lee 2015

Research in Autism Spectrum Disorders xxx (2015) xxx–xxx

G Model

RASD-1040; No. of Pages 10

Contents lists available at ScienceDirect

Research in Autism Spectrum Disorders

Journal homepage: ht tp : / /ees .e lsev ier .com/RASD/defaul t .asp

A further evaluation of the effects of listener trainingon derived categorization and speaker behavior in childrenwith autism§

Greg P. Lee, Caio F. Miguel *, Emily K. Darcey, Adrienne M. Jennings

California State University, Sacramento, United States

A R T I C L E I N F O

Keywords:

Autism

Categorization

Derived relations

Naming

Verbal behavior

A B S T R A C T

Previous research has shown that children with autism may accurately categorize visual

stimuli[2_TD$DIFF] after learning to both tact (i.e., speaker behavior) and receptively discriminate (i.e.,

listener behavior) them using common category names. The purpose of the current study

was to further evaluate the effects of category listener training alone on the development

of Visual Categorization and Category Tacts with four children diagnosed with autism. We

administered standardized language assessments to evaluate participants’ skills prior to

beginning the study and used a non-concurrent multiple-baseline design across

participants. Two of the participants whose language assessments identified both speaker

and listener scores greater than 36 months, passed Visual Categorization and Category

Tact testing with previously unfamiliar stimuli following Category Listener training. The

two participants whose language assessments identified a deficit in either speaker or

listener scores failed Visual Categorization and Category Tact testing following Category

Listener training. These results [14_TD$DIFF]suggest that both speaker and listener behavior may be

required for the emergence of untrained categorization and tacting following listener

training.

� 2015 Published by Elsevier Ltd.

Categorization plays an important role in a wide variety of skills, including identifying objects and their similarities,recalling and learning new information, and problem-solving skills. Thus, much of our understanding of human cognitiondepends on our understanding of categorization (Markman, 1989). One way of sorting or classifying objects is based onshared similarities, whether these are physical or functional. Behavior analysts have investigated taxonomical categorizationfor years without the need for relying on concepts as the mental representations of categories (e.g., Hernstein & Loveland,1964; Lowe, Horne, Harris, & Randle, 2002; Miguel, Petursdottir, & Carr, 2005; Petursdottir, Carr, Lechago, & Almason, 2008).A concept could be easily defined as a group of objects (stimuli, actions, etc.) that control similar responses. When anindividual behaves similarly in response to a group of objects, these objects are said to form a class that can be called a

§ This study is based on a thesis submitted by the first author under the supervision of the second author to the Department of Psychology at California

State University, Sacramento in partial fulfillment of the requirements for an M.A. degree in Psychology: Applied Behavior Analysis. We thank the staff at

Applied Behavior Consultants, Inc. in Sacramento, CA for their invaluable on-site support, as well as, Charisse Lantaya, Jasmine Rychard and Ashley Dawn

Mitchell for their assistance with data collection, as well as Danielle LaFrance and Jonathan Tarbox for their editorial assistance.* Corresponding author at: Department of Psychology, California State University, Sacramento, 6000 J. Street, Sacramento, CA 95819-6007, United States.

Tel.: +1 916 278 6813; fax: +1 916 278 6820.

E-mail address: [email protected] (C.F. Miguel).

Please cite this article in press as: Lee, G. P., et al. A further evaluation of the effects of listener training on derivedcategorization and speaker behavior in children with autism. Research in Autism Spectrum Disorders (2015), http://dx.doi.org/10.1016/j.rasd.2015.04.007

http://dx.doi.org/10.1016/j.rasd.2015.04.007

1750-9467/� 2015 Published by Elsevier Ltd.


mailto:[email protected]



http://www.sciencedirect.com/science/journal/17509467

http://ees.elsevier.com/RASD/default.asp


G.P. Lee et al. / Research in Autism Spectrum Disorders xxx (2015) xxx–xxx2

G Model


concept. As a result, concepts may be equated to stimulus classes. The concept of ‘‘bird,’’ for example, involves stimulusgeneralization among all stimuli we label as birds. The same concept involves discriminating between these stimuli andothers we do not call birds (Keller & Schoenfeld, 1950). Most of these concepts are acquired through formal or informaleducation, when the same response (e.g., saying ‘‘bird’’) is reinforced in the presence of several different stimuli (e.g., birds),but not in the presence of other animals or objects. Over time, the verbal topography ‘‘bird’’ comes under control of thephysical characteristics common to all of these stimuli. When a new stimulus that shares the same or some of thesecharacteristics is presented, induction is observed when the novel stimulus evokes the same behavior as the other birds. Inbehavior analysis, this form of induction is referred to as stimulus generalization or in the case of verbal induction, genericextension (Skinner, 1957).

