|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200385Ka 4500
controlfield tag 001 002069311
007 cr mnu|||uuuuu
008 100421s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003192
An assessment of a naturalistic in-home training protocol to establish joint attention responding with children diagnosed with autism spectrum disorders
h [electronic resource] /
by Heather Burris.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 68 pages.
Thesis (M.A.)--University of South Florida, 2009.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
ABSTRACT: Children with autism have deficits in social interactions and verbal and nonverbal communication and engagement in rigid and repetitive activities and/or interests (ASA, 2008). A joint attention (JA) repertoire has been identified as a behavioral cusp for later social development and thus, JA deficits serve as an early indicator for diagnosis of autism spectrum disorder (Vismara & Lyons, 2007; Whalen & Schbreibman, 2003). A JA repertoire consists of both responses to- and initiations for-bids for coordinated attention. Previous research has shown teaching strategies such as pivotal response and discrete trial training for joint attention skills to be effective (Vismara & Lyons, 2007; Whalen & Schreibman, 2003). The purpose of the current study was to evaluate a mixture of pivotal response and discrete trial training as an intervention method for training joint attention behaviors with children diagnosed with autism spectrum disorder in a home setting. In addition, this study evaluated the effects of interspersing targets during training and incorporated generalization probes to assess JA initiations in the form of shifting eye gaze and pointing. Lastly this study examined a parent training procedure to determine if it would promote maintenance after skill acquisition. Results show that all targets were acquired when taught simultaneously. The results show that 2 of the 3 participants made JA initiations during probes throughout intervention. Lastly, the results indicated that parent training did not help maintain JA responding for participant one.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Timothy Weil, Ph.D.
Discrete trial training
Applied behavior analysis
Pivotal response training
x Child and Family Studies
t USF Electronic Theses and Dissertations.
An Assessment of a Naturalistic In H ome Training Protocol to Establish Joint Attention Responding w ith Children Diagnosed w ith Autism Spectrum Disorders by H eather Burris A thesis submitted in partial fulfillment of the requirements for the degr ee of Master of Arts Department of Child and Family Studies College of Behavioral and Community Sciences University of South Florida Major Professor: Timothy Weil, Ph.D. Raymond Miltenberger, Ph.D. Danielle Sutton, Ph.D. Date of Approval: Novem ber 16, 2009 Keywords: Discrete trial training autism, applied behavior analysis, pivotal response training, parent training Copyright 2009, Heather Burri s
Dedication I dedicate this manuscript to my Mom. It was her continuous love and support that encouraged me to pursue and achieve my goals. It was a tremendous journey and I am grateful to have had her with me along the way.
Acknowledgements I would like to acknowledge my advisor, Dr. Timothy Weil for his willingness to support my research interests and his direction, assistance, and guidance with this project. I would also like to thank my committee member and wonderful professor, Dr. Raymond Miltenberger for being on my committee and for providing me with the opportuni ty to attend the Applied Behavior Analys is program under his di rection; for this I will always be grateful. My deepest appreciate goes out to Dr. Danielle Sutton for sharing her knowledge of joint attention, for being a wonderful mentor and friend, and fo r always believing in my abilities as a behavior analyst. I would also like to express my gratitude to Victoria Fogel, who served so many roles during this project. She always provided invaluable feedback and offered me encouraging words which aided in my confidence as a researcher. Lastly, I would like to thank my research assistants Shannon Koehler and Laura Kulikowski for their time, dedication and thoughtful observation on my thesis.
i Table of Contents List of Tables iii iii List of Fi gures i v iv Abstract v i v Introduction 1 1 Behavior Analysis of Joint Attention 3 Methods for Training Joint Attention Skills 4 Pivotal response training 4 Discrete trial training 5 Relevant Research on Joint Attention Training 6 Method 13 Participants Settings, & Materials 13 Target Behaviors 15 Interobserver Agreement 16 Social Validity 17 Fidelity of Implementation 18 Design 18 Procedure 19 Preference assessment 19 Baseline 20 Response training 21 Init iation probes 23 Follow up 24 Parent training 24 Results 26 Mean of Sessions 26 Participant 1 28 Participant 2 32 Participant 3 36 Multiple Baseline and Nonconcurrent Replication 40 Initiation Probes 42 Social Validity 44 Interobserver A greement 49 Participant 1 49 Participant 2 49 Participant 3 49 Fidelity of Implementation 49 Participant 1 49
ii Participant 2 49 Participant 3 50 Discussion 51 References 5 9 Appendices 63 Appendix A: Parent Scoring Survey 64 Appendix B: Paren t Questionnaire for Pre Training 65 Appendix C: Parent Questionnaire for Post Training 66 Appendix D: Data Sheet for Trainer Behavior 6 7 Appendix E: Data Sheet for Child Behavior 68
iii List of Tables Table 1. Parent Questionnaire Results 42 Table 2. Parent Scoring Survey Results 43
iv List of Figures Figure 1. The average percentage of all six targets per session for all three participants 27 27 Figure 2. The percent of opportunities with correct responding for targets 1 3 for participant 1. Circles= change in stimuli. 30 Figure 3. The percent of opportunities with correct responding for targets 4 6 for participant 1. Circles= change in stimuli and arrows=fading of prompts. 31 Figure 4. The percent of opportunities with correct responding for targets 1 3 for participant 2. Circles= change in stimuli. 34 34 Figure 5. The percent of opportunities with correct responding for targets 4 6 for participant 2. Circles=change in stimuli and arrows=fading of prompts. 35 35 Figure 6. The percent of opportunities with correct responding for targets 1 3 for participant 3. Circles=change in stimuli. 38 38 Figure 7. The percent of opportunities with correct responding for targets 4 6 for participant 3. Circles =change in stimuli and arrows=fading of prompts. 39 39 Figure 8. The percent of opportunities with correct responding for all six targets in a multiple baseline across 2 participants and a non concurrent replication with participant 3. 41
v Fig ure 9. The percent of initiations made during probes for all three participants. Blue=baseline, red=intervention, and green= follow up. 43
vi An Assessment of a Naturalistic In Home Training Protocol to Establish Joint Attention Re sponding with Children Diagnosed with Autism Spectrum Disorders Heather Burris ABSTRACT Children with autism have deficits in social interactions and verbal and nonverbal communication and engagement in rigid and repetitive activities and/or interests (ASA, 2008). A joint attention (JA) repertoire has been identified as a behavioral cusp for later social development and thus, JA deficits serve as an early indicator for diagnosis of autism spectrum disorder (Vismara & Lyons, 2007; Whalen & Schbreibman, 2003). A JA repertoire consists of both responses to and initiations for bids for coordinated attention. Previous research has shown teaching strategies such as pivotal response and discrete trial training for joint attention skills to be effective (Vism ara & Lyons, 2007; Whalen & Schreibman, 2003). The purpose of the current study was to evaluate a mixture of pivotal response and discrete trial training as an intervention method for training joint attention behaviors with children diagnosed with autism spectrum disorder in a home setting. In addition, this study evaluated the effects of interspersing targets during training and incorporate d generalization probes to assess JA initiations in the form of shifting eye gaze and pointing. Lastly this study e xamined a parent training procedure t o determine if it would promote maintenance after skill acquisition Results show that all targets were acquired when taught simultaneously. The results show that 2 of the 3 participants made JA initiations during pro bes throughout intervention. Lastly, the results indicated that parent training did not help maintain JA responding for participant one.
