Verbal behavior in young children with autism spectrum disorders at the onset of an early behavioral intervention program.
Abstract: The scope of this study was direct observation of verbal behaviors of 14 children with autism spectrum disorders at the onset of an early behavioral intervention (EBI) program delivered in a public services agency. Objectives were to (1) describe frequencies of vocal, verbal, and listener behaviors; (2) evaluate the relationship between the behaviors and the children's level of functioning (IQ and severity of autistic symptoms); and (3) describe the types of consequences provided by EBI therapists following the children's behaviors. The descriptive and statistical analysis of the data showed that 3 profiles of children were distinctively different in vocal, verbal, and listener behaviors. The results concerning the consequences contingently delivered by therapists to those behaviors show that 65% were followed by positive attention, 5% by negative attention, and 30% by no attention. Results led to recommendations for planning EBI programs, where therapists are not formally trained in verbal behavior analysis.

Key words: verbal behaviors, autism spectrum disorders, early behavioral intervention
Article Type: Report
Subject: Autistic children (Psychological aspects)
Behavioral assessment (Methods)
Children (Behavior)
Children (Research)
Authors: Melina, Rivard
Jacques, Forget
Pub Date: 03/22/2012
Publication: Name: The Psychological Record Publisher: The Psychological Record Audience: Academic Format: Magazine/Journal Subject: Psychology and mental health Copyright: COPYRIGHT 2012 The Psychological Record ISSN: 0033-2933
Issue: Date: Spring, 2012 Source Volume: 62 Source Issue: 2
Topic: Event Code: 310 Science & research Canadian Subject Form: Child behaviour; Behavioural assessment
Geographic: Geographic Scope: Canada Geographic Code: 1CANA Canada
Accession Number: 288689694
Full Text: One of the characteristic manifestations of autism spectrum disorders (ASDs) is the presence of a language deficit (Tager-Flusberg, Paul, & Lord, 2005). Some children with ASDs develop language more or less normally but have a delay in socially functional language and rarely use language to initiate social interaction (Roge, 2003; Stone & CaroMartinez, 1990). Furthermore, their speech seems lacking in spontaneity (Greer & Ross, 2004; William & Greer, 1993), and these children do not spontaneously use other appropriate methods of communication that could compensate for language limitations (Roge, 2003). As mentioned by Tager-Flusberg, Paul, and Lord (2005), even among children with ASDs who acquire some functional language, there is significant variability in their rate of progress. Nevertheless, although language deficits may vary in ASDs, there are also similarities in language limitations among individuals with ASDs, such as echolalia and palilalia (Greer & Ross, 2004).

The assessment tools used to measure language competencies, variations, and similarities among children with ASDs are mainly standardized instruments. These molar assessments present some advantages but also some disadvantages, such as a lack of sensitivity required to capture subtle improvements during treatment (Granpeesheh, Dixon, Tarbox, Kaplan, & Wilke, 2009). These tests may also not be reflective of the language behavior demonstrated by children in everyday situations, where interactions with significant people might affect their language use. Test performance may not be constructive in assisting the instructional purposes on a day-to-day basis (Greer & Ross, 2008). Another challenge in ASD language research lies in the fact that the language abilities of many preschoolers with ASDs are so poor that formal language tests cannot capture the competencies of these children or the variations among them (Charman, 2004).

Descriptive analysis and direct observation are methods that appear to provide a good ecological validity compared to evaluation through standardized testing and may therefore be more sensitive and representative of real behavior patterns (Bijou, Peterson, & Ault, 1968; Bloom, Fisher, & Orme, 2003; Cooper, Heron, & Heward, 2007). Direct measures of behavior give a very fine-grained picture of what children can do in a particular context (Granpeesheh et al., 2009). Different groups of clinicians and researchers have used Skinner's analysis of language to observe, describe, and teach the functions of verbal behaviors for children and adults with ASDs (McGreevy, 2009). As mentioned by Greer and Ross (2008), verbal behavior analysis is suited for identifying verbal developmental capabilities, creating curricula for children with or without language deficits, and providing environmental or teaching interventions to establish functional verbal repertoires.

Skinner's Verbal Behavior

In his book Verbal Behavior, Skinner (1957) defined language as a learned behavior that is established by its antecedents and its consequences in the same manner that nonverbal behavior is acquired. In examining the function of language, verbal behavior analysis focuses on the effect of the speaker on the behavior of the listener, and the effect of the listener on the behavior of the speaker, rather than on language structures alone (Greer, 2008; Sundberg & Michael, 2001). A single word may serve different functions. For example, "apple" could be vocalized to ask for food, to label an image, or to answer the question, "What is your favorite food?" The analysis of language by its functions is different from the analysis of language by its structures or parts of speech (Greer & Ross, 2008). Also, recent studies on verbal behavior show that although speaker responses (production response) and listener responses (observing response) are initially developmentally independent, their integration plays a crucial role in the successful language development of children (Greer & Speckman, 2009).

Different verbal behaviors are acquired during the course of development, each with a specific function, which Skinner termed verbal operants. Skinner initially identified five primary verbal operants: mand, tact, textual, intraverbal, and echoic. These verbal operants may occur in various topographies (e.g., sign language, logographic symbols, vocal language) and are not limited to vocal or oral language (Greer, 2008; Greer & Speckman, 2009). In the verbal behavior analysis, there are also basic listener responses that the child has to integrate to be truly verbal, such as responding to vocal verbal instructions (Greer & Speckman, 2009).

The starting point in the development of a verbal behavior intervention plan for children with ASDs is the mand (Brady, Saunders, & Spradlin, 1994; Drash, High, & Tudor, 1999; Sundberg & Michael, 2001). This functional class of verbal operant includes demands, requests, orders, instructions, questions, warnings, and appeals. Mands are controlled by establishing operations, which are environmental events, operations, or stimuli that affect the speaker by momentarily altering the reinforcing or punitive value of other events and affect the probability of eliciting a behavior in the speaker's repertoire that has already been reinforced by these other events (Hersen, 2005; Michael, 1982, 1985, 1988, 1993). The mand allows the speaker to benefit directly from his behavior (Stafford, Sundberg, & Braam, 1988). For example, a child who is hungry (an establishing operation) asks for an apple (a mand) and receives an apple (a specific reinforcer). This operation increases the probability that the child will emit a mand to obtain food when he is hungry in the future.

Unlike mands, the four other primary verbal operants are controlled by discriminative stimuli and are maintained by nonspecific consequences, such as social attention and tangible reinforcers (Skinner, 1957). Tacts refer to the use of words as labels and are controlled by natural discriminative stimuli or concrete components of the environment, such as objects or images. However, there is some evidence that there are establishing operations for tacts (Tsiouri & Greer, 2003). Tacts are reinforced by generalized reinforcers, such as attention or confirmation (Greer & Ross, 2008).

Textuals refer to reading behaviors and are controlled by symbolic discriminative stimuli, such as written words. Here, there is no formal similarity, but there is point-to-point correspondence between the discriminative stimulus and the verbal response. Textuals are under the control of printed words (Greer & Ross, 2008).

Intraverbals are controlled by the verbal behavior of other speakers. They are verbal responses related to, but not identical to, the verbal antecedents presented by another speaker. Intraverbals can occur as exchanges between two speakers or as part of a verbal chain (Greer & Ross, 2008). Conversation and responding verbally to questions are examples of intraverbals.

Echoics are also controlled by the verbal behavior of other speakers but have point-to-point correspondence and formal similarity with the verbal stimuli (Greer & Speckman, 2009; Skinner, 1957). A typical example of an appropriate echoic is when a young child repeats a word that is hard to pronounce after being prompted by her parent. An example of an inappropriate echoic is if a child repeats words of an instruction several times without actually responding to the instruction itself. Here lies the distinction between vocal verbal behavior and vocal nonverbal behavior. According to Skinner (1957), a pathological echoic is seen in echolalia, in which a bit of speech is repeated once or many times. Echolalia, defined by Greer and Ross (2008), is an inappropriate immediate repetition of a proximate verbal antecedent. Other examples of nonverbal vocal behaviors are parroting (repeating verbal stimuli) and babbling (reproducing phonemic sounds of caregivers' speech), which occur in the early stages of speaker development and automatically reinforce themselves (Greer, 2008). It is clear that vocal language is not necessarily verbal.