Some classes or categories are comprised of objects or events that although topographically distinct, may share the samename. Having the same name means these objects occasion the same listener and speaker behavior (Horne & Lowe, 1996).Apples, bananas, and oranges, for instance, all evoke the vocal topography, ‘‘fruit’’ (i.e., speaker behavior) and can all beselected when someone asks for ‘‘fruits’’ (i.e., listener behavior). When encountering a new item (e.g., a pineapple) alsolabeled as a ‘‘fruit,’’ a child may[4_TD$DIFF] respond to it the same way as the other fruits, including placing it in the fruit basket and notplacing it [15_TD$DIFF] in some other basket, for example, containing utensils (i.e., categorization). This form of taxonomical induction orcategorization can be considered a novel behavior since it has not been directly reinforced. Moreover, this form of inductionmay depend on the child’s verbal behavior, more specifically, her listener and speaker repertories. Sorting or relating visuallydissimilar objects for the first time could emerge as a function of the child’s ability to label or tact the pineapple as a ‘‘fruit’’and respond to her (sub) vocalizations as a discriminative stimulus. The discriminative stimulus would evoke the (listener)behavior of searching for other fruits or the fruit basket that, of course, she had already learned to identify. In other words,apples, bananas, oranges, and pineapples may become related to one another simply because they have the same name(Horne & Lowe, 1996; Miguel & Petursdottir, 2009).

Naming itself can be described as a higher-order operant acquired through a history of multiple exemplar instruction inwhich words and their referents are related by cues of sameness (e.g., ‘‘this is a cup,’’ ‘‘this object is called a brush,’’). InRelational Frame Theory (RFT) literature, naming is conceptualized as a generalized operant class of deriving mutuallyentailed relations of coordination between two stimuli (Hayes, Barnes-Holmes, & Roche, 2001; Miguel & Petursdottir, 2009).This is a complicated way of simply saying that the ability to derive either object-name or name-object relations is ageneralized operant. This generalized operant is thought to be acquired through a history of reinforcement for multipleexemplars of responding bidirectionally between an object and its name (Greer, Stolfi, & Pistoljevic, 2007). Despite itsorigin,1 naming has been shown to be a fundamental (verbal) skill responsible for facilitating the incidental learning of words(i.e., language explosion), as well as reading, writing and, of course, categorization (Greer & Longano, 2010).

A series of studies with typically developing children have found that participants could only sort new objects andpictures into categories (i.e., visual categorization) when they could name them as defined by the bidirectional relationbetween categorical listener and speaker behaviors (i.e., tact the category name of an object and receptively identify theobject when given the category name). So far, a number of studies have shown that in the absence of this skill, VisualCategorization does not occur (Horne, Hughes, & Lowe, 2006; Horne, Lowe, & Harris, 2007; Horne, Lowe, & Randle, 2004;Lowe et al., 2002; Lowe, Horne, & Hughes, 2005; Miguel, Petursdottir, Carr, & Michael, 2008). These studies’ typicalpreparation involved teaching participants to either tact novel stimuli by their category name (e.g., calling all stimuli in ClassA as ‘‘VEK’’ and in Class B as ‘‘ZOG’’), or receptively discriminate them by category (e.g., selecting stimuli from Class A whenhearing ‘‘VEK’’ and selecting stimuli from Class B when hearing ‘‘ZOG’’), with subsequent tests to see if participants wouldsort these stimuli into classes/groups (i.e., Visual Categorization). Findings have consistently shown that when participantsname the stimuli by category, they also sort them accurately. In other words, equivalence classes or frames of coordination[5_TD$DIFF]can be established via verbal behavior training (i.e., category tact training) alone, as long as participants respond as bothspeaker and listener.

Until recently, this preparation had not been used with children with disabilities. Instead, many of these children havebeen taught to categorize objects directly, one by one, as opposed to being taught the pre-requisite skills for novel or derivedcategorization (Miguel & Petursdottir, 2009). In the first assessment of this preparation with children with autism, Migueland Kobari-Wright (2013) evaluated whether two preschool-aged children would categorize pictures after learning how totact them with a common category name. Visual Categorization was measured via a visual–visual matching-to-sample task.Each participant learned to tact nine pictures belonging to three unfamiliar categories with their common category name(e.g., hound dog, work dog, and toy dog). Each participant received Visual Categorization and Category Listener pretests priorto Category Tact training. In Visual Categorization testing, the experimenter presented a sample stimulus (e.g., card),prompted or asked for an observing response (e.g., looking at, and touching the card), and then presented an array of threecomparisons paired with the instruction, ‘‘Match.’’ A correct response was recorded if the participant matched the sample tothe correct corresponding stimulus. During Category Listener testing, the experimenter dictated the sample (e.g., ‘‘toy dog’’)and presented three visual comparisons. A correct response was recorded if the participant selected the positive comparisonout of the three-stimulus array. Following Category Tact training, one participant successfully passed both VisualCategorization and Category Listener tests. The other participant initially failed Visual Categorization posttests despite

1 See Horne and Lowe (1996) for a detailed description of how naming is acquired in typically developing children.




G.P. Lee et al. / Research in Autism Spectrum Disorders xxx (2015) xxx–xxx 3

G Model


passing Category Listener posttests. The Visual Categorization posttest was presented a second time with an additionalrequirement to tact the samples by category before responding, which resulted in correct categorizations. These resultssuggest that failure to tact the sample by category may lead to failures in Visual Categorization, supporting the notion thatboth speaker and listener behavior may often be emitted as a problem-solving strategy during categorization trials.

