Keywords

Social skills are a complex class of responses that relate to one’s communication and interactions with other people in a common environment (Little et al., 2017; Ulrich et al., 1966). Although there is no explicit definition, many responses may be categorized as social. This class of behavior may be best conceived of as those responses that increase the likelihood of producing positive consequences with other people, while minimizing negative consequences (Gresham & Elliott, 1984). For example, eye contact, prosody of speech, gestures, affect, language and interactions, joint attention, and play behavior are all important responses that can affect environmental outcomes. The extent to which one displays each, or a combination of these responses can contribute to the likelihood with which that person will function competently in society (Vener et al., 2017). Social skills are vital to one’s development and correspond with the likelihood of success in various avenues of life including academics, vocations, and peer relationships (Mayer et al., 2012). Many individuals with developmental disabilities demonstrate deficits in one or more areas of social behavior.

When a person displays a deficit in one or more of these responses, the probability of a positive outcome may be compromised. Nevertheless, evidence-based practices in the field of applied behavior analysis have demonstrated success in teaching social skills and increasing the likelihood of social reinforcement. Effective teaching strategies have included differential reinforcement, modeling, incidental teaching, discrete-trial teaching, script fading, and self-management, among many others. Although not an exhaustive review of the literature, this chapter does address the different components of social behavior and training procedures used in behavior-analytic research.

Research on Social Skills

As outlined by Baer et al. (1968), one of the core dimensions of applied behavior analysis is that the research is applied and relates to behavior that is important and socially significant to the person and consumers of the behavior. By definition, social skills satisfy these criteria. Applied behavior analysts have dedicated a vast amount of research to determining effective interventions to teach these skills.

Some of the earliest research on social skills analyzed the relationship between the broad concept of social competence and treatment outcomes of patients with various diagnoses including schizophrenia, manic depression, and personality disorders. These studies concluded that patients with lower social competency were associated with a “poorer prognosis,” or less desirable outcomes, than those with higher social competency skills. Patients of the latter were associated with more favorable psychiatric outcomes, such as shorter hospitalization times or a lesser number of hospital readmissions (Zigler & Phillips, 1961; Zigler et al., 1979).

As researchers continued to focus on the study of social skills, the components of social behavior were operationalized, quantified, and measured in a more discrete manner. For example, Bornstein et al. (1977) conducted a study that taught social skills to four unassertive children. The children were between the ages of 8 and 11 years, and displayed deficits in eye contact, speech duration, speech volume, and requests. The authors presented role-play scenarios to the participants that were similar to their typical encounters in school. For example, a scenario may have been that a classmate cut the participant in the lunch line, or that a classmate broke the participant’s pencil after borrowing it. The authors used instructions, modeling, feedback, and rehearsals to teach the four participants to engage in social skills appropriate for each scenario. The results indicated a systematic change in overall assertiveness across the four participants. Moreover, the findings demonstrated generalization across untrained scenarios.

Subsequent behavior-analytic studies have targeted social behavior such as eye contact, prosody of speech, gestures, affect, joint attention, language and interactions, and play behavior. Moreover, studies have included analyses of the effectiveness of specific independent variables on such social behavior, such as discrete-trial teaching, in vivo and video modeling, incidental teaching, script-fading procedures, and self-management, among many others. The results of such research have provided important information that can be used to improve social competency.

Eye Contact

One of the most fundamental responses crucial to social behavior is eye contact with other people. Eye contact is a pivotal response that is an important component in conversational behavior, direction-following skills, and in responding to social cues (Jeffries et al., 2016). One of the earlier studies on eye contact was conducted by Foxx (1977) to increase the extent to which three children with autism displayed eye contact with an instructor. The authors used an overcorrection avoidance procedure to teach participants to engage in eye contact when provided the instruction, “Look at me.” When the participant emitted eye contact within 5 seconds of the instruction, an edible reinforcer and praise were delivered. If the participant did not engage in eye contact within 5 seconds, then the overcorrection procedure was implemented, during which instructors would manually prompt the participant to engage in repeated head orientations. The authors found a systematic increase in the percentage of trials with eye contact. Essentially, participants increased eye-contact responses to avoid the aversive alternative. The authors did indicate that this strategy should be considered as a last-resort option when other positive-reinforcement procedures prove ineffective.

Eye contact also may be taught by using differential reinforcement. Differential reinforcement consists of the reinforcement of responses that adhere to a specified property, while withholding reinforcement for those responses that do not (Catania, 1998). To differentially reinforce eye contact, instructors would deliver reinforcers contingent upon the occurrence of appropriate eye contact, and withhold reinforcer delivery when the learner is looking away from the instructor. Several researchers have demonstrated the effectiveness of differential reinforcement in modifying eye-contact behavior. For example, Carbone et al. (2013) conducted a study in which differential reinforcement was used to teach eye contact to a 3-year-old boy with autism. Access to preferred stimuli was granted when the participant emitted appropriate mands in conjunction with eye contact to the instructor, whereas access was denied when mands were emitted without eye contact. This use of differential reinforcement produced an increase in the percentage of mands accompanied by eye contact.

Jeffries et al. (2016) conducted a similar study and found the use of differential reinforcement to be a more effective strategy to teach eye contact to three children with autism than the use of a tablet application. By withholding reinforcers until eye contact was established during manding opportunities, the authors observed a systematic increase across all three participants.