While Skinner (1957) emphasized that the verbal operants were independent from each other (Greer & Speckman, 2009; Lamarre & Holland, 1985), more recent research has shown that the acquisition of one operant can facilitate the acquisition of other operants (Egan & Barnes-Holmes, 2009; see Grow & Kodak, 2010). Even so, learning a specific verbal operant does not necessarily lead to learning other verbal operants (Shafer, 1994), and separate instructional operations may be required (Greer & Ross, 2008). In addition, verbal operants acquired under the control of a discriminative stimulus are not necessarily transferred to the control of establishing operations (Lamarre & Holland, 1985; Sundberg, Loeb, Hale, & Eigenheer, 2002). For example, the acquisition of listener responses, such as a response to the demand of a therapist in an early intervention program based on discrete trial teaching (e.g., S-R-C sequence), is not necessarily transferred to speaker responses that are consequents of establishing operations. Intervention programs that target these various language functions must therefore take into consideration every determining variable for a specific operant and explicitly teach those functions according to those variables. This also highlights the importance of considering a combination of early intervention strategies, such as discrete trial teaching and other applied behavior analysis (ABA) verbal behavior techniques (Fava & Strauss, 2011). As mentioned by Hall and Sundberg (1987), the first behavioral studies on language assumed that teaching the denomination of words and sentences led to appropriate, functional language. However, this method did not necessarily lead to spontaneous communication or to the initiation of interaction.

Applications of Verbal Behavior Analysis

Several authors have applied Skinner's analysis and adapted his language classification to a variety of client groups or different learning contexts. Sundberg and Partington (1998), for example, translated the elementary verbal behaviors into several operational language skills and created an evaluation and intervention tool from this model called the Assessment of Basic Language and Learning Skills (ABLLS). The ABA-Verbal Behavior Program of McGreevy (2009) is based on teaching appropriate verbal behaviors as alternatives to socially mediated problem behaviors. In this sense, language is a pivotal behavior in social interactions and can be an appropriate alternative to inappropriate behaviors, such as physical aggression toward others or self-injurious behaviors, which serve the function of communicating something to people in the environment (Vismara & Rogers, 2010).Many empirical studies have focused on the impact of functional communication training, particularly mand training, as a treatment for challenging behaviors (Hanley, Iwata, & Thompson, 2001; see McGreevy, 2009; Roane, Fisher, Sgro, Falcomata, & Pabico, 2004). A study by Chu (1998) showed that teaching children with ASDs how to mand decreased the frequencies of physical assaults. Another example of an application of Skinner's theory is the Comprehensive Application of Behavior Analysis to Schooling (CABAS) model, in which the findings from naming, relational frame theory, stimulus equivalence, developmental psychology, and verbal development research are integrated into a theory of verbal development (Greer, 2008; Greer & Speckman, 2009). CABAS provides a program of research in this field that identifies ways to join Skinner's method, incidental teaching, and other strategies based on research (Greer & Keohane, 2005; Greer & Ross, 2004, 2008).

In conclusion, the contributions and benefits of verbal behavior intervention for children with language disabilities, such as ASDs, are well documented (see Greer, 2008) and could be strategically used in combination with other forms of intervention, such as early intensive behavioral intervention (EIBI; e.g., Lovaas, 1981), in order to help a child become truly verbal.

Contributions of the Present Study

Despite the scientific and clinical scope of verbal behavior studies, many of those studies are typically interested in describing a limited number of verbal operants in Skinner's model or the effect of one specific intervention on the development of a certain function. A paper by Dixon, Small, and Rosales (2007) analyzing recent work on verbal operants revealed that empirical applications of Skinner's analysis of verbal behavior tended to focus mainly on the study of mands and tacts. However, more recent research in the field, in particular the CABAS model (Greer, 2008; Greer & Ross, 2008), has combined all the verbal operants into a more comprehensive approach to the language acquisition of children with ASDs. Yet, to our knowledge, no studies have attempted a descriptive analysis of vocal behavior in a sample of young children with ASDs to determine how the behaviors correspond to all five primary verbal operants described by Skinner and whether they are contextually appropriate or inappropriate.

As mentioned by Greer and Ross (2008), studies on EIBI (e.g., Lovaas, 1987) have shown that while this form of intervention provides effective training procedures for children with ASDs, EIBI shows fewer gains in obtaining generalization of trained language to natural and everyday settings. Though studies have shown that EIBI programs (e.g., Lovaas, 1987) lead to significant gains in children's development and learning in areas such as intellectual functioning, adaptive behaviors, independence, social skills, and academic performance (Eikeseth, Smith, Jahr, & Eldevik, 2002; Makrygianni & Reed, 2010; Ospina et al., 2008; Perry et al., 2011; Remington et al., 2007), communication behaviors are the most resistant to intervention. To our knowledge, no studies have attempted to describe the language of children with ASDs using observation of verbal behaviors during an EIBI context, to determine if verbal operants are promoted in discrete trial teaching, or to describe the listener (EIBI therapist) responses that mediate the children's verbal behavior in an EIBI context. This kind of data could help to suggest ways of integrating both discrete trial training and incidental or naturalistic (e.g., verbal behavior) language interventions. In addition, there is little known about the relationship between the acquisition of verbal operants in ASDs and the individual characteristics of children, such as intellectual functioning.

The present empirical study set out to fill these gaps in the literature on verbal behavior analysis. It presents a description of the five primary verbal behaviors in 14 young children with ASDs occurring naturally in an early behavioral intervention (EBI) context. It analyzes the relations between verbal behavior patterns and the children's various characteristics and describes the behaviors of the therapists occurring naturally and contingently to the verbal behavior of the children, as contingencies that affect the speaker behaviors of the children.

The first objective of the study was to analyze the vocal behavior of 14 children with ASDs ages 3 to 5 years at the onset of an EBI program, along the following parameters: (a) describe the frequencies of the five primary verbal operants proposed by Skinner (1957), the vocal nonverbal behaviors, and the listener behaviors; (b) determine whether the target behaviors were emitted in contextually appropriate or inappropriate ways; and (c) identify the similarities and the differences among the participants' verbal behavior patterns. The second objective was to evaluate the correlations between the vocal, verbal, and listener behaviors with the severity of the autistic profile and the children's intellectual functioning. The third objective was to describe the behaviors of the therapists occurring naturally and contingently to verbal, vocal, and listener behaviors of the children during the community-based EBI program.


Research Design

This was an observational and single case study. The target behaviors of the participants were observed in their naturally occurring EBI contexts. There were no manipulations or modification of the intervention program of the public developmental services agency for the study. The EBI therapists were not given instructions with regard to the research project and did not know the specific objectives of the study. Intraindividual analyses of the observational data, post hoc statistical analyses of group data, and correlations were used.

Participant Characteristics

The 14 children participating in this study were starting an EBI program delivered a by a public developmental service agency in a French province of Canada (Quebec). The language of instruction for all participants in the program was French. Each participant had a prior diagnosis of ASD, and a multidisciplinary team that included a child psychiatrist confirmed each diagnosis. All participants were being served by the same developmental service agency, in the same EI31 program. None of the children had received any language intervention or other intervention services prior to the study. The participants were between 3 and 5 years of age at the onset of the study. Table 1 gives each participant's age; gender; Wechsler Preschool and Primary Scale of Intelligence Full Scale IQ (FSIQ), Verbal IQ (VIQ), and Performance IQ (PIQ); and scores on three indices of severity of autistic symptoms, from the Childhood Autism Rating Scale, the Gilliam Autism Rating Scale, and the Gilliam Asperger Disorder Scale. All children starting the EBI program at the developmental service agency in September 2007 were solicited for this study, with 87.5% acceptance.


The community-based EBI program consisted of one-on-one (one child to one therapist ratio) therapy in each participant's public kindergarten. Each participant had two therapists during the year of the EBI program: a principal therapist who gave two thirds of the sessions and another therapist who gave one third of the sessions. The number of hours of intervention per week per participant varied from 10.5 to 17.38 hr, with a mean of 14.46 hr.

Observation sessions for the present study were carried out during the first 7 weeks of the EBI program and took place in each participant's EBI room: a small room containing a child-sized table and chair and few stimuli other than the material used for the intervention. During these sessions, the therapist worked one-on-one with the participant, with the exception of the presence of a research observer with a handheld camera in the corner of the room. This observer did not interact with the therapist or the participant during the observation sessions.