In a follow-up study, Kobari-Wright and Miguel (2014) exposed four children diagnosed with autism (4–5 years old) toCategory Listener training and tested for the emergence of Category Tacts and Visual Categorization. Visual Categorizationtests were the same as described in Miguel and Kobari-Wright (2013). In this experiment, each participant was exposed toCategory Tact pre and posttests. Category Listener training consisted of presenting a three-stimulus array with one picturefrom each category (same as in Miguel & Kobari-Wright, 2013) with the instruction, ‘‘Give me a hound/work/toy dog.’’ Acorrect response was scored if the participant selected the corresponding comparison during each trial. Three out of fourparticipants, who did not visually categorize or emit tacts of categories[6_TD$DIFF] did so following Category Listener training. Oneparticipant failed both Visual Categorization and Category Tact posttests. This participant was then taught to tact eachstimulus individually by its category name (i.e., Category Tact training) in a condition identical to Miguel and Kobari-Wright(2013). This was followed by Visual Categorization tests in which the participant was required to tact each sample by itscategory[7_TD$DIFF] before matching. During testing, the participant tacted all samples and categorized the pictures accurately.

Taken together, both studies conducted with children diagnosed with autism (Kobari-Wright & Miguel, 2014; Miguel &Kobari-Wright, 2013) suggest that tact or listener training alone might produce derived categorization, as long asparticipants have a naming repertoire. This supports the teaching curriculum proposed by Miguel and Petursdottir (2009) onhow to develop derived categorization in children with autism. However, these studies were not without their limitations.First, only six participants were evaluated in total, limiting the generality of the results. Second, participant language skillswere not assessed in the first study. In the second study, each participant was assessed using idiosyncratic criterion-basedassessments, making it difficult to compare them. By assessing participants’ verbal skills, clinicians may be able to predictwhether this type of intervention would be successful or whether pre-requisite naming skills would first need to be taught(Greer & Ross, 2008).

Given the necessity of replications to increase the generality of findings in single-subject research (Kazdin, 2011), thepurpose of the current study was to further evaluate the effects of teaching listener behavior on the emergence of speakerbehavior (i.e., tacts) and categorization (i.e., visual–visual matching) with children diagnosed with autism, and to evaluateparticipants’ language skills prior to training using standardized assessments. The use of these assessments may serve toidentify language characteristics that could potentially influence teaching outcomes.

1. Method

1.1. Participants, setting, and materials

Participants included four children, Charli (5 years), Marcus (3 years), Teresa (3 years), and Quincy (3 years) who hadreceived diagnoses of autism and were receiving behavioral services for an average of 23 h per week. The Peabody PictureVocabulary Test (PPVT), Fourth Edition (Dunn & Dunn, 2007) and the Expressive Vocabulary Test (EVT; Williams, 2007) wereadministered prior to intervention to evaluate participant comprehension of the spoken word, expressive vocabulary, andword retrieval.

Sessions consisted of one to four nine-trial blocks conducted a minimum of three[8_TD$DIFF], and a maximum of five days per week.Sessions for Charli and Marcus were conducted in their homes, at tables in designated therapy areas; sessions for Quincy andTeresa were conducted in a [16_TD$DIFF]secluded area in a center-based program. Sessions lasted approximately 20–30 min with theexperimenter and, sometimes, a secondary observer for interobserver agreement and treatment integrity purposes (seebelow). The experimenter was positioned at a small table on the same side as the participant, in order to avoid inadvertentcueing. The second observer was positioned so that he or she could see the stimuli presented and hear the instructions fromthe experimenter, but so neither the experimenter or second observer could see each other’s data sheets.

Materials included the PPVT and EVT assessments; nine laminated, colored pictures (7.62 cm� 7.62 cm) from familiarcategories (e.g., vehicles, animals and clothing) and nine (7.62 cm� 7.62 cm) laminated, colored pictures with imagesdepicting three categories of dog breeds (e.g., hound, work and toy dog) as used in previous studies (Kobari-Wright & Miguel,2014; Miguel & Kobari-Wright, 2013; see Fig. 1). Images were divided into three sets (A, B, and C), with each consisting of onestimulus from each category. The pictures were presented on a plain white poster board (20.3� 41.9 cm) attached to a stripof white Velcro. Toys, activities, and edibles identified via preference assessments were used during training and contingentupon completion of each session.

1.2. Dependent measures and experimental design

A correct response was defined as (a) selecting the target comparison from a three-stimulus array in the presence of aspoken sample for a Category Listener response, (b) selecting the target comparison from a three-stimulus array in thepresence of a visual sample for Visual Categorization, and (c) saying the category name corresponding with the picture for aCategory Tact response.




[(Fig._1)TD$FIG]

Fig. 1. Unfamiliar pictures used during training [1_TD$DIFF].


G Model


We used a single-subject non-concurrent multiple-baseline design across participants (Watson & Workman, 1981) toassess the effects of listener training on the emergence of categorization and tacts. This type of design allows forexperimental control to be achieved with a limited number of participants.

1.3. Interobserver agreement and treatment integrity

Interobserver agreement (IOA) was assessed in 72% of all sessions. We calculated IOA using point-by-point agreement bydividing the number of agreements by the total number of trials and converting the results to a percentage (Kazdin, 2011).Average agreement for tact and categorization tests was 100% for Teresa and Quincy, and 99% for Marcus and Charli (range,89–100%). Average IOA for listener training was 99% for each participant (range, 89–100%).