1 Introduction Autism is primarily characterized by an im pairment in social interactions and both ver bal and nonve rbal communication (ASA, 2008). O ne of the earliest social impairments to be detected in children with autism is their deficit in joint attention (JA) skills which is thought to serve as a foundation for all subsequent language and social development (Mu ndy, Sigman, & Kasari, 199 0 ) I t is possible to use this repertoire as a marker for potential social/language deficits that have yet to reveal themselves. Difficulties in joint attention skills may include lack of eye gaze to coordinate attention, inabili ty to follow the attention of another person, and an inability to direct attention to an object of interest (Wetherby et al 2004). Joint attention shares various definitions in the literature. Mundy Sigman, and Kasari e capacity of the young child to use gestures and eye contact to coordinate attention with another person in order to share the experience of an while Dawson et al. interactive social partners with respect to objects or events, or to share an awareness of he constructs provided in these definitions have come from the developmental literature and may be difficult for a behavior analysis due to the imprecision of their d efinitions. For example, analysis (for more on this see p. 198). Fortunately, a parsimonious definition is offered
2 by Whalen and Schreibman (2003) that includes a focus on the behavioral event. They describe JA social context Joint attention is broken into two distinct repertoires : a child can respond to joint or eye gaze, or by point ing /gestur ing toward an object or event The second repertoire involv es initiating a bid for JA (IJA), such as eye contact and gestures with another person in order to coordinate attending to an object or event ( Bono, Daley, & Sigman, 2004 ; Mundy & Thorp, 2007 ). These responses to or solicitations for social interaction ha ve been labeled protodeclarative. A clear distinction is made between these bids, which function to solicit social interaction, and protoimperatives which are simply requests for aid in obtaining an object or event ( Bruinsma, R.L. Koegel, & Koegel, 2004; Mundy & Thorp, 2007). Joint attention develops between 6 and 18 months of age and is seen as an important milestone related to promoting language and social development. At 6 months of age babies begin to move eye gaze between a person and an object and soon after use gestures to share attention to something external to the social interaction. Coordinated gaze shift between a person and an interesting object or event in the environment is the earliest topography of joint attention (Tomasello, 1995). Coor dinated attention is observed at approximately 13 months of age, thus setting up conditions that facilitate nonverbal communication (Bakeman & Adamson, 1984). In typically developing children, the emergence of joint attention skills is intertwined with t he development of intentional communication (Bruinsma et al. 2004) as can be seen when children begin learning their first words during joint attention
3 interactions with objects and adults. For example, when a child is playing with a car an adult might touches or picks up the car. This labeling of the car would serve as a prompt for the child to orient to both the object and the parent and would thus be considered a JA response on the part of the child (Vismara & Lyons, 2007). Joint attention has become important to the autism community because a deficit in joint attention can be discovered before language emerges. Thus when a skills deficit is observed in JA ability, it is sensible to target it for remediation in an effort to impact future social and language ability. (Whalen & Schreibman 2003). Behavior Analysis of Joint Attention Responses to joint attention can be described as discriminative control of the tention and are possibly maintained by generalized reinforcers, such as social attention (Dube et al. 2004; Taylor & Hoch, 2008). Similarly, an item or event in the environment. A history of adult attending stimuli as a consequence for JA initiations often results in an increased probability of these behaviors. Dube et al. (2004) describes adult aware of t attending becomes a conditioned reinforcer because the child has a learning history that the adult will react to the event and that this reaction has been related to increased social reinforcement. The adult attending stimuli not only or with respect to the
4 Dube et al., p. 199). The interesting object or event in an environment that includes an adult serves as a n establishing operation. The value altering effect of an interesting event incr eases the reinforcing effectiveness of adult attending stimuli and the evocative effect i ncrease s the frequency of JA initiations that produce adult attending stimuli as consequences. Dube et al. stated that three possible operant classes of initiations for joint attention exist : 1) positive reinforcement when another individual participates in the event or engages with the item; 2) positive reinforcement in which another individual maximizes the reinforcing properties of an item or event; and 3) negative reinforcement by alleviating fear or distress about an item or event. Methods for Training Joint Attention Skills There are two methods for training joint attention skills; these include Pivotal response training (PRT) and Discrete trial training (DTT) Pivotal response training. Behavioral procedures have been used over the years to train several skills for children and adults with varying behavior excesses and deficits. Pivotal response training (PRT) is an intervention method derived from the princ iples of ABA and expand ed to incorporate child development and contextual fit (R.L. Koegel, Koegel, & McNerney, 2001). L.K. Koegel, Koegel, Harrower, and Carter (1999) discuss that PRT focuses on core deficits and excesses of autism that are considered piv otal areas. Early research suggested that targeting specific core areas may result in widespread effects across non targeted behaviors (R.L. Koegel et al. 2001). From this, PRT has developed to focus on these core pivotal behaviors in an attempt to have a great affect in building a behavior repertoire.
5 The components of PRT that were used by Whalen & Schreibman: Pivotal Response Components 1 ) Allowing the child to choose the activity 2 ) Using prompts that are appropriate to the task 3 ) Interspersing maintenance tasks with acquisition tasks 4 ) Taking turns with the child 5 ) reinforcement of alternative (DRA) behaviors was used for approximations 6 ) Prompted responses and unprompted correct responses was reinforced 7 ) Functional response reinforcer relationships was used to ensure a naturalistic training scenario Discrete trial training. Discrete trial training (DTT) is a widely used behavior analytic training procedure that is used to teach a variety of sk ills, such as communication, social interaction, and self care, to children with autism (Smith, 2001). A discrete trial is a three term contingency that begins with a delivery of a discriminative stimulus followed by a prompt (if necessary) that is fade d over time, and then the correct response is followed by a consequence. If the child gives a correct response a reinforcer is immediately delivered, however, if the child gives an incorrect response a variety of procedures can be used, such as error corr ection which involves providing prompting in order for the child to get the correct response and receive reinforcement. After a consequence has been provided the trial is complete and another trial will begin within 5 10 seconds (Smith, 2001). Important components of DTT: Discrete Trial Training 1 ) (when appropriate) 2 ) Training will be one to one (also used with PRT ) 3 ) Training will consist of distinct and repeated trials during a ses sion 4 ) Data collection will occur after each trial
6 Relevant Research on Joint Attention Training A study conducted by Bono et al. (2004) examined the number and variety of intervention s and subsequent language development in children with autism. The interv ention methods used varied among participants and included the following: ABA/ d iscrete trials, e arly intervention programs, so cialization classes, m usic therapy, o ccupational therapy, s ensory integration, s ocial skills training, and s peech language patholo gy. The three most frequent intervention methods in the sample of children w ere ABA, occupational therapy, and speech therapy. The average number of hours per week for intervention was 24, this also included school instruction time. Bono et al. (2004) f ound a correlation between joint attention skills and language development. They found the more frequent ly children engaged in initiations for joint attention, the greater gain in language ability was seen for those children at when joint attention skills are the target of interventions, behavior modification procedures can increase these skills to the extent that even a nave observer can notice a positive change p. 504 ). A limita tion discussed in Bono et al. is the design of the study, in that it was not designed to evaluate or compare the different interventions. Therefore, the efficacy of each type of intervention is not evaluated in this study. Taylor and Hoch (2008) provide us with the opportunity to assess the prompting procedures and social reinforcement contingencies to each child with autism to engage in three components of joint attention: to shift their gaze between an object y respond to bids for joint attention, and to initiate vocal bids for joint attention. Four dependent measures were collected during probe
7 point within 5 seconds; 2) the percentage of trials the child made an appropriate comment about the target item; 3) the percentage of trials the child looked back at the instructor after making a comment about the target item; and 4) the number of joint attention initiat ion bids made by the child. A multiple baseline across participants was used to assess the effects of the intervention on responses to and initiations of joint attention bids. During baseline and probe sessions the environment was manipulated to includ e (Taylor et al., 2008 p. 379 ). The researcher waited 1 minute for the participant to initiate a bid for joint attention, if the child made an initiation the researcher responded with an appropriate social c omment and if the participant did not make an initiation then the researcher start ed a leisure activity with the child and beg a n to make joint attention bids every 30 seconds. No additional prompts or reinforcement w ere provided during these sessions. Du ring training responding to JA bids the trainer initiate d a joint attention bid and then used least to a comment, and to look back at the instructor. During the i nitiating bids training a prompt delay procedure was used to teach the children to initiate JA bids. The instructor walked the child close to a target item and waited 5 seconds to see if the child initiate d a joint attention bid. If no bid was made, the instructor prompted the child using most to least physical and gestural prompts to point and used an echoic prompt to make a comment about the target item. One participant used an index card that had textual prompts and boxes to check off because she wa s not responding to the prompt delay after
8 3 days of training. Generalization was assessed during pre and posttest probe sessions in and around the school where training took place. For a ll participants in the study responding to JA initiations increa sed to 100% within one to five probe sessions after beginning training. Initiating JA bids were rarely made during baseline and training for responding to JA bids. This indicates that training joint attention responding does not lead to collateral change s in initiating JA bids. One potential reason for this could be that any child initiations about items that were not considered important (outside of desired targets) were redirected and not scored. Perhaps ons. In an earlier study, Whalen and Schreibman (2003) utilized behavioral intervention procedure s to specifically target joint attention behaviors to children with autism. They attempted to train JA responding and initiating with a combination of disc rete trial and pivotal response training (PRT) methodologies in an analog setting. The components of PRT include d : providing choices to influence child motivation, reinforcing attempts, using reinforcers related to the task, and interspersing maintenance t asks. It is unclear what specific components of discrete trial training were used during the study. A multiple baseline across participants design was implemented w ith data obtained during baseline, treatment, post treatment, and at a three month follow up. Treatment consisted of two phases: 1) Response training and 2) Initiation training. The specific target behaviors include d : 1) responding to showing, pointing, and gaze shifting of adults; 2) coordinat ion of joint attention; and 3) p ointing. All fiv e children made gains in responding to joint attention and four of the 5 subjects ( one participant did not complete the study ) engaged in variable rates of joint attention initiations. However,