The structured activities used by the therapists with the participants during the observa-tion periods were those recommended in the developmental service agency's EBI program. The basic structure of the program was based on the intensive behavioral intervention (IBI) curriculum of Lovaas (1987) and principally focused on discrete trial training. The term IBI is not used in the present study because the researchers could not control the various IBI parameters, namely the intensity of teaching. The EBI program followed the criteria of the National Research Council (2001) for early intervention programs: an early start, active engagement of the child, use of a structured curriculum that includes specific objectives, intervention priority on direct teaching of basic skills, procedures for generalizing learning, individualized intervention, low therapist child ratio, parental involvement, specialized therapist training, and ongoing assessment of the child's progress. The therapists were not formally trained in applied verbal behavior analysis, such as in the programs of Greer and Ross (2008), McGreevy (2009), or Sundberg and Partington (1998).

Two research ethics committees evaluated this study: the University Research Ethics Board at the researchers' university (University du Quebec a Montreal) and the Joint Research Evaluation Committee at the public developmental services agency in Quebec, the Canadian province where the research was carried out.


Observations of vocal, verbal, and listener behaviors. The target behaviors were selected via a direct observation grid created and validated for the current study. A committee of five experts in ASDs, language, and ABA completed a content analysis and validation of the grid, which was trialed on the videotaped observations of nine children with ASDs (not the participants in the present study) in their family settings. This process enabled the researchers to identify 13 target behaviors for the participants of the current study and nine target behaviors for the therapists, which served as antecedents or consequences for the participants' behavior.

The target verbal operants for each participant as a speaker were mand, tact, intraverbal, echoic, and textual. Each mand, tact, intraverbal, or textual operant was coded as either contextually appropriate or contextually inappropriate. Contextually appropriate verbal behavior occurred when the participant's speech was socially appropriate (e.g., the child manded for milk during an activity by saying, "Burrhus, I want milk") and/or was emitted in a correct manner (e.g., the participant said "bear" when he saw a bear in his book). Contextually inappropriate verbal behavior occurred when the participant's speech was socially inappropriate (e.g., if the participant manded "stop the activity" by pushing the therapist) and/or was emitted in an incorrect way (e.g., the participant said "cat" when she saw a snake). Two other kinds of vocal behaviors that do not correspond to verbal functions were also targeted: babbling and vocal stereotypy. In order to record each participant's capacity to understand the verbal instructions/demands of her therapist and her ability to respond as a listener, two forms of listener behaviors were also targeted: appropriate nonverbal response to a request and inappropriate nonverbal response to a request. Table 2 presents and briefly defines the 13 target behaviors for the participants.

It is important to note that in the case of the mand, the observation system did not include the distinction between pure and impure mands. In the context of this observational study, the researchers had no control over establishing operations, and some of these variables were difficult to measure (e.g., thirst, hunger). To make the distinction between mands and other operants, the observers noted the antecedents and consequences for each behavior, recorded all words or sentences emitted, and then made a decision about which code should be accorded.

The nine target behaviors for the therapists as listeners and speakers (i.e., the antecedents and consequences of the participants' behaviors) were request/demand (D: the therapist made a request to the participant to obtain a verbal or nonverbal response; e.g., giving instructions, asking the participant to sit on her chair, asking a question); environment (E: a change in environment preceded or followed the participant's target behavior, or the participant's target behavior generated a change in the environment; e.g., a fork fell on the floor before the child said, "The fork fell"); verbal reprimand (Rv: the therapist reprimanded the participant or made a negative comment about what the participant was doing; e.g., the therapist raised his voice when the participant made a mistake, the therapist threat-ened to withdraw a toy if the child would not stop screaming); withdrawal of privilege (Rp: the therapist withdrew something [e.g., a toy, food, an activity] because the participant was doing something particular. This category was also used when the participant was in a time-out); no attention (: when the participant engaged in a target behavior, the therapist gave no attention to the child before or after the behavior); social reinforcer (Rs: before or after the participant's target behavior or in response to a request, the therapist gave her a social reinforcer, such as smiling, a hug, or a positive comment); material reinforcer (Rm: before or after the participant's target behavior or in response to a request, the therapist gave him something that was known to be an object of reinforcement, such as food, a toy, or a token); verbal behavior (Yb: the therapist engaged in verbal behavior other than Rs or Rv before or after the child's target behavior; e.g., the participant asked about the next activity and the therapist responded, the therapist made a comment about what she saw in a book and the participant agreed).

Seven 60-min sessions of continuous observation were completed for 13 of the participants, and six were carried out for one participant (3). Each participant was directly observed for a total of 420 min (360 for Participant 3) by the same observer. All observation sessions were videotaped and were carried out with the same conditions for each participant--the time of the observation, the ratio of people present in the setting, the day of the week, the context and the structure of the observation environment, and the size of the recording window. During the observation sessions, each participant's behaviors and corresponding antecedents and consequences were noted by the observer. All the words, sentences, and sounds (frequencies and latencies) were detailed in the direct observation grid to help the observers properly code the responses and clearly distinguish the verbal functions according to their operational definitions.

Seventeen undergraduate psychology students functioned as research assistants who carried out the direct observations. These assistants underwent a 60-hr training program that included theoretical courses, a reading program, oral exams, and practices with numerous videos of children with ASDs in interaction with their caregivers. Their training was considered successful when they achieved an interobserver agreement of at least 85%. The principal researcher supervised their work frequently, based on the needs of the individual assistants. The interobserver agreement was established through observations of the videotaped sessions and performed for 14.3% (60 min/420 min) of the observation sessions for 13 children and for 16.7% (60 min/360 min) for one child. The range of interobserver agreement was between 85% and 95% (percentage agreement = agreements/[agreements + disagreements] x 100).

Assessment of autistic severity and intellectual functioning. To obtain standardized measures of the participants' autistic severity, three tools were used. The Childhood Autism Rating Scale (CARS; Schopler, Reichler, & Rochen Renner, 1988) assesses the child's severity level on a continuum of "non-autistic" to "severely autistic." It includes 15 items assessed on a 4-point scale from 1 (no sign) to 4 (severe sign) based on the frequency and intensity of the behaviors. These items are social relations; imitation; emotional responses; use of body; use of objects; adaptation to change; visual responses; auditory responses; taste, smell, and touch; fear and anxiety; verbal communication; nonverbal communication; level of activity; intellectual functioning; and general impression. The Gilliam Autism Rating Scale (GARS; Gilliam, 1995) assigns a level of autistic severity based on the evaluation of stereotypical behaviors, communication, social interactions, and development. This tool has 42 items, and each is rated on a scale of 0 (never observed) to 3 (frequently observed). The Gilliam Asperger Disorder Scale (GADS; Gilliam, 2001) was used to validate the ASD diagnosis in children who did not score high on the CARS or the GARS. This test covers social interactions, restricted behavior patterns, cognitive patterns, and pragmatic skills. It is composed of 38 items, including several questions that the observer must answer on a scale of 0 (never observed) to 3 (frequently observed). The four research assistants (graduate psychology students) who carried out these tests received theoretical and practical training as well as supervision. In addition, these assistants administered the tests after having observed the participants for seven different 1-hr periods.

The Wechsler Preschool and Primary Scale of Intelligence (WPPSI III; Wechsler, 2002) was used to evaluate the participants' intellectual functioning. Three of the WPPSI-III indices were retained: Full Scale IQ (FSIQ), Verbal IQ (VIQ), and Performance IQ (PIQ). This intellectual evaluation was completed for 11 participants only, because of parental refusal for three of the participants to take the test. The WPPSI III testing was performed by the same assistant for all the participants. This assistant was a graduate stu-dent in psychology and an ABA therapist. She received thorough training and was extensively supervised by the principal researcher (psychologist).


Objective 1

The frequencies of target behaviors for each participant across the seven observation sessions (six for Participant 3) are presented in Figure 1. The visual analysis reveals that there are intraindividual and interindividual differences in the type and frequency of the target behaviors emitted.