Treatment integrity (TI) was assessed in 72% of all sessions. A trial was implemented correctly when the experimenterutilized the correct prompt delay, presented stimuli according to what was specified on the data sheet, and provided theappropriate consequence. We considered the entire trial incorrect if one or more of these components was not executedcorrectly. We calculated TI by dividing the number of correctly implemented trials by the total number of trials andconverting the result to a percentage. Treatment integrity[17_TD$DIFF]was 100% for Teresa, Marcus, and Quincy[9_TD$DIFF] and 99% for Charli (range,89–100%) for Category Tact and Visual Categorization. Average TI was 100% for Quincy and 99% for Teresa, Marcus, and Charli(range, 89–100%) for listener training.

1.4. Procedure

The following training conditions were presented to each participant: Pretraining, Category Tact pretraining,Category Listener pretraining, Visual Categorization pretest, Category Tact pretest, Category Listener training, VisualCategorization posttest, and Category Tact posttest (see Table 1). These conditions are explained in detail below. Fig. 2depicts the trained and tested relations using two stimuli from one of the three classes as an example. The mainindependent variable consisted of Category Listener training, while dependent variables included derived CategoryTacts and Visual Categorizations.

1.5. Stimulus preference assessment

We administered a modified version of the Reinforcer Assessment for Individuals with Severe DevelopmentalDisabilities (RAISD: Fisher, Piazza, Bowman, & Amari, 1996) to the caregiver for each participant at the start of the study.The items identified on the RAISD were used to conduct a brief multiple-stimulus-without-replacement (MSWO)preference assessment (Carr, Nicolson, & Higbee, 2000) before the start of each session. This was done to identify whichstimulus, edible, and/or activity the participant could earn contingent upon correct responses and participation in thesession.




Table 1

Order of experimental conditions.

Order Condition Stimuli per trial Training criterion

1 Visual Categorization Pretraining 3 2 blocks at 89%

2 Category Tact Pretraining 1 2 blocks at 89%

3 Category Listener Pretraining 3 2 blocks at 89%

4 Visual Categorization Pretest 3 N/A

5 Category Tact Pretest 1 N/A

6 Category Listener Training 3 2 blocks at 89%

7 Visual Categorization Posttest 3 N/A

8 Category Tact Posttest 1 N/A

9* Category Tact Posttest 2 3 N/A

10** Category Tact Training 1 2 blocks at 89%

11** Visual Categorization Posttest 3 N/A

* Children were exposed to this condition only if they failed the previous tact posttest.

** Children were exposed to this condition if they fail Visual Categorization Posttest, Category Tact Posttest 1, and Category Tact Posttest 2.


G Model


1.6. Pretraining

As in previous studies (Miguel & Kobari-Wright, 2013; Kobari-Wright & Miguel, 2014), we taught participants how torespond to the procedures by exposing them to all experimental conditions (e.g., Visual Categorization, Category Tacting andCategory Listener training) using pictures of familiar categories (i.e., clothing, vehicles, animals). Mastery criterion consistedof two consecutive blocks at 89%. Pictures were pre-arranged so that each set appeared three times per block. There was oneblock for each set and one mixed-block with stimuli from all three sets.

1.7. Language assessments

The PPVT folded easel was placed on the table between the participant and the experimenter. We exposed theparticipants to four images per page and the experimenter asked, ‘‘Show me/Point to/Touch <name of the testing item>.’’Testing was conducted in 12-trial sets. Each participant’s response was marked as either: 1, 2, 3, or 4, corresponding withpicture placement on the easel. Incorrect responses were indicated by circling an ‘‘E’’ for the item. All responses (correct andincorrect) were followed by praise (i.e., ‘‘Good!’’, ‘‘Nice try!’’). Testing continued until eight or more errors occurred within aset, referred to as the ceiling. Ceiling items and total errors were used to calculate each participant’s raw score and the rawscore was used to derive the standard score and age equivalencies.

The EVT folded easel was placed on the table between the participant and the experimenter. There was one image perpage and the experimenter presented the stimulus question (e.g., ‘‘What do you see?’’, ‘‘What is this?’’, ‘‘How many balls doyou see?’’) as specified by the EVT protocol. All responses (correct and incorrect) were followed by praise (i.e., ‘‘Good!’’, ‘‘Nicetry!’’). Testing was conducted in 10–20 trial blocks, as indicated by the EVT protocol. Testing continued until the participantresponded with five consecutive incorrect responses (i.e., the ceiling). Ceiling items and total errors were used to calculate

[(Fig._2)TD$FIG]

Fig. 2. Example of trained (solid lines) and derived relations (dotted lines) for two members of the ‘‘Toy dog’’ class.





G Model


each participant’s raw score and the raw score was used to derive the standard score and age equivalencies. The PPVT wasalways administered first, as recommended by EVT protocols (Williams, 2007).