Similarly, shaping can be used to effectively modify eye-contact behavior. Shaping consists of the modification of behavior by differentially reinforcing successive approximations to a target response (Catania, 1998). With successive shifts in reinforcement contingencies, responding gradually transforms until it resembles the desired terminal response (Cooper et al., 2020). With eye contact as the targeted terminal response, the instructor may initially reinforce a learner’s orientation to the general direction of the instructor. As this behavior is reliably emitted, the instructor may then decide to withhold reinforcer delivery for these responses, and only deliver for learner orientation to the instructor’s face. Finally, as this form of responding is steadily emitted, only eye contact to the instructor’s eyes may occasion reinforcer delivery. This procedure, of course, may involve many more detailed steps in the shaping process, as well as careful and systematic measures. For example, Fonger and Malott (2019) used a shaping procedure to teach three young children with autism to emit eye contact with an instructor for a duration of 3 seconds. The researchers presented and removed preferred stimuli within the participants’ environment. Initially, reinforcers were delivered contingent upon the orientation of the participants’ eyes to the instructor’s body. Once this response occurred reliably, reinforcers were only delivered contingent upon eye orientation to the instructor’s face. Subsequently, any occurrence of eye-to-eye contact occasioned reinforcer delivery, regardless of duration. Thereafter, the criterion for reinforcement shifted such that eye contact was required to occur for a duration of 1 second, then 2 seconds, and finally 3 seconds. The data demonstrated a functional relation between the shaping procedure and the occurrence of eye contact across the three participants. Moreover, the authors found that initial occurrence of eye contact and the use of prompting were not required to produce this behavior.

Prosody

Prosody consists of various characteristics of speech, including volume, intonation, and rhythm. People with developmental disabilities may display speech that is monotonic, lacks volume control, and presents inappropriate quality or stress patterns (Brown & Poulson, 2009; Paul et al., 2005). All of these aspects of speech play an important role in one’s interactions with other people, and a deficit in any single one can impact the outcome of a given interaction. Fortunately, behavior-analytic strategies have proven effective in modifying prosodic responses.

Voice volume is one important characteristic of prosody that can clearly impact social behavior. Researchers have successfully modified voice volume by employing various behavior-analytic procedures. Schwartz and Hawkins (1970) used a delayed reinforcement procedure to increase the voice volume of a 12-year-old “maladjusted” girl. The participant was videotaped during her regularly occurring math and spelling classes at school. At the end of the school day, she reviewed this video with the experimenter. If her voice volume had satisfied a specified criterion for volume, the participant would earn poker chips to be exchanged for preferred items. The authors found that this delayed-reinforcement strategy caused an increase in her voice volume and generalized to an untrained class in school.

Fleece et al. (1981) used a shaping procedure to increase the voice volume of two 4-to-5-year-old children with developmental delays. The participants recited nursery rhymes with the experimenter in a sound-proof room. A voice-activated relay with adjustable sensitivity was used to activate a display of red and green lights contingent upon the occurrence of a predetermined voice volume. During intervention, the sensitivity of the voice-activated relay was modified to shape increases in volume. During baseline, participant voice volumes ranged from inaudible to below-normal levels. With the introduction of the shaping procedure, participant voice volume increased systematically to levels which would be considered “normal.” Furthermore, the authors observed a maintenance of increased voice volume at one- and four-month follow-up sessions.

Another important component of prosody is intonation, or pitch. Intonation refers to the rise and fall of one’s speech and has been systemically modified using behavior-analytic strategies. For example, Daou et al. (2014) used a discrete-trial teaching package to teach children with autism to emit the appropriate affective behavior, one component of which consisted of vocal intonation. When presented with scenarios from various affective categories, including absurdities, empathy, excitement, and gratitude, the participants were taught to emit appropriate verbal responses in conjunction with appropriate vocal intonation. Through the use of in vivo modeling, instructions, and reinforcement, the authors observed a systematic increase in the percentage of correctly emitted vocal intonation for all three participants. Responses also generalized across untrained scenarios presented by the experimenter.

Similarly, Charlop et al. (2010) used video modeling to modify intonation among three boys with autism as one of four components targeted in teaching socially expressive behavior. The participants were presented with video scenarios which occasioned forms of intonation such as loud/authoritative or exclamatory. During baseline, none of the participants displayed appropriate intonation. With the introduction of the video-modeling procedure, however, the percentage of correct intonation responses increased systematically across participants, and demonstrated generalization across untrained settings, stimuli, and people.

Although incidental teaching (Hart & Risley, 1975, 1980) and script-fading procedures (Krantz & McClannahan, 1993, 1998) are primarily used to teach language and conversational skills, these strategies also may embed the modeling of prosodic characteristics of speech. Reinforcement contingencies may be modified such that appropriate voice volume, intonation, and/or rhythm are required to occasion the delivery of reinforcers.

Gestural Behavior

Gestures are an important aspect of social and communicative behavior. They are hand movements that add meaning to speech and convey information that otherwise would not be present during speech alone (de Marchena et al., 2019; McNeill, 1992). For example, a person who is describing a toy might hold up two hands with palms facing each other, two feet apart. The conversation partner would understand that the toy was quite large without the speaker having to say it. Similarly, gestures can link the physical environment to spoken language, such as when a parent says, “it’s almost time for dessert,” while pointing to the empty dinner dishes on the table.

In addition to supporting and enhancing information conveyed by speech, greater comprehension is found when speech is accompanied by gestures than when speech is not accompanied by gestures (Hostetter, 2011). Gestures can signal emphasis, uncertainty, and convey meaning. A friend who says, “the movie is still on,” while placing her index finger to her lips is clearly stating that one should stop talking. Gestures are also beneficial in situations in which the quality of spoken language is weakened and/or the listener is less verbally proficient. For example, if the environment is noisy, a listener has less verbal skills than the speaker, and/or speech is unclear, gestures can perhaps clarify spoken meaning.

Evidence suggests that gestures are important in the initial stages of the development of communication (Volterra et al., 2006). Gestures develop prior to the development of speech and are later used in conjunction with spoken language (Morford & Goldin-Meadow, 1992). Between the ages of nine and 13 months, gestures such as give, show me, and point emerge and begin to express communicative intent. For example, a baby might reach toward an object or point to a desired toy. Pointing is the most frequently used gesture and is later correlated with the onset of one’s first word (Volterra et al., 2006).

McNeill (1992) described gestures as iconic, metaphorical, deictic, beat, and/or emblem. Iconic gestures are gestures that describe physical characteristics, such as size or shape or size. Metaphorical gestures illustrate information in a more abstract manner, such as moving a hand outward and upward to describe a skill. Deictic gestures are pointing gestures that identify a specific object in the environment. Beat gestures are rhythmic hand movements presented in close temporal synchrony with targeted words or phrases. Emblem gestures are gestures that are culturally defined. For example, one’s palm facing away from speaker’s body, pointer and thumb in a circle, with the other three fingers pointing up to indicate, “OK.” Other culturally defined gestures include clapping hands to show approval, bringing hand to mouth to express hunger, or bringing the hand to the ear to represent a telephone. These gestures represent reliably consistent meaning and can be interpreted without contextual reference (Volterra et al., 2006).