9 s ome of the initiation s such as coordinated gaze shifting, d id not maintain at follow up. Whalen and Schreibman (2003) associate this with the fact that the parents received no training on how to maintain the skills their children had learned. In fact, the training setting (lab) may have contributed to the lack of In a follow up study Whalen, Schreibman, and Ingersoll (2006) investigated whether training joint attention skills may lead to collateral increases in other social behaviors. The procedure similar to the 2003 st udy incorporated response training in the first phase, and initiation training in the second phase The dependent variables in this study were social initiations, positive affect, play, empathetic response, imitation, and language, rather than the joint attention behaviors targeted during the two phases of training. Whalen et al. included language and play probes to establish context separate from the training sessions throughout baseline and treatment in order to assess the social changes. Results of the study suggest that the two phase training procedure produced increases in social initiations for all four participants and the initiations observed post training were at comparable levels of typically developing children. All participants showed incre ases in positive affect empathetic responses and spontaneous language Although Whalen et al. (2006) show ed increases in language, play, and social behaviors, the exact components necessary for these changes are unidentifiable from their research. In a ddition, m any of the collateral behaviors did not maintain from post treatment to follow up. Th is lack of maintenance identifies a need to go beyond acquisition in analog settings with trainers as the primary change agent. It is sensible to conclude that deficits
10 in both generalization and maintenance may be impacted by incorporating the parents in training of these skills and to implement the training in familiar settings Rocha, Schreibman, and Stahmer (2007) furthered joint attention training resear ch by examining a parent implemented intervention program that targeted joint attention responding. The investigators examined the following questions: a) C an parents, with children with autism, be taught to implement joint attention training in a clinica l setting? set generalize to the home? d) Will changes in both the parent and the child maintain over time? a progress? (Rocha et al., 2007). Parents received training in the PRT and DTT teaching methods utilized in Whalen and Schreibman (2003) to t each their child to respond to joint attention bids The DTT components used by Rocha et al. (2007) included using complete trials (discriminative stimulus, response and consequence) teaching the parent to provide a specific instruction, distinguishing b etween appropriate and inappropriate responses, and The PRT components included the use of child choice and motivating toys. At the beginning of each training phase (5 total phases as seen in Whalen & Schreibman 2003; Loveland & Landry, 1986) 20 minutes of didactic teaching took place to describe the specific techniques to be used and the type of joint attention that would be introduced during that phase. A discussion of the benefits of parent train ing, the importance of joint attention skills, examples of social games, and how the parents can
11 create a better joint attention environment in the home took place during this time Handouts were provided for the parents and an opportunity to ask question s was provided. The parents were also asked to give good and bad examples of the use of the procedures. Once procedures were clear to the parents, the techniques were modeled by the trainer with the child for about 10 minutes. After observ ing the skills, the parent implemented the procedures while the trainer provided prompting for each procedure. If a bid for joint attention was not given once every 2 minutes the trainer prompt ed the parent to provide a bid. During subsequent sessions prompting was fa ded to providing examples and explanations every 2 minutes, then 5 minutes, and then only when the parents made a mistake in implementing the procedures. All 3 participants received at least 17 hours of intervention over 6 weeks and the child parent dyads progressed through the phases as the child mastered each phase. Video was scored during the middle 10 minutes of a 20 minute session for the following child behaviors: 1) coordinated joint attention; 2) joint attention initiations; and 3) joint attenti on responding. Generalization probes were conducted during baseline and use the procedures they were learning but no specific instruction was given on how to do this specifically in the home. R ocha et al. (2007) found that t he parents effectively implemented the joint attention training with their children increase in responses during training and generalization sessions The findings fro m this study contributed to the research on JA by expanding joint attention training implementation to parents. The Unstructured Joint Attention Assessment showed that all participants responded to a higher percentage of joint attention initiations after
12 intervention and all children showed an increase in initiation of coordinated joint attention across phases however initiations were not specifically targeted in this study. The design of the study does not allow for specification on how initiations deve loped. All parents were encouraged to implement their new skill set in the home, however, no additional instruction was given on how to do so and as a result none of the parents maintained an increase in joint attention initiation bids during follow up at home (Rocha et al., 2007). This finding establishes a need to conduct a parent training within the home to ensure that the parent is able to exhibit joint attention bids in their natural environment. The previous studies identify the need to examine t he developmental relationship between joint attention responding and initiations specifically looking at how targeting one type of joint attention (responding) might generalize to the other type (initiating) without explicit training By examining the tr aining procedures for joint attention skills it may be possible to discover more efficient and effective methods for training responding and initiating. It is also important for researchers to look at the development of responding to joint attention and h ow that a ffects the acquisition of other, non targeted social behaviors. The purpose of the current study was to examine: 1) the effects of working on all JA responding targets each session as opposed to sequentially ; 2 ) e ffects of training JA respondin g in home as opposed to clinic settings ; 3 ) the effects of specific training on JA responding and how that will facilitate the emission of non targeted JA initiations, and 4 ) the effects of maintenance of JA skills when parents receive a short training ses sion at the end of treatment.
13 Method Participants, Setting & Materials Three children with an autism diagnosis between the ages of 2 and 3 years old were selected for this study. The children all received an independent diagnosis of autism by a psy chologist or physician using the Autism Diagnostic Observat ion Schedule (ADOS) and the Autism Diagnostic Interview (AD I) prior to starting the study. All children in the study were recruited from the Silver Child Development Center at the University of So uth Florida. An initial interview was conducted with each family prior to being asked to participate. During this interview the first author gathered the following information from the parents: types and frequency of therapies, identification of problem 15 minutes at a table. Acceptance into the study required that the child be between the ages of 24 and 48 months have an autism diagnosis from a psychologist or physician, have little to no problem behaviors, and have the ability to sit for 10 15 minutes at a table. During this initial interview the researcher also played with the child in order to assess if the child had the skills that were going to be targeted in the study, several trials of the 6 Participant 1, Dakota, was a 42 month old, non verbal girl who had little to no problem behaviors. Prior to beginning the study she was receiving pivotal response training with a psychologist, however, these services were temporarily put on hold during the study. During the study she received no services and was not attending school. In
14 her initial interview she had good eye contact and was responding to her name (target 4) with the trainer, although she had this skill she was not responding to the other targets and therefore qualified for the study. Participant 2, Riley, was a 30 month old, non verbal boy who had little to no problem behaviors. During the stud y Riley was receiving the follow therapies: applied behavior analysis for 10 hours per week, speech therapy for 2 hours per week, occupational therapy for 1 hour per week, physical therapy for 1 hour per week, and feeding therapy for 1 hour per week. Prio r to starting the study the therapists were informed of the study and requested to not work on responding to name and other joint attention programs. targets and presented as a good par ticipant for the study. Participant 3, Connor, was a 41 month old boy with limited verbalizations who had little to no problem behaviors. During the study Connor was receiving applied behavior analysis therapy for 6 hours per week. He attended school 5 days a wee k and received group speech therapy with 3 other kids for 1.5 hours per week. Prior to starting the study the behavior analyst was informed of the study and requested to not work on responding to name and other joint attention programs. During interview, he had good eye contact and was responding to his name (target 4) ; however, he was not responding to any of the other targets and therefore was still included in the study. The study was conducted in the homes of the particip ants/guardians. Baseline, treatment, and follow up sessions took place in The bedroom includ ed a small table with a chair for the child, baskets with high preference
15 toys, stimuli that are typically present in the room and a video camera (during a majority of the sessions). The following materials were used in this study: camera recorder, pen, tragus tripod, three ring notebooks, plastic bins, joint attention data sheets for child behav ior and for trainer behavior, Sony USB flash drives, and an assortment of high preference toys (varying for each participant). Target Behaviors Targets 1 3 are considered prerequisite skills for responding to joint attention bids and are thought to in crease responding to JA, however, because it is unclear in the literature whether they serve as prerequisites all of the targets were trained simultaneously. Targets 4 6 are responses to joint attention bids. The definitions of the target behaviors are a s follows: 1 Target 1 Response to ha nd on object: While the child was playing with one t oy, the trainer placed rent toy. A correct response include d taking the toy within 5 seconds and manipulating the newly presented toy for a t least five seconds. 2 Target 2 Response to object being tapped: The trainer tapped a newly presented toy while th e child was engaged with another toy (same protocol as target 1). A correct response include d taking the toy within 5 seconds and manipula ting the newly presented toy for at least five seconds. 3 Target 3 Reponse to showing of object: The trainer showed the child a new toy while the child was engaged in another activity (same protocol as target 1