1.8. Visual Categorization tests

During Visual Categorization pretest and posttest 1 trials, the experimenter presented a sample card to the participantand obtained an observing response (i.e., looking at the card) to ensure attending (Dinsmoor, 1985). If the participant did notattend to the card, a more intrusive prompt was used (i.e., participant was asked to point to the card). After the participantresponded to the sample, the experimenter presented the three comparisons affixed to the stimulus board and said, ‘‘Match.’’Regardless of accuracy, the experimenter signaled the end of the trial by saying, ‘‘Okay.’’ Visual Categorization posttest 2trials were the same as Visual Categorization posttest 1, except that the experimenter started by presenting the sample andasking the participant, ‘‘What is it?’’ Then the experimenter presented the comparison stimuli within 5 s regardless of theparticipant’s response. Sessions consisted of nine-trial blocks in which the correct choice (i.e., comparison) appeared once onthe left, middle, and right positions in the array across three consecutive nine-trial blocks. To maintain responding, onereinforced listener trial with familiar stimuli was presented after three unreinforced trials (for a total of 12 trials). At the endof each trial block, the participant was allowed to engage in the activity identified by the MSWO for 2–5 min. The criterion foremergence during posttests was a minimum of two consecutive blocks at 89% (8 out of 9[18_TD$DIFF] test trials) or above.

1.9. Category Tact tests

During Category Tact pre and posttest 1, the experimenter presented one picture attached to the stimulus board andasked, ‘‘What is it?’’ A correct response was scored when the participant labeled the picture by its category name (i.e., hounddog, toy dog, work dog). An incorrect response consisted of mislabeling the category name (e.g., emitting a different tact forthat picture such as, ‘‘dog’’ or ‘‘puppy’’) or not responding within 5 s. For participants who failed the first tact posttest, asecond posttest (Category Tact posttest 2) was conducted in which three stimuli were presented simultaneously on thestimulus board. During each trial, the experimenter pointed to one of the three cards and asked the child, ‘‘What is it?’’ Thiscondition was conducted to control for the possibility that failure in testing could be due to the type of discriminationrequired. More specifically, posttest 1 required successive discrimination, while Listener training and Category Tact posttest2 required simultaneous discrimination (Miguel et al., 2008). All other variables remained the same as in the first tactposttest.

1.10. Category Listener training

During Category Listener training, each trial consisted of the experimenter instructing the participant to select a specificcategory (i.e., ‘‘Give me the hound/toy/work dog.’’), followed by the presentation of the stimulus board with threehorizontally placed comparisons. Training was conducted using a progressive prompt delay (i.e., 0 s, 1 s, 2 s, 3 s, 4 s, and noprompt) in which the experimenter pointed to the correct comparison after the designated period as specified by the promptdelay. Criterion for increasing the prompt delay was two consecutive blocks with a score of 89% or better (independent andprompted responses). If the participant made three consecutive errors during one block, the prompt delay was regressed tothe previous one. The experimenter trained all three sets of stimuli and a mixed set (i.e., all sets combined) until mastery.Each set was considered mastered when the participant completed two consecutive blocks at 89% accuracy (independentresponses only), at which point training began with the next set. If at any prompt level (i.e., 1 s) the participant engaged inindependent correct responses with 89% accuracy, the prompt delay was removed on the next block (no prompt). Promptedresponses were praised and independent correct responses were praised and followed by an edible. Incorrect responses or noresponses after 5 s were followed by a [10_TD$DIFF] correction procedure: the experimenter pointed to the correct stimulus and then thestimulus board was removed for 3–5 s. The same stimuli and instruction were given again with a 0 s prompt delay, whichwas conducted without feedback and the board was removed again. Finally the trial was represented with the previousprompt delay.

1.11. Category Tact training

This condition was only implemented with participants who failed both Visual Categorization and Category Tactposttests. The experimenter presented the participant with one card at a time and asked, ‘‘What is it?’’ Sets were taughtconsecutively, using a progressive one-second prompt delay in nine-trial blocks as described above. The same [11_TD$DIFF] errorcorrection detailed in Category Listener training was used for Category Tact training. Mastery criterion was two consecutiveblocks of independent, correct responding at 89% (or higher) for each set.

2. Results

Table 2 shows each participant’s chronological age, PPVT & EVT scores, the number of trials required to reach masterycriterion during pretraining and Category Listener training, and whether each participant passed or failed Visual




Table 2

Summary of participant chronological age, scores and results.

Participant

Quincy Marcus Charli Teresa

Chronological age (months) 46 47 69 41

PPVT (Listener) age (months) [13_TD$DIFF]N/A 49 54 30

EVT (Speaker) age (months) 24 45 65 [13_TD$DIFF]N/A

# of Pretraining Blocks 66 44 24 55

# of Training Blocks 94 31 27 121

Visual Categorization FAIL PASS PASS FAIL

Category Tact FAIL PASS PASS FAIL


G Model


Categorization and Category Tact posttests. Charli’s (69 months) listener age (PPVT) equivalent was 54 months and speakerage (EVT) equivalent was 65 months. Marcus’ (47 months) listener age equivalent was 49 months and speaker age equivalentwas 45 months. Quincy’s (46 months) listener age equivalent was below 24 months and speaker age equivalent was 24months. Teresa’s (41 months) listener age equivalent was 30 months and speaker age equivalent was below 24 months.Teresa and Quincy’s scores are considered extrapolations since they fell below 24 months on standardized measures (Dunn &Dunn, 2007; Williams, 2007).