Bavelas et al. (1992) included interactive gestures to the list of gestures presented by McNeill (1992). An interactive gesture makes reference to the conversational partner or the conversation. For example, a speaker might point to an individual when referencing something that he or she said earlier, or gesture along with specific word or phrase to signal importance.

Individuals with autism rarely use gestures as a means of communication (Ingersoll et al., 2007; Loveland et al., 1988). Nevertheless, if they do use gestures, they use iconic more than other gesture types (Medeiros & Winsler, 2014). This finding suggests that individuals with autism predominantly use gestures for concrete as opposed to abstract function. In addition to an insufficient use of co-speech gestures, individuals with autism speak at a rate slower that typically developing peers, are slower to respond than typically developing peers, and use more unusual vocalizations during pauses more often than typically developing peers (Parish-Morris et al., 2016).

Because individuals with autism are less likely to use gestures during social interactions, they have difficulty participating in social situations. Buffington et al. (1998) conducted a study with four children with autism who used little or no gestural communication prior to intervention. Following the introduction of a modeling, prompting, and reinforcement treatment package, all four children emitted verbal and gestural responses during social interaction opportunities. These findings suggest that by learning to use gestures in combination with vocal behavior, children with autism can begin to develop a repertoire needed to engage in social interactions. The authors assessed generalization in the presence of novel stimuli and novel settings. The results showed an increase in the use of gestural communication during probe assessments. Social validity data suggested that the children’s communication was rated as more expressive during treatment than baseline, suggesting that teaching children with autism to use gestures leads to improvements in the quality of their social communication.

Similarly, Duker and Van Lent (1991) stated that individuals with developmental disabilities may engage in a low variation in gestures. Five individuals with severe to profound mental retardation between the ages of 13 and 30 participated in the study. During baseline, each gesture emitted was reinforced with access to the requested object. Gestures were ranked according to frequency of use. During treatment, access was withheld from 2 to 3 of the high-frequency gestures. Low-frequency gestures were differentially reinforced. Results showed an increase in previously used gesture requests. By increasing the variation in gestures used, the ability to engage in meaningful social communication increases.

In an attempt to increase gestures in five young boys with autism, ages 3–4, Ingersoll et al. (2007) used a reciprocal imitation training strategy to increase spontaneous use of descriptive gestures during play. During free play during baseline, the experimenter modeled a descriptive gesture accompanied by a corresponding verbal statement. For example, while placing a doll in a bed, the experimenter placed her finger over her lips and said, “Sh, baby sleeping.” During intervention, the experimenter followed the child’s lead, imitated child actions and vocalizations, and commented on the actions that the child performed. In addition, the experimenter modeled gestures and verbal statements corresponding to the child’s actions. Rather than targeting specific gestures to criterion, multiple gestures were taught that were relevant to the play actions initiated by the child. If the child imitated the gesture, praise and access to play materials were provided. If the child did not imitate the gesture, manual prompts were provided and praise delivered. The authors found that all children increased their imitation of descriptive gestures in the training setting and in an untrained setting.

Overall, co-speech gestures assist the speaker in conveying meaning. People with developmental disabilities may display considerable impairment in their spontaneous use of meaningful gestures. Across the studies reviewed, imitation training, modeling, contingent reinforcement, and prompting are promising procedures to address these deficiencies.

Affective Behavior

Affective behavior consists of various categories of responding such as facial, verbal, gestural, and postural responses (Gena et al., 1996). From as early as infancy, affective behavior serves as an important social behavior and form of communication with others. For example, the reciprocation of smiles with parents (Dawson et al., 1990) and demonstration of empathic responding to others’ distress (McDonald & Messinger, 2012) are notable differences between children with developmental delays and their typically developing peers. The display of appropriate affect, however, has a large impact on the development of relationships with others across a person’s lifespan.

Several studies have demonstrated that components of affective behavior can be successfully modified using operant conditioning procedures. For example, DeQuinzio et al. (2007) used a treatment package consisting of modeling, manual prompts, reinforcement, and error correction to teach three children with autism to imitate various facial models. Facial models included responses such as smiling, frowning, and a surprised face. Prior to teaching, the participants did display some imitative responding, although performance was variable and inconsistent. The introduction of the treatment package, however, produced a systematic increase in the imitation of facial models, and responding generalized across an untrained angry facial model.

Gena et al. (1996) used modeling, verbal prompting, and reinforcement to teach four youths with autism to emit appropriate affective behavior across various categories. The authors taught the participants to emit contextually appropriate verbal and facial responses to scenarios concerning sympathy, appreciation, dislike, preferred items, and absurdities. For example, the experimenter may have stated, “I have a headache,” to which an appropriate affective response would include a serious facial expression and a statement of sympathy, such as, “I’m sorry to hear that.” The data demonstrated that the intervention produced a systematic change in affective behavior across the participants for three or four of the targeted categories. Moreover, responding generalized across untrained scenarios within a given affective category.

In a study by Schrandt et al. (2009), a treatment package consisting of pretend play with manual prompts, auditory prompts, behavior rehearsals, and reinforcement was used to teach four children with autism to emit vocal and motor responses of empathy in the presence of a verbal stimulus. During a given session, the experimenter presented a vignette by using a doll or puppet. For example, the experimenter may have had the doll bump its leg on a table and exclaim, “Ouch!” Empathic responding was taught across three affective categories including sadness/pain, happiness/excitement, and frustration. The experimenters used manual prompts to teach the participants to emit empathic motor responses, such as patting the doll’s arm. An auditory prompt, consisting of a recorded script on a Language Master © card, was used to teach participants to emit empathic verbal responses, such as, “Are you okay?” During the baseline condition, participants emitted few or none of the appropriate empathic responses. With the introduction of the treatment package, however, empathic responding increased systematically across the affective categories for all participants. Moreover, responding generalized for two of the participants from the pretend-play stimuli to untrained people in an untrained setting.