16 and 2). A correct response include d taking the toy within 5 seconds and manipulating the newly presented toy for at least five seconds. 4 Target 4 Eye contact: The trainer said simultaneo usly present ed an object near her eyes and then faded to nose until not object w as necessary to establish eye contact with the child. Ma king eye contact within 5 seconds was scored as a correct response. 5 Target 5 Following a point: When the child was engaged with an object, the trainer establish ed eye contact with the child by fol lowing the steps in target 4 fading to not at all). Following the establishment of eye contact, the trainer turned her head and point ed to another object in the room. The goal was for th e child to turn his or her head in the same direction as the traine r. A correct response included the child looking at the trainer and then shifting his or her gaze to the object specified by the trainer. 6 Target 6 Following a gaze: Target 6 looks simi lar to target 5. The trainer shift ed her gaze, rather than point. The same topography of response was scored. Interobserver Agreement Interobserver agreement was scored by two trained research assistants by independently viewing videoed sessions. For targets 1 3 the observers needed to agree with the first author on whether or not the child was engaged with a toy, whether or not the child manipulated the newly presented toy for approximately 5 seconds and then whether or not the trials were scored as correct. For target 4 the observers needed to
17 agree whether or not the child made eye contact and whether or not the trial was scored as correct. For targets 5 and 6 the observers needed to agree on whether or not the child was engaged with a toy, made e ye contact, turned head in direction of point or eye gaze, and whether or not the trial was correct. Interobserver agreement was calculated by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100. Soci al Validity The parents in the study were asked to fill out a pre intervention and post intervention survey to assess the percentage of time they feel their child : 1) makes eye contact while playing; 2) initiates social interaction using eye contact while playing; 3) responds to social initiations using eye contact while playing; 4) initiates social interactions using verbalizations while playing; and 5) responds to social interactions using verbalizations while playing. See Appendix A. All of the questions were explained and questions were answered prior to the parents filling it out to ensure there was no intervention survey and therefore a post intervention survey was not adminis tered. The parents also filled out a parent questionnaire prior to starting training. The following questions were asked: 1) do you feel this training is important to your child; 2) nt do you feel it is you to work with your child; and 5) how many days per week is it possible for you to work with your child. The parents were given a follow up quest ionnaire after their child completed the study, the following questions were asked: 1) do you feel that your child
18 benefitted from this training; 2) do you feel that you benefitted from receiving training; 3) approximately how much time did you spend a day working on joint attention with your child; 4) approximately how many days per week did you spend working on joint attention with your child; 5) would you want your child to participate in future studies, similar to this one; and 6) how do you feel about having a research er in your home. At the bottom of the questionnaire was a spot for the parents to insert additional comments and thoughts about the study. See Appendix B. Fidelity of Implementation Fidelity of implementation was scored by reviewing v ideo and scoring the s behavior for a percentage of the sessions. This measure assess ed correct execution of joint attention bids by the researcher and correct application of See Appendix C for an exa mple of the data sheet. Design A concurrent multiple probe across two participants was utilized. A nonconcurrent replication was used for the third participant. A baseline probe was used to save time overall, and to avoid extensive baseline sessions for the children. Session data was collected during baseline, response training intervention, and at follow up See Appendix D for example of data sheet used. Baselines were staggered across participants in order to control for extraneous variables, mat uration and exposur e to the treatment. Dakota was in baseline for 6 sessions or 20 days She was in intervention for 2 4 sessions or 60 days Riley was in baseline for 70 days and had 14 probe session s during that time. He was in intervention for 24 se ssions which was 68 days Connor was in baseline for 6
19 sessions, lasting 14 days He had 13 intervention sessions which was 42 days. Initiation probes took place at the end of every session to evaluate whether or not the child would point or shift eye g aze with the trainer without any previous training. took place 2 months after the completion of intervention, due to a family situation which prevented her from receiving fol low up sessions to ok place 1 month after the completion of intervention and sess ions are expected to take place approxima tely 1 month after intervention. Procedure Preference a ssessment A multiple stimulus without replacement, preference assessment (MS WO) (DeLeon & Iwata, 1996) was conducted with all participants. The procedures for the MSWO follow ed protocol as described in Carr, Nicolson, and Higbee (2000). At the beginning of each preference assessmen t session the investigator place d a linear array of eight stimuli on the table in front of the participant. The eight stim uli included in the array were picked using parents of what may serve a reinforcing function (visual display, sound, etc). The children wer e All participants picked a stimulus with the first instruction. After the stimulus was selected, the child was allowed access to the item for 10 s and then it was removed from the array. Any attempt to select more than one stimulus at a time was blocked and the initial verbal instruction was repeated. Following the selection of a stimulus, the remaining stimuli were repositioned in a quasi rando mized manner. This process continue d until all stimuli were selecte d and was were ranked from first picked which was recorded as a 1 (highest) to last pick which was recorded as a n 8
20 (lowest). Overall ranking of stimuli was computed from the aggregat e of the three rankings. For instance, if a bear was chosen 1 st 3 rd and 2 nd the average of these scores was 2. If, during the same assessment a car was chosen 3 rd 1 st and 1 st the average of these rankings was 1.66. Thus, the car would be ranked as most preferred (1 st ) and the bear would be ranked second. The stimuli were considered high preference if ranked in the top five and w ere utilized during training. The preference assessments were conducted before the first baseline session An addition al MSWO was conducted for Riley and Connor, due to decreases in data which suggested that the stimuli were no longer exhibiting reinforcing properties. The toys were placed in a basket and the children did not have access to the stimuli outside of trainin g sessions with a researcher. During the preference assessment the trainer worked on building r apport with the child. The trainer used t his opportunity to pair herself with the preferred items in order to establish them as a conditioned reinforcer. Bui lding rapport was important in order to Baseline Baseline measures took place approximately three times a week (with the exception of the participants being sick or going out of town in which 1 or 2 sessions occurred), with each session lasting 15 minutes researcher asked was then directed to the table and given toys to play with while the train er set up the stimuli and video camera and then the session began. The child sat on one side of the table and the therapist sat on the opposite side. Next to the trainer was a basket with the toys identified as highly preferred during the MSWO. The chil d was assessed on the six target behaviors listed previously.
21 The trainer provided five opportunities for the child to respond to each of the 6 targets resulting in 30 trial presentations during all sessions. These targets were randomly dispersed through out the session; no more than 2 trails of the same target were conducted consecutively. If an error occurred, the trainer marked the trial as incorrect and presented the next trial. was scored as correct or inc orrect, however, no con sequence was provided for correct, incorrect, or no response. Prior to presenting the discriminative stimulus for targets 5 and 6, the researcher gained saying his or her name ntion (this was later faded out) and then turned and point ed or shifted eye gaze to an item to the left or right of the child. An example of target 1 look ed as follows: the child w as already playing with a toy and then the trainer placed on a di fferent toy than the child was currently engaged with (discriminative stimulus) and wait ed to see if the child st art ed to manipulate the new toy within 5 seconds. The child was given 5 seconds to emit a target response (such as engaging with the toy or making eye contact ). Inclusion in this study required that the participants did not have an average above 40% correct for all responses during baseline. Once baseline was stable for the Dakota response training was implemented. Riley continued in baseline until desirable changes in level and trend were observed for the first child. nonconcurrent he started independent of Dakota and Riley. Response training Intervention consist ed of Response Training of 15 20 minute sessions in which the participants were taught to respond to protodeclarative joint attention bids from a researcher trained in both DTT and PRT methodologi es. Response training involved teaching the child to respond appropriately to joint attention bid s
22 provided by the researcher. For example, when attempting to evoke a JA response from the child (orienting, engaging the item), the trainer may have toy; the appropriate response was for the child to engage with the newly present ed toy for at least 5 seconds. A trial start ed with the delivery of the discriminative stimulus (SD) and an opportunity for the child to respond. During the first session for all participants the first 3 trials were presented and an initial prompt was pro vided followed by praise for engaging in the desired response. When the child respond ed correctly (manipulate d or engage d with an object for at least 5 seconds g ave eye contact to trainer, shift ed gaze in direction ) the tra iner provide d a variety of social positive statements for approximately 5 seconds. If the child d id not emit the desired response within 5 seconds, a correction procedure was implemented using a least to most prompting hierarchy. The trainer provided a gestural prompt, such as m odeling the appropriate response, blocking another stimulus, and/or pointing and then a physical prompt, such as Following a correction procedure than when the skill was demonstrated independent of error correction. Implementation with Dakota varied from the above procedure slightly, in that every time error correction was used for a target, the follow ing trial for that target was started with a prompt. Starting with a prompt resulted in limiting the number of opportunities she had to respond correctly (this will be discussed further in the discussion section). Following each trial (prompted or unprompted) an intertrial interval of 5 10 seconds elapse d prior to the start of the next trial. During this time the trainer recorded response as correct or
23 incorrect. The targets were randomly dispersed throughout the session; thus, all five opportunities for one target were not given in sequence and no more than 2 consecutive trials for a target were given Mastery criteria for all targets were met when each child g ot 80% for three consecutive session s. After a skill was mast ered, it was continued throughout until all targets were mastered and all prompts were faded. Training sessions were video recorded in order for scoring and interobserver agreement to be calculated later. Initiation Probes Initiation probes were conduc ted at the end of every second session throughout the training period and at follow up. The trainer place d a n attention grabbing and/or obscure object out of reach from the child (Taylor & Hoch, 2008). An example of a stimulus that was used is a Phlat ba ll, which is pressed down into a flat disc and after 5 10 seconds pops open into a ball. The trainer pretend ed to collect data and watched to see if the child made any joint attention initiation s such as looking at the trainer and then pointing to the sti mulus or looking at the stimulus and then at the trainer. If the child d id initiate joint attention the trainer provide d verbal praise and a statement about the and the chil d d id not make an initiation then the stimulus was removed and the session was ended. I n the event that a child engaged in a JA initiation during response training (outside of the initiation probes), the trainer provided social positive attention in an attempt to make this response more probable in the future while not explicitly ta rgeting/training initiations. It was noted in the daily notes if initiations occurred during the session.
