This item is only available as the following downloads:
1 THE RELATIONSHIP BETWEEN PRAGMATIC LANGUAGE SKILLS AND DEPRESSIVE SYMPTOMS IN CHILDREN AND ADOLESCENTS WITH AUTISM SPECTRUM DISORDER By EMILY A. WRAY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
2 2011 Emily A. Wray
3 To my family
4 ACKNOWLEDGMENTS I thank the members of my committee for their guidance in completing this project. This I also thank my husband, family, and friends for their support and encouragement. It has been wonderful to have you cheering me on!
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 ABSTRACT ................................ ................................ ................................ ................................ ..... 9 CHAPTER 1 LITERATURE REVIEW ................................ ................................ ................................ ....... 11 Introduction ................................ ................................ ................................ ............................. 11 Autism Spectrum Disorder (ASD) ................................ ................................ .......................... 13 Core Characteristics ................................ ................................ ................................ ......... 13 Communication ................................ ................................ ................................ ........ 15 Social interactions ................................ ................................ ................................ .... 17 Restricted and repetitive behaviors and interests ................................ ..................... 20 Life Outcomes ................................ ................................ ................................ ................. 20 Intellectual Functioning ................................ ................................ ................................ ... 22 Comorbid Psychiatric Conditions ................................ ................................ .................... 23 Pragmatic Language ................................ ................................ ................................ ............... 24 What is Pr agmatic Language? ................................ ................................ ......................... 24 Pragmatic Language Impairment ................................ ................................ .................... 25 Pragmatic Language Difficulties in ASD ................................ ................................ ........ 27 Pragmatic Language Assessment ................................ ................................ .................... 29 Depression ................................ ................................ ................................ .............................. 30 ASD and Comorbid Depression ................................ ................................ ............................. 31 Prevalence ................................ ................................ ................................ ........................ 31 Impact on Functioning and Life Outcomes ................................ ................................ ..... 33 Factors Contributing to Depression in ASD ................................ ................................ .... 34 Biological and environmental factors ................................ ................................ ....... 35 Cognitive ability ................................ ................................ ................................ ....... 35 General language ability ................................ ................................ ........................... 36 Pragmatic language ability ................................ ................................ ....................... 40 Li mitations of Previous Research ................................ ................................ .................... 47 Purpose of the Current Investigation ................................ ................................ ...................... 48 2 METHODS ................................ ................................ ................................ ............................. 50 Participants ................................ ................................ ................................ ............................. 50 Inclusion Criteria ................................ ................................ ................................ ............. 50 Age ................................ ................................ ................................ ........................... 50 ASD diagnosis ................................ ................................ ................................ .......... 51 Nonverbal cognitive ability ................................ ................................ ...................... 52 Exclusionary Criteria ................................ ................................ ................................ ....... 52
6 Measures ................................ ................................ ................................ ................................ 52 Outcome Variable ................................ ................................ ................................ ............ 52 Predictor Variables ................................ ................................ ................................ .......... 54 Biological and environmental risk factors ................................ ................................ 54 Nonverbal cognitive ability ................................ ................................ ...................... 55 General language ability ................................ ................................ ........................... 56 Pragmatic language ability ................................ ................................ ....................... 57 Procedure ................................ ................................ ................................ ................................ 59 Statistical Analyses ................................ ................................ ................................ ................. 60 Descript ive Statistics ................................ ................................ ................................ ....... 61 Correlational Analyses ................................ ................................ ................................ .... 61 Hierarchical Multiple Regression ................................ ................................ .................... 61 3 RESULTS ................................ ................................ ................................ ............................... 64 Descriptive Statis tics ................................ ................................ ................................ .............. 64 Demographic Variables ................................ ................................ ................................ ... 64 Outcome Variables ................................ ................................ ................................ .......... 65 Predictor Variables ................................ ................................ ................................ .......... 66 Correlational Analyses ................................ ................................ ................................ ............ 68 ................................ ................................ ......... 68 Partial Correlations ................................ ................................ ................................ .......... 70 Hierarchical Multiple Regression ................................ ................................ ........................... 70 Parent Report of Depression ................................ ................................ ............................ 71 Regression 1 ................................ ................................ ................................ ............. 71 Unique contribution of Pragmatics Profile: Regression 2 ................................ ........ 72 Self Report of Depression ................................ ................................ ............................... 72 Re gression 3 ................................ ................................ ................................ ............. 72 Unique contribution of Pragmatics Profile: Regression 4 ................................ ........ 73 4 DISCUSSION ................................ ................................ ................................ ......................... 83 Key Findings ................................ ................................ ................................ ........................... 83 Do Parent and S elf Reports Agree Regarding the Presence of Depressive Symptoms in Youth with ASD? ................................ ................................ .................. 84 Is Depression More Prevalent among Children with ASD than among Typically Developing Children? ................................ ................................ ................................ .. 85 Are the Three Measures Believed to Represent Pragmatic Language Ability Sign ificantly Associated? ................................ ................................ ............................. 86 What Predictor Cariables are Significantly Correlated with Depressive Symptomatology? ................................ ................................ ................................ ........ 87 Research Question 1: To What Degree does the Inclusion of Pragmatic Language Ability Increase the Predictive Power of the Previously Investigate d Factors? .......... 89 Research Question 2: For What Portion of the Variance in Depressive Symptomatology does Pragmatic Language Ability Uniquely Account? ................... 91 Implications ................................ ................................ ................................ ............................ 92 Limitations ................................ ................................ ................................ .............................. 94
7 LIST OF REFERENCES ................................ ................................ ................................ ............... 97 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 105
8 LIST OF TABLES Table page 2 1 Variables included in four hierarchical regression analyses ................................ .............. 63 3 1 Frequencies for demographic variables: Gender ................................ ............................... 74 3 2 Frequencies for demographic variables: Clinical diagnosis ................................ .............. 74 3 3 Frequencies for demographic variables: Autism Diagnostic Observation Schedule (ADOS) classification. ................................ ................................ ................................ ....... 74 3 4 Descriptive statistics for outcome variables ................................ ................................ ...... 74 3 5 Parent (CDI P) .... 74 3 6 Self (CDI S) ........ 75 3 7 Descriptive statistics for predictor variables ................................ ................................ ...... 75 3 8 Frequencies for outcome variable: Family history of mood disorders .............................. 75 3 9 Pearson product moment correlations between outcome and predictor variables ............. 76 3 10 Pearson product moment correlations between outcome and pragmatic language variables ................................ ................................ ................................ ............................. 77 3 11 Partial correlations between CDI P and variab les measuring pragmatic language .......... 77 3 12 Partial correlations between CDI S and variables measuring pragmatic languag e .......... 78 3 13 Regression 1: Hierarchical regression predicting CDI P total score ................................ 79 3 14 Regression 2: Hierarchical regression predicting CDI P to tal score ................................ 80 3 15 Regression 3: Hierarchical regression predicting CDI S total scor e ................................ 81 3 16 Regression 4: Hierarchical regression predicting CDI S total score ................................ 82
9 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE RELATIONSHIP BETWEEN PRAGMATIC LANGUAGE SKILLS AND DEPRESSIVE SYMPTOMS IN CHILDREN AND ADOLESCENTS WITH AUTISM SPECTRUM DISORDER By Emily A. Wray August 2011 Chair: John H. Kranzler Major: School Psychology Children and adolescents with Autism Spectrum Disorder (ASD) are at increased risk for developing symptoms of depression. To date, few research studies have examined factors likely to contribute to t he depressive symptomatology experienced by children with ASD. Previous investigations have suggested that language deficits, particularly deficits in pragmatic language, may be associated with socioemotional problems, such as depression, in typically deve loping children. However, this relationship has never been investigated among children with ASD. The current study utilized hierarchical multiple regression to investigate the relationship between pragmatic language ability and depressive symptoms among 50 children with ASD between the ages of 7 and 17. The researchers hypothesized that pragmatic language ability would account for a significant portion of the variance in depressive symptomatology exhibited by the sample of participants. The investigators pr edicted that children with better pragmatic language ability would exhibit fewer symptoms of depression. Results of the study provided partial support for this hypothesis. When predicting parent reported symptoms of depression, including a parent rating sc ale of pragmatic language ability significantly increased the predictive value of the model. This variable uniquely accounted for a significant portion of the variance in parent
10 reported depressive symptoms. In contrast, a performance based indicator of pr agmatic language skills was not a useful contributor to the model.