A cursory analysis of the graphs included in Figure 1 reveals that most participants emitted more appropriate than inappropriate target behaviors. This was not the case, however, for Participants 1, 6, 7, and 11. These four participants emitted more inappropriate behaviors, including babbling and vocal stereotypy, than appropriate behaviors, and they presented a higher rate of inappropriate behaviors in comparison to the other participants in the study. For these four participants, the mean of inappropriate behaviors for 1 hr of observation was higher than the group mean of 161.45. These participants also emitted more vocal stereotypy, and for three of them (6, 7, and 11), vocal stereotypy was the most frequently coded type of behavior. For Participant 1, vocal stereotypy was the third most frequent behavior, after babbling and appropriate nonverbal response to a demand. Five participants (3, 4, 5, 8, and 13) emitted more appropriate target behaviors than their peers. These five participants had a higher mean frequency of appropriate behaviors for 1 hr of observation than the group mean of 278.7. Participants 7 and 14 also had a higher mean of appropriate behavior than the group mean, but these participants were not included with the previous five because they also presented a high number of inappropriate behaviors. Of the five participants who emitted more appropriate behaviors, four (Participants 3, 4, 5, and 13) emitted more appropriate mands and appropriate intraverbals than others. For two of these participants (4, 5), appropriate intraverbals were the most frequent behavior, and for two of them (3, 13), this behavior was the second most frequent behavior after appropriate nonverbal responses to a request. For Participant 8, appropriate response to a demand was the most coded behavior, followed by echoics. In general, these five participants emitted few vocal stereotypies.

Figure 2 divides the participants into three groups, based on frequency of appropriate and inappropriate target behaviors: (1) those who emitted more inappropriate behaviors (IVB group; Participants 1, 6, 7, and 11), (2) those with a mixed pattern of behaviors in the middle of the continuum (IVB-AVB group; Participants 2, 9, 10, 12, and 14), and (3) and those who emitted more appropriate target behaviors (AVB group; Participants 3, 4, 5, 8, and 13).


A post hoc analysis of variance (ANOVA) across the three groups was completed to corroborate whether a significant difference existed between the frequencies of appropriate and inappropriate target behaviors. The data confirm that there were significant differences between the three groups. The difference in the mean appropriate behaviors for the three groups was P(2, 11) = 14.48, p = .05. The q2 of 72% indicated a major difference between these groups. The contrast showed a linear trend that explained 77% of the effect but also a quadratic trend that explained 23% of the effect. The AVB group was significantly different from the other two groups in terms of the frequency of appropriate behaviors. The data also showed that there was a significant difference between the IVB and IVB AVB groups and between the IVB and AVB groups based on the frequency of inappropriate behaviors but not between the IVB AVB and AVB groups. The difference in the mean of inappropriate behaviors for the three groups was F(2, 11) = 11.16, p = .05. The q' of 67% showed a major difference between the groups. The contrast showed a linear trend that explained 84% of the effect but also a quadratic trend that explained 16% of the effect.

Figures 3 and 4 show the mean frequencies of the three groups for all of the target behaviors: (1) the IVB group presented more vocal stereotypy and inappropriate nonverbal responses to a request, (2) the IVB AVB group presented fewer behaviors in general and seemed to emit less extreme behavior frequencies or were more in the middle of the continuum, and (3) the AVB group emitted the most appropriate intraverbals, mands, and tacts. Some behaviors (i.e., echoic, appropriate textual, appropriate response to a request, inappropriate mand, inappropriate textual, inappropriate intraverbal, inappropriate tact, and babbling) seemed to be relatively equivalent in all three groups. A post hoc ANOVA across the groups showed that there were no significant differences for those eight behaviors. Nevertheless, post hoc ANOVA data showed that there were some differences between the groups for the other five behaviors. The AVB group presented significantly more appropriate mands, intraverbals, and tacts than did the two others. The difference in the mean of appropriate mands was F(2, 11) = 4.39, p = .05. The of 44% showed a major difference among the groups. The contrasts showed a linear trend that explained 67% of the effect. The difference in the mean of appropriate intraverbal was F(2, 11) = 6.11, p = .05. The [[eta].sup.2] of 35% showed a difference among the groups. The contrasts showed a linear trend that explained 66% of the effect. The difference in the mean of appropriate tact was F(2, 11) -L.--7.55, p = .05. The [[eta].sup.2] of 58% showed a difference among the groups. The contrasts showed a linear trend that explained 67% of the effect. The IVB group presented significantly more vocal stereotypy and inappropriate responses to a request than did the other two groups. The difference in the mean for vocal stereotypy was F(2, 11) = 9.11, p = .05. The [[eta].sup.2] of 62% showed a major difference among the groups. The contrasts showed a linear trend that explained 87% of the effect. The difference in the mean for inappropriate response was F(2, 11) = 5.09, p = .05. The of 48% showed a major difference among the groups. The contrasts showed a linear trend that explained 94% of the effect.



Cumulative basic descriptive statistics for each target behavior were calculated and are described in Table 3. The mean frequency per 60-min period, the standard deviations, and the percentage of each target behavior are shown in Table 3, as are the means and standard deviations for each target behavior. The most frequent behavior was appropriate nonverbal response to a request; it represented 30.7% of the total interactions between the participants and the therapists. This behavior was followed by appropriate intraverbal response, which represented 15.6% of the target behaviors, vocal stereotypy (14.3%), babbling (13.6%), and echoic behavior (10.7%). The other target behaviors each represented less than 10% of the behaviors observed, and the least frequent were textual (appropriate, 0.17%, and inappropriate, 0.03%), inappropriate tact (1.1%), inappropriate intraverbal (1.7%), and inappropriate mand (2.6%). The behaviors that varied the most among the participants were appropriate intraverbal and vocal stereotypy.

Objective 2

The second objective of this study was to evaluate whether the frequencies of verbal, vocal, and listener behaviors of the participants were related to their level of functioning. Correlations, presented in Table 4, were calculated between the frequency of the target behaviors and the CARS, GARS, and WPPSI III results. The analysis showed that the frequency of appropriate mands was positively correlated with the three IQ indices and negatively correlated with the two autistic profile severity indices. The frequency of appropriate intraverbals was also positively correlated with the three IQ indices, negatively correlated with the two autistic profile severity indices, and positively correlated with the frequency of appropriate mands (r = .71, p < .5). The frequency of inappropriate intraverbals was negatively correlated with the severity of the autistic profile. The frequency of appropriate tacts was positively correlated with Full Scale IQ and Performance IQ. The frequency of inappropriate tacts was positively correlated with Full Scale IQ and Performance IQ and negatively correlated with the severity of the autistic profile. The frequency of inappropriate nonverbal responses to a request was negatively correlated with the three IQ indices and positively correlated with the severity of the autistic profile and the frequency of inappropriate mands (r = .55, p < .5) and vocal stereotypies (r = .57, p < .5). Inappropriate mands, echoics, appropriate textuals, babbling, vocal stereotypies, and appropriate nonverbal responses to a request have no significant link with the IQ indices and the measures of severity of autistic profile. The total frequency of appropriate behaviors was significantly and positively linked to the three IQ indices and negatively linked to the two measures of severity of autistic profile.

Objective 3

A central focus of this study concerned how often the participants in each group engaged in one of the target behaviors and obtained consequences from the therapist for doing so. Table 5 presents the mean frequency of target consequences for all participants and for all their target behaviors. This table also shows the relative percentage of a target consequence within all consequences noted in the observations. The data show that on average 30% of the verbal behaviors emitted by the participants were not followed by attention from the therapist. The most frequent consequences of the participants' verbal behaviors were requests from the therapist (22.2%), social reinforcers (20.3%), and verbal behaviors (19.9%). The other types of consequences each represented less than 5% of the total number of consequences noted during the observations. Table 5 also presents the correlations between the delivered consequences and the three intellectual functioning indices as evaluated by the WPPSI III. The use of requests and verbal behaviors was positively correlated with a high score on these three indexes. The use of verbal reprimands was negatively correlated with the Verbal IQ index.

