Teaching Children with Autism to Tact the Private Events of Others

Abstract

We evaluated the efficacy of a most-to-least intrusive error-correction/prompting procedure for teaching 3 children with autism to tact the private events of others through publicly accompanying stimuli. Participants did not reliably demonstrate correct tact responses to publicly observable stimuli that accompany common emotions reported by others (e.g., bandage = hurt) in baseline. Procedures were taken from the Promoting the Emergence of Advanced Knowledge Relational Training System: Direct Training Module (PEAK-DT) to aid in clinical replication, and training was introduced in a multiple-baseline design. Results showed that the procedures were efficacious in teaching this skill to each of the participants, and fast rates of acquisition were observed.

Teaching children to tact private stimuli experienced by others can be both theoretically and practically challenging, especially when this skill does not develop naturally,1 as may be the case for many children with neurodevelopmental disabilities such as autism. Private stimuli are those that an individual alone reacts to, such as a toothache, fatigue, or excitement, where other individuals are not able to react to such stimuli (Skinner, 1957). Tacting the private events of others is involved in many social behaviors, such as saying, “She is hurt!” which necessarily involves inferring the presence or nonpresence of private events. The challenge in teaching children to tact private stimuli is that they are not able to observe in any way the private events of others. For example, if a child sees a person who is likely to feel sad, the child is not able to see “sadness,” and therefore the tact “That person is sad” is unlikely to come under the discriminative control of the private stimuli to which the tact refers. Skinner (1957) described four ways in which verbally able individuals learn to tact their own private events, where one way that could be used to teach children to tact the private events of others is through public accompaniment. Public accompaniment occurs when a person uses co-occurring stimuli in the environment to infer the nature of private events. Skinner (1957) gives an example of teaching a child to tact “That hurts” in the presence of observable tissue damage. Several studies have evaluated procedures for teaching children to recognize co-occurring facial expressions (e.g., Bölte et al., 2002), as well as to tact emotions given scenarios (e.g., McHugh, Bobarnac, & Reed, 2011); however, no studies to date have evaluated a method for teaching children to tact the private stimuli of others through public accompaniment.

The purpose of the current study was to evaluate the efficacy of a discrete-trial training procedure for teaching three children with autism to tact the private events of others through public accompaniment. To aid in clinical replication, the procedures were taken directly from the Promoting the Emergence of Advanced Knowledge Relational Training System: Direct Training Module (PEAK-DT; Dixon, 2014). The first of four PEAK modules, PEAK-DT uses direct training in a discrete-trial arrangement to teach simple and complex forms of verbal operant behavior. The procedures in PEAK-DT are described in such a way as to allow for implementation by therapists, teachers, caregivers, and parents, and the procedures have demonstrated effectiveness both in between-group (McKeel, Dixon, Daar, Rowsey, & Szekely, 2015) and in single-subject (e.g., McKeel, Rowsey, Belisle, Dixon, & Szekely, 2015) evaluations.

Method

Participants, Setting, and Materials

Three children with diagnoses of autism participated in the study: Andrew, a 10-year-old male with an IQ score of 48 (Weschler Intelligence Scales for Children, Fourth Edition, Short Form); Oliver, a 5-year-old male with an IQ score of 73 (Weschler Preschool and Primary Scale of Intelligence, Third Edition [WPPSI-III], Short Form); and Jimmy, a 5-year-old male with an IQ score of 48 (WPPSI-III Short Form). Each of the participants had previously demonstrated several of the elementary verbal operants, such as generalized echoics, tacting common objects, manding for preferred objects, and basic intraverbals (e.g., answering the question “What is your name?”). Oliver had additionally demonstrated early transcription and textual behavior with directly trained three-letter sight words. The participants each attended a specialized school for children with autism and related disorders. Sessions were conducted in the children’s regular classroom setting, in an area separate from the other students, or in a separate classroom designated for direct assessment and training. All experimental areas contained a single desk, two chairs, and the preferred stimuli. Sessions included four to five trial blocks of eight trials and lasted between 30 and 45 min. A 1-min break was given in between trial blocks where participants were given access to nonpreferred classroom materials (e.g., pencil and paper). The materials used in this study included a standard-size pillow (20 in. × 26 in.; 50.8 cm × 66.04 cm), a bandage, wooden blocks, and an eye drop container (2 in. × 1 in.; 5.08 cm × 2.54 cm). In addition, the PEAK-DT program used in this study was “14R–Public Accompaniment: Expressive,” which provided a one-page instruction document that described each step of the training procedure to guide the experimenter (which included the program goal, required materials, step-by-step discrete-trial instructions, and example stimuli). The instruction sheet was reviewed prior to each session. Token boards were also present throughout all trials, where eight tokens were exchangeable for 5-min access to preferred items during the training phase of the investigation. The 5-min access was selected due to the young age of the participants in the study. The participants had prior experience completing classroom work with a token-economy system prior to the study. Preferred items were determined through interviews with the classroom staff, where all preferred items were presented to the participants once the tokens were obtained, and the student was provided access to the first preferred item selected (i.e., touched). Preferred items included blocks, spinning tops, and toy cars, as well as other items within the classrooms.