11 CHAPTER 1 LITERATURE REVIEW Introduction Autism Spectrum Disorder (ASD) is characterized by impairments in three broad areas of functioning: communication, social interactions, and restricted, repetitive behaviors and interests (Volkmar & Klin, 2005). These core domains interact. For example, social success is associated with com munication ability (Tager Flusberg, Paul, & Lord, 2005). Pervasive and complex impairments in these primary areas lead to disruptions in many facets of life, including establishing and maintaining social relationships, achieving educational and vocational success, and living independently (Eaves & Ho, 2008). Additionally, co occurring psychiatric conditions are common among individuals with ASD. Empirical investigations suggest that, among comorbid disorders, depression is the most common, affecting up to 3 0% of individuals with ASD (Matson & Nebel Schwalm, 2007; Vickerstaff, Heriot, Wong, Lopes, & Dossetor, 2007). Depression can lead to negative long term outcomes and may put a person at risk for withdrawal, oppositional behaviors, aggression, and suicide. Also, depression and the resulting behaviors can lead to more strain for affected individuals and their families, causing increased stress and conflict (Matson & Nebel Schwalm, 2007). Comorbid psychiatric conditions are often perceived by individuals with ASD and their caregivers as the most debilitating feature of ASD; however, these difficulties are rarely the primary focus of intervention efforts (Sterling, Dawson, Estes, & Greenson, 2008). Research in the field of ASD is limited by methodological issues such as the heterogeneity of most research samples, the difficulty of assessing internalizing behaviors of subjects with communication impairments, and the overlap between symptoms of psychiatric disorders and core features of ASD (Leyfer et al., 2006; M atson & Nebel Schwalm, 2007; Tager Flusberg, 2004). As a result, clinicians may not receive adequate information about proper
12 screening, diagnosis, and treatment of co occurring psychiatric disorders when working with individuals with ASD. Thus, more resea rch in the area of comorbid depression among children with ASD is critical to the development of effective interventions. Research indicates that pragmatic language ability, or social communication ability, is a primary area of deficit for children and ado lescents with ASD (Tager Flusberg, 2004; Tager Flusberg et al., 2005; Travis & Sigman, 1998). Many children with ASD, especially those with diagnoses of Asperger syndrome (AS) or high functioning autism, seek social interactions with peers and family membe rs; however, limitations in pragmatic language ability are likely to result in unsuccessful interactions. Repeated failures in social interactions can lead to social isolation, withdrawal, and depression (Fujiki, Brinton, Morgan, & Hart, 1999). Thus, commu nication impairments, especially in the area of pragmatics, may be an important contributor to co occurring social and emotional impairments. It is likely that acquiring a better understanding of the relationship between pragmatic language and socioemotion al outcomes, specifically depressive symptoms, will enable researchers and clinicians to implement more effective treatment programs. This review will begin with an introduction to ASD, including core characteristics, diagnostic criteria, and common comor bidities. Next, pragmatic language will be discussed, along with a description of numerous pragmatic language limitations with which individuals with ASD often contend. Finally, a discussion of the co occurrence of depression in the ASD population will be presented, including prevalence estimates, impact on life outcomes, and proposed risk factors. In conclusion, evidence for pragmatic language impairment as a
13 Autism Spectrum Disorder (ASD) Core Ch aracteristics Kanner (1943) described the syndrome of autism based on his observations of a group of children who demonstrated impaired or nonexistent social relationships from infancy and who had very deviant language. The primary distinguishing features noted by Kanner were impaired social and affective development. Typically developing infants are born exhibiting marked interest in social interactions; however, Kanner identified a group of children who lacked the typical motivation for social relationshi in reference to peculiar, egocentric thinking to indicate that these children seemed to live in a world of their own. Kanner also identified communication impairments in his clinical sample, includi ng mutism, echolalia, and the use of overly literal language (Kanner, 1943; Volkmar & Klin, 2005). Decades later, Rutter (1978) reviewed extensive research on autism and identified four core characteristics of the disorder: impairment in the development of so cial relationships, delayed or deviant language development or both resistance to change, and onset before the age of three years (Rutter, 1978). These criteria formed the basis for the diagnostic category of Autistic Disorder put forth by the American Psychological Association in 1980 (Tager Flusberg, Baron Cohen, & Cohen, 1993). These observations have proved to be robust and, although the definition has expanded in the years to follow, the basic premise has been maintained (Tager Flusberg et al., 199 3). According to the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (Text Revision; 2000) ( DSM IV TR ), a widely used classification system used to diagnose individuals with mental illnesses, individuals with Pervasive Developmental Disorders (also known as ASD) exhibit impairments in three broad categories: social interaction abilities,
14 communication, and restricted and repetitive behavior and interests (RRBI). Impairment in social interaction abilities is reflected by difficulties in nonverbal behaviors, failure to develop peer relationships, tendency to avoid sharing enjoyment with others, and lack of social reciprocity. Lack of social reciprocity may be manifested as an avoidance of social activities in favor of solitary tasks or a lack of awareness of others. Communication impairments can be manifested by a delay or lack of spoken language, the inability to initiate or sustain conversation, the presence of repetitive or idiosyncratic language, or the lack of age appropriate imita tive or imaginative play. The presence of RRBI is reflected in a preoccupation with a restricted interest abnormal in its intensity or focus, strict adherence to rituals or routines, repetitive motor mannerisms, or preoccupation with parts of objects. In o rder to be classified with ASD, a child must have exhibited impairments in these areas prior to the age of three years (American Psychiatric Association, 2000). Today, the World Health Organization estimates that six in every 1000 children are living with ASD (Vickerstaff et al., 2007). In 2006, the Centers for Disease Control and Prevention (CDC) concluded that, across study samples, an average of one child in every 110 was classified as having an ASD. One in every 70 males was affected, as compared to one in every 315 females (CDC, 2009). This trend indicates that the disorder is four to five times more common in males than females (CDC, 2009; Volkmar & Klin, 2005). Although the general characteristics of autism reported by Kanner (1943) and Rutter (1978) have proven to be relatively stable over time, the definition of autism has broadened since its inception in the early 1940s. Many researchers and clinicians in the field adhere to the concept of ASD, which proposes that autism and its related disabilities fall on a spectrum Disorder, Childhood Disintegrative Disorder, Asperger Syndrome (AS), and Pervasive
15 Developmental Disorder Not Otherwise Specified (PDD NOS) (V olkmar & Klin, 2005). The pattern of skill deficits exhibited by an individual determines where he or she is placed on the autism spectrum. Individuals with Autistic Disorder must demonstrate six total diagnostic indicators, including two social skill impa irments, one communication impairment, and one quality indicative of RRBI (APA, 2000; Volkmar & Klin, 2005). Individuals who meet diagnostic criteria for Autistic Disorder and have intellectual ability in the average range are considered to have high func tioning Autism. Individuals with AS must have average intellectual functioning, no delay in language acquisition, two social skill impairments, and one indicator of RRBI. To receive a diagnosis of PDD NOS, an individual must demonstrate two total diagnosti c indicators, including one social impairment and one impairment in either communication or RRBI (APA, 2000; Volkmar & Klin, 2005). The flexibility and change that has occurred in the diagnostic criteria for ASD have far reaching implications for the iden tification and treatment of individuals with autism and related disorders. Based on broad interpretation of diagnostic criteria, individuals may be classified as having ASD for whom some clinicians feel this label is inappropriate. Such issues have led to the development of a group of individuals, all identified and treated under the same categorical label, who are perhaps more different than they are similar. Individuals with ASD represent a broad range of abilities and behaviors (Sterling et al., 2008; Ta ger Flusberg et al., 1993). Communication Understanding language and communication is essential to the study of autism. A delay in language development or an unexplained loss of language skills is often the first sign that parents perceive indicating that language deficits and facilitating language based interventions for children with ASD is crucial, as better language skills are associated with better long term outcomes (Tager Flusber g et al.,
16 2005). Common language deficits reported in individuals with ASD include delayed language acquisition and deficient pragmatic language. Language abilities are highly variable among individuals with ASD. Most children with ASD acquire language lat er than their peers and at a much slower rate. Retrospective studies report that, in infancy, children with ASD are less responsive to the vocalizations of others, includin g their mothers (Tager Flusberg et al., 2005). A small portion of children with ASD, often those with AS, exhibit no marked language deficit or delay in young childhood. In contrast, Lord, Shulman, and DiLavore (2004) found that about 20% of the ASD population never acquires verbal language. However, this figure is expected to decrease wi th time, as early diagnoses and effective interventions become more likely (Lord, Shulman, & DiLavore, 2004). Approximately 25% of parents report their children with ASD acquiring some language by 12 to 18 months, and then experiencing a regression of lang uage skills (Tager Flusberg et al., 2005). Children with ASD, and especially those with AS, are likely to demonstrate impairments in more advanced communication skills, such as pragmatics, or social communication. Pragmatic language abilities have been emp hasized in autism research for decades because impairments in this area are found across nearly all individuals with ASD (Tager Flusberg et al., 2005). Numerous investigations have yielded consistent results: pragmatic language is deviant in children with ASD relative to their peers and relative to other domains of language (Tager Flusberg, 2004; Travis & Sigman, 1998). Pragmatic deficits observed in individuals with ASD lead them to abruptly shift topics, select inappropriate content, and fail to tailor i nteractions according to their speaking partner. They are reported to engage in persistent questioning, repetitive talk about a narrow range of topics, and failure to adapt speech or behavior to their social situation (Paul, Orlovski, Marcinko, & Volkmar, 2009; Reichow, Salamak, Paul, Volkmar,