Table 6 presents the mean target consequences per targest behavior, for all participants. It shows the principal consequences for each category of a participant's target behaviors. Appropriate mands were followed mainly by verbal behaviors (43.72%), requests from the therapist (24.64%), or lack of attention (17.2%). Inappropriate mands were followed by no attention in 35.7% of the cases, by requests in 31.05%, and by verbal behaviors in 21%. Echoics were followed by no attention 37.98% of the time, by social reinforcers 22.2% of the time, by requests 20.7% of the time, and by verbal behaviors 15.6% of the time. Appropriate and inappropriate textuals were very rarely emitted, but when they were, they were usually followed by no attention. Appropriate intraverbals were followed by verbal behavior from the adult 36.9% of the time, by requests 28.9% of the time, by a lack of attention 19.9% of the time, and by social reinforcers 11.9% of the time. Inappropriate intraverbals were followed mainly by the following consequences: requests (38.9%), lack of attention (31.3%), and verbal behaviors (26.6%). Appropriate tacts were followed by social reinforcers (39.1%), requests (27.5%), verbal behaviors (16.7%), and lack of attention (11.7%). Inappropriate tacts were followed by requests (45.6%), verbal behaviors (32.73%), and lack of attention (12%). Babbling behaviors were mainly followed by lack of attention from the therapist (42.6%), requests from the therapist (22.8%), verbal behaviors from the therapist (21%), and social reinforcers (9.5%). Vocal stereotypy was followed by a lack of attention from the therapist (82.1%), verbal behaviors from the therapist (6.8%), and requests from the therapist (5.2%). Appropriate responses to the therapist's requests were mainly followed by social reinforcers (37.5%), requests (21.3%), lack of attention (18%), verbal behaviors (11.5%), and material reinforcers (5.9%). Though this response form was followed by a verbal reprimand or a withdrawal of privilege only 5.3% of the time, it was the target behavior most likely to elicit these types of consequences compared to the other target behaviors. Inappropriate responses to the therapist's requests were mainly followed by requests (40.2%), lack of attention (30.8%), and verbal behaviors (13.6%).


EIBI is a well-established treatment form for children with ASDs (Granpeesheh et al., 2009), yet it presents some limitations in the generalization of language learned in the discrete trial teaching context to regular and everyday settings (Greer & Ross, 2008). Verbal behavior analysis has become a dominant intervention form to promote language development in people with ASDs (Vismara & Rogers, 2010), but there are no empirical studies to show if the contributions of ABA verbal behavior techniques of teaching in those programs can optimize the language gains in children. However, a recent study of Fava and Strauss (2011), in Italy, shows that "it is possible to make further gains by comprehending several key elements of effective ABA programs and by including elements of verbal behavior analysis as a basis for assessment and intervention programs" (p. 520). Fava and Strauss concluded that one key component shared by effective EIBI programs is the combination of discrete trial teaching and natural environment teaching. The current study sought to empirically observe if the units of language described in verbal behavior analysis, verbal operants, were emitted and promoted in an EIBI-based program that focused on discrete trial teaching.

Individual analyses of the vocal, verbal, and listener responses of the 14 participants in this study showed that in general contextually appropriate behaviors related to language were emitted more frequently than inappropriate ones, even at a young age and at the onset of an early intervention program. Four participants who emitted inappropriate behaviors more frequently than appropriate ones were exceptions to this observation. These participants also presented a lower level of functioning, more severe autistic symptoms, and more vocal stereotypy. The data reveal the existence of three separate groups of participants, based on verbal behavior patterns. One group of children (IVB), made up of those four participants, demonstrated a high frequency of inappropriate verbal behaviors, vocal stereotypy, and inappropriate responses to requests, and a lower frequency of appropriate verbal behaviors. A second group (IVB AVB) emitted fewer verbal behaviors and could be considered to be more verbally "passive." The third group (AVB) emitted a high frequency of appropriate verbal behaviors, especially mands, intraverbals, and tacts, and a relatively lower frequency of inappropriate verbal behaviors. This group also demonstrated a higher profile (intellectual functioning, less severe autistic symptoms). Thus, the appropriate emission of target verbal behaviors was differentially related to other measures of the autistic profile of the partici-pants, according to the three different groups. The existence of three distinctive groups at the onset of this specific program demonstrates how evaluating the emission of verbal operants could be useful for assessing the prognosis of children with ASDs and could help guide the creation of individualized intervention goals.

The analysis of the verbal behaviors of the participants as a whole group showed that an appropriate nonverbal response to the therapist's request was the most common behavior (30.7%). This result could have been influenced by the fact that the intervention context was focused on discrete trial teaching, which may promote mand compliance but not necessarily the verbal functions as a speaker (McGreevy, 2009). The second most frequent whole-group behavior was appropriate intraverbal behavior (15.6%). This result was influenced by the means of intraverbal behaviors from the AVB group, which emitted this behavior frequently. In the two other groups, intraverbal behaviors were rarely emitted. Furthermore, the standard deviation for this type of behavior was very pronounced. The third most frequent behavior was vocal stereotypy (14.3%). These data were also influenced by one group the--IVB group--and there was a wide variation among the participants in the emission of vocal stereotypy. The two other most commonly emitted behaviors were babbling (13.6%) and echoics (10.7%). In short, the most commonly emitted behaviors for all participants were response to a request, babbling, and echoics. Appropriate mands and tacts appeared to be less frequent for all participants. These results suggest that the acquisition of effective mands and tacts requires well-planned, specific interventions and should be a part of the intervention goals.

There is little known about the relationship between the acquisition of verbal operants in children with ASDs and the individual characteristics of the children. The second objective of the study was to evaluate the correlations between verbal behavior patterns and both the severity of the autistic profile and intellectual functioning. As discussed earlier; some behaviors seemed to be associated with the participant's level of functioning. A high-level profile, demonstrated by high results on intelligence tests and low results in terms of the severity of autistic symptoms, was related to a higher frequency of appropriate mands, intraverbals, and tacts, and to all appropriate behaviors regrouped together. This profile was also linked to a lower occurrence of inappropriate responses to requests and to a higher frequency of inappropriate tacts and intraverbals. A profile with a lower level of functioning, measured by low scores on the intelligence test and high scores on autism symptom severity indices, presented the inverse correlations with these various behaviors.

These data suggest that identifying vocal responses that were truly verbal, from the perspective of verbal behavior analysis, was a key component in measuring the children's functioning and severity of autism. As such, the data give supplemental support for the existence of relations between verbal operants and the profiles of children with ASDs. Furthermore, these data highlight the importance of the initial assessment of the profile of the children and the direct observation of their verbal behavior patterns before intervention, in order to allow for a rigorously planned and individualized intervention to promote functional language.

This study also sought to describe how therapists in this specific program using discrete trial teaching, untrained in verbal behavioral analysis, manage their own behaviors and attention as consequences to the verbal behaviors of children. The researchers were interested in empirically observing the therapists' responses to evaluate if they promoted the verbal functions described by Skinner (1957). The therapists' responses were observed occurring naturally and contingently to verbal behaviors of the participants in the discrete trial teaching sessions. The data reveal that the majority of verbal behaviors (65%) were followed by positive attention (e.g., social reinforcers, verbal behaviors) from the therapist, while a fraction of behaviors were followed by negative attention (e.g., verbal reprimand, withdrawal of a privilege; 5%). The data also show that 30% of the participants' verbal behaviors received no attention or consequences from the therapists approximately 20% of appropriate mands and intraverbals and 15% of appropriate tacts and responses to requests were not followed by attention from the therapists. These results suggest that this specific program would benefit from including functional language goals, to ensure that therapists are responding to children's appropriate use of functional language. Furthermore, functional language goals could be easily targeted in other learning contexts, such as at home or day care, in order to reinforce such important verbal behaviors. The data also show that most of the behaviors were followed by nonspecific consequences such as social reinforcers, requests, or verbal behaviors. Skinner's theoretical model stipulates that mands are motivated by the speaker's establishing operations, which specify the nature of the desired consequence, that is, a specific reinforcer. In this study, the specific reinforcer was labeled as the "material reinforcer" consequence. The data show, however, that this type of consequence follows appropriate mands only 12.2% of the time. As mentioned by Goetz, Schuler, and Sailer (1983), a verbal response produced by a child must be followed by an immediate and logical consequence in a natural interaction between the learner and the environment in order to transfer the control of the response from the discriminative stimulus (instruction) to naturally occurring events. Given the importance of mands in communicating the wants and needs of children and the impact of this early intervention program on the reduction of other nonfunctional behaviors such as self-injury and physical aggression toward others (McGreevy, 2009), the program would benefit from placing greater emphasis on specific reinforcers for mands in order to appropriately reinforce them and thereby increase their frequency. This would involve contriving occasions where the child must request preferred objects and activities in order to gain access to them. These results concerning therapist behavior also support the idea of including an initial assessment of verbal behavior of children before beginning an EIBI program, to ensure that the program is planned in a way that improves those behaviors by providing the relevant educational consequences. Many ABA verbal behavior teaching techniques have been proposed in the literature, such as interrupted behavior chains, brief deprivation of preferred items, or speaker immersion (i.e., where the child must emit vocal verbal behaviors; Goetz, Schuler, & Sailer, 1983; Greer & Ross, 2008), which could be combined with discrete trial teaching. More recent studies of verbal behavior intervention have highlighted new key areas of focus, such as that tact may be the critical verbal operant in connecting listener and speaker responses (Greer & Speckman, 2009).