Gena et al. (2005) used in vivo modeling, video modeling, verbal prompts, and gestural prompts to teach three children with autism to emit appropriate affective behavior during pretend play. Affective categories included sympathy, appreciation, and disapproval. Affective behavior was scored as correct if the participant emitted verbal responses, vocal intonation, and facial expressions that were appropriate for the given category. The authors observed a systematic increase in the percentage of appropriate affective responding emitted by the three participants with the introduction of the treatment package. Furthermore, responding generalized to untrained stimuli and people.

Similarly, Charlop et al. (2010) taught three children with autism to emit affective behavior across scenarios during a play session, such as denying access to tickling, being shown a preferred toy, making a basket, and being shown a large toy. For each scenario, the researchers used video modeling to teach the children to emit the appropriate verbal statement, intonation, gesture, and facial expression. For example, when shown a preferred toy, the children may have been taught to say, “That’s cool!” while using an exclamatory voice with positive valence, raising their eyebrows at least ¼” from resting position, and pointing with their index finger or full hand toward the object. The results demonstrated that the video-modeling procedure led to a systematic increase in the percentage of correctly emitted affective components. Moreover, results generalized across people, settings, and stimuli.

A study by Daou et al. (2014) used a discrete-trial teaching package to teach children with autism to emit the appropriate verbal responses, vocal intonation, and facial expressions when presented with scenarios from various affective categories including absurdities, empathy, excitement, and gratitude or appreciation. Through the use of in vivo modeling, instructions, verbal prompts, script fading, and reinforcement, the authors observed a systematic increase in the percentage of correctly emitted affective components for all three participants. Responses also generalized across untrained scenarios presented by the experimenter.

Finally, Argott et al. (2017) taught four children with autism to display empathetic responding across the affective categories of joy, frustration, and pain. The authors used a prompting sequence that progressed from video modeling, to in vivo modeling, to manual prompts, and then to verbal prompts. During sessions, the instructor presented an affective stimulus from the given category. For example, within the category of joy, the instructor might have held up a puzzle and exclaimed, “I finished my puzzle!” A correct empathetic response consisted of a statement of empathy emitted with appropriate intonation, an appropriate gesture, and a corresponding facial expression. In this example, a correct response included a statement such as, “That’s great!” emitted in a high pitch while also smiling and offering a high five to the instructor. During baseline, the participants displayed none, or inconsistent levels, of the targeted empathetic responses. Correct empathetic responding increased systematically across participants for the three affective categories when treatment was introduced. In addition, responding generalized across untrained people and across untrained affective stimuli within each of the categories.

Joint Attention

Joint attention is a significant developmental milestone and an underlying component in language acquisition and social skills development. Joint attention refers to the mutual attention between two people toward an object in the environment (Bakeman & Adamson, 1984). It is a skill that typically develops during infancy and is a pivotal response that can lead to the acquisition of other important behavior (Jones & Carr, 2004). A deficit in joint attention is one of the core criteria in the diagnosis of autism (American Psychiatric Association, 2013). Given the impact of joint attention on various other social skills, intervention aims at teaching this skill as early as possible and has been successful through the use of behavior-analytic methods.

Whalen and Schreibman (2003) used shaping and prompt-fading procedures to teach five children with autism to initiate and to respond to bids for joint attention. To teach responding to joint-attention bids, the experimenters conducted six levels of training. During all levels, the participant was playing with a toy, but the response requirement successively shifted. At the first level, the experimenter manually prompted one of the participant’s hands onto a different toy. If the participant looked at or played with this different toy, then the response was scored as correct. During the second level of training, the experimenter tapped on a different toy, and responding to joint attention was scored as correct if the participant looked at or engaged with that toy. During the third level, the experimenter showed a toy to the participant, and responding was correct if the participant looked at or engaged with the toy being shown. At the fourth level of training, the participant was now required to engage in eye contact with the experimenter. During the fifth level, the participant was required to first engage in eye contact with the experimenter and to orient their head toward an object at which the experimenter pointed. Finally, at the sixth level of training, the participant was required to orient to an object toward which the experimenter shifted their gaze, without the presentation of a pointing response. To teach initiations for joint attention, Whalen and Schreibman (2003) using a prompt-fading procedure while the participant was playing with a toy. Initially, the experimenters provided manual prompts to teach the participant to hold the toy with which they were playing and to orient their head until eye contact was established. The experimenters delivered the verbal prompt, “Show” in conjunction with gestural prompts to their eyes to teach a gaze shift between the toy and the experimenter. Experimenters faded prompts from full physical with verbal, to partial physical with verbal, to gestural with verbal, to verbal prompts only, and finally by removing all prompts. The results demonstrated that their shaping and prompt-fading procedures systematically increased the percentage of correct responses to joint attention for all five participants and increased the percentage of correct initiations for joint attention for four of the five participants.

Taylor and Hoch (2008) used a least-to-most prompting procedure to teach three children with autism to initiate and respond to bids for joint attention. The researchers targeted three components of joint attention including a shift in eye gaze between an object in the environment and an adult’s eyes, a vocal response to a bid for joint attention, and a vocal initiation as a bid for joint attention. Sessions were conducted in a room containing novel or unusually placed toys. To teach responding to bids for joint attention, the instructor presented a bid for joint attention, such as pointing to a toy and exclaiming, “Wow!” Initially, if the participant did not orient to the item, the instructor provided a gestural prompt from the participant’s orientation to the toy. If the participant still did not gaze with the delivery of this prompt, then the instructor used physical prompts to turn the participant’s head toward the toy. The instructor then modeled a comment for the participant to emit about the toy. Gestural and verbal prompts were then used to teach the participant to shift their orientation from the toy to the instructor. A similar prompting procedure in conjunction with verbal prompts was used to teach the participants to initiate bids for joint attention. The results demonstrated that their least-to-most prompting procedure was effective in teaching the participants to engage in the three targeted components of joint attention.