24 Follow up Follow up measures were collected 2 month s after comple ting response training for Dakota and w ere collected 1 month after completing response training for Riley. ve yet to be collected but will be collected approximately 1 month after the completion of his training. Follow up adhere d to the same protocol as baseline. Three, 15 minutes sessions were conducted at follow up using the same stimuli that was used during baseline/intervention. Parent Training The parents of Dakota and Connor received parent training within 1 week of th eir child complet ing intervention. parents received parent trai ning after their child completed follow up i n order to assess whethe r or not the parent training had any affect on maintenance of the responding to joint attention skills. The traini ng took place in home after the completion of responding to joint attention training. The training sessions lasted between 1.5 and 2 hours and followed behavior skills training (BST) protocol. Behavior skills training consist ed of giving instructions, mo deling the desired behavior, having the parent rehearse the skills with the trainer and with their child and providing feedback to the parent. Training bega n with providing details to the parents about responding to joint attention, including why it is important and what it looks like. The parents were given definitions on the targeted responses (targets 1 6 above). Next the trainer model ed how the parent should initiate joint attention in order to get the desired responding from their child. The prim ary researcher model ed th e targets with the parent playing the role of the therapist and the trainer playing the role of the child were
25 used as the stimuli during the modeling and rehearsal phases. The researcher modeled target 1, then the parent rehearsed target 1, and the trainer provided feedback. This process was completed for all 6 targets. After completing all 6 targets the parent s went through all targets with their child and the trainer provide d feedback and answered a ny questions that the parents had.
26 Results Mean of Sessions The use of PRT and DTT to train responding to joint attention skills proved to be effective for all 6 targets across all 3 participants. The averages of all 6 targets for ea ch session were calculated in order to determine the average for a session and the overall progress for all of the targets. For example, if the participant r esponded correctly on 20% of opportunities for target 1, 40% for target 2, 20% for target 3, 60% for target 4, 0% for target 5 and 0% for target 6, these percentages were added together (totaling 140) and then divided by 6 resulting in a score of 23.3% for the session. During baseline Dakota had a mean of 28.7% for all targets (see Figure 1). Her m ean during intervention was 68.5% and her mean during follow up fell to levels similar to baselin e with a mean of 24.4% during baseline was 26.7% and his mean increased to 76.9% during intervention. His follow up mean was 16.6% (see Figure 1). baseline was consistent with Dakota and Riley with 29.4%, only slightly higher than Dakota. His mean during intervention was also consistent with the other two participants with 73.8% (see Figure 1). No follow up data is availabl e for Connor at this time.
27 Figure 1. Average of all six targets per session for Dakota, Riley, and Connor. BL=Baseline, RT (A)= Response training with stringent criteria and RT= Response training as described in the procedure section. Dakota Riley Connor Average Percentage for Each Session Sessions B/L Response Training (A) RT Follow up B/L RT
28 Participant 1 Data for Dakota show ed that she was able to acquire all six targets by the conclusion of the intervention. During baseline target 1 display ed a downward trend. At the start of intervention target 1 was at 0% (see Figure 2). Targets 2 and 3 leveled out at 40% prior to the start of intervention Target 4 was on an upward trend during baseline, the study (see Figure 3). During baseline targets 5 and 6 stayed at 0% for all six sessions The procedure protocol required initial prompting for the first three trials for all targets. The stringent procedure resulted in low percentages of independent correct responding during the first session for all participants. No initial prompting was conducted for target 4 because this was a skill that Dakota was exhibiting during baseline, however, a drop in the data occurs during session 9 (3 rd interve ntion session) from 80% to 20%. It is not clear why Dakota responded poorly during this session. Dakota mastered targets 1 and 2 after 12 intervention sessions. She mastered target s 3 and 4 after 7 sessions. Targets 5 and 6 took much longer for Dakota to ma ster. She mastered target 5 after 19 sessions and target 6 after 21 sessions (see Figure 3 ). A second phase line was put in for Dakota due to a slight procedure change. The change is referred to as response training A in the graphs. The data displayed a drop to 60% during session 21 for target 2 and during session 22 for target 3 The data for target 5 w ere variable, between 60% and 20% and target 6 was leveled at 0%. The drops and variability in the data w ere determined to be a result of the restrict iveness of the procedure. The procedure required that a fter an error occurred the trainer immediately provid ed a prompt with the presentation of the SD for the next trial for that target. For
29 example, if an error occurred for target 2 the next presentat ion of target 2 the trainer would tap the item and then immediately put her hand on the item to begin playing with the toy. If Dakota made o ne error with one target then the highest percentage she could obtain would be 60% for that session. Two errors w ould result in the participant getting 20% for the session. A phase line was put in and the procedure described above was used for the rema ining sessions. This change allowed Dakota to have more independent opportunities to respond, which resulted in he r m astering target 5 in 3 sessions and target 6 in 5 sessions. After the change in procedure targets 1 3 stayed at 80% or higher until the end of intervention (see Figure 2). During session 2 4 Dakota drop ped to 60% for target 4 Dakota was observed to attend little during this session. She look ed out the window quiet often and her little brother knocked on the door, which was seemingly distracting to her. intervention and 2 mo nths later follow up data were taken. During follow up Dakota presented levels similar or higher than baseline for targets 1 3 and 5 6 and f or target 4, she presented levels at follow up that were lower than what was seen during baseline (see Figure 2 & 3 ). For target 1 she averaged 30% during baseline and for follow up she averaged 46.7%, slightly higher than baseline. For target 2 she averaged 26.7%% during baseline and averaged 33.3% during follow up. For target 3 she averaged 35% and during follow up she averaged 33.3%. Targets 3 and 4 were on a downward trend at the end of intervention. The average for target 4 during baseline was 80% and during follow up it was 33.3%. The average for targets 5 and 6 during baseline was 0% and during follow up wa s also 0%. This will be discussed further in the discussion.
30 Figure 2. Targets 1 3 for Participant 1. T he circles indicate when stimuli w ere changed. Target 1 Target 2 Target 3 Percent of Opportunities with Correct Responding Sessions B/L Response Training (A) RT Follow up up
31 Figure 3. Ta rgets 4 6 for Participant 1. The arrows indicate when prompts were faded and the circles indicate when stimuli were changed. Target 4 Target 5 Target 6 Percent of Opportunities with Correct Responding Target 5 Target 4 Percent of Opportunities with Correct Responding B/L Response Training (A) RT Follow up Sessions
32 Participant 2 Probe s were conducted during baseline for Riley in an effort to shorten the exposure to baseline testing Durin g baseline t arget 1was between 40 % and 60%, with the exception of session 1 where it was 100% and was on a slight downward trend (see Figure 4 ). Targets 2 and 4 showed quite a bit of variability with a range between 80% and 0% during baseline. Although th e data points for target 2 were variable they were on a downward trend (see Figure 4 ). Target 4 displays a slight an upward trend at the start of intervention Target 3 was moderately variable during baseline with a range between 60% and 0% and displayed a downward trend Target 5 had one session at 20% and all other sessions were at 0% and target 6 maint ained at 0% throughout baseline (see Figure 5 ). As with participant 1, t he first 3 trials of each target were prompted during the first session of inte rvention. D uring intervention target 1 ranged from 40% to 100%, showing only moderate variability. Target 2 had dips during session 34 and 37 to 60% but maintained between and 80% and 100% after that (see Figure 4 ). Qui te a bit of v ariability was see n for target 3, ranging between 40% and 100%. After mastering target 3 he dropped to 60% during sessions 41 and 48. Target 4 had one drop to 60% during session 48. Targets 5 and 6 alternated between 80% and 60% after mastery for several sessions These dips in data are thought to have been due to a motivational issue because he had already mastered the targets and variability was still seen. Riley mastered targets 1 and 2 after 4 sessions. He mastered targets 3 and 4 after 8 sessions. Similar to Dako master targets 5 and 6. He mastered target 5 after 14 sessions and target 6 after 12
33 sessions. The sessions continued after mastery in order to fade out the embedded prompts (objects near eyes and then nose) for targets 4, 5 and 6. No parent training was conducted with following completion of the intervention phase. Target 1 maintained at 40% throughout follow up which is well below the average for baseline of 52.9%. Targets 2 and 3 show a downward t rend during follow up. Riley averaged 35.7% during baseline for target 2 and average d 26.7% during follow up. He averaged 34.3% during baseline for target 3 and average d 26.7% during follow up. Target 4 shows a slight increasing trend, however, the data points are well below what was seen during baseline. He averaged 35.7% during baseline and had an average of 6.7% during follow up. Targets 5 and 6 returned to 0% and maintained at 0% throughout follow up.