This study does present certain limitations. One major limitation is the small sample size, which may limit the generalization of the results. The study has to be extended to corroborate the results, particularly those of the three different verbal behavior/functioning groups. Another pitfall was the thoroughness and time required to train the research staff on the use of the observation grid. These concepts were fairly complex for psychology undergraduates, which increased the need for and length of training and the level of supervision required. Another limitation was the duration required for the scoring of observational data, which was between 4 and 8 hr for every 60-min observation. The total data scoring time was between 28 and 56 hr for each child and between 392 and 784 hr for all the children together. After scoring, the statistical analysis of the data required approximately 10 hr per child. This investment of time and research staff resources may limit the possibility of replicating the study. Two other limitations may be that there was no distinction between the pure and impure verbal operants and no control of the opportunities for verbal behaviors or the antecedents such as establishing operations. Pure operants are controlled by one variable, such as an establishing operation (e.g., the child is hungry). Impure operants are those controlled by more than one variable, such as an establishing operation and a question (e.g., the child is thirsty, and the therapist asks him what he would like to drink). However, by standardizing the event settings and the observation conditions, we also standardized the antecedent variables, which may have played a role in the occurrence and frequency of the behaviors (e.g., always the same hour of intervention in the participants' daily routine). Finally, interobserver agreement data were collected for only 14.3% of sessions, due to the limited access to human resources and the time required for observations. The standard is to collect interobserver agreement on 30% of the data.

One contribution of this study was the breadth of the empirical analysis on all primary verbal behaviors proposed in Skinner's (1957) model. To our knowledge, this study is the first that has analyzed the five basic verbal operants in an EBI context to determine if they were emitted in contextually appropriate or inappropriate ways, and to perform a comprehensive intraindividual study of vocal, verbal, and listener behaviors in children with ASDs being exposed to EBI intervention. Another of the study's contributions was to propose a French tool for direct observation of verbal operants in young children with ASDs, which is significant for French-speaking autism communities, since very little work is available in French on ASDs and ABA. The study presented interesting clinical observations of the relationships between children's profiles and verbal patterns. The importance of direct observation of verbal behaviors in the initial assessment and of including specific verbal intervention components in EBI programs is highlighted. The study suggests intervention options with regard to consequences emitted following different types of verbal operants, to promote their acquisition and maintenance. Finally, the study questions the place of language in EBI programs and proposes that therapists should be trained in verbal behavior analysis.

In conclusion, the integration of verbal behavior analysis in EIBI programs may facilitate the acquisition of language in children with ASDs. As mentioned by Greer and Ross (2008), the role of applied behavior analysts "is to select the appropriate research-based tactic needed at a given time for the child from the growing research base" (p. 6). The current study was realized in a French-speaking community, where the autism programs are new and largely based on discrete trial teaching without verbal behavior components. The present study has demonstrated that discrete trial teaching and other teaching techniques proposed by verbal behavior analysis could be integrated to respond to this goal, described by Greer and Ross (2008), and to optimize gains in language for children. A follow-up study was conducted to observe the progression of the verbal behaviors of the same participants, based on the three different verbal behavior groups, after several months of EBI intervention (Rivard, Forget, Kerr, Regli, & Giroux, 2011).


BIJOU, S. W., PETERSON, R. F., & AULT, M. H. (1968). A method to integrate descriptive and experimental field studies at the level of data and empirical concepts. Journal of Applied Behavior Analysis, 1, 175-191.

BLOOM, M., FISHER, J., & ORME, J. G. (2003). Evaluating practice: Guidelines for the accountable professional (4th ed.). Boston, MA: Allyn & Bacon, Pearson Education, Inc.

BRADY. N. C., SAUNDERS, K. J., & SPRADLIN, J. E. (1994). A conceptual analysis of request teaching procedures for individuals with severely limited verbal repertoires. The Analysis of Verbal Behavior, 12, 43-52.

CHARMAN, T. (2004). Matching preschool children with autism spectrum disorders and comparison children for language ability: Methodological challenges. Journal of Autism and Developmental Disorders, 34, 59-64.

CHU, H. C. (1998). A comparison of verbal behavior and social skills approaches for development of social interaction skills and concurrent reduction of aberrant behaviors of children with developmental disabilities in the context of matching theory (Doctoral dissertation). Available from UMI Proquest Digital Dissertations. (Item AAT 9838900)

COOPER, J. O., HERON, T. E., & HEWARD, W. L. (2007). Applied behavior analysis (2nd ed.). Upper Saddle River, NJ: Prentice-Hall, Inc.

DIXON, M. R., SMALL, S. L., & ROSALES, R. (2007). Extended analysis of empirical citations with Skinner's verbal behavior: 1984-2004. The Behavior Analyst, 30, 197-209.

DR ASH, P. W., HIGH, J. W., & TUDOR, R. M. (1999). Using mand training to establish an echoic repertoire in young children with autism. The Analysis of Verbal Behavior, 16, 29-44.

EGAN, C. E., & BARNES-HOLLMES, D. (2009). Emergence of tacts following mand training in young children with autism. Journal of Applied Behavior Analysis, 42, 691-696.

EIKESETH, S., SMITH, T., JAHR, E., & ELDEVIK, S. (2002). Intensive behavioral treatment at school for 4 to 7 year old children with autism: A 1 year comparison controlled study. Behavior Modifcation, 26, 49-68.

FAVA, L., & STRAUSS, K. (2011). Cross-setting complementary staff-and parent-mediated early intensive behavioral intervention for young children with autism: A research-based comprehensive approach. Research in Autism Spectrum Disorders, 5, 512-522.

GILLIAM. J. E. (1995). Gilliam autism rating scale. Austin, TX: PRO-ED.

GILLIAM, J. E. (2001). Gilliam asperger disorder scale. Austin, TX: PRO-ED.

GOETZ, L., SCHULER, A., & SAILER, W. (1983). Motivational considerations in teaching language to severely handicapped students. In M. Hersen, V. van Hassett, & I Matson (Eds.), Behavior therapy for the developmentally and physically disabled (pp. 57-77). New York, NY: Academic Press.

GRANPEESHEH, D., DIXON, D. R., TARBOX, J., KAPLAN, A. M., & WILKE, A. E. (2009). The effects of age and treatment intensity on behavioral intervention outcomes for children with autism spectrum disorders. Research in Autism Spectrum Disorders, 3, 1014-1022.

GREER, R. D. (2008). The ontogenetic selection of verbal capabilities: Contributions of Skinner's verbal behaviour theory to a more comprehensive understanding of language. International Journal of Psychological Therapy, 3, 363-386.

GREER, R. a, & KEOHANE, D. D. (2005). The evolutions of verbal behavior in children. Behavioral Development Bulletin, 1, 31-47.

GREER, R. D., & ROSS, D. E. (2004). Verbal behavior analysis: A program of research in the induction and expansion of complex verbal behavior. Journal of Early Intensive Behavioral Intervention, 1, 141-165.

GREER, R. D., & ROSS, D. E. (2008). Verbal behavior analysis: Developing and expanding verbal capabilities in children with language delays. Boston, MA: Allyn & Bacon/Merrill.

GREER, R. D., & SPECKMAN, J. M. (2009). The integration of speaker and listener responses: A theory of verbal development. The Psychological Record, 59, 449-488.

GROW, L. L., & KODAK, T. (2010). Recent research on emergent verbal behavior: Clinical applications and future directions. Journal of Applied Behavior Analysis, 43, 775-778.

HALL, G., & SUNDBERG, M. L. (1987). Teaching mands by manipulating conditioned establishing operations. The Analysis of Verbal Behavior, 5, 41-53.

HANLEY, G. P., IWATA, B. A., & THOMPSON, R. (2001). Reinforcement schedule thinning following treatment with functional communication training. Journal of Applied Behavior Analysis, 34, 17-38.

HERSEN, M. (2005). Encyclopedia of behavior modification and cognitive behavior therapy: Vol. Ill. Educational applications. New York, NY: Sage.

LAMARRE, J., & HOLLAND, J. (1985). The functional independance of mands and tacts. Journal of the Experimental Analysis of Behavior, 43, 5-19.