Kryzak et al. (2013) taught three children with autism to respond to bids for joint attention by using the participants’ intense interests in specific topics in conjunction with a most-to-least verbal-prompt-fading procedure. While the participants were engaged with their object of interest, the experimenter gazed and pointed at the object while emitting an exclamatory statement (e.g., “That’s cool!”). The experimenter used verbal prompts to teach the participant to shift their eye gaze to the object by stating the participant’s name or the direction, “Look.” After a specified criterion of responding was achieved, the verbal prompt was faded to the initial sound of either the participant’s name of the direction to look. Subsequently, a constant 4-second prompt delay was used to teach responding. Experimenters reinforced correct responses to joint attention with verbal praise statements. The findings demonstrated a systematic relation between the prompt-fading procedure and the percentage of correct responses to bids for joint attention. Moreover, data suggested generalization across untrained stimuli. Kryzak et al. (2013) suggested that the use of the participants’ intense interests with their prompt-fading procedure may be a successful starting point for teaching joint-attention skills.

MacDuff et al. (2007) used a script-fading procedure and manual prompts to teach three young children to emit bids for joint attention. The experimenter walked the participant through a hallway containing various different pictures and toys. Voice-over-recording devices with the recorded script, “See” were placed upon the stimuli. Instructors used manual prompts to teach the participants to press the recorders and to orient to the experimenter while emitting the scripted statement. Correct responding was reinforced with a token motivational system and edible snacks. The recorded script and voice-over-recording device were faded systematically as the participants reliably emitted independent bids for joint attention. The data demonstrated a systematic increase in number of correct bids for joint attention emitted across the three participants. In addition, as the scripts were faded, participants began to emit unscripted initiations, such as object labeling. The findings also demonstrated generalization across untrained settings and stimuli.

Language and Interactions

There is an expansive breadth research on behavioral interventions that have demonstrated effective changes in language and interaction skills. The review that follows consists of some of the fundamental studies in this area in the field of applied behavior analysis.

One of the earliest studies on language and behavior-analytic methodologies was that conducted by Rheingold et al. (1959). The experimenters analyzed the effects of reinforcement on the vocal behavior of 3-month-old infants. Prior to intervention, the experimenter displayed a neutral affect while orienting to the infant over the crib. During intervention, the experimenter delivered social praise in the form of smiling and tickles to the infant when the infant had vocalized. The experimenters then withdrew intervention and returned to a baseline condition. They found that the infants vocalized at a higher rate during the experimental condition than in the baseline conditions. Nevertheless, it was possible that the delivery of social praise served to elicit, rather than evoke, the observed vocal behavior.

Todd and Palmer (1968) compared the effects of auditory social reinforcers (e.g., a human voice) with the adult present versus the adult being absent from the environment on the frequency of infant babbling behavior. Although the authors observed an increase in babbling under both conditions, they noted a markedly higher frequency of babbling by infants for whom an adult was present when delivering auditory reinforcers.

Other research on language and reinforcement by Ramey and Ourth (1971) found that the rate of infant vocalization behavior increased only with the presentation of immediate reinforcer delivery, rather than reinforcer delays of 3 or 6 seconds. A later study by Reeve et al. (1992) found a systematic relation between infant vocalizations and delayed social reinforcement. These studies helped to establish that language, as early as in the infant stages of life, can be modified using operant conditioning procedures.

Incidental Teaching

Evidence-based research continued to demonstrate the operant nature of language and effectiveness of behavior-analytic methods on language acquisition. Incidental teaching is the process by which language is acquired in a naturalistic setting. A child initiates a verbal or nonverbal interaction with an adult that indicates a want or need for a given stimulus. For example, a child may reach for an item that is too high, or may emit a request for a desired item, activity, or information. The child’s want or need is then used as an opportunity to teach language. The adult interrupts access to the desired stimulus and waits for the child to emit appropriate language, or provides a verbal prompt to emit appropriate language in the given context. Access to the desired stimulus is then granted contingent upon the occurrence of the targeted language (Hart & Risley, 1975). Hart and Risley (1975) demonstrated the effectiveness of incidental teaching in their seminal study that increased the use of compound sentences by preschool age children to their teachers.

McGee et al. (1985) conducted a study that compared the effectiveness of traditional and incidental teaching procedures to teach the use of prepositions by three children with autism. During the traditional procedure, an instructor used items at a desk to demonstrate the targeted preposition and asked, “Where is the (item)?” The instructor used verbal prompts to teach correct preposition responses. During the incidental teaching procedure, items used to teach prepositions were placed upon shelves in a manner to demonstrate targeted prepositions. When the participant emitted a request for a given item on the shelf, the instructor presented the question, “Where is the (item)?” Access to the requested item was granted contingent upon responses that contained the correct preposition (e.g., “The car is under the box”). The authors found that the incidental-teaching procedure produced a greater extent of generalization across settings and stimuli, and promoted a greater likelihood of spontaneous speech.

McGee et al. (1992) also used incidental teaching to promote reciprocal peer interactions to three children with autism. In this study, the authors trained three typically developing children to serve as peer tutors for incidental teaching to the participants in a socially integrated classroom. The peer tutor was seated with a given participant in the classroom as well as a bucket containing items identified as highly preferred by the participant. The child peer tutor was taught to wait for the participant to initiate to the bucket, to provide a verbal prompt for language (e.g., “Say duck”), to provide access to the toy contingent on the occurrence of the language, and to deliver behavior-specific praise for the language response. Data were collected on the percentage of 10-s intervals within a 5-min observation period that each participant emitted a reciprocal peer interaction. During the baseline condition, the mean percentage of intervals during which the participants emitted the targeted interactions ranged from 0% to 7%. With the introduction of the incidental-teaching procedure, data increased systematically across the three participants, with the percentage of intervals during which reciprocal peer interactions occurred ranging from 13% to 35%.