17 & Klin, 2008). These conversational limitations decrease the likelihood of children with ASD creating meaningful social relationships. While many children with ASD experience the pragmatic skill deficits discussed pre viously, those with average or above average intellectual abilities may be particularly vulnerable to pragmatic language impairments. Although these children may attempt to communicate with others, they do so incorrectly, leading to social skill impairment s (Downs & Smith, 2004). Concurrently, language problems impede the ability of children with ASD to learn appropriate social skills through verbal communication, as do their typically developing peers. Thus, impairments in social and communication skills a re reciprocal (Travis & Sigman, 1998). Social interactions original description of the disorder (Carter, Davis, Klin, & Volkmar, 2005). Vickerstaff et al. (2007) studied 30 children with ASD between the ages of 7 and 13. According to parent and teacher reports of social competence, only one child was within one standard deviation of the mean (Vickerstaff et al., 2007). They found that 97% of the sample demonstrated signifi cant social skill impairments. Common social skill deficits in individuals with ASD included engaging in fewer social initiations, difficulty creating social relationships, and difficulty interpreting the emotions of others. Social characteristics of indi viduals with ASD are extremely divergent. Many young children with autism appear to have relatively little interest in social interactions, even with caregivers and family members. This quality differentiates infants with ASD from typically developing chil dren, who demonstrate an intense interest in their interactions and relationships with others (Carter et al., 2005). When children with ASD develop social relationships, they are often less intimate than those created by normally developing children. As com pared with
18 typically developing children, who usually initiate social contact and cooperative play with peers, children with autism typically engage in more solitary play activities (Carter et al., 2005). Many children with autism only initiate interaction s with others in order to meet a nonsocial goal, such as accessing a desired object or activity. However, higher functioning children and adolescents with ASD are often highly interested in social interactions, but their social impairments make it difficul t for them to interact with others appropriately. They may not Deficits in core social skills, such as empathy and assertiveness, are associated with an increa sed likelihood of social anxiety for children with ASD. Consequently, the high prevalence of social anxiety among individuals with autism may lead to withdrawal from social interactions (Bellini, 2006). Although many social skills emerge over time, problem s establishing and maintaining social relationships continue across the lifespan, even for individuals with AS and high functioning autism (Carter et al., 2005). High functioning children and teens with ASD report difficulty forming friendships and increas ed loneliness. A study of 235 adolescents and adults with autism indicated that, according to maternal ratings on the Autism Diagnostic Interview, only 8% of the sample had a reciprocal friendship that involved shared activities and 46.4% were reported to have no peer relationships (Orsmond, Krauss, & Seltzer, 2004). Only 33% of subjects reported participating in group recreational activities and, of those, very few engaged in informal socialization. Further analysis indicated that participants who engaged in more frequent social activities were more likely to have better daily living skills, fewer internalizing behaviors, and better overall social skills. In contrast participants with more significant language impairments demonstrated fewer social initiati ons (Orsmond, Krauss, & Seltzer, 2004). Researchers who have
19 interviewed high functioning teens and adults with ASD noted that respondents were likely to endorse feelings of inadequacy and isolation (Carter et al., 2005). Additionally, current research fin dings suggest that children and adolescents with ASD may have difficulty perceiving the emotions of others and that they are usually not responsive to responding to emotions. Individuals with ASD demonstrate abnormal expressions of emotions, which is likely to affect relationships with caregivers and peers (Travis & Sigman, 1998). Children with ASD are often characterized by a limited range of emotions, uncharacte ristic displays of emotion (quantity and quality), and failure to integrate multiple displays of emotion, such as eye gaze and smile. Children with ASD often struggle to correctly identify the emotional states exhibited by others (Carter et al., 2005). Th ese areas of social difficulty may result from an impairment in theory of mind. Theory of mind has been proposed to be a precursor to adequate social development that many children with ASD struggle to acquire. Theory of mind refers to the ability of child ren to recognize the mental states of themselves and others. Mental states include beliefs, desires, and intentions; the behavior (Tager Flusberg et al., 1993 ). Proponents of the theory of mind hypothesis believe that a failure to develop a theory of mind underlies the social and communicative impairments observed in children with autism. According to the theory of mind, children with ASD may understand themsel ves as distinct beings, but may not understand others and their interactions leading to lost learning opportunities and poor socialization (Travis & Sigman, 1998).
20 Restricted and repetitive behaviors and interests Restricted and repetitive behaviors and interests (RRBI) are one of the defining features of ASD. RRBI is generally defined as a behavior that is performed with no clear goal. The class of behaviors known a s RRBI vary among individuals and can include dyskinesias, stereotyped movements, tics, repetitive use of objects in non functional ways, self injurious behavior, object attachments, rigid reliance on particular routines, and circumscribed interests (Turne r, 1999). Circumscribed interests are particularly common among individuals with high functioning autism and AS. These can include marked preoccupations that are unusual in nature, such as a fascination with public restrooms, or interests in common activit ies th at are unusual in their degree (e.g., an intense interest in superheroes ) Szatmari et al. (1989) concluded that 86% of their sample of individuals with high functioning autism and 37% of individuals with AS demonstrated circumscribed interests, as c ompared with 9% of typically developing control participants (Szatmari, Bartolucci, & Bremner, 1989, as cited in Turner, 1999). Researchers suggest that the demonstration of RRBI is persistent and stable over time. Seltzer et al. (2004) investigated the de velopmental trajectory of ASD characteristics across the lifespan. The authors concluded that RRBI symptoms appeared to ameliorate only minimally over time, and did not improve as much as social and communication symptoms (Seltzer, Shattuck, Abbeduto, & Gr eenberg, 2004). Life Outcomes The majority of ASD research examines the manifestation of the disorder in childhood; thus, little is known about the transition from adolescence to adulthood or the long term life outcomes for individuals with ASD. Due to the paucity of empirically derived information, parents of individuals with ASD are unsure of what to expect as their children age. Many parents experience anxiety regarding how their children will fare in the future (Eaves & Ho, 2008).
21 In the 1960s and 70s, outcome investigations suggested that most individuals with ASD had poor or very poor adult adjustment. Few experienced normal social relationships and exhibited adequate educational and vocational functioning. Although some researchers claimed that indiv iduals with ASD who had IQ in the average range experienced better life outcomes, most high functioning individuals had no close friends, were dependent on their families, and experienced limited employment success (Eaves & Ho, 2008). As early as 1970, Rut ter noted that ASD was often accompanied by an increased risk for the development of comorbid psychiatric disorders, most commonly anxiety and depression (Rutter, 1970). More recently, Billstedt, Gillberg, and Gillberg (2005) conducted a follow up study of 120 individuals with ASD to assess long term life outcomes. Results indicated that outcomes were poorer than expected: 57% of participants had very poor outcome, 21% had poor, 13% had restricted, 8% had fair, and none had good outcome. Very few participan ts were living independently at the time of the follow up. Self injury and violence were common and were considered among the most difficult problems experienced by participants (Billstedt et al., 2005). In 2008, Eaves and Ho conducted phone interviews wit h 48 individuals with ASD between the ages of 19 and 31 years and their caregivers. During the interviews, the authors assessed term outcomes in the areas of current health, physical activity, education, public supports, work, friendship global outcome, derived from work, friendships, and level of independence, was fair. Of the total sample, 21% of participants reported good or very good outcome and 0% reported very poor outcome (Eaves & Ho, 2008). These findings indicate improvement in overall life outcomes for individuals with ASD, as compared with previous studies. However, many participants reported troubling health and mental health conditions. Sixty two percent of participa nts reported
22 experiencing emotional difficulty, 50% reported experiencing obsessive c ompulsive disorder, and 50% reported experiencing anxiety. In addition, participants reported elevated rates of obesity, vision problems, and epilepsy. Self injury and agg ression were also commonly experienced by individuals with ASD. Caregivers identified unmet needs in the area of social development (Eaves & Ho, 2008). In summary, although long term life outcomes may be improving for individuals with ASD, comorbid health and psychiatric conditions continue to negatively impact quality of life and impede treatment effectiveness. As noted previously, superior language skills are predictive of better life outcomes for individuals with ASD. Kanner (1973) followed up on his ori ginal work by investigating social outcomes of individuals with autism in adulthood. Of his original sample, those with better communication skills evidenced a greater degree of improvement over time (Kanner, 1973). More recently, Eaves and Ho (2008) ident ified verbal IQ as the most important predictor of overall life outcome rating among their sample of young adults with ASD. Furthermore, Seltzer et al. (2004) proposed that improved social language skills may be predictive of better life outcomes for indiv iduals with ASD. Intellectual Functioning Although individuals with ASD vary widely in terms of cognitive abilities, intellectual disability frequently co occurs with ASD and may exacerbate ASD symptoms. AS does not co occur with intellectual disability; h owever, many individuals with PDD NOS and autism also suffer from intellectual disability. According to a 2005 meta analysis, 70.4% of individuals with autism had mild to severe intellectual disability (Fombonne, 2005). Individuals with ASD and below avera ge intellectual functioning are more likely to exhibit profound language delays, more stereotypic behaviors, and higher levels of self injury (Matson & Nebel Schwalm, 2007).