LOVAAS, O. i. (1987). Behavioral treatment and normal educational and intellectual functioning in young autistic children. Journal of Consulting and Clinical Psychology, 55, 3-9.

MAKRYGIANNI, M. K., & REED, P. (2010). A meta-analytic review of the effectiveness of behavioural early intervention programs for children with autism spectrum disorders. Research in Autism Spectrum Disorders, 4, 577-593.

MCGREEVY, P. (2009). Teaching verbal behavior to children and adults with developmental disabilities including autism spectrum disorders. In P. Reed (Eds.), Behavioral theories and intervention for autism (pp. 133-168). New York, NY: Nova Science.

MICHAEL, J. (1982). Skinner's elementary verbal relations: Some new categories. The Analysis of Verbal Behavior, 1, 1-3.

MICHAEL, J. (1985). Two kinds of verbal behavior plus a possible third. The Analysis of Verbal Behavior, 3, 2-5.

MICHAEL, J. (1988). Establishing operation and the mand. The Analysis of Verbal Behavior, 6,3-9.

MICHAEL, J. (1993). Establishing operations. The Behavior Analyst, 16, 191-207.

NATIONAL RESEARCH COUNCIL. (2001). Educating children with autism. Washington, DC: National Academy Press.

OSPI N A, M. B., SEIDA, J. K., CLARK, B., KARKHANEH, M., HARTLING, L., TJOSVOLD, L., ... SMITH, V. (2008). Behavioural and developmental interventions for autistic spectrum disorders: A clinical systematic review. PloS One, 3(11), e3755.

PERRY, A., CUMMINGS, A., DUNN GEIER, J., FREEMAN, L., HUGUES, S., MANAGHAN, T., ... WILLIAMS, J. (2011). Predictors of outcome for children receiving intensive behavioral intervention in a large, community-based program. Research in Autism Spectrum Disorders, 5, 592-603.

REMINGTON, B., HASTINGS, R. P., KOVSHOFF, H., ESPINOSA, F., JAHR, E., BROWN, T., ... MACLEAN, W. E. (2007). Early intensive behavioral intervention: Outcomes for children with autism and their parents after two years. American Journal on Mental Retardation, 112, 418-438.

RIVARD, M., FORGET, J., KERR, K., REGL1, G., & GIROUX, N. (2011). Children with autism spectrum disorders: Social and verbal behavior changes during intervention. Manuscript submitted for publication.

ROANE, H. S., FISHER, W. W., SGRO, G. M., FALCOMATA, T. S., & PABICO, R. R. (2004). An alternative method of thinning reinforcer delivery during differential reinforcement. Journal of Applied Behavior Analysis, 37, 213-218.

ROGE., B. (2003). Autisme, comprendre et agir. Paris, France: Dunod.

SCHOPLER, E., REICHLER, R. J., & ROCHEN RENNER, B. (1988). The childhood autism rating scale. Los Angeles, CA: Western Psychological Services.

SHAFER, E. (1994). A review of intervention to teach a and repertoire. The Analysis of Verbal Behavior, 12, 53-66.

SKINNER, B. F. (1957). Verbal behavior. New York, NY: Prentice Hall.

STAFFORD, M. W., SUNDBERG, M. L., & BRAAM, S. J. (1988). A prelimary investigation of the consequences that define the mand and the tact. The Analysis of Verbal Behavior, 6, 61-67.

STONE, W. L., & CARO-MARTINEZ, L. M. (1990). Naturalistic observation of spontaneous communication in autistic children. Journal of Autism and Developmental Disorders, 20, 437-453.

SUNDBERG, M. L., LOEB, M., HALE, L., & EIGENHEER, P. (2002). Contriving establishing operations to teach mands for information. The Analysis of Verbal Behavior, 18, 15-29.

SUNDBERG, M. L., & MICHAEL, J. (2001). The benefits of Skinner's analysis of verbal behavior for children with autism. Behavior Modification, 25, 698-724.

SUNDBERG, M. L., & PARTINGTON, J. W. (1998). Teaching language to children with autism or other developmental disabilities. Pleasant Hill, CA: Behavior Analysts, Inc.

TAGER-FLUSBERG, H., PAUL, R., & LORD, C. (2005). Language and communication in autism. In F. R. Volkmar, R. Paul, A. Klin, & D. Cohen (Eds.), Handbook of autism and pervasive developmental disorders: Vol. 1. Diagnosis, development, neurobiology and behavior (3rd ed., pp. 335-364). Hoboken, NJ: John Wiley & Sons.

TSIOURI, I., & GREER, R. D. (2003). Inducing vocal verbal behaviour in children with severe language delays throught rapid motor imitation responding. Journal of Behavioral Education, 12, 185-206.

VISM ARA, L. A., & ROGERS, S. J. (2010). Behavioral treatments in autism spectrum disorders: What do we know? Annual Review of Clinical Psychology, 6,447-468.

WECHSLER, D. (2002). Wechsler preschool and primary scale of intelligence. San Antonio, TX: Psychological Corp.

WILLIAM, G., & GREER, R. D. (1993). A comparison of verbal-behavior and linguistic-communication curricula for training developmentally delayed adolescents to acquire and maintain vocal speech. Behaviorology, 1,31-46.

Melina Rivard and Jacques Forget

Departement de Psychologie, Universite die Quebec a Montreal

We would like to acknowledge and thank the financial support that we have received during this doctoral dissertation. We extend our thanks to the Social Sciences and Humanities Research Council of Canada, the Fonds Quebecois de Recherche sur Ia Societe et Ia Culture, the Eleanor Cote Foundation, and the Consortium National de Recherche sur l'lntegration Sociale. We also want to thank Dr. Normand Giroux, Kelly Kerr, and Gisela Regli for their support, and we extend our gratitude to the Centre de Readaptation Monteregie-Est, especially to Gilles Lemaire, Sylvie Gladu, and the therapists, parents, and children who participated in this study.

Correspondence concerning this article shoutd be addressed to Melina Rivard, Universite du Quebec a Montreal, C.P. 8888, Succursale Centre-ville, Montreal (Quebec), Canada H3C 3P8. E-mail:
Table 1
Participants' Descriptions

Participant  Age   Gender  FSIQ  VIQ  PIQ  CARS  GARS  GADS

1             3-5       M    45   52   55    44   103   N/A
2             4-6       M    60   60   69    36    91   N/A
3             3-2       F    68   66   77    30    81   N/A
4             3-4       M   111  100  121    25    74    52
5             3-1       M   117  111  118  22.5    64    72
6             4-6       M    51   52   67    38    96   N/A
7             4-8       M    67   54   83    34    72   N/A
8             4-5       M   N/A  N/A  N/A  36.5    91   N/A
9            4-11       M   N/A  N/A  N/A  36.5    76   N/A
10            4-0       F    57   58   77  37.5   100   N/A
11            4-2       M    44   52   53  44.5   117   N/A
12            4-5       M  4  4   52   53    31    83   N/A
13            3-1       M    79   64   89    31    83   N/A
14            4-3       M   N/A  N/A  N/A  48.5   115   N/A

Note. FSIQ = Full Scale IQ from Wechsler Preschool and Primary Scale
of Intelligence (WPPSI-III; Wechsler, 2002); VIQ = Verbal IQ from
WPPSI-III; PIQ = Performance IQ from WPPSI-III; CARS = Childhood
Autism Rating Scale (Schopler, Reichler, & Rochen Renner, 1988);
GARS = Gilliam Autism Rating Scale (Gilliam, 1995); GADS = Gilliam
Asperger Disorder Scale (Gilliam, 2001); N/A = not available.

Table 2
Operational Definitions of Target Verbal Behaviors for Participants

Behavior               Operational definition

Verbal operant

Ma: appropriate mand   Makes a contextually appropriate request that
                       specifies the reinforcer delivered by the
                       listener (e.g., verbally asks to have an
                       apple, gives his therapist a picture of his
                       favorite toy to request it).

Mi: inappropriate      Makes a contextually inappropriate request
mand                   (e.g., requests to stop an activity by biting
                       her hand, requests to play with his therapist
                       by hitting the therapist's leg).

Ec: echoic             Repeats words or sentences after a verbal
                       stimuli. The vocal response of the participant
                       has point-to-point correspondence and formal
                       similarity with the verbal stimuli (e.g.,
                       repeats the word dog when his therapist
                       presents a picture of dog).