Similarly, Farmer-Dougan (1994) used incidental teaching delivered by peers to increase the requesting behavior of adults with developmental disabilities. The experimenter used incidental teaching during the typically occurring lunch-making routine in the participants’ group home. When a participant attempted to gain access to one of the necessary lunch-making items, the item was placed out of reach, and the participant was prompted to emit a request for the item. Participants were granted access to the items contingent upon the appropriate request. The results demonstrated a functional relation between the number of appropriate requests and the use of the incidental-teaching procedure.

Video Modeling

Researchers have also used video-modeling procedures to teach language and social interaction skills. Video modeling is a strategy by which desirable responses are displayed through video presentation. The participant views the modeled behavior and it is anticipated that observational learning and an imitative repertoire contribute to the successful performance of the desired behavior (Bellini & Akullian, 2007). Video modeling can include peers, family members, and adults, and the medium of this technique allows for it to be practiced in virtually any context.

Charlop and Milstein (1989) used a video-modeling procedure to teach conversational skills to three children with autism. Each participant viewed a videotaped conversation three times. The conversation consisted of a series of questions and statements between two familiar adults about a given topic, such as school, toys, or sports. Subsequently, the participant was presented with the opportunity to engage in the same conversation with an instructor. Correct conversational targets included the three lines presented in the video and were reinforced with praise and edibles. If the participant did not emit the three targeted lines from the video, then the video was again presented for a single viewing, and the conversation was again attempted. This sequence repeated until the participant emitted the three targeted lines for a specified topic. The findings of the study indicated a functional relation between the video-modeling procedure and the display of the targeted conversational behavior. Moreover, performance demonstrated generalization across untrained topics, conversation recipients, and settings.

In a study by Maione and Mirenda (2006), the authors used a video-modeling procedure to teach a young boy with autism to emit peer interactions during three play activities. Three exemplars of 1-minute vignettes were recorded for each of the play activities, and depicted two adults talking to one another while playing. The participant viewed the three vignettes for a given activity 30–60 minutes prior to a play session with a peer. The authors analyzed the effect of their video-modeling procedure on the total number of verbalizations emitted by the participant, on the content of his verbalizations, including whether the verbalizations matched or varied from the model, and on the frequency of initiations and responses. During baseline, there was considerable variability in the verbalizations emitted by the participant. With the introduction of the video-modeling procedure, the authors observed an increase in social language across two of the three activities. For the third activity, the authors introduced feedback on the participant’s recorded sessions as well as verbal prompts to increase the targeted language responses.

Finally, Scattone (2008) used Social Stories ™ in conjunction with video modeling to modify the conversational skills of a boy with Asperger’s Disorder. In addition to other dependent measures such as eye contact and smiling, the authors analyzed the effectiveness of their video-modeling procedure on the percentage of 10-second intervals during which the participant emitted initiations to a conversation partner in a 5-minute session. The video-modeling intervention consisted of three different narrated Social Stories ™ that targeted eye contact, smiling, and initiations, as well as the modeling of an appropriate conversation by two adults. The authors found that the video-modeling intervention effectively increased the percentage of 10-second intervals with initiations from about 0.8% during the baseline condition to about 33% after intervention. Moreover, the authors observed generalized behavior change across untrained peers in school.

Scripts and Script Fading

Script fading is another behavior-analytic procedure that can be used to teach language and interaction skills through the process of stimulus fading. A script is a written or recorded word, phrase, or sentence that serves as a model to teach conversation. The core dependent measures when using this procedure include both scripted and unscripted interactions. Scripted interactions are defined as verbal productions that match the original scripts provided in teaching, whereas unscripted interactions are those that differ from the original scripts by more than verb tense, conjunctions, articles, prepositions, or pronouns (Krantz & McClannahan, 1993). As the participant reliably emits the scripted interactions, the scripts are systematically faded, and the number of unscripted interactions should increase correspondingly as scripted interactions decrease (McClannahan & Krantz, 2005). This procedure has proven effective in many studies aimed at teaching language and interaction skills.

In their preliminary study on the script-fading procedure, Krantz and McClannahan (1993) used written scripts to teach four children with autism to emit peer initiations during art activities. The authors provided each participant with a sheet of paper that contained a cue to talk and a list of ten different scripts pertaining to the stimuli present in the environment. The authors faded scripts by deleting words one at a time from the end of each script, until all scripts were blank. At the final fading step, only the blank sheet of paper with the cue to talk was present. The number of unscripted initiations emitted by the participants increased systematically as scripts were faded out. Moreover, responding generalized across an untrained setting and stimuli.

Krantz and McClannahan (1998) embedded Language Master© cards into activity schedules to teach three children with autism to emit initiations to adults. The authors attached written scripts, such as “Look” and “Watch me,” to photographs of objects in the activity schedules. The children were taught to obtain the object and to initiate a response to an adult by emitting the script that was attached to the pertinent photograph. Scripts were faded out by removing words one at a time from the end of each script. The authors observed a systematic increase in the number of unscripted social initiations as the scripts were faded. In addition, the authors indicated the behavior change generalized across untrained activities and in the presence of an untrained conversation recipient.

Similarly, Betz et al. (2008) also used script-fading procedures within activity schedules to teach social interactions among children with autism. The researchers presented an activity schedule to be shared between two children. The activity schedule consisted of various games for the children to play together. On each page of the schedule, a textual script was superimposed upon the game to be played. Scripts consisted of statements that would initiate the game-playing behavior, such as “Let’s play ____.” As scripts were faded, the authors observed a systematic increase in the number of unscripted social interactions emitted by the children.

Sarokoff et al. (2001) used script fading with stimuli that naturally contained textual print to teach two children with autism to engage in conversation with peers. For example, the authors used the script, “Gummi Savers© are my favorite” with a package of Gummi Savers©. The text that naturally appeared on the package was incorporated into the script. As scripts were systematically faded, the text that naturally appeared on the package remained at the final fading step. As a result, the package acquired discriminative control for engaging in conversation. The authors observed a systematic increase in the number of unscripted initiations to peers with the introduction of their script-fading procedure.