23 Comorbid Psychiatric Conditions Rutter (1970) noted an increased likelihood of co morbid behavioral difficulties among adolescents with autism. Current estimates indicate that between 65 and 80% of individuals with ASD experience comorbid psychiatric disorders (Sterling et al., 2008). Common comorbid conditions include depression, hyper activity, inattention, aggression, Obsessive Compulsive Disorder (OCD), Tourette syndrome, phobias, and anxiety. These co occurring conditions increase the burden of illness experienced by individuals with ASD and add to the heterogeneity of the group, res ulting in treatment limitations (Volkmar & Klin, 2005). Numerous methodological limitations impede research studies investigating comorbid psychiatric conditions among individuals with ASD (Volkmar & Klin, 2005). In the past, information about comorbidit y was gathered from case studies, which do not indicate whether levels of comorbidity are greater than would be expected based on chance. Additionally, the nature of autism often impedes comorbidity research, as the symptoms of ASD change with age and deve lopmental level, leading to diagnoses that change during the lifespan based on communication and social abilities. Additionally, controversy has erupted regarding whether symptoms of other disorders should be considered to be core features of autism or ind icative of an additional condition (Volkmar & Klin, 2005). For example, unusual motor movements that suggest a comorbid diagnosis of tics or stereotyped movement disorder, though not necessary for a diagnosis of ASD, are linked to RRBI diagnostic criteria. Language impairments typically found in ASD individuals also complicate comorbid diagnoses by making it difficult for individuals to report cognitive and affective symptoms indicative of a co occurring condition (Volkmar & Klin, 2005). Numerous empirical investigations have indicated that depression is the most common comorbid condition experienced by individuals with ASD (Ghaziuddin, Ghaziuddin, & Greden,
24 2002; Matson & Nebel Schwalm, 2007; Sterling et al., 2008). A review investigating the developmental trajectory of ASD in adulthood concluded that depression affects up to 28% of adolescents and adults with ASD (Seltzer et al., 2004). In summary, ASD is a developmental disability characterized by deficits in social interaction skills, communication, and restricted and repetitive behaviors. Individuals with ASD may present with a spectrum of symptoms ranging from mild to quite severe. Although evidence indicates that long term life outcomes for those with ASD are improving, co occurring psychiatric condit ions, such as depression, present a significant challenge for affected individuals and their families. Pragmatic Language Most children and adolescents with ASD experience deficits in the pragmatic aspects of communication. The current investigation hypoth esizes that, for those with ASD, pragmatic following discussion will provide a definition of pragmatic language, discuss difficulties faced by individuals with pr agmatic language impairment, and identify the pragmatic skills that are most commonly deficient am ong individuals with ASD. Last issues related to the valid and reliable assessment of pragmatic language ability are discussed. What is Pragmatic Language? Among humans, language is a common code of self expression (Benner, Nelson, & psychological tool. Understanding and producing language facilitates thought and behavior. As such, it is widely considered an essential component to successful interactions with others and the achievement of positive life outcomes.
25 There are three general categories of language ability. Receptive language involves an eceive and interpret language; receptive language deficits impede listening and comprehension. Deficits in producing language, generally through speech, are referred to as expressive language problems. Pragmatic language skills are those that allow the use of language in social contexts to serve a social function, for example, successfully navigating a conversation (Donahue & Cole, 1994). Pragmatic language skills are based on the ability to integrate contextual information with language to produce communic ation (Chaban, 1996). More specifically, pragmatics includes the abilities to produce connected and organized texts or units of speech, such as conversations or narratives; to u se language for different purposes; to develop methods for participating in conversations; and to understand the needs of Adkinson, Self, Simmons, & Jantz, 2001). Unlike traditional language theories that focus on the structure and form of language, theories that Pragmatic Language Impairment Appropriate pragmatic skills are as important as language itself the production of correctly structured s entences that are pragmatically inappropriate can cause communication to fail (Johnston, Weinrich, & Glaser, 1991). Typically developing children acquire pragmatic language skills early in developmen t. For instance, unimpaired two year olds can consistentl y adapt their language use based on the background knowledge possessed by the listener (Baron Cohen, 1988). In contrast, the majority of children and adolescents with ASD are at risk for pragmatic language impairment (Tager Flusberg et al., 2005). Tannock
26 There are three primary types of pragmatic language abilities: emotional processi ng skills, discourse strategies, and goal directed language. Individual s with pragmatic language impairment demonstrate problems in one or more of these domains (Tannock & Schachar, 1996). Emotional processing refers to the ability to understand and commun icate emotions. These skills are essential to the development of positive interpersonal relationships. According to functionalist theories of emotion, emotions have both interpersonal and intrapersonal capacities. The interpersonal qualities of emotions in of emotions include the abilities to assess the meaning of events and use that knowledge to guide sub sequent behavior (Bretherton, Fritz, Zahn Waxler, & Ridgeway, 1986). Discourse strategies result in the production of coherent conversation and cohesive narration. These skills include adequate vocabulary, creating grammatically sound sentences, linking s maintenance, understanding appropriate body space and movement in conversation, and recognizing and repairing miscommunications (Johnston et al., 1991). Goal directed language is the ability to make language serve a desired purpose (Johnston et al., 1991; Larson & McKinley, 1987). Goal directed language consists of two distinct areas: using language appropriately and making language perform various functio ns. In order to demonstrate appropriate language use, a child must: provide enough information, but not too much; provide information that is appropriate to the person(s), place, and situation; include enough truth value to be acceptable to other participa nts; and be acceptably polite. Another prerequisite to effective language use is the ability to make language perform various functions. Language is a tool and, as such, it can be used to elicit desired outcomes. Young children use
27 language to accomplish b asic tasks, such as labeling, requesting, and protesting. As children mature, they learn to perform a variety of complex functions with language, including imagining, informing, and negotiating. Adults utilize language to accomplish even more sophisticated tasks, such as projecting, defending, and coercing (Johnston et al., 1991; Larson & McKinley, 1987). Impairment in any of these areas may lead to pragmatic language disorder (Tannock & Schachar, 1996). Furthermore, the failure to develop pragmatic skill s can lead to academic failure and social and cultural exclusion (Chaban, 1996). Pragmatic Language Difficulties in ASD Pragmatic language difficulties have been widely reported in individuals with ASD. Individuals with ASD typically have pragmatic langua ge abilities that are impaired relative to their peers and relative to other language domains (Paul et al., 2009; Reichow et al., 2008; Tager Flusberg, 1989; Tager Flusberg et al., 2005). A synthesis of the research on the linguistic profiles of children w ith autism suggests that, although children with autism are likely to have deficits in all areas of language, pragmatic language is the only area in which their deficits are unique to the autism population. Although many children with ASD will acquire lin guistic structure, such as phonology, semantics, and grammar, in a fashion that is similar to other children, their social use of language is unique. Typically, structural components of language are more advanced than pragmatics in individuals with ASD (Ta ger Flusberg, 1989; Baron Cohen, 1988). In fact, children with high functioning autism, AS, and PDD NOS often demonstrate well developed structural language, which leads to under identification of pragmatic deficits by parents and teachers (Tager Flusberg, 1989). Communicative functions can be divided into two types: nonsocial functions, which are used to achieve environmental outcomes, such as access to a desired object or activity, and social functions, which are used to attract attention to oneself or t o call attention to an object or activity.