Ta: appropriate        Adequately reads words or sentences. The
textual                textual is under the control of printed words
                       (e.g., reads her name on a ticket, reads a
                       word on a flash card).

Ti: inappropriate      Inadequately reads words or sentences (e.g.,
textual                does not read the right letters on a flash
                       card, does not pronounce the right word on a

la: appropriate        Has a contextually appropriate and coherent
intraverbal            conversation or makes a contextually
                       appropriate remark. There is no point-to-point
                       correspondence with the previous verbal
                       stimuli of the therapist (e.g., explains to
                       the therapist what he did on the weekend,
                       makes a comment about the game that he is
                       playing with the therapist).

li: inappropriate      Has a contextually inappropriate conversation,
intraverbal            has an incoherent conversation, or makes
                       inappropriate remarks (e.g., gives a monologue
                       on a topic of her interest during a
                       conversation on something else, interrupts
                       therapist's conversation to talk about her own

Ca: appropriate tact   Accurately labels something (i.e., the prior
                       controlling stimulus, which could be an object
                       or a picture) with or without requesting it
                       (e.g., says "duck" when the therapist shows a
                       picture of a duck, says "chair" when he comes
                       near a chair).

Ci: Inappropriate      Inaccurately labels something with or without
tact                   requesting for it (e.g., says "cat" when the
                       therapist shows a picture of a duck, says
                       table when she points to a bed).

Other vocal behavior

Ub: babbling           Vocalizations and different sounds that were
                       emitted but do not serve any function for
                       communication. Vocalizations labeled as
                       babbling were not repetitive or

Ua: vocal stereotypy   Repetitive sounds, intonations, words, or
                       sentences. A word or sentence is labeled Ua if
                       it is emitted more than three times or is
                       known to be vocal stereotypy tor the
                       participant. Echolalia and palilalia were
                       included in this category.

Listener response

R: appropriate         Does not emit a verbal behavior but responds
nonverbal response to  to the therapist's verbal instruction (e.g.,
a request              the therapist asks the participant to come to
                       see her, and the participant comes).

Ri: inappropriate      The participant does not provide the
nonverbal response to  appropriate nonverbal response to the
a request              therapist's verbal instruction (e.g., the
                       therapist asks the participant to put the red
                       block in the bag, and the participant takes
                       the green block).

Table 3
Cumulative Basic Descriptive Statistics for Each Target Behavior

Tarqet behavior                          M       SD    % TB/T

Appropriate mand                        21.70   22.64   4.90%

Inappropriate mand                      10.85   15.34   2.50%

Echoic                                  43.20   29.10   9.80%

Appropriate textual                      0.68    1.58   0.15%

Inappropriate textual                    0.12    0.23   0.03%

Appropriate intraverbal                 64.57   96.78  14.70%

Inappropriate intraverbal                7.04   10.76   1.60%

Appropriate tact                        24.53   23.23   5,60%

Inappropriate tact                       4.49    4.17   1.01%

Babbling                                55.10   27.33  12.50%

Vocal stereotypy                        57.80   79.34  13.10%

Appropriate nonverbal response to a    124.03   37.61  28.20%
therapist's request

Inappropriate nonverbal response to a   26.03   15.26   5.90%
therapist's request

Appropriate verbal behaviors           278.70  125.60

Inappropriate verbal behaviors         161.50  101.30

Note. Table shows mean of frequencies of each target behavior for 14
children with ASD, the standard deviation of each target behavior for
the 14 children with ASD, and the percentage of each target behavior
based on the 13 target behaviors observed in seven 60-min sessions

Table 4
Correlations Between Levels of Functioning and Verbal Behaviors

Behavior                      FSIQ     VIQ     P\Q     CARS      GARS

Appropriate mand            0.66 *   0.55 *   0.62 *  -0.66 *  -0.55 *

Inappropriate mand           -0.23    -0.36    -0.18     0.27     0.08

Echoic                       -0.01    -0.17     0.03     0.20     0.00

Appropriate textual           0.07    -0.07     0.04    -0.13    -0.09

Inappropriate textual         0.26     0.21     0.24    -0.22    -0.12

Appropriate intraverbal     0.98 *   0.96 *   0.95 *  -0.80 *  -0.67 *

Inappropriate intraverbal     0.45     0.28     0.48    -0.45    -0.60

Appropriate tact            0.59 *     0.52   0.57 *    -0.40    -0.40

Inappropriate tact          0.57 *     0.48    .58 *    -0.50  -0.53 *

Babbling                      0.04    -0.12     0.09     0.11    -0.01

Vocal stereotypy             -0.40    -0.42    -0.35     0.48     0.60

Appropriate nonverbal        -0.27    -0.38    -0.25     0.15     0.14
response to a therapist's

Inappropriate nonverbal    -0.61 *  -0.63 *  -0.64 *   0.57 *     0.50
response to a therapist's

Appropriate verbal          0.87 *   0.77 *   0.85 *  -0.72 *  -0.65 *

Inappropriate verbal         -0.36    -0.47    -0.30     0.46     0.28

Note. FSIQ = Full Scale IQ from Wechsler Preschool and Primary Scale of
Intelligence (WPPSI-III; Wechsler, 2002); VIQ = Verbal IQ from
WPPSI-III; PIQ = Performance IQ from WPPSI-lfl; CARS = Childhood Autism
Rating Scale (Schopler, Reichler, & Rochen Renner, 1988); GARS = Gilliam
Autism Rating Scale (Gilliam, 1995). *p < .05.

Table 5
Cumulative Basic Descriptive Statistics for Each Target Consequence

Target consequence         M     % TC/T     FSIQ     VIQ    PIQ

Request                  104.58   22.20    0.72 *   0.71 *  0.66 *

Environment                3.61    0,80      0.11     0.10    0.06

Verbal reprimand           7.18    1.50     -0.43  -0.56 *   -0.39

Withdrawal of privilege    0.52    0.10     -0.26    -0.40   -0.21

No attention             144.25   30.60     -0.09    -0.26   -0.03

Social reinforcer         95.48   20.30     -0.02    -0.01    0.07

Material reinforcer       21.58    4.60     -0.50    -0.46   -0.39

Verbal behavior           93.95   19.90    0.91 *   0.00 *  0.86 *

Note. TC/T = percentage of this targeted consequence per total
consequences; FSIQ = Full Scale IQ from Wechsler Preschool and Primary
Scale of Intelligence (WPPSl-lli; Wechsler, 2002); VIQ = Verbal IQ from
WPPSI-III; PIQ = performance IQ from WPPSI-III. *p [less than or
equal to] 0.5.

Table 6
Mean of Target Consequences Delivered by Therapist for Target Behaviors
of AH Participants

                  Therapist's consequences to verbal behavior

Target            D      E     Rv    Rs     Rm     Rp    -      Vb

Appropriate     5.53   0.31  0.17   1.54   1.20  0.03   3.86   9.81

Inappropriate   3.27   0.07  0.58   0.52  0.254  0.05   3.75   2.02

Echoic          8.46   0.31  0.36   9.09   0.79  0.02  15.53   6.36

Appropriate     0.11   0.00  0.00   0.19   0.05  0.00   0.22   0.08

Inappropriate   0.01   0.00  0.00   0.00   0.00  0.00   0.01   0.01

Appropriate    20.68   0.58  0.20   8.52   0.84  0.03  14.23  26.37

Inappropriate   2.27   0.03  0.22   0.20   0.09  0.00   2.13   1.78

Appropriate     7.06   0.15  0.02  10.06   1.02  0.03   3.07   4.30

Inappropriate   2.02   0.01  0.30   0.08   0.03  0.00   0.53   1.45

Babbling       12.74   0.25  0.90   5.30   1.10  0.08  23.81  11.73

Vocal           3.08   0.09  0.37   2.27   0.65  0.04  48.30   4.03

Appropriate    29.86   0.70  2.39  52.69   8.29  5.11  25.30  16.12
response to a

Inappropriate  12.11   0.26  2.60   1.35   0.34  0.11   9.31   4.10
response to a

Note. D = demand or request; E =: environment; Rv = verbal reprimand;
Rs = social reinforcer; Rm = material reinforcer; Rp = withdrawal
privilege; - = no attention; Vb = verbal behavior.
Gale Copyright: Copyright 2012 Gale, Cengage Learning. All rights reserved.