was used. The game included bingo chips, question cards, and several bingo boards. There are six different types of boards that correspond to wh- questions. Each board consists of 25 squares. Each square has a cartoon picture and printed text of an answer that corresponds to a wh- question (e.g., “What do you wear on your head?”).

Dependent Variable

The dependent variable was the number of correct, independent vocal-verbal intraverbals per nine-trial session. A correct response was defined as a vocal-verbal response that answered the question and occurred within 6 s of the question. For example, if the participant answered “hat” within 6 s of the researcher asking, “What do you wear on your head?” the response was scored as correct. A response was scored as incorrect if the participant made a vocal-verbal response that did not relate to the question (e.g., said “bunny” to the previous question) or said nothing.

Selection of Targets

Participants were taught to answer three different sets of wh- questions (i.e., what, where, and when). Each set was composed of three questions from the same category of questions. The first author carried out several systematic steps to make the questions and responses age appropriate and similarly difficult within and across sets. Each question was determined to be age appropriate, was one syllable (what questions) or between two and five syllables (when and where questions), and was randomly selected from a greater list of questions. A detailed explanation of these steps is available upon request.

Experimental Design

A multiple-probe design with both nonconcurrent (across participants) and concurrent (across stimuli sets) baselines was used to evaluate the effects of the treatment package on the emergent intraverbals. This design was carried out by conducting a series of consecutive intraverbal probes for each set of stimuli. Treatment began with Set A, whereas Sets B and C remained in baseline. Once participants demonstrated mastery-level responding during the bingo game (i.e., listener response), intraverbal probes were conducted without the bingo board present (same as baseline). Probes for previously mastered sets were also conducted to assess the maintenance of emergent responses (and were part of the center’s general practice). Follow-up maintenance probes were conducted 1 month (Esau) and 2 months (Jed) after Set C probes were conducted.

Independent Variable

The treatment package consisted of three components: (a) teaching participants to play wh- bingo using graduate guidance, (b) providing verbal praise that contained the question and response, and (c) modeling a pretend play action related to the target response.

Procedure

During all sessions, the participant and researcher sat next to or across from each other at a child-sized table. Intraverbal probes and treatment sessions were composed of nine trials (each question was presented three times).

Baseline

Baseline sessions were conducted to determine if participants could vocally answer wh- questions. During previous intraverbal direct training, participants engaged in high rates of problematic behavior (e.g., falling out of their chair, screaming, vocal stereotypy), and it had become difficult to conduct these trainings. These same problematic behaviors were observed during the current study’s baseline sessions, so mastered imitation trials were interspersed in order to provide reinforcement and complete sessions. Each trial consisted of the researcher obtaining attending, subsequently asking one of the wh- questions from the set, and waiting 6 s for a response. The next trial was presented after a 5-s intertrial interval. Imitation trials were interspersed on a variable ratio 2 schedule, and correct imitation responses were reinforced on a continuous schedule of reinforcement with social praise, high fives, and tickles.

Treatment Package

At the beginning of a treatment session, the researcher enthusiastically showed the participant the bingo board and provided a small pile of bingo chips. The researcher said, “We’re going to play bingo! I’ll read a question, and you put a bingo chip on the answer. OK, here we go! Put a chip on ‘free space.’” The researcher then modeled doing so. Then the researcher removed the chip and encouraged the participant to place the bingo chip on the free space. Next, the researcher asked a question (e.g., “What do you wear on your head?”) and prompted the participant to pick up a bingo chip and place it on the correct square using graduated guidance. The graduated guidance sequence included (a) hand over hand, (b) elbow, and (c) shoulder. The researchers faded prompts both within and across sessions. These prompt-fading procedures were used in various instructional programs at the EIBI centers. When the bingo chip was placed on the correct square, the researcher provided verbal praise that contained the question and the answer (e.g., “You do wear a hat on your head!”), engaged in an exaggerated pretend play action related to the answer (e.g., the researcher pretended to put a fake hat on her head), and narrated the play action (e.g., “Look at my hat!”). Participants were not required to engage in the pretend play action, but if they did, social praise was provided (e.g., “Wow! You have a hat too!”). No other stimuli were presented to the participant after a prompted or independent response. The researcher continued to engage in the pretend play action for a short time (between 5 and 10 s) and then presented the next question by saying something like “OK, are you ready? Here’s the next one.” After each question was presented once, the researcher modeled saying “bingo”; said, “OK, let’s clear the board”; and prompted the participant to pick up the bingo board so the chips fell off. None of the other questions on the board were asked. The researcher then said, “OK, let’s play again. Put a chip on ‘free space,’” and presented each question again but in a different, predetermined, and randomized order. This continued until each question was presented three times for a total of nine trials. Mastery criteria for the listener response (i.e., playing bingo) was placing a bingo chip on the correct square for eight out of nine questions for two (Set A) or three (Sets B and C) consecutive sessions.