28 Children with ASD are much more proficient with nonsocial communication (Tager Flusberg, 1989). Children with ASD often use language for instrumental purposes, such as requesting food or toys, rather than for pure ly social reasons, such as obtaining attention, commenting, acknowledging their speaking partner, initiating social interactions, or requesting information. Although children with autism can develop social communication, they do so at a later developmental period and they use it less frequently and less appropriately (Baron Cohen, 1988 ; Tager Flusberg, 1989 ). Pragmatic language difficulties are associated with a wide range of communication limitations in individuals with ASD. Tager Flusberg et al. (2005) reported that, among children and adolescents with ASD, rates of spontaneous initiation of communication are low. Additionally, evidence indicates that individuals with ASD direct most of their communication attempts to adults, rather than s ame age peers (Tager Flusberg et al., 2005). Additional pragmatic difficulties noted in individuals with ASD include talking to oneself, problems listening, difficulty following norms for politeness, making irrelevant statements, providing too much or too little information to listeners, and terminating topics inappropriately. Individuals with ASD may also demonstrate inappropriate eye gaze and vocal intonation (Tager Flusberg et al., 2005). According to Baron Cohen (1988), the social impairments observed i n individuals with ASD may instead be a reflection of underlying pragmatic language deficit. Individuals who are described as aloof or socially avoidant may in fact desire interactions with others; however, they display inappropriate social communicative a pproach. Paul et al. (2009) used a parent rating scale to evaluate the pragmatic language skills of adolescents of normal intelligence diagnosed with ASD. Their sample consisted of 29 individuals between the ages of 12 and 18 with clinical diagnoses of hig h functioning autism,
29 AS, and PDD NOS and 26 typically developing peers. Results indicated that the ASD group did not perform more poorly than the TD group on all conversational behaviors; however, they did perform significantly worse on the following pra gmatic domains: providing the appropriate amount of information to meet the needs of the listener, managing topics adequately, responding appropriate conversatio n (Paul et al., 2009). Pragmatic Language Assessment socioemotional development has been well established, many traditional language assessment instruments do not include a measure of pragmatic ability (Russell & Grizzle, 2008). Traditional assessment of pragmatic language ability has relied on subjective parent reports and observations, neither of which demonstrate the psychometric rigor of standardized, norm referenced tests. Furt her assessments that rely only on performance based pragmatic measures may be limited in reliability, as they are based on a single sample of behavior that is occurring in a standardized, laboratory setting. This method is incongruent with the spontaneous social nature of pragmatic language skills (Volden, Coolican, Garon, White, & Bryson, 2009). These methodological issues often lead to pragmatic language evaluations that are inconsistent and lack validity. The ability to conduct appropriate communicat ive exchanges in various social contexts is crucial to the ease with which children navigate the social world. Pragmatic language deficits are particularly evident in children and youth with ASD, regardless of their level of functioning in other areas, inc luding cognitive ability and receptive and expressive language ability. Valid and reliable assessment of pragmatic ability is a difficult but necessary task for clinicians working with individuals with ASD.
30 Depression Depression in childhood and adolescen ce is an urgent issue for children, caregivers, clinicians, and educators. Depression is characterized by depressed or irritable mood, loss of interest in activities, weight gain or weight loss, insomnia or hypersomnia, feelings of worthlessness, difficult y concentrating, and suicidal ideation (APA, 2000). In contrast with the pervasive sadness reported by adults with depression, depression in childhood and adolescence is often manifested as irritability, boredom, and anhedonia (McCarthy, Downes, & Sherman, 2008). According to the National Comorbidity Study, the only nationally representative study that includes adolescents, the lifetime prevalence rate of Major Depression in teenagers is 14%, with an additional 11% reporting minor depression (Hammen & Rudo lph, 2003). Subsyndromal depressive symptoms, those symptoms of depression which are not sufficiently severe to meet diagnostic criteria, still portend a high level of distress. According to Cooper and Goodyer ( 1993 ), 20.7% of females between the ages of 1 1 and 16 years demonstrate subsyndromal depressive symptoms. Depressive symptoms increase steadily between the ages of 13 and 15 years, after which they become relatively stable (Montague et al., 2008). Williamson et al. (2008) speculated that depression b ecomes more common in adolescence because social Slinger, 2008). Females are at a greater risk for depression than males (Montague et al., 2008). Without proper treatmen t, a major depressive episode typically lasts about eight months (Hammen & Rudolph, 2003). Forty percent of individuals who have experienced a major depressive episode will experience a recurrence within two years and 72% will experience a recurrence withi n five years (McCarthy et al., 2008; Montague et al., 2008).
31 School based mental health professionals frequently rate depression as one of the most pressing challenges they face. In fact, they report depression to be a more urgent issue than teen violence Unfortunately, many schools do not have clear procedures established for the identification and treatment of students with these mental health issues (McCarthy et al., 2008). ASD and Comorbid Depression As previously discussed, depression is among the mo st commonly reported comorbid disorders experienced by individuals with ASD. Depressive symptoms negatively impact quality of life for those with ASD. Consequently, the treatment and prevention of co occurring depressive symptoms in the ASD population is o f critical importance. The following discussion will address the prevalence of depressive symptoms among children and youth with ASD, the impact of depression on life outcomes, and suspected risk factors for depression in the ASD population. Prevalence Num erous experts in the ASD research community contend that depression among children and teenagers with ASD is a critical issue requiring further investigation (Ghazuiddin, Ghazuiddin, & Greden, 2002; Sterling et al., 2008; Vickerstaff et al., 2007). Rates o f depression among children with ASD are markedly higher than those reported in the general population. Vickerstaff et al. (2007) measured the depressive symptomatology of 30 children with ASD between the ages of 7 and 13 and found that 24% were classified of experiencing depressive symptoms for youth with ASD, as compared to the general population. Leyfer et al. (2006) reported that 10% of their sample of children with ASD had at least one Major Depressive Episode meeting DSM IV criteria. An additional 24% had
32 subsyndromal symptoms. These results are particularly powerful when considering that the mean age of their participants was n ine years (Leyfer et al., 2006). Rates of depression among unimpaired individuals and those with ASD increase with age; therefore, the prevalence of depression is higher among adolescents than children (Ghazuiddin et al., 2002; Sterling et al., 2008; Vicke rstaff et al., 2007). The likelihood of experiencing depression for those with ASD may increase with age because of the perceived pressure to meet societal expectations, such as marriage, job success, educational attainment, and independence. Additionally, with advancing age, the developmental gap between individuals with ASD and their typically developing peers becomes more apparent (Sterling et al., 2008). The increasing rates of depressive symptoms and other co occurring psychiatric conditions experience d by teenagers with ASD might explain the regression in behavioral development and skills that is often reported by parents of adolescents with ASD (Sterling et al., 2008). Empirical evidence suggests that higher functioning individuals, especially adolesc ents, are at the highest risk for developing depressive symptoms (Ghazuiddin et al., 2002; Vickerstaff et al., 2007). Samples of individuals with high functioning autism and AS have consistently yielded higher estimates of depression than are found in indi viduals with ASD of lower cognitive ability. In particular, adolescents with AS may be particularly vulnerable to depression (Klin, Volkmar, & Sparrow, 2000; Volkmar & Klin, 2005). For example, Wing (1981) found that approximately one third of her sample o f individuals with AS were experiencing depression (Wing, 1981, as cited in Ghazuiddin et al., 2002). Further investigation in the area of comorbid depression in ASD is warranted, as prevalence estimates indicate that a large portion of individuals with AS D experience depressive symptoms. Furthermore, emotional and behavioral problems are stable over time; children and
33 adolescents who exhibit depressive symptomatology are likely to also experience depression later in life (Montague et al., 2008). In particu lar, researchers and clinicians will benefit from depressive symptoms (Sterling et al., 2008). Impact on Functioning and Life Outcomes Experiencing depression d uring childhood and adolescence is associated with a host of negative long term outcomes. Research based on depression in the general population indicates being, relationships with peers and family members, and academic performance. Furthermore, child and adolescent depression is associated with suicide, the third most frequent cause of death for young adults (McCarthy et al., 2008). Unfortunately, many sufferers of childhood depression a re never diagnosed or remediated. In fact, according to Evans, van Velsor, and Schumacher (2002), (p. Based on the impact of depres sion on typically developing children and teens, it is not surprising that depression may lead to poorer outcomes in those with ASD. Outcome studies of adults and children with ASD have revealed that few develop social relationships and loneliness is a com mon experience (Vickerstaff et al., 2007). Depression can lead to aggression and anxiety and can exacerbate the social difficulties with which children and adolescents with ASD already contend. The effects of depression on children with ASD may include red uced attempts at communication, increased social withdrawal, oppositional and aggressive behavior, and suicidal behavior. Among more severely impaired individuals with ASD, depression can lead to increased motor agitation, self injurious behavior, stereoty pic behaviors, and obsessions (Sterling et al., 2008). Comorbid depression is also likely to generate a negative impact on family and
34 social relationships (Ghazuiddin et al., 2002). Kim and colleagues (2000) concluded that children with high functioning AS D who exhibited symptoms of anxiety and mood disorder were more aggressive and had poorer relationships with peers, parents, and teachers. Additionally, their identifying and treating comorbid mood problems among children with ASD. Factors Contributing to Depression in ASD Although elevated rates of depression among children and adolescents with ASD have been consistently documented, very few empirical investi gations have provided insight regarding risk factors for the development of depressive symptoms in this population (Sterling et al., 2008). This gap in the literature has hindered the development of effective prevention and intervention plans for this grou p of individuals (Sterling et al., 2008). In the following discussion, three risk factors for depression among children with ASD will be postulated based on previous research findings: biological and environmental risk, cognitive ability, and general langu age ability. Furthermore, evidence in support of each relationship will be presented. These three factors have been proposed by previous investigators as primary predictors of depressive symptomatology among individuals with ASD (Benner et al., 2002; Galla gher, 1996; Geurts & Embrechts, 2008). Finally, the possible contribution of pragmatic language deficits to depressive symptoms will be discussed. Emerging evidence implicates the role of pragmatics, or social communication skills, in the development of in ternalizing and externalizing behavior disorders. However, past investigations have frequently ignored the likely effects of pragmatic ability on socioemotional outcomes. Instead, they have emphasized the role of general language ability in the development of emotional and behavioral problems. The role of pragmatic language has not been examined as it pertains to depressive symptomatology in youth with ASD.