Procedure

To evaluate the efficacy of the training procedures, a multiple-baseline across-subjects experimental design was used. A mastery criterion was set at five consecutive sessions with 100% independent correct responding (i.e., where no prompts were required to evoke the correct response). The dependent variable was the percentage of independent correct responses within a trial block (number of correct responses, divided by 8, and multiplied by 100), where a correct response occurred when the participant emitted the target vocal tact independently. Data were collected by one of two experimenters on every trial, and a second observer was present during 66% of the trials to calculate interobserver agreement (IOA; agreements, divided by the total number of trials, multiplied by 100). IOA was calculated at 100%.

Baseline

The experimenter began each trial by presenting a discriminative publicly accompanying stimulus (e.g., by placing a bandage on his or her left hand) to the participant. The experimenters refrained from also engaging in any collateral responses when presenting the discriminative stimulus (e.g., closing eyes, saying “Ouch”). Once the stimulus was visible to the participant, the experimenter delivered the auditory verbal stimulus, “How might I be feeling right now?” If the participant demonstrated the correct response within 3–5 s, then the trial was scored as correct. In baseline, any single one-word vocal utterance that (a) included an emotion name and (b) was relevant to the given stimulus was considered correct. If the utterance contained more than one word, the response was considered incorrect. This criterion was established to ensure that response chains containing multiple emotion words (e.g., “sad, angry, or happy”) or mands for information (e.g., “Are you feeling happy?”) were not considered correct. A correct response, for example, occurred if the experimenter had a tear in his or her eye, and the participant tacted “sad” or “upset”; however, a tact of “silly” would be considered incorrect. Each of four stimuli was used within each block, and their presentation order was randomized; each stimulus was presented twice in each block. Table 1 provides a summary of all stimuli. The emotions tested and subsequently trained in the current study were “tired,” “sad,” “hurt,” and “mad.” For “tired,” the experimenter would hold the pillow in one hand and place his or her head on the pillow. For “hurt,” the experimenter would place a bandage on the top of his or her left hand just above the thumb. For “sad,” the experimenter would turn away from the participant and use the eye drop container to place drops under his or her left eye then turn back around to the participant. Last, for “mad,” the experimenter would stack up three to four wooden blocks and subsequently knock them down using the palm of his or her right hand. A modification was made for Jimmy in that the word “angry” was targeted instead of the word “mad” due to difficulties pronouncing the latter word. Reinforcement items were presented at the end of each trial block for 5 min, and no tokens were presented or used in the baseline phase. In addition, neither prompts nor feedback was provided during this phase.

Table 1.

List of stimuli used in the study

Program Name	Stimuli Used	Exemplar Correct Responses
14R–Public Accompaniment: Expressive	Set 1 (All)	Set 1 (All)
	• Knocking blocks over	• “Mad” or “angry”
	• Laying head on pillow	• “Tired”
	• Bandage on arm	• “Hurt”
	• Eye drop in eye	• “Sad”

Training

The delivery of the publicly accompanying stimulus was identical to the baseline condition, whereby the experimenter presented a stimulus that corresponded with a given feeling and asked, “How might I be feeling right now?” If the participant demonstrated the correct response independently, he received verbal praise from the experimenter and was given a single token. If the participant failed to demonstrate the correct response, a most-to-least intrusive error-correction/prompting procedure was used. After the verbal antecedent stimulus was delivered, the experimenter waited 5 s for the participant to demonstrate the correct target response. If, following 5 s, the participant either failed to produce a response or produced a response that was incorrect, the experimenter prompted the participant using a first-level prompt. This level of prompting involved the experimenter vocally modeling the entire correct response (e.g., “tired”), where the participant was required to echo the response. After the participant demonstrated the correct response (i.e., the echoic response), subsequent incorrect responses were prompted using the second-level prompt. The second-level prompt involved the experimenter vocally providing only the first sound of the correct response (e.g., “ti” if the correct response was “tired”). If the participant failed to demonstrate the correct response following the second-level prompt, then the experimenter delivered the first-level prompt. When prompts were used, verbal praise was provided only after the correct response was observed; however, a token was not provided. If the participant had demonstrated the correct response in previous trials requiring a lower prompt level (e.g., if the participant required only a second-level prompt on the previous trial but required a first-level prompt on the current trial), then verbal praise was also withheld following the current correct response, and the experimenter progressed to the next trial.