Brown et al. (2008) used a similar stimulus-fading strategy to transfer the discriminative control of social interactions from scripts to naturally occurring stimuli in the environment. The authors used a textual script-fading procedure to increase the number of social interactions among three children with autism. The authors attached textual scripts directly onto stimuli in the participants’ environment. Stimuli included items commonly found in a convenience store, such as candy bars, in a sporting-goods store, such as a football, and in a video store, such as videotapes. These stimuli were presented in a mock store display, as they might typically appear in the community. During script fading, scripts were deleted one word at a time, and were removed progressively from the training stimuli. The authors also changed the location of the scripts on the stimuli with each fading step. As scripts were removed, the environmental stimuli evoked the social interactions. This script-fading procedure produced a systematic increase in unscripted interactions. Moreover, the effects of the script-fading procedure generalized to untrained stimuli.

Finally, Wichnick-Gillis et al. (2019) used a script-fading package to teach three adolescents with autism to engage in conversation with one another during leisure activities. The activities selected including computer games, model building, and a lunch period. These activities were deemed as appropriate situations during which conversation might naturally occur for their typically developing peers. The authors superimposed five different textual scripts onto various stimuli such as a laptop computer, Lego© model building sets, and items used during lunch (e.g., utensils, cups). Scripts were pertinent to the given activity, and included statements such as, “Check out what I’m building,” or “I love this game.” The authors observed a systematic increase in the number of unscripted initiations emitted across participants as the scripts were faded. Moreover, generalization was demonstrated across all three participants while engaging in the targeted activities with their siblings.

Play Behavior

Play can enhance mutual communication with others. The American Academy of Pediatrics (2013) acknowledged the importance of play in establishing relationships with others. Play provides opportunity for children to attend to directions, resolve disputes with others, behave appropriately in the presence of others, focus on activities without consistent supervision, and collaborate (Yogman et al., 2018).

There are four types of play: object play, physical play, outdoor play, and social or pretend play. Object play occurs when a child explores an object. An infant may place an object in his mouth. An older child may use a stick as a microphone. Physical play may begin with wheels-of-the-bus, and move to playing catch. Outdoor play provides opportunity for physical exercise. Social or pretend play encourages cooperation, turn taking, and negotiations of guidelines. Imaginary play, dressing up, and make believe are included as social play (Yogman et al., 2018).

Walsh et al. (2006) reviewed “the quality of learning” in 70 classes for children ages 4–5, comparing traditional classroom curriculum with play-based enhanced curriculum. Thirty-eight of the classrooms experienced traditional curriculum that involved structured play, writing, painting, listening to stories, watching television, and song. The curriculum was subject oriented, and emphasized reading, writing, and arithmetic. Thirty-two of the classrooms experienced play-based enriched activities that allowed children the opportunity to choose from a range of play stations such as the sand corner, water station, construction area, and house corner. Activities were structured by the teacher and of shorter duration. Reading activities were conducted using big books, math using puppets and games, and writing focused on creative content as opposed to writing technique. Overall, the authors found that the enriched play-based curriculum led to more opportunities for making choices and engaging in more challenging activities, and an increase in motivation, independence, and social interaction.

Stahmer et al. (2003) identified numerous behavioral approaches used to promote play with children with autism. The following is a sampling of the research-based techniques reviewed, and is by no means an exhaustive review.

Discrete-Trial Teaching

Discrete-trial teaching is a highly structured behavioral strategy that focuses on the presentation of clear discriminative stimuli by the instructor. In play-based discrete-trial teaching, play materials are selected by the instructor, and a direction is presented. Reinforcement is delivered contingent upon the production of a correct response. Incorrect responses are followed by specified consequences. Rate of responding is controlled by the given opportunity to emit the response (Cooper et al., 2020).

Discrete-trial instruction is often presented with repetitive, one-step actions (e.g., rolling a ball; pushing a car). In two experiments conducted by Nuzzolo-Gomez et al. (2002), discrete-trial teaching was used to teach play behavior. In the first experiment, a three-year-old boy with autism was presented with a book and given the instruction to look at the book. If the child oriented toward the book, verbal praise and an edible reinforcer were provided. If the child did not orient toward the book, the instructor modeled the desired behavior and/or provided manual guidance. The results showed a functional relationship between the discrete-trial teaching procedure and an increase in orientation toward a book.

In the second experiment by Nuzzolo-Gomez et al. (2002), discrete-trial teaching was used to teach three children with autism to play with toys. Toys were presented and the instruction to play was provided. Correct play responses were following by reinforcer delivery. Incorrect responses were following by modeling and/or physical prompts. The results of the second experiment showed an increase in appropriate toy-play behavior and a decrease in dysfunctional, stereotypical hand movements.

Similarly, when a more complex play sequence is targeted, the discrete-trial procedure can focus on smaller component parts of the play behavior. For example, feeding a baby might involve holding a baby, getting a bottle, placing the bottle in the baby’s mouth, and saying “Drink milk.” Discrete-trial teaching can target the different components of the sequence in succession using verbal instruction, modeling, and/or manual guidance.

Despite the success of discrete-trial teaching, it does present several limitations. Teachings in highly structured environments: (a) do not necessarily generalize to the natural environment (Lovaas, 1977 as cited in Ingersoll & Schreibman, 2006), (b) are often dependent upon the use of continued reinforcement to maintain newly acquired play responses, and (c) are taught as isolated skills, rather than as part of a social context (Schreibman et al., 1991).

Reciprocal Imitation Training

Reciprocal imitation training is another strategy used to promote play. In a study conducted by Ingersoll and Schreibman (2006), five children with autism participated. Prior to intervention, all of the participants displayed deficits in spontaneous imitation of object use by others during play. During intervention, the instructor (a) modeled actions with familiar and novel toys, (b) imitated child vocalization and actions with toys, and (c) commented on child actions. The authors found that imitation training resulted in imitative pretend play. Play responses were also found to generalized across stimuli, locations, and adults.

Stereotyped Behavior to Increase Play

It is not uncommon for children with autism to engage in perseverative behavior. In a study conducted by Baker et al. (1998), three children with autism were selected to participate due to their engagement in perseverative behavior with particular items. For example, one child perseverated on Disney© characters. As a result, Disney© characters were incorporated into a playground game. Another child perseverated on maps. As a result, a game of tag was played on a giant outline of a map of the United States. A third child perseverated on movies and movie characters. Laminated photographs of movie characters were placed around a play area and used as “bases” in a game of tag. The authors found that the inclusion of perseverative themes into play activities resulted in an increase in social interactions during the targeted activities and generalized across novel activities.