35 Biological and environmental f actors Family and environmental factors affect the likelihood of a child or adolescent with ASD experiencing depression. Individuals with ASD who experience depression are more likely to have family members with a history of mood disorder ( Ghazuiddin et al., 2002; Vickerstaff et al., 2007 ). In addition, social and environ mental stressors may make one more susceptible to the development of depressive symptomatology (Vickerstaff et al., 2007). Ghazuiddin and colleagues concluded that children with ASD who develop clinical depression are more likely to experience negative lif e events (Leyfer et al., 2006). Additionally, experiencing co occurring medical conditions in addition to ASD, such as seizure disorders or intellectual disability, may make one more likely to experience depression (Ghazuiddin et al., 2002). Cognitive a bi lity One avenue of research suggests that awareness of their social limitations may make individuals with ASD susceptible to experiencing depression (Vickerstaff et al., 2007). Social skill deficits are a core characteristic of ASD and are of particular c oncern among children and adolescents with high functioning autism and AS. In these cases, individuals may have strong language and cognitive skills, but exhibit impairments in the ability to navigate social interactions successfully. In fact, researchers have noted that children with higher intellectual functioning may be at greater risk for depression than their lower functioning peers because they are aware of their limited social skills (Vickerstaff et al., 2007; Volkmar & Klin, 2005). Vickerstaff et a l. (2007) investigated this claim by administering an assessment battery to 30 children and adolescents with high functioning autism. Outcome measures included ratings scales designed to evaluate social skills, self perceived social competence, and depress ive symptomatology. Results indicated that older children and children with higher IQ perceived themselves as having lower social competence. Additionally, those with lower self perceived
36 social competence demonstrated a higher level of depressive symptoms The authors suggested that this trend may be explained by the fact that these individuals are more aware of their limitations in comparison with typically developing peers, whereas individuals with developmental disabilities and low IQ may not notice tha t they are different (Vickerstaff et al., 2007). Sterling et al. (2008) evaluated 46 adults with ASD between the ages of 18 and 44 years. Among their sample, higher IQ was correlated with more depressive symptoms. Furthermore, participants with better soc ial and communication skills, as measured by the Autism Diagnostic Observation Schedule (ADOS), experienced more symptoms of depression, suggesting that higher functioning individuals may be more vulnerable to depression (Sterling et al., 2008). This findi ng is of great importance to teachers, parents, and clinicians working with children with ASD, because it suggests that children and teenagers with average intelligence should be assessed for low self esteem. Additionally, interventions that focus on impro ving social skills may assist these children in maintaining high levels of self confidence. General language a bility The relationship between language and behavioral outcomes has been well documented among children without ASD. Numerous empirical investig ations have found evidence to support a link between impaired language functioning and emotional/behavioral difficulties (Benner et al., 2002; Gallagher, 1999; Prizant et al., 1990; Redmond & Rice, 1998). Upon evaluating numerous studies reporting overlaps between children with emotional/behavioral disorders (EBD) and language impairments, Gallagher (1999) found moderate to severe language deficits in 62 to 95% of children with EBD. Camarata, Hughes, and Ruhl (1998) administered a test of language developme nt to 38 students with mild or moderate behavior disorders in a school
37 setting. Of these, 71% performed two standard deviations below the mean and 97% performed one standard deviation below the mean. Baker and Cantwell (1987) and Baltaxe and Simmons (1988 ) were among the first groups of researchers to investigate the occurrence of comorbid psychiatric disorders among children with language impairments. Baker and Cantwell (1987) administered psychiatric assessments to 600 children between the ages of 2 and 16 years who were receiving treatment at a community speech/language clinic. The authors reported that 50% of the sample had a diagnosable psychiatric illness. They concluded that speech and language dysfunction appeared to have a significant impact on psy chiatric wellness. Further scrutiny revealed that, although cognitive ability and psychosocial stressors were also associated with psychiatric illness, the biggest differentiating factor between the psychiatrically ill and well groups was speech and langua ge status (Baker & Cantwell, 1987). Similarly, after evaluating 125 preschool children with language impairment, Baltaxe and Simmons (1988) concluded that 10% also met diagnostic criteria for attention deficit/hyperactivity disorder, 5% for conduct disorde r, 5% for oppositional defiant disorder, and 6% for emotional disorders, such as anxiety and depression. Further research on the overlap between language ability and internalizing behaviors, such as depression, is limited; however, based on these findings, it is likely that depression frequently co occurs with language impairment. This proposition is further supported by Fujiki et al. (1999), who examined teacher behavior ratings of language impaired students. They concluded that language impaired students were more likely than unimpaired children to display withdrawn behaviors. In particular, these children demonstrated reticent withdrawal, a type of withdrawal characterized by a desire to interact with others impeded by social anxiety (Fujiki et al., 1999)
38 Although a causal role has yet to be confirmed, research findings implicate language deficit as an etiological factor for emotional/behavioral disorders. In an investigation of the socioemotional behavior and verbal abilities of a sample of language impa ired and age matched children, Redmond and Rice (1998) found support for the Social Adaptation Model According to this model, emotional and behavioral problems experienced by children with communication deficits are a direct consequence of their language impairment. This model proposes that increased social and behavioral problems are a product of the relationship between the underlying language difficulty, the social context, and the biased perceptions of others regarding individuals with language impairm ents. This model proposes that problematic socioemotional example, children with language problems typically initiate and respond to peer interactions less freq uently than do typically developing peers. B ased on the Social Adaptation Model, children with language impairments were rated as having more severe behavioral problems by teachers than by parents. Given these results, the authors concluded that the childr difficulties, which caused increased stress in the school environment as compared to the home, resulted in problem behaviors (Redmond & Rice, 1998). Effective language skills foster successful social interactions and positive academic outcom es (Donahue & Cole, 1994; Nelson, Benner, & Cheney, 2005). Prizant and colleagues (1990) proposed that expressive language disorders may make it difficult for children and adolescents to express their ideas, feelings, fears, and needs. Thus, they may appea r impulsive, restless, immature, or even aggressive. Receptive language difficulties may cause individuals to misinterpret the verbalizations of others; furthermore, expressive deficits may impair their ability to request clarification, resulting in frustr ation. This frustration could lead to externalizing
39 behaviors, internalizing behaviors or both (Prizant et al., 1990). Poor communication skills may lead to rejection and isolation from peers. In addition, the academic problems faced by youth with language impairment may lead to rejection and low self esteem (Howlin, Mawhood, & Rutter, 2000). Lastly, Gallagher (1993) reported that children with language difficulties are more likely than peers to withdraw or use aggression to solve conflicts. These factors c ontribute to the development of depression. Results of the Ottawa Longitudinal Study (OLS) support the theory that speech and language problems cause emotional and behavioral difficulties. The OLS w as an investigation of 142 five year old children with la nguage impairment and 142 unimpaired peers. Participants were assessed at 5 years of age and re evaluated at 12.5 years of age (Beitchman, Brownlie, & Wilson, 1996). The investigators found that language impairment was significantly associated with emotion al and behavioral problems, even after controlling for socioeconomic status. In fact, speech and language function at age 5 proved to be the most important predictor of psychiatric outcome at age 12.5. Children with language impairment at 5 years of age ha d higher ratings of psychopathology and lower ratings of global functioning at age 12 (Beitchman, Brownlie, & Wilson, 1996). These findings support the theory that language deficit is associated with and precedes the development of emotional/behavioral pro blems. Children with language deficits are more likely to develop both internalizing and externalizing behavior disorders than are their typically developing peers. Externalizing behaviors, the demonstration of overt behaviors, are more common in younger a ge groups (Gallagher, 1999; Prizant et al., 1990; Stevenson, 1996). Commonly comorbid externalizing problems include aggression, immaturity, Attention Deficit/Hyperactivity Disorder (ADHD), Oppositional Defiant Disorder, and Conduct Disorder (Gallagher, 19 99, Prizant et al., 1990).