34 Figure 4 Targets 1 3 for Participant 2. T he circles indicate when stimuli were changed. Percent of Opportunities with Correct Responding Target 1 Target 2 Target 3 B/L RT Follow up Sessions
35 Figure 5 Targets 4 6 for Participant 2. The arrows indicate when prompts were faded and the circles indicate when sti muli were changed. Percent of Opportunities w ith Correct Responding Target 4 Target 5 Target 6 B/L RT Follow up Sessions
36 Participant 3 Connor had slight variability for target 1, rang ing between 80% and 40% during baselin e, but he was on a downward trend prior to intervention (see Figure 6 ). Target 2 displayed moderate variability and ranged between 60 % and 0% and was on a downward trend at the start of intervention (see Figure 6 ). Target 3 ranged between 40% and 0% and was on a downward trend at the s tart of intervention. Target 4 was already in the art of the study but i t was on a downward trend at the start of intervention and ranged between 100% and 60% (see Figure 7 ). Targets 5 and 6 were at 0% throughout baseline. The first 3 trials of each target, with the exception of target 4, were prompted during the first ses sion of intervention, so it was expected that the first data points of intervention would be low. Data for targets 1 3 are qui te variable dipping to 60% even after mastery (see Figure 6 ). It is hypothesized that this was a motivational issues rather than lacking skill acquisition based upon observation of the first author. He had already mastered the skills but was observed not wanting to take the newly presented toy. Connor had one dip during intervention for target 4, due to prompt fading, and then m ainta ined at 100% (see Figure 7 ). The data for target 5 displayed a dip to 20% during session 12 but maintain ed an upward trend throughout. Target 6 was on an upward trend throug hout intervention and maintained at 100% after mastery (see Figure 7 ). Conn or mastered target 1 in 5 sessions. It took 12 sessions for Connor to master target 2 Connor mastered targets 3 and 4 in 6 sessions. Target 4 required 6 sessions due to the fading of prompts (from eye level to nose level) which occurred during the seco nd session (before mastery). Target 5 was mastered after 11 sessions and target 6 was mastered after 9
37 sessions. With the exception of the second target targets 5 and 6 were acquired after the other targets met mastery. was consistent wi th the performance seen with participants 1 and 2. No data h ave been collected on follow up for Connor at this time.
38 Figure 6 Targets 1 3 for Participant 3 T he circles indicate when stimuli were changed. Target 1 Target 2 Target 3 Percent of Opportunities with Correct Responding Target 1 Sessions B/L RT
39 Figure 7 Targets 4 6 for Participant 3. The arrows indicated when prompts were faded and the circles indicate when stimuli were changed. Target 4 Target 5 Target 6 Percent of Opportunities with C orrect Responding Sessions B/L RT
40 Multiple Baseline and Nonconurrent Replication Dakota and Riley completed interventi on in 2 5 sessions and Connor completed intervention in 1 4 sessions. Dakota and Riley mastered their skills in a very similar order. Dakota mastered target 4 (eye contact was strong when she started the study) then target 3, next targets 1 and 2, and fina lly targets 5 and 6 (see Figure 8 ). Riley mastered targets 3 and 4, then targets 1 and 2, and finally targets 5 and 6 (see Figure 8 ). Connor mastered targets in a slightly different order, he mastered targets 1 and 3, then target 4 (eye contact was strong when he started the study and prompts were faded out early), then targets 6 and 5, and finally target 2 (see Figure 8 ). For all three participants targets 5 and 6 were acquired after most of the other targets, although it is less dramatic with Connor. I n figure 5 you can see dips in the data to 60% during intervention. These dips were more prevalent for Riley and Connor and occurred one time for Dakota during the second phase of intervention (same procedure used for Riley and Connor). The decreases for Dakota during the first phase of intervention can be related to the stringent procedure and artificial ceiling that was put in place. The variability that is seen throughout intervention will be discussed further in the next section.
41 Figure 8 Percentage of opportunities for correct responding for Dakota, Riley, and Connor during baseline, intervention and follow up. Percent of Opportunities with Correct Responding B/L B/L Response Training (A) RT RT Follow up Sessions
42 Initiation Probes Initiation probes were taken at the end of every session to see whether or not the child wo uld begin making initiations, although this was never specifically trained. The purpose of the initiation probes was to compare with previous studies which state that initiations must be trained separately and do not develop during responding to joint att ention training (Taylor et al. 2008). Dakota made eye gaze initiations; (no pointing initiations were made), during the intervention probes 41.7% of the time (see Figure 9). During baseline and follow up she made no eye gaze or point initiations. Riley made no initiations during the probes throughout baseline, intervention, and follow up. Connor made eye gaze initiations; no pointing initiations were made, during the probes in intervention 14.3% of the time (see Figure 9 ). No follow up data on initiat ion probes are provide d for Connor because he has not completed follow up.
43 Figure 9. Percentage of initiation s made during probes in baseline (blue) intervention (red) and follow up (green) for all three participants.
44 Social Validity The parents for all three participants filled out a questionnaire prior to starting the study. In the questionnaire they all felt as if this training was very important for their e parents said that they could spend greater than 30 minutes a day working with their child. hild 3 4 days per week. In the post questionnaire administered after the completion of follow up, ed that they spent less than 10 minutes per day working with her and spent 3 4 days working with her per week. The results from t he social validity involved in the study and that they benefitted from receiving parent training. The to be involved in a study similar to this one and also liked having the researcher in their home (see T able 1). much from this training. She now makes eye contact most of the time. She is willing to understand what the instructions are. She responds to her name and follows simple instructions with hand gestures. With more training similar t o this, Dakota will thrive in during this study. His parents also reported that they spent 20 minutes a day working on
45 joint attention skills for more than 5 days a week. Working with their child 5 days a week is consistent with the pre questionnaire but they spent less time than reported during the pre questionnaire (see Table 1). His parents reported that they would like their child to participate i n studies similar to this one in the future and that they liked having a has improved with following a point and responding to his name we could see the The post questionnaire at this time because he has not yet completed follow up. A parent scoring survey was also given to the parents of Dakota and Riley, not return the scoring survey and therefore no data were collected. The parent scoring survey allowed the parents to score the average percentage of time they felt their child engaged in joint attention initiating and responding. The survey enabled the researchers to determine if th e parents saw a change in their child initiations or responding to initiations during play time. progress in the average percent of time she spent engaging in joint attention behaviors for 4 of the 5 questions. She made progress in the average percentage of time she made eye contact while playing, initiated social interaction using eye contact, responding to social initiations using eye contact, and responded to social interactions using verbalizations while playing They reported that prior to intervention, the average percentage of time Dakota initiate d social interactions using verbalizations while playing was 10% and post intervention they reported that this was at 0%. The increase in ratings ranged from a 40% increase to a 10% increase.
46 pre to post intervention. They reported a 30% increase for 4 of the 5 questions. The other question was reported as a 20% increase (see Table 2).
47 Table 1 Note. The post questionnaire was not filled out by P3 because he has yet to complete raining.
48 Table 2 Parent Scoring Survey Results P1 P2 Pre Post Pre Post Avg % of time child makes eye contact with you while playing 80% 90% 10% 30% Avg % of time your child initiates social interaction using eye contact while playing 50% 90% 10% 40% Avg % of time your child responds to social initiations using eye contact while playing 50% 70% 10% 40% Avg % of time your child initiates social interactions using verbalizations while playing 10% 0% 0% 3 0% Avg % of time your child responds to social interactions using verbalizations while playing 0% 20% 0% 30% Note. The parent scoring survey was administered before the participants started baseline and after they completed intervention (before par ent training and follow up).