Intraverbal and Maintenance Probes

The intraverbal probes were identical to baseline, except mastered imitation trials were eliminated. The imitation trials were removed because participants were no longer engaging in problematic behavior (see Discussion). The bingo boards were not presented during baseline, intraverbal, or maintenance probes. Maintenance probes were conducted 1 month (Esau) and 2 months (Jed) after intraverbal probes for Set C. Esau’s maintenance probe occurred after 1 month because the therapy year ended, and he graduated from the program.

Interobserver Agreement and Procedural Integrity

The first author observed 50% of Jed’s intraverbal probes and 50% of his treatment sessions and 75% of Esau’s intraverbal probes and 30% of his treatment sessions. Interobserver agreement (IOA) and procedural integrity were taken at the same time. IOA was determined by dividing agreements by agreements plus disagreements and multiplying by 100%. Data were compared trial by trial, and IOA was calculated using the point-by-point method. IOA for Jed’s intraverbal probes was 100% and 99% for treatment sessions. IOA for Esau’s intraverbal probes was 100% and 100% for treatment sessions.

Procedural integrity was determined using two checklists: one for intraverbal probes and one for treatment sessions. Each discrete behavior of the researcher was scored as correct or incorrect. If the researcher presented a treatment component out of order, the step was scored as incorrect. Checklists are available upon request. Procedural integrity was calculated by dividing the number of correct behaviors by the total number of behaviors and multiplying the total by 100%. Procedural integrity during Jed’s intraverbal probes was 100% and 97.6% (range 90%–100%) during treatment sessions. Procedural integrity during Esau’s intraverbal probes was 100% and 98% (range 95%–100%) during treatment sessions.

Results

Figure 1 depicts the number of correct intraverbals for Jed and Esau. Each panel represents a set of three wh- questions. During baseline sessions, Jed did not correctly answer any questions for Set A (what), Set B (where), or Set C (when). During the posttraining probe for Set A, Jed answered one question correctly, so treatment was reinstated. When the posttreatment probe was conducted again for Set A, he answered each question correctly. During posttraining probes for Sets B and C, he answered each question correctly and maintained responses for the previously trained sets. During the 2-month follow-up, he answered nine, eight, and two questions correctly for Sets A, B, and C, respectively.

Fig. 1.

Number of correct intraverbals during baseline, posttreatment, and follow-up sessions for Jed (left panel series) and Esau (right panel series). Follow-up session are indicated by asterisks.

Esau did not correctly answer any questions for Set A (when), Set B (what), or Set C (where) during baseline. During the posttreatment probe for Set A, he answered three questions correctly. He answered two questions correctly at least once, but for the third question, he never made a correct response, so treatment was reinstated. When the posttreatment probe was conducted again for Set A, he answered seven questions correctly. During the posttreatment probes for Set B, he answered seven questions correctly and answered eight questions correctly for Set A (maintenance). During the posttreatment probe for Set C, he answered each question correctly and answered nine and six questions correctly for Sets A and B, respectively. During the 1-month posttreatment probes, he answered nine, eight, and nine questions correctly for Sets A, B, and C, respectively.

Discussion

The present study evaluated a treatment package on emergent intraverbals in young children with ASD. Results of the study suggest that the treatment package may be a tenable option to teach young children with ASD to answer wh- questions. There are a few possible reasons why intraverbals may have emerged after listener training. First, participants may have been covertly or overtly tacting the cartoon picture or reading the printed text on the bingo board when they placed the bingo chip on the correct square. It is also possible that participants associated the pretend play gesture with the target response. It has been shown that imitating gestures during play increases language skills (Ingersoll & LaLonde, 2010). It is also possible that the game format created an establishing operation to covertly or overtly tact or make an intraverbal response during the bingo game. It appeared the treatment package acted as an abolishing operation for problematic behaviors because those behaviors were not observed during the bingo game or posttreatment probes. This is further supported by participants asking to play the game, smiling and laughing during sessions, and engaging in the pretend play actions during the game and probes. It is difficult to determine why using a board game to teach intraverbals would decrease problematic behaviors because the physical arrangement was similar to the traditional DTT procedure except for the inclusion of the game board and bingo chips. It is possible the game board was an S^D and signaled that reinforcement for listener responses was available, and it is probable that participants viewed this as a less aversive task compared to making vocal-verbal responses with no visual supports.