40 Internalizing problems, such as low self esteem, social withdrawal, depression, phobias, and anxiety, are more likely to occur among adolescents (Gallagher, 1999; Stevenson, 1996). Initially, children with co occurring language i mpairments demonstrate frustration and aggression. These behaviors tend to decline rapidly, while internalizing symptoms, such as depression and low self esteem, remain (Stevenson, 1996). As children age, their comorbid behavior problems are rated as incre asingly severe, according to caregiver report. Additionally, children become more adept at masking their communication impairments at older ages (Cohen, 1996). These trends emphasize the need for continuous monitoring and intervention throughout adolescenc e. Many investigators in the fields of special education and speech/language pathology have provided consistent evidence that general language ability is related to emotional and behavioral outcomes (Beitchman et al., 1996; Howlin et al., 2000; Prizant et al., 1990). Few studies, however, have systematically investigated the contribution of general language ability in predicting socioemotional outcomes in ASD. Howlin, Mawhood, and Rutter (2000) investigated factors influencing the social outcomes of adults with ASD. The authors found that expressive language, as measured by a direct assessment of expressive vocabulary was the most significant predictor of social success. As compared with nonverbal cognitive ability which accounted for 6%, language ability accounted for 32% of the variance in social outcome. Pragmatic language a bility Correlations between pragmatic language ability and socioemotional functioning have been documented among children with and without developmental disabilities. In fact, past r esearch findings have suggested that pragmatic language is more closely related to socioemotional functioning than is general language ability. Benner et al. (2002) conducted a meta analysis in which they evaluated 26 studies of 2,358 children with emotion al/behavioral
41 disorder and 438 without. Results indicated that, of those with EBD, 71% exhibited pragmatic language impairments, 64% exhibited expressive language impairments, and 56% exhibited receptive language impairments (Benner et al., 2002). McDonoug h (1989) evaluated the pragmatic language use of children with and without behavior disorders. The majority of subjects with EBD showed deficits in conversational skills, including the tendency to use shorter utterances, lack of ability to maintain convers ational topic, and failure to appropriately respond to questions. Children with EBD did not display problems with general language, such as syntax or semantics (McDonough, 1989). These trends implicate pragmatic language impairment as a contributing factor to emotional and behavioral maladjustment. Few empirical studies are available which investigate the role of pragmatic language in the development of specific psychiatric disorders, such as depression. This is likely due to the paucity of commercially a vailable tools for adequately assessing pragmatic language skills. Furthermore, an emphasis on assessment and intervention of pragmatic language skills is a relatively recent development in the field of speech/language pathology (Kaiser, 1993). Prior to th e 1980s, the majority of communication research focused on structural aspects of language, such as phonology, syntax, and semantics (Kaiser, 1993). The current review identified only one empirical investigation examining the relationship between pragmatic language ability and symptoms of depressio n. Mupawose, Katijah, and Naran (2007) evaluated the association between pragmatic language and depression in a sample of adults with traumatic brain injury. The authors concluded that participants with depression demonstrated greater pragmatic language deficit (Mupawose, Katijah, & Naran, 2007). In support of this conclusion, Gravell and France (1991) reported that depressed patients often demonstrate social communication difficulties, particularly reduced nonverb al communication
42 and reduced initiations. Additionally, Baltaxe and Simmons (1988) conducted case studies on six children with EBD and found that all had pragmatic language deficits, including three subjects whose primary psychiatric condition was depressi on. The authors concluded that, for those with affective disorders such as depression, pragmatics is often the only impaired language domain. likelihood of experiencing symptoms of depression. Despite the paucity of research studies investigating pragmatic ability and its impact on depression, current research has suggested that pragmatic language impairment is related to other psychiatric problems, such as inattention and hyperactivity. Geurts and Embrechts (2008) found that children with ADHD had pragmatic problems similar to those of children with ASD; however, they had no significant structural language deficits. This implies that their psychiatric condition was rela ted to their pragmatic impairment, rather than their overall language abilities. Tannock and Schachar (1996) also reported pragmatic difficulties among children with ADHD, including problems with conversation skills, difficulty clearly expressing informati on, and an inability to adjust language output to meet the needs of the listener. Mack and Warr Leeper (1992) investigated the language abilities of 20 boys with chronic behavior disorders and average cognitive functioning. They found that 16 of the 20 par ticipants had significant impairments on at least four language measures, a prevalence of language impairment ten times greater than that found in the general population. The language skills found to be most commonly deficient were higher order pragmatic a bilities, including understanding ambiguity, making inferences, and extracting meaning from context cues (Mack & Warr Leeper, 1992). The current study hypothesizes that a similar relationship exists between pragmatic language deficit and internalizing beha viors, such as depressive symptoms.
43 through its impact on peer relationships. Deficits in pragmatic language often lead to rejection by peers. Among children, conversation al competence is associated with social status. Children who demonstrate communication deficits are less sought after as conversational partners (Rice, Sell, & Hadley, 1991; Rice, 1993). In turn, impaired children are less likely to initiate conversations with peers, likely in an attempt to avoid contexts with high language demands. According to Rice (1993), these trends contribute to a negative social spiral in which already impaired children fail to further develop their communication abilities due to lim ited opportunities for practice. According to Gallagher (1996), as early as preschool, children prefer playmates who talk more, communicate more successfully, express coherent messages, produce responses relevant to the topic, and engage in positive social interactions using language. The conversational skills known to best facilitate peer acceptance and the development of appropriate social skills are precisely those found to be deficient in pragmatic language impaired children, such as those with ASD. Th e concept of a negative social spiral is supported by behavior patterns documented by Rice et al. (1991). The authors during unstructured playtime. During the observation sessions, children were able to freely choose activities and interaction partners. Results indicated that typically developing children were the preferred playmates for all interactions, suggesting that the children were aware of the relative communication skills of their p eers and adjusted their choices accordingly. Children with language deficits were more likely to be ignored by their peers. In fact, as compared to 12% of with language impairments were less responsive to the initiations of others. Children with impaired language did not respond to 30% of initiations directed to them; typically developing
44 children did not respond to 7%. Lastly, children with communication defici ts were less likely to initiate interactions with typically developing peers (Rice, Sell, & Hadley, 1991). These findings suggest that pragmatic language impairment can lead to social isolation. Additionally, the tendency of language impaired children to a void communication interactions is likely to result in further pragmatic language difficulties, as these children will have fewer opportunities to learn and practice new social communication skills. The relationship between pragmatic language skill and soc ial development has not been investigated among individuals with ASD. The current study proposes that a similar mechanism accounts for the impact of pragmatic difficulties on the negative social interactions and peer rejection of children with ASD. Impaire d social relationships lead to the development of socioemotional difficulties, such as depression (Fujiki et al., 1999). Elder et al. (2006) stated that a lack of close friends may contribute to depression. Children who experience social isolation and peer rejection are more likely to experience emotional and behavioral problems than are their well accepted peers (Sandstrom, Cillessen, & Eisenhower, 2003). Investigations have found that children who are rejected by their peers are at increased risk for inte rnalizing and externalizing problems, according to parent and teacher ratings. Peer rejection is associated with greater risk for aggression, depression, learned helplessness in social situations, and anxiety (Sandstrom et al., 2003). More specifically, nu merous researchers have concluded that peer rejection is predictive of later depressive symptoms (Nolan, Flynn, & Garber, 2003 ; Reijntjes, Stegge, & Terwogt, 2006 ). Based on these findings, the current investigation proposes that the underlying pragmatic l anguage difficulties experienced by children with ASD lead to an increased likelihood of social problems and peer rejection, which result in a greater risk for depressive symptomatology.
45 On the contrary, Sterling et al. (2008) concluded that better social communication skills, as measured by the ADOS social communication index, may predict more depressive symptoms in adults with ASD. However, their group selection was very broad and may have included some pressed. Also, their findings may have been exaggerated due to methodological limitations in their statistical analyses. The authors conducted three t tests, but did not divide the alpha level by three to account for multiple comparisons. Had this been don e, their findings would not have been significant. Alternatively, the findings presented by Sterling and colleagues may reflect the likelihood that participants with better social communication skills were also more aware of their differences and difficult ies, leading to higher rates of depression. Finally, it is possible that language factors do not play as important a role in predicting depression for adults due to the salience of other factors, such as life stressors (Sterling et al., 2008). Readers may speculate that pragmatic language and general language skills are too closely depression. In fact, language research indicates that a positive relationship exists between overall language ability and pragmatics. Although the development of different aspects of language may be divergent, general language ability is associated with the development of pragmatic skills (Geurts & Embrechts, 2008). On the contrary, severa l investigations have documented that pragmatic language ability contributes uniquely to socioemotional outcomes, after accounting for the contribution made by general language skills. Volden et al. (2009) administered a test of general language skills and a test of pragmatic language to 37 participants with ASD. They found that pragmatic language disability was the most significant predictor of social impairment for individuals with ASD. This variable accounted for 11% of the variance in social impairment.