During pretraining, Charli required 216 trials, Marcus required 396 trials, Teresa required 495 trials, and Quincy required594 trials to reach mastery. The participants with the higher age equivalent scores reached mastery criterion in fewer trialsduring pretraining.

Fig. 3 shows the results for Charli and Marcus, with pre and posttest scores for Visual Categorization and Category Tacttests with experimental stimuli. Charli and Marcus both responded below chance levels (M = 18% and 22% respectively) inVisual Categorization pretests. During Category Tact pretests neither Charli nor Marcus responded correctly to any of thestimuli. During Category Listener training, Charli required 243 trials and Marcus required 279 trials to reach masterycriterion. After this training, they both accurately categorized the stimuli visually (M = 81.6% for Charli, and M = 89% forMarcus) and passed the Category Tact posttest 1 (M = 96.3% for Charli and M = 100% for Marcus).

Fig. 4 shows the results for Qunicy and Teresa, with pre and posttest scores for Visual Categorization and Category Tacttests with unfamiliar images. Qunicy and Teresa responded below mastery levels (M = 44% and M = 26.7%, respectively) andneither participant tacted the stimuli in pretesting. Both Quincy and Teresa failed Visual Categorization and Category Tactposttests. Quincy required 846 trials to reach mastery criterion during Category Listener training, after which he failed tovisually categorize the stimuli (M = 29.3%), failed Category Tact posttest 1 (M = 56%) and failed Category Tact posttest 2(M = 48%). Quincy required 108 trials of Category Tact training to meet mastery criterion and [19_TD$DIFF] subsequently passed the VisualCategorization posttest (M = 100%). Teresa required 1089 trials to reach mastery criterion during Category Listener training,after which she failed to visually categorize the stimuli (M = 51.6%), failed Category Tact posttest 1 (M = 63.3%) and failedCategory Tact posttest 2 (M = 25.6%). Teresa required 747 trials of Category Tact training to meet mastery criterion, afterwhich she passed the Visual Categorization posttest 1 (M = 76.2%).

3. Discussion

The purpose of this study was to further evaluate the effects of teaching (categorical) Listener behavior on the emergenceof Category Tacts (i.e., speaker behavior) and Visual Categorization (i.e., visual–visual matching) with children diagnosedwith autism while measuring their language skills using standardized language assessments. The results of the current studyreplicate previous findings with children with autism (Kobari-Wright & Miguel, 2014; Miguel & Kobari-Wright, 2013), aswell as the many studies conducted with typically developing children (Horne et al., 2004, 2006, 2007; Lowe et al., 2002,2005; Mahoney, Miguel, Ahearn, & Bell, 2011; Miguel et al., 2008) by showing that both listener and speaker behaviors werenecessary for the emergence of Visual Categorization. Thus, Category Listener training alone may be sufficient to producenovel Visual Categorization with children with autism, as long as these children also demonstrate speaker behavior (i.e., tactthe pictures by category), and already have bidirectional naming repertoires in place.

For Charli and Marcus, Category Listener training was sufficient to produce accurate Category Tacts and VisualCategorization responses. Category Listener training was not sufficient for Quincy and Teresa, who failed VisualCategorization and Category Tact posttests. In other words, even though these two participants demonstrated CategoryListener behavior, neither showed derived speaker behavior in the form of Category Tacts or Visual Categorization. However,with direct training of Category Tacts, both participants passed Visual Categorization tests. These data suggest that thebidirectional components of naming were required for participants to pass Visual Categorization tests, leading us to inferthat these two repertoires may have also played a role in Charli and Marcus’ categorization performances.

While the results from all four participants support previous research on the role of naming in derived VisualCategorization, the comparison of participants’ PPVT (listener) and EVT (speaker) scores may provide additional insight intothe development of naming. Although previous research (Kobari-Wright & Miguel, 2014) did not find any consistent relation




[(Fig._3)TD$FIG]

Fig. 3. The percentage of correct Category Tacts and Visual Categorizations for Charli and Marcus. Information on chronological and equivalent ages [12_TD$DIFF] in

months is included (S = speaker/expressive, L = listener/receptive).


G Model


between age and the development of naming, Greer and Longano (2010) suggested that naming is related to the ‘‘explosionof vocabulary’’ (Hart & Risley, 1995), which occurs around the chronological age of three. This indicates that an accurateassessment of participants’ speaker and listener behavior may predict the likelihood of a naming repertoire, and thus theeffectiveness of our teaching procedure. The PPVT and EVT scores for Charli and Marcus were greater than 36 months andboth demonstrated untrained categorization and speaker behavior. Quincy and Teresa’s PPVT and EVT scores were 30months or less and both participants required more trials to meet mastery criterion for listener training and passedcategorization only after the tact was taught directly. In summary, the two participants whose PPVT and EVT scores wereassessed above 36 months, visually categorized and tacted the stimuli by category, after completing category listenertraining. The two participants whose PPVT and EVT scores were assessed at 30 months or below, failed to visually categorizeand tact the stimuli by category after completing Category Listener training. These results suggest that standardizedassessments may serve to predict the presence of a naming repertoire. Future research should directly test this possibility, aswell as compare the efficiency of using standardized assessments versus directly assessing the naming repertoire by trainingand testing word-object bidirectionality.