Similarly, in a study conducted by Koegel, Fredeen, Kim, Danial, Rubinstein, and Koegel et al. (2012), social clubs were based on the perseverative interests of young adolescents with autism. For one participant, the perseverative interest in movies was transformed into a Movie Trivia Club. For the second participant, the perseverative interest in comic books and cartooning was turned into a Comic Book and Gaming Club. For the third participant, the perseverative interest in card games was turned into a Card Game Club. Prior to baseline, all participants remained socially isolated. Following the introduction of themed social clubs, social initiations and appropriate engagement with peers increased.

Pivotal Response Training

Pivotal response training (PRT) is a behavioral strategy that has been used to increase motivation to engage in play behavior. In PRT, (a) the child selects materials from a set presented by the instructor (e.g., blocks), (b) discriminative stimuli are presented by the instructor (e.g., instructor gives block to instructor and says, “What do you want to build?”), (c) maintenance tasks are interspersed among tasks targeted for acquisition, (d) modeling is used to encourage turn taking and appropriate pace of interaction (e.g., instructor begins to stack blocks), (e) natural consequences are used to reinforce correct responses (e.g., instructor gives the child the remaining blocks in an effort to reinforce correct responding), and (f) reinforcers are delivered contingent upon attempts to respond (Stahmer, 1999).

Stahmer (1995) conducted a PRT study with seven children with autism to increase play behavior. The children were presented with toys identified as desirable (e.g., the child reached for or oriented toward the toy). If the child did not engage in play, the experimenter played with the toys and modeled symbolic actions. Simplified play responses with the toys were also interspersed and modeled to increase the likelihood of variation in responding. The author found PRT effective in increasing positive responding during play. Moreover, the findings generalized across novel toys and sustained over time.

Self-Management

Self-management strategies have been used to teach children with autism to engage in independent play in the absence of a treatment provider. Self-management may rely on self-monitoring, self-evaluation of play performance, and self-reinforcement (Stahmer & Schreibman, 1992). In a study conducted by Newman et al. (2000), two, six-year-old children and one preschool child with a diagnosis of autism were taught to engage in varied play responses with preferred toys. Children were instructed to take a token contingent upon the display of a variation in play behavior. All three children were successfully able to use the self-management system and showed an increase in variability in responding.

In 1992, Stahmer and Schreibman targeted appropriate toy-play in three children with autism using a self-management training package. During training, the experimenter modeled both appropriate and inappropriate toy-play behavior and asked the question, “Is this right?” Following this training, the participants were taught to self-manage their own behavior and to mark a check in a box if they engaged in appropriate play behavior during a specified interval. Access to a reinforcer was provided contingent upon appropriate self-monitoring a toy-play behavior. Over time, the experimenter systematically faded his or her presence from the child. The results showed a functional relationship between the self-management system and an increase in appropriate play in unsupervised settings, along with a reduction in self-stimulatory behavior.

In Vivo Modeling and Play Scripts

Researchers have found that children with autism can increase play skills via observation of predictable and repeated sequences. Jahr et al. (2000) compared two in vivo modeling procedures to teach children with autism to engage in play. One procedure provided opportunity for the child to observe two models engage in a scripted cooperative play episode. Following the episode, the child took the place of one of the models and the play episode was repeated. The second procedure was the same as the previously described procedure with one exception. The child participant was taught to verbally describe the modeled play episode prior to taking the place of one of the models. The authors found that modeling, in conjunction with training with verbal description, was effective in evoking play behavior across settings, time, and play partners.

Video Modeling

Similar to in vivo modeling, video modeling allows a child with autism to observe repeated and predictable play episodes presented in a video format. In 2002, Schwandt, Pieropan, Glesne, Lundahl, Foley, and Larsson (as cited in Stahmer et al., 2003) used video modeling to teach three- and four-year-old children to engage in play for extended periods of time, to increase variation in play responses, and to emit play-related verbal statements. Each child was provided three repeated opportunities to observe play with target toys. Following the third video presentation, the child was presented with the same toys depicted in the video. Variation in toy play responses and engagement in play were differentially reinforced throughout the play activity. Over time, the videos were removed. The authors found that the children engaged in toy play responses in novel settings with novel toys. Some of the participants emitted verbal toy-related responses similar to those on the video.

Charlop-Christy et al. (2000) conducted a study designed to compare in vivo modeling with video modeling with five children with autism. After viewing the video model twice, the child was provided the instruction, “Let’s do the same, just like on TV.” After viewing the in vivo model twice, the child was provided the instruction, “Let’s do the same, just like they did.” The authors found that video modeling was effective in teaching different behavior, including independent and cooperative and social play. Although both in vivo and video modeling were effective in increasing play, video modeling led to faster acquisition and was more time effective than in vivo modeling.

Overall, the importance of play has been well established. The sampling of research reviewed suggests that there are a variety of behavioral approaches used to increase play behavior. Whether one uses discrete-trial teaching, video-modeling with play scripts, or self-management techniques, a systematic approach to the application of the teaching procedure should result in an increase in the desired play behavior.

Conclusion

Social skills are essential in creating relationships with others. Clear verbal expression of one’s self, attention to the verbal behavior of others, and cooperation and interest in games and stories with peers are just some of the necessary components of a positive relationship. Over the years, behavioral research has continued to analyze the effectiveness of strategies that teach social skills to those who display deficits. The research reviewed in this chapter demonstrates the operant nature of various components of social skills behavior. Through techniques including differential reinforcement, modeling, incidental teaching, discrete-trial teaching, script fading, and self-management, behavior analysts have been able to teach social skills responses such as eye contact, prosody, gestures, affect, language and interactions, joint attention, and play behavior. As applied behavior analysts continue to examine these functional relations, the lives of many can continue to improve and develop successfully.