46 Pragmatic language accounted for a greater portion of the variance in social outcome than did general expressive language ability (Volden et al., 2009). Furthermore, the authors found that a large portion of the variability in pragmatic language ability w as not associated with general language skills. In addition, while weaker pragmatic language skills predicted social skill impairment, higher expressive language scores predicted more social skill problems (Volden et al., 2009). In conclusion, it is likely that having a relatively high degree of verbal fluency that is not supported by strong pragmatic skills could put a child with ASD at increased risk for depression. Although these results have yet to be replicate d, they provide support for the hypothesis that adequate pragmatic ability, rather than expressive language ability, is crucial to successful socioemotional development. Based on these findings, it is likely that adding pragmatic language ability to the ex isting model will significantly increase the usefulness of the model in predicting depressive symptoms. In summary, within the language domain, impaired pragmatic abilities appear to be the most closely associated with the development of co occurring psyc hiatric conditions (Hyter et al., 2001). Deficits in pragmatic language are common among individuals with ASD (Paul et al., 2009; Reichow et al., 2008; Tager Flusberg, 1989; Tager Flusberg et al., 2005). Pragmatic impairments reported among those with ASD include poor conversational skills, inappropriate expression of emotions, and the inability to use language to meet social functions. Pragmatic skills are critical for establishing and maintaining relationships with peers, teachers, and family members (V ickerstaff et al., 2007). Pragmatic difficulties can lead to failed social interactions and, consequently, depression. Individuals with ASD, particularly those with high functioning autism and AS, may have interest in social interactions but lack the skill s to participate successfully. A history of peer rejection paired with the inability to remediate their
47 communication impairment may lead to depression (Sterling et al., 2008). It is critical that the current study determines the unique contribution of pra gmatic language to the model, while controlling for the effects of general language ability. It is hypothesized that pragmatic ability will be related to, but independent from, general language ability. Furthermore, it is expected that pragmatic language w ill be a better predictor of depressive symptomatology than will general language ability. Limitations of Previous Research A review of the literature indicated that studies of depression in ASD have demonstrated numerous limitations. Some investigators h ave failed to account for confounding factors related Additionally, some researchers choose to rely on strict diagnostic criteria when identifying participants with depression. Using a screening tool such as the CDI that detects subclinical levels of depression may be advantageous over more rigorous criteria, such as clinical interviews or clinical diagnoses, because reliance on stricter criteria could limit the samp le and prevent researchers from identifying less severe cases. Also, investigating a broader range of depressive symptoms allows investigators to determine the degree to which certain factors contribute to depression, even when the depression may not yet b e fully manifested. There is a high likelihood of subclinical symptoms resulting in Major Depressive Disorder later in life; thus, when studying young children, it may be more useful to select for preliminary symptoms. report. In order ies, it is critical to collect data from multiple sources using multiple methods. The current study seeks to contribute uniquely to the field of related research by investigating a novel hypothesis, the contributing role of pragmatic
48 language skills in the depressive symptoms of children with ASD, while avoiding the limitations commonly noted in similar studies. Purpose of the Current Investigation In the area of communication, research indicates that pragmatic language ability is a primary area of deficit for children and adolescents with ASD. Many children with ASD desire to establish positive social relationships with peers and family members; however, due to limitations in social communication ability, they do not engage in successful interactions. Recur ring negative social interactions are likely to result in social isolation, withdrawal, and depression (Sterling et al., 2008). Furthermore, unsuccessful and infrequent social interactions can exacerbate existing pragmatic language deficits (Gallagher, 199 6; Sterling et al., 2008). Thus, communication impairments, especially in the area of pragmatics, may be an important contributor to co occurring socioemotional impairments. Acquiring a better understanding of the relationship between pragmatic language an d socioemotional outcomes, specifically depressive symptoms, will enable researchers and clinicians to implement more effective treatment programs The current investigation will determine the degree to which the inclusion of pragmatic language among a gro up of predictor variables will improve the usefulness of the model in predicting depressive symptoms. The addition of pragmatic language to the predictor set is likely to improve the predictive value of the model, as it accounts for emerging evidence that pragmatic ability is related to social and psychiatric outcomes for children with ASD. If pragmatic language limitations are found to be uniquely associated with increased susceptibility to depression in this population, then remediation of pragmatic skill s could lead to improved long term life outcomes for children and adolescents with ASD.
49 In addition, this investigation attempted to overcome methodological limitations inherent in the assessment of pragmatic language ability by assessing this skill throug h the use of two assessment tools: an observational rating scale and a standardized, performance based assessment of pragmatic behavior. These instruments provided valuable information about cessing skills, discourse strategies, and goal directed language (Russell & Grizzle, 2008). The current study seeks to shed more light on factors associated with co occurring depressive symptoms in ASD. This will be accomplished through an investigation of the degree of association between pragmatic language ability and depressive symptoms among children and adolescents with high functioning ASD. Specifically, the study will seek to answer the following research questions: Question 1: Previous research fi ndings indicate that genetic and environmental factors, general language ability, and nonverbal cognitive ability influence the risk for depressive symptomatology among children with ASD. To what degree does the inclusion of pragmatic language ability incr ease the predictive power of these factors? Question 2: For what portion of the variance in depressive symptomatology does pragmatic language ability uniquely account?
50 CHAPTER 2 METHODS Participants Subjects who had participated in research at the University of Florida involving children and adolescents with ASD were recruited for the current investigation. Participants were rmed Consent Form (IRB# 171 2008, entitled If such a clause was signed, a member of the research team contacted the family by phone and described the current study. The research team also recruited patients being seen at the Child and Adol escent Psychiatry Outpatient Clinic at University of Florida and the families enrolled in the University of Florida/Gainesville Center for Autism and Related Disabilities (CARD). In these cases, families learned about the study through word of mouth and th rough a recruitment flyer. Families who were interested in participating in the study contacted the research team by phone. Participants were recruited without regard to their likelihood of experiencing depressive symptoms. Within this group, potential par ticipants were identified for the current study based on the inclusion and exclusion criteria described below. Fifty children participated in the investigation. Inclusion Criteria Age To achieve eligibility for the current investigation, individuals were b etween the ages of 7 and 17 years. The age at which comorbid psychiatric conditions are likely to arise in ASD individuals is not currently known. In the typically developing population, depressive symptoms
51 peak between the ages of 13 and 15. However, Vick ersta ff et al. (2007) measured the depressive symptomatology of 30 children with ASD between the ages of 7 and 13 and found that 24% (Vickerstaff et al., 2007). This finding suggests that symptoms of depression may arise in children with ASD very early in development. The outcome measures selected for this investigation also restrict the age range for eligible participants, which further guided the age related inclusi on criteria. ASD diagnosis Individuals were also required to meet the criteria for ASD based on the diagnostic criteria set forth in the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (DSM IV). In order to demonstrate that this stan dard was met, participants were required to have a clinical diagnosis of ASD, made by a pediatrician, child psychiatrist, or licensed clinical psychologist with experience in ASD diagnosis. The following diagnoses were accepted: ASD, Autistic Disorder, AS, and PDD NOS. Due to the symptom heterogeneity observed among individuals with ASD, diagnosis is clinical diagnoses, the Autism Diagnostic Observation Schedule ( ADOS) was administered. The ADOS is a semi structured, standardized observation instrument used to assess social and communication behaviors in children suspected of having autism or ASD. It is widely used by clinicians and researchers to determine whether an individual is likely to meet diagnostic criteria for ASD (Lord, Rutter, DiLavore, & Risi, 2008). The ADOS yields a Social Communication score and a Total Score. Higher scores indicate a greater degree of impairment. The value of the ADOS Total Score is compared with cut off criteria to place an individual into one of three diagnostic categories: Autistic Disorder, Autism Spectrum Disorder, or Non spectrum.
52 Nonverbal cognitive ability Prior to selection, all potential participants were administered a mea sure of nonverbal cognitive ability: the Leiter International Performance Scale Revised (Leiter R; Roid & Miller, 2002). Participants whose nonverbal IQ was greater than or equal to 70, as measured by the Leiter R Brief IQ, were recruited to participate in the current research program. Exclusion ary C riteria Potential participants who demonstrated any of the following exclusion criteria were deemed ineligible to participate in the investigation: ASD that is secondary to some other genetic condition. Any severe sensory impairment, such as vision or hearing impairment, which could impede his or her performance on study related tasks. Adopted child whose family psychiatric history was not available for review. Measures Outcome V ariable Depressive symptoms through the use of two complementary norm Inventory Parent report (CDI Self report (CDI Self ; Kovacs, 2003). The CDI is an instrument designed to measure the presence of depressive symptomatology in children and adolescents ranging from age 7 to 17. The CDI Self includes 27 items which are scored with a 0, 1, or 2, indicating the degree to which a symptom is present. The CDI Parent includes 17 items presented in a similar format. Higher scores indicate a greater level of reported depressive symptomatology. Test items for the CDI Self are written at a first grade reading level. Items evaluate cogn itive, affective, and behavioral indicators of depression. Participant responses
53 on the CDI Self yield five factor scores: Negative Mood, Interpersonal Problems, Ineffectiveness, Anhedonia, and Negative Self Esteem. The CDI Parent yields two factor scores: Functional Problems and Emotional Problems. An overall measure of depressive symptomatology, called the Total Score, is also generated by both CDI forms. The CDI has adequate psychometric properties, including test retest reliabilities that range from 0. 54 to 0.87 and coefficient alpha ranging from 0.71 to 0.86 (Kovacs, 2003; Vickerstaff et al., 2007). Validity for the CDI is satisfactory; evidence of construct validity is acceptable, particularly for the Total Score (Sattler & Hoge, 2006). The CDI posses ses good discriminative validity; when a cut off score of 17 is used, sensitivity was found to be 80% and specificity was found to