49 Interobserver Agreement Interobserver agreement was calculated by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100. Participant 1 R e liability data were taken for 52% of all videoed sessions Agreements averaged 97.3% across all observations, with a range of 92% to 100%. Participant 2. R eliability data were taken for 49% of all videoed sessions Agreements averaged 94.6% across all observations, with a range of 88% to 99%. Participant 3. R eliability data were taken for 56.3% of all videoed sessions Agreements averaged 96.9% across all observations, with a range of 91% to100%. Fidelity of Implementation This measure assess ed correct execution of joint attention bids by the researcher See Appendix C for an example of the data sheet. Participant 1. Treatment integrity was scored for 72% of the video s that were recorded and interobserver agreement was taken for 62% of those videos. Treatment integrity for implementation of all six targets averaged 97.6%, with a range of 64% to a range of 95% to 100%. Participant 2. Treatment integrity was scored for 56.3% of the videos that were recorded and interobserver agreement was taken for 72.2% of those videos. Treatment integrity for implementation of all six targets averaged 99.9%, with a range of 93% to 97% to 100%.
50 Participant 3. Treatment integrity was scored for 56.2% of the videos that were recorded and interobserver agreement was taken for 50% of those videos. Treatment integrity for implementation of all six targets averaged 99.5% with a range of 91% to 100%. 98.5% with a range of 96% to 100% This data does not include anything fr om follow up, as follow up has yet to be collected.
51 Discussion The purpose of the present study was to evaluate an in home training protocol that taught the above six targets non sequentially, to examine the effectiveness of providing a short parent training and what effects that may have on maintenance results, and to determine if training responding to joint attention would generalize to joint attention initiations. The r esults suggest that the targets may develop in a systematic way (5 and 6 com ing on board at the end of intervention ) but training concurrently is an effective way of training. Participants 1 and 2 acquired targets 5 and 6 last and participant 3 acquired targets 5 and 6 near the end of his training. The results also show that pa rticipant 1 and 3 made JA initiations during probes throughout intervention. Lastly, the results indicated that the short parent training provided at the end of intervention did not help maintain JA responding for participant 1. Previous research taught the above six targets in a sequential order, however prior to this study there was no research to support that this was necessary. This study provides support to the previous literature, in that the participants acquired targets 1 4 prior to targets 5 an d 6, with the exception of Connor Connor acquired target 1, 3, and 4, then 5 and 6, and finally target 2. It is hypothesized that Connor took longer to acquire target 2, tapping object, not because he did not have the skill but because he was not inter ested in the materials being presented to him. Although preference assessments were conducted during the study (when the researcher felt the child was becoming satiated
52 with the materials) the items being presented were always competing with the item that the child was already engaged with Whalen and Schreibman (2003) also discuss that after the child responded to a joint attention initiation the child was given the item that he or she attended to. The current study provided only social attention in th e form of praise and tickles. The definition of joint attention is the ability to coordinate attention between an object and a person in a social context (Whalen & Schreibman, 2003). By providing a toy (tangible) to the child for engaging in joint attent ion responding or initiating it seems you may be reinforcing a request rather than joint attention (a shared interaction). The results of parent training for Dakota indicates that the parent training did not aid e in maintenance of the six targets, howeve r, follow up was taken 2 months after the one month after he completed intervention. His results conclude that he did not maintain the joint attention responding, as w as seen with Dakota. His parents received no training until after follow up was taken, therefore, no parent training was in place to influence maintenance performance. During the post worked with him for app roximately 20 minutes a day, 5 days a week on joint attention skills. Because they had no training after intervention it is unclear what exactly was work ed on and maintenance data show that what the parents did work on did not help maintain the 6 targets that were taught during the study. Initiation probes were conducted in order to determine if training responding to joint attention would generalize to initiation of joint attention bids. Dakota had the most initiations during these probes and also dur ing her sessions. She pointed along with the
53 trainer three times during her intervention sessions and shifted eye gaze with the trainer five times during one session. Dakota also began saying novel words including her name, bubbles, and lion during her s essions. pointing and saying new words, such as cracker and saying numbers 1 5. This is a significant observation because Dakota was no t receiving any other intervention throughout the study. Connor also made e ye gaze initiations during the initiation probes but had no observed initiations during his sessions. Connor began saying novel words during the study and increased his length of utterances from 1 word to 2 3 words. Because Connor was also in therapies d uring the study, it is unclear whether these increases in language are linked with the study. Riley had no initiations during his probes but was observed pointing during one session. It is important to note that during the probe sessions, both Connor and Dakota only made eye gaze initiations. Pointing initiations were only made by Dakota and Riley during sessions. Pointing initiations never occurred during the probe sessions. It is unclear why eye gaze initiations would occur and pointing initiations wo uld not. The social validity questionnaire given from pre to post intervention showed benefitted from the parent training and that they would want to be involved in similar 10 minutes per day 3 4 days per week. In their initial questionnaire they reported that they were able to work with her for more than 30 minutes per day for more than 5 days per week. The researcher was unable to determine why the parents spent less time and fewer days working on joint attention with their child.
54 During this study a short training was conducted to train the parents to implement the pro cedures discussed above, however, after this training the researcher provided no other training or follow up with the parents. Also, the parents were not given any specific instructions on how often or how much they should work with their child. The pa skills (responding and initiating) during play from pre to post intervention. Dakota and they engaged in joint attention skills while playing (see Table 2), with the exception of Dakota making initiations with verbalizations during play (which the parents indicated a decrease from 10% to 0%). No survey was filled out for Connor, as it was giv en to his parents but never returned during pre intervention. There are several limitations to th is study that center on method and design. One limitation of this study is the change in procedure for Dakota. The strict procedure resulted in placing a n artificial ceiling on her scores due to the inclusion of a prompt, to aide responding on trials following errors. Because of inclusion of the prompt, Dakota was not able to respond independently and thus, her % measure for the session was capped. For e xample, o ne error automatically resulted in getting 60% for that target. After removing the initial prompt (after an error occurred) from the training procedure an increase in all targets was seen Another limitation of this study is that the third pa rticipant was conducted as a non concurrent replication. It would hav e been preferred for him to be included as part of the multiple baseline across participants. During recruitment of participants it was difficult to find participants that qualified for the study who were able to sit in a chair for
55 10 15 minutes and who had little to no problem behaviors. The first three participants were found; however, the first participant was dropped prior to starting baseline because he was engaging in tantrum like behaviors, such as kicking, screaming, and crying, when left alone with the primary researcher. The other participants had already started their MSWO and were ready to start the study, therefore, the research team decided to go ahead and start them and f ind a third participant to be conducted non concurrent of the other two. The last identified limitation of this study was the participants involvement in other therapies. Only 1 of the 3 participants did not have any other therapies during the study. When examining collateral effects it is difficult to identify if the changes were a result of the current study or a result of other therapies they were receiving. Although the researcher communicated with the therapy providers of the participants, there is no way of knowing whether or not they refrained from working on joint attention skills during their sessions. Future research should examine the effects of establishing and abolishing operations in relation to joint attention training. Throughout th e study it was observed, although not specifically targeted, that the participants would not orient to the presented items or the items that the trainer wanted them to look at if it was competing with a higher preferred item. For example, if the child was engaged with a high preferred item and another high but less preferred at the moment was presented, then it was likely that the child would not respond (pick up the new toy and manipulate it). It was also observed that attention grabbing items, such as i tems that move and/or make sounds, needed to be used for targets 5 and 6 in order to gain their attention and for them to orient
56 to the item (Taylor & Hoch, 2008). For Dakota, the trainer had been pointing and shifting eye gaze (target 5 and 6) to an Eey ore stuffed animal and was getting little to no response from her. Once the trainer changed the stimuli to bubbles and other toys that Future research should also ex amine the number of s essions or amount of time it takes for participants to acquire joint attention responding when the skills are taught sequentially in an in home setting. More research is necessary to determine whether or not training simultaneously or sequentially is the more efficient way of training. Participant 1 and 2 acquired the targets in 25 sessions (approximately 2 months) and particip ant 3 acquired the targets in 15 sessions (approximately 1.5 months). Whalen and Schreibman (2003) report that response trainin g took between 16 and 26 days when t argets 1 6 were taught sequentially h owever, their training took place three days a week for 1.5 hours. The current study took place three days a week for 15 20 minutes. Twenty five sessions broken down into total nu mber of hours equals 8.3 hours, this is the most time that intervention took for the current study. For the Whalen and Schreibman (2003) study the most time it took was 39 hours, provided the number of days is number of sessions. They provide no interven tion data that would allow us to know if the number of days it took to acquire the skills is equal to number of sessions. Substantially less time was spent providing intervention to the participants in the current study and future research should expand o n this by examining how much time is really necessary for children to acquire joint attention responding. Further research should evaluate the collateral effects of training responding to joint attention bids. The study found that 2 of 3 participants we re making joint attention
57 initiations during probes; however, it is unclear why they o nly made eye gaze initiations during the probes. During sessions 2 of 3 participants were also making point initiations. F urther evaluation is also necessary to determi ne why 2 of 3 participants made pointing initiations during sessions but never during the initiation probes. Another collateral effect of training responding to joint attention bids that should be examined is an increase in verbalizations. The primary re searcher noted an increase in verbalizations for all three of the participants as they went through intervention. Dakota seemed to have the biggest increase; however, no data were collected on her verbalizations Her caregivers reported that she was sayi ng several novel words and she was starting to imitate words. Connor increased from saying 1 word to saying 2 3 words throughout his session s however, it is unclear whether or not t he increase in verbalizations is a result of the study or a result of the therapies he received during the study. Riley started imitating more sounds during the study and his caregivers reported that he was saying more sounds and had began pointing. Like Connor, it is unclear nting is a result of the study or a result of the therapies he received during the study. A final recommendation for future research is to develop a more effective parent training protocol and to examine potential factors that would inhibit parents from working with their child. The results of this study determi ne that parent training did not maintenance of responding to joint attention skills. Future research should examine the motivation of parents to work on a skill set with his or h er child It is unclear whether or not the parents worked on joint attention at all, as we rely on their verbal report and did not observe the behaviors taking place. Future research should also
58 examine in situ training with the parent while they are wor king with the child which would allow the researcher to provide immediate feedback to help shape the joint attention initiations Perhaps setting up a schedule or a contract with the parents for how often they will work with the child would evok e an increase in responding on the part of the parents. The findings of this study found that although the targets can be taught simultaneously, targets 1 4 were acquired before targets 5 and 6 (with the exception of Connor). All of the participants w ere able to acquire all six targets in between 1.5 months and 2 months. Previous research shows assessment data pre and post training but does not show the actual intervention data. The intervention data in this study allows us to see what acquisition l ooks like for all six targets across all three participants. Previous research has trained responding to joint attention and then has train ed initiations without looking to see if any initiations started to occur during the response training. This study the initiation probes. Also, 2 of the 3 participants made eye gaze and/or pointing initiations during sessions. The current study also provides evidence that more efforts need to be taken when providing parent training in order to help the children maintain the responding to joint attention skills.