This study supports previous research, which posits that, in the presence of a naming repertoire, listener training alonemay be sufficient to produce essentially three skills at once: listener behavior, speaker behavior, and categorization (e.g.,Sprinkle & Miguel, 2012). While additional research is required to more precisely identify a benchmark standard for thepresence of naming, findings of this study are in line with the assertion that naming develops around the age of 36 months(Greer & Longano, 2010).

The current study trained relations of coordination between category names (e.g., ‘‘toy dog’’) and their exemplars (e.g.,pictures of a Yorkie, an Italian greyhound, and a Chihuahua; i.e., Category Listener training), and tested for derived relations




[(Fig._4)TD$FIG]

Fig. 4. The percentage of correct Category Tacts and Visual Categorizations for Quincy and Teresa. Information on chronological and equivalent ages [12_TD$DIFF] in

months is included (S = speaker/expressive, L = listener/receptive).


G Model


between the pictures and their category names (i.e., Category Tacts) and among the pictures themselves (i.e., VisualCategorization). Results suggested that derived Visual Categorization seemed to be dependent upon participants’ repertoireof relating pictures to their category names and vice versa. Thus, it is possible that for those participants who passed VisualCategorization tests after Category Listener training and scored high on verbal assessments, a generalized operant ofresponding in terms of coordination had already been established. In summary, our results support both a naming (Horne &Lowe, 1996), as well as a relational responding analysis (Hayes et al., 2001) since both consider naming as a generalizedoperant brought about by a history of multiple exemplar instruction.

Categorization is often discussed in the RFT literature as hierarchical relational responding (Hayes et al., 2001).Hierarchical relating consists of responding to the relation between two or more stimuli where one stimulus belongs to or ispart of another. Children learn that work, toy, and hounds are all dogs, and that dogs are pets, and pets are animals, and so on.It would be interesting for future research to further expand the categorizing behavior of children with autism byprogressing from relations of coordination as was done in the current study to relations of hierarchy, and then extendingthose to greater and greater complexity.

A possible limitation of this study is that the PPVT and EVT assessments are not normed for children with developmentaldisabilities (Dunn & Dunn, 2007; Williams, 2007). However, we were still able to infer valuable information from theirresults. According to Esch, LaLonde, and Esch (2011), both the PPVT and EVT may serve as accurate measures of listener andspeaker behaviors, respectively. Thus, these assessments seem adequate for measuring the two components of naming.However, it may be wise to analyze deficit scores, such as Teresa’s (chronological age: 41 months; PPVT score: 30); andQuincy’s (chronological age: 46; EVT score: 24) with more caution due to potential confounds, such as lack of familiarity andreinforcement during testing, and possible deficits of prerequisite skills (e.g., attending to auditory stimuli for the PPVT).Future research should continue to examine the relation between standardized language assessment scores, naming, and theeffects of derived relations and behavior. Since the PPVT and EVT are not normed for children with language delays,





G Model


additional language assessments should be compared to them in order to identify similarities and/or differences in speakerand listener scores predictive of a naming repertoire. Future research should also investigate the effects of varied scores onthe PPVT and EVT assessments and the relation between speaker and listener training on the development of derivedcategorization. As hierarchical naming seems to be required for the emergence of Visual Categorization and possibly otherskills, it may be important to assess whether normed language tools can predict the presence of this type of naming, andwhether certain scores can predict the use of listener or speaker training to generate derived categorization. If test scoresshow the absence of naming, then clinicians should make efforts to directly assess (Delfs & Frampton, 2014), and if needed,establish a generalized naming repertoire prior to teaching categorization (Greer & Ross, 2008; LaFrance & Miguel, 2014).

Although additional replications are necessary, the current study describes a practical way in which clinicians may teachnovel categorization to children with autism (see Miguel & Petursdottir, 2009). As long as children’s receptive and expressiveage equivalencies are of at least 36 months, then teaching them to select pictures by their category names (listener) may besufficient to produce tacts (speaker) and sorting by category.

References

Carr, J. E., Nicolson, A. C., & Higbee, T. S. (2000). Evaluation of a brief multiple-stimulus preference assessment in a naturalistic context. Journal of Applied BehaviorAnalysis, 33, 353–357.

Delfs, C., & Frampton, S. (2014). Practical implications of evaluating the efficiency of listener and tact instruction for children with autism. Journal of AppliedBehavior Analysis, 47, 810–813.