This game-based treatment package may be both an effective and efficient intervention to teach intraverbals to children with ASD. The number of trials required until intraverbals emerged decreased across sets of stimuli for both participants. Jed demonstrated emergent responses in 63 (what), 45 (where), and 54 (when) trials for Sets A, B, and C, respectively. Esau demonstrated emergent responses in 99 (when), 90 (what), and 36 (where) trials across Sets A, B, and C, respectively. The downward trend that occurs across sets does not appear to be related to the type of wh- question. The behavioral mechanisms responsible for the downard trend of the number of trials needed until emergence was observed remains unclear, however it is possible that participants may have learned a generative repertiore of intraverbal responding that was facilitated by the listener respose.

There are several limitations to the current study. Most notably, baseline and posttreatment probes differed. The inclusion of positive reinforcement for imitation trials during baseline could have reduced the likelihood of participants responding correctly to the intraverbal questions. The decision to omit imitation trials during posttreatment probes was made in the moment based on participant behavior and the clinic’s mission to make therapy sessions naturalistic. Each participant spent part of the day in an inclusive preschool classroom, and a goal was to have participants answer wh- questions when asked by the general education teachers. The general education teachers asked these questions in a small-group format and did not intersperse imitation trials. Imitation trials were only included during baseline probes because participants engaged in problematic behaviors and sessions were difficult to complete. Problematic behavior data were collected as part of the participants’ behavior plans but were not analyzed in the context of this study. Participants were engaged during posttreatment probes and often engaged in the pretend play action (e.g., pretending to brush their hair) associated with the correct response. Other interesting behaviors were observed. For example, during one of the posttreatment probes, Jed answered “broom” when asked, “What do you use to sweep?”; got up from his chair; went to the corner of the room; and picked up the classroom’s broom and started sweeping. The researcher provided social praise for making the association.

Another limitation is that it is unclear which components of the treatment are responsible for the emergent intraverbals. At this time, we do not know which treatment components, either singularly or in combination, produced emergent intraverbals because no component analysis was conducted. Correct responses were consequated in two ways. First, the researcher provided verbal praise that contained the auditory S^D and the target response. Second, the researcher modeled a pretend play action related to the correct response. The social praise may have functioned similarly to instructional feedback (IF). IF is a method of presenting extra nontarget stimuli in the consequence of instructional trials (e.g., during praise statements). For example, during intraverbal training using IF, the researcher provides a primary target that is an auditory S^D (e.g., “Name a fruit.”). Contingent on correct responses (e.g., “apple”), the researcher provides praise, a preferred item, and a secondary target (e.g., “Mango and pineapple are also fruits.”). In this example, “apple” is the primary target and “mango and pineapple” are secondary targets. Previous research shows that most participants acquire some or all of the secondary targets in the absence of direct reinforcement. In the current study, the verbal praise statement did not contain a secondary target but contained both the auditory S^D and the target response. The behavioral mechanisms responsible for the acquisition of secondary targets during IF are unclear, but one possibility is that participants acquire the IF targets through echoic responding. Although they are not required to echo the secondary targets, participants may overtly or covertly repeat the IF targets before the delivery of the reinforcer for the primary target (Loughrey, Betz, Majdalany, & Nicholson, 2014). If participants in the current study did covertly or overtly echo the verbal praise statement, it was followed by the researcher modeling the pretend play action, which may have reinforced the echoic, thus implementing a teaching procedure. Data were not collected on the number of trials in which participants overtly echoed the praise statement. Future researchers could collect these data and also record the number of times when participants move their lips. Lip movement may correlate with engaging in a covert response (Esch, Mahoney, Kestner, LaLonde, & Esch, 2013).

The second manner in which the researcher consequated prompted or independent responses was by modeling a pretend play action related to the target response. Ingersoll and LaLonde (2010) showed that children with ASD engaged in more prompted and unprompted language during gesture imitation trials than they did in object imitation trials. The behavioral mechanisms responsible for this increase in language are unclear. In the current study, it is possible that the participants overtly or covertly imitated the researcher’s statements that were presented during the play action. For example, when the researcher was pretending to put a hat on her head, she said something like “I have a hat!” If the participants imitated (covertly or overtly) the verbal statements (e.g., said “hat”), it was likely followed by the researcher continuing to engage in the pretend play action or other possible reinforcers, including the researcher smiling or laughing. In such a case, the verbal response “hat” would have been strengthened and more likely to occur when the auditory wh- question was presented in the future.

Compliance with Ethical Standards

Conflict of Interest

Kate B. LaLonde declares that she has no conflict of interest. Shantinique Jones declares that she has no conflict of interest. Laura West declares that she has no conflict of interest. Cora Santman declares that she has no conflict of interest.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.