59 References Autism Society of America (2008, January 21). About autism. Retrieved August 2, 2008, from http://www.autism society.org/site/PageServer?pagename=about_home Bakeman, R., & Adamson, L.B. (1984). Coordinating attention to people and objects in mother infant and peer infant interaction. Child Development, 55 1278 1289. Bono, M.A., Daley, T., & Sigman, M. (2004). Relations among joint attention, amount of intervention and language gain in autism. Journal of Autism and Developmental Disorders, 34 495 505. Bruinsma, Y., Koegel, R.L., & Koegel, L.K. (2004). Joint attention and children with autism: A review of the literature. Mental Retardation and Developmental Disabilities, 10 169 175. Carr, J.E., Nicolson, A.C., & Higbee, T.S. (2000). Evaluation of a brief multiple stimulus preferenc e assessment in a naturalistic context. Journal of Applied Behavior Analysis, 33, 353 357. Dawson, G., Munson, J., Estes, A., Osterling, J., McParland, J., Toth, et al. (2002). Neurocognitive function and joint attention ability in young children with a utism spectrum disorder versus developmental delay. Child Development, 73, 345 358. DeLeon, I.G. & Iwata, B.A. (1996). Evaluation of a multiple stimulus presentation format for assessing reinforcer preferences. Journal of Applied Behavior Analysis, 29, 51 9 533.
60 Dube, W. V., MacDonald, R.P.F., Mansfield, R.C., Holcomb, W.L., & Ahearn, W.H. (2004). Toward a behavioral analysis of joint attention. The Behavior Analyst, 27 197 207. Fenson, L., Dale, P.S., Reznick, J.S., Thal, D., Bates, E., Hartung, J. P., Pethick, S., & Reilly, J.S.. (1993). MacArthur communicative development inventories. San Diego, CA: Singular Publishing Group, Inc. Koegel, L.K., Koegel, R.L., Harrower, J.K., & Carter, C.M. (1999). Pivotal response intervention 1: Overview of appr oach. The Journal of the Association for Persons with Severe Handicaps, 24 174 185. Koegel, R.L., Koegel, L.K., & McNerney, E.K. (2001). Pivotal areas in intervention for autism. Journal of Clinical Child & Adolescent Psychology, 30 19 32. Loveland, K., & Landry, S. (1986). Joint attention and language in autism and developmental language delay. Journal of Autism and Developmental Disorders, 16, 335 349. Mundy, P., Sigman, M., & Kasari, C. (1990). A longitudinal study of joint attention and language development in autistic children. Journal of Autism and Developmental Disorders, 20, 115 128. Mundy, P., Sigman, M., & Kasari, C. (1994). Joint attention, developmental level, and symptom presentation in autism. Development and Psychopathology, 6 389 401. Mundy, P., & Thorp, D. (2007). Joint attention and autism: Theory, assessment and neurodevelopment. In Martos Perez, J., Gonzalez, P.M., Llorente Comi, M., & Nieto, C., (Eds.). New Developments in Autism: The Future is Today (pp. 104 138). Jess ica Kingsley Publishers : New York, NY.
61 Rocha, M.L., Schreibman, L., & Stahmer, A.C. (2007). Effectiveness of training parents to teach joint attention in children with autism. Journal of Early Intervention 29 154 171. Smith, T. (2001). Discrete trial training in the treatment of autism. Focus on Autism and other Developmental Disabilities, 16 86 92. Taylor, B.A. & Hoch,H. (2008). Teaching children with autism to respond to and initiate bids for joint attention. Journal of Applied Behavior Analysi s, 41 377 391. Tomasello, M. (1995). Joint attention as social cognition. In C. Moore & P.J. Dunham (Eds.). Joint attention: Its origins and role in development (pp. 103 130). Hillsdale, NJ: Erlbaum. Vismara, L.A. & Lyons, G.L. (2007). Using perseve rative interests to elicit joint attention behaviors in young children with autism: Theoretical and clinical implications for understanding motivation. Journal of Positive Behavior Support, 9 214 228. Whalen, C., & Schreibman, L. (2003). Joint attentio n training for children with autism using behavior modification procedures. Journal of Child Psychology and Psychiatry, 44 456 468. Whalen, C., Schreibman, L., & Ingersoll, B. (2006). The collateral effects of joint attention training on social initiati ons, positive affect, imitation, and spontaneous speech for young children with autism. Journal of Autism and Developmental Disorders, 36 655 664.
62 Whetherby, A.M., Woods, J., Allen, L., Cleary, J., Dickinson, H., & Lord, C. (2004). Early indicators of autism spectrum disorders in the second year of life. Journal of Autism and Developmental Disorders, 34 473 493.
64 Appendix A: Parent Scoring Survey Adapted from V. Fogel & R.G. Miltenberger, 2008 Pre or Post (Circle one) Date Admini s tered: On a scale of 0% to 100%, please rate the following statements: The average percentage of time your child makes eye c ontact with you while pl aying 0 10 20 30 40 50 60 70 80 90 100 The average percentage of time your child initiates social interaction using eye contact while playing 0 10 20 30 40 50 60 70 80 90 100 The average percentage of time your child responds to social initiations using eye contact while playing 0 10 20 30 40 50 60 70 80 90 100 The average percentage of time your child initiates social interactions using verbalizations while playing 0 10 20 30 40 50 60 70 80 90 100 The average percentage of time your child responds to social interactions using verbalizations while playing 0 10 20 30 40 50 60 70 80 90 100
65 Appendix B : Parent Questionnaire for Pre Training Name: Date: Child's Name: Does your child currently rec eive therapy? If so, what therapy and for how long? Do you feel that this training is important for your child? Y N Would you like to be involved in your chil d's training? Y N How important do you feel it is to be involved in your child's training? Not at all Somewhat Important Very important How many minutes per day is it possible for you to work with your child? >10 mins 30 mins <30 mins How many days per week is it possible for you to work with your child? >2 days 3 4 days <5 days Please list your child's most preferred toys/activities 1 6 2 7 3 8 4 9 5 10
66 Appendix C : Parent Questionniare for Post Training Name: Date: Child's Name: Do you feel that your child benefitted from this tr aining? Y N Don't Know Do you feel that you benefitted from receiving training? Y N Don't know Approximately how much time did you spend a day working on joint attention with your child? >10 mins 30 mins <3 0 mins Approximately how many days per week did you spend working on joint attention with your child? >2 days 3 4 days <5 days Would you want your child to participate in future studies, similar to this one? Y N Don't know How did you feel about having researchers in your home? I liked it It was difficult/inconvenient No opinion Additional Comments:
67 Appendix D: Data Sheet for Trainer Behavior
68 Appendix E : Data Sheet for Child Behavior