Results and Discussion

The results of the present study are summarized in Fig. 1. In the baseline phases for Andrew, Oliver, and Jimmy, the mean baseline scores were 3.57% (range 0%–25%). Stability was defined as three consecutive trial blocks with scores within 25% of each other. Stability was achieved for all participants in the baseline phase. In the training phase, Andrew’s mean correct responding increased to 89.06% (range 37.5%–100%), and he achieved the mastery criterion in eight trial blocks. Oliver’s mean correct responding increased to 93.75% (range 62.5%–100%), and he achieved the mastery criterion in six trial blocks. Finally, in the training phase for Jimmy, his mean correct responding increased to 85.71% (range 37.5%–100%), and the mastery criterion was achieved in seven trial blocks.

Fig. 1.

Percentage of correct independent responding within trial blocks during baseline and training conditions across three participants; DTT = discrete-trial training; MTL = most to least

The results of the present study demonstrated the efficacy of a procedure for teaching children with autism to tact the private events of others. Because individuals are not able to react to the private stimuli of others, appropriate social responses must come under the discriminative control of public stimuli if individuals are to engage in such behavior. Tacting private events of others may be required to engage in common social interactions, as well as in communicating how a peer might be feeling to a caregiver (e.g., reporting that a peer is “hurt”). Skinner (1957) discussed four ways in which individuals learn to tact their own private events, and public accompaniment is one of two ways that may be the most easily applied with children who cannot tact the private events of others. The other is through collateral responses, such as smiling or laughing, that co-occur with private stimuli, and previous research has demonstrated that tacting can come under the discriminative control of these responses in the case of facial expressions (e.g., Bölte et al., 2002). Collateral responses and publicly accompanying stimuli in isolation are both unlikely to provide all information required for accurate tacting of others’ private stimuli; however, further research is needed on a variety of methods for teaching this early social skill. Undoubtedly, once tacting of others’ private stimuli based on these readily available external stimuli comes under appropriate discriminative control, tacting private stimuli may come under the control of other stimuli discussed by Skinner (1957). The other two ways in which individuals learn to tact their own private stimuli are through metaphorical extension or when private stimuli are similar (except in magnitude) to those that have previously accompanied publicly accompanying stimuli (Skinner, 1957). Although discussion of these conditions is outside the ambit of the present article, tacting private stimuli of others based on these events may more closely resemble empathetic responding.

The results of the present study should be interpreted with caution due to the similar age and diagnoses of the participants in the study, limiting the external validity of the obtained results. In addition, because the publicly accompanying stimulus (e.g., a tear) was presented with a verbal antecedent stimulus (i.e., “How might I be feeling right now?”), it is unclear if the tact was a pure tact or a partial intraverbal/tact. Given it is unlikely in a natural context that another child, for example, will provide this verbal antecedent stimulus, future research may fade this stimulus and test for naturally occurring stimulus control over the tact response. A final limitation is that treatment fidelity data were not reported. Future research may extend the obtained results by evaluating other ways to further increase the accuracy of tacting the private stimuli of others by bringing the tact response under multiple control of several external stimuli. For example, the tact of “tired” or “bored” when a person is seen closing his or her eyes (collateral response) could be either correct or incorrect in the absence of publicly accompanying stimuli, such as a pillow or a schoolbook. Future research may also evaluate how the ability to tact the private stimuli of others influences social behavior outside of the discrete-trial training arrangement used in the present study.

Compliance with Ethical Standards

Conflict of Interest

Mark R. Dixon discloses that he receives royalties from sales of the PEAK curriculum. All remaining authors declare they have no conflict of interest.

Ethical Approval

All procedures performed in this study that involved human participants were in accordance with the ethical standards of the institution and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained for participation in this research.