Dinsmoor, J. A. (1985). The role of observing and attention in establishing stimulus control. Journal of the Experimental Analysis of Behavior, 43, 365–381.Dunn, L., & Dunn, D. (2007). Peabody Picture Vocabulary Test (4th ed., p. 55437). Bloomington, MN: Pearson.Esch, B. E., LaBonde, K. B., & Esch, J. W. (2011). Speech and Language assessment: A verbal behavior analysis. Journal of Speech and Language Pathology – Applied

Behavior Analysis, 5, 166–190.Fisher, W. W., Piazza, C. C., Bowman, L. G., & Amari, A. (1996). Integrating caregiver report with a systematic choice assessment. American Journal on Mental

Retardation, 101, 15–25.Greer, R. D., & Longano, J. (2010). A rose by naming: How we may learn how to do it. The Analysis of Verbal Behavior, 26, 73–106.Greer, R. D., & Ross, D. E. (2008). Verbal behavior analysis: Introducing and expanding new verbal capabilities in children with language delays. Boston, MA: Pearson

Education Inc.Greer, R. D., Stolfi, L., & Pistoljevic, N. (2007). Emergence of naming in preschoolers: A comparison of multiple and single exemplar instruction. European Journal of

Behavior Analysis, 8, 109–131.Hayes, S. C., Barnes-Holmes, D., & Roche, B. (2001). Relational Frame Theory: A Post-Skinnerian account of human language and cognition. New York: Plenum Press.Hart, B. M., & Risley, T. R. (1995). Meaningful differences in the everyday life of America’s children. Baltimore: Brookes.Hernstein, R. J., & Loveland, D. H. (1964). Complex visual concepts in the pigeon. Science, 146, 549–551.Horne, P. J., & Lowe, C. F. (1996). On the origins of naming and other symbolic behavior. Journal of the Experimental Analysis of Behavior, 65, 185–241.Horne, P. J., Lowe, C. F., & Randle, V. R. L. (2004). Naming and categorization in young children: II. Listener behavior training. Journal of the Experimental Analysis of

Behavior, 81, 267–288.Horne, P. J., Hughes, J. C., & Lowe, C. F. (2006). Naming and categorization in young children: IV: Listener behavior training and transfer of function. Journal of the

Experimental Analysis of Behavior, 85, 247–273.Horne, P. J., Lowe, C. F., & Harris, F. D. A. (2007). Naming and categorization in young children: V. Manual sign training. Journal of the Experimental Analysis of

Behavior, 87, 367–381.Kazdin, A. E. (2011). Single-case research designs: Methods for clinical and applied settings (2nd ed.). Oxford University Press Inc.Keller, F. S., & Schoenfeld, W. N. (1950). Principles of psychology. New York: Appleton-Century-Crofts.Kobari-Wright, V. V., & Miguel, C. F. (2014). The effects of listener training on the emergence of categorization and speaker behavior in children with autism.

Journal of Applied Behavior Analysis, 47, 1–6.LaFrance, D. L., & Miguel, C. F. (2014). Teaching language to children with autism spectrum disorder. In P. Sturmey, J. Tarbox, D. R. Dixon, & J. L. Matson (Eds.),

Handbook of early intervention for autism spectrum disorders: Research, practice, and policy (pp. 403–436). New York: Springer.Lowe, C. F., Horne, P. J., Harris, F. D. A., & Randle, V. R. L. (2002). Naming and categorization in young children: Vocal tact training. Journal of the Experimental Analysis

of Behavior, 78, 527–549.Lowe, C. F., Horne, P. J., & Hughes, J. C. (2005). Naming and categorization in young children: III. Vocal tact training and transfer of function. Journal of the

Experimental Analysis of Behavior, 83, 47–65.Markman, E. M. (1989). Categorization and naming in children: Problems of induction. Cambridge, MA: MIT Press.Mahoney, A. M., Miguel, C. F., Ahearn, W. H., & Bell, J. (2011). The role of common motor responses in stimulus categorization by preschool children. Journal of the

Experimental Analysis of Behavior, 95, 237–262.Miguel, C. F., Petursdottir, A. I., & Carr, J. E. (2005). The effects of multiple-tact and receptive-discrimination training on the acquisition of intraverbal behavior. The

Analysis of Verbal Behavior, 21, 27–41.Miguel, C. F., Petursdottir, A. I., Carr, J. E., & Michael, J. (2008). The role of naming in stimulus categorization by preschool children. Journal of the Experimental

Analysis of Behavior, 89, 383–405.Miguel, C. F., & Petursdottir, A. I. (2009). Naming and frames of coordination. In R. A. Rehfeldt & Y. Barnes-Holmes (Eds.), Derived relational responding, applications

for learners with autism and other developmental disabilities (pp. 128–144). Oakland, CA: New Harbinger Publications, Inc.Miguel, C. F., & Kobari-Wright, V. V. (2013). The effects of tact training on the emergence of categorization and listener behavior in children with autism. Journal of

Applied Behavior Analysis, 46, 669–773.Petursdottir, A. I., Carr, J. E., Lechago, S. A., & Almason, S. M. (2008). An evaluation of intraverbal training and listener training for teaching categorization skills.

Journal of Applied Behavior Analysis, 41, 53–68.Skinner, B. F. (1957). Verbal behavior. Acton, MA: Copley.Sprinkle, E. C., & Miguel, C. F. (2012). The effects of listener and speaker training on emergent relations in children with autism. The Analysis of Verbal Behavior, 28,

111–117.Watson, P. J., & Workman, E. A. (1981). The nonconcurrent multiple baseline across individuals design: An extension of the traditional multiple baseline design.

Journal of Behavior Therapy and Experimental Psychiatry, 12, 257–259.Williams, K. (2007). Expressive vocabulary test (2nd ed., p. 55437). Bloomington, MN: Pearson.


http://refhub.elsevier.com/S1750-9467(15)00047-1/sbref0005
























































lee 2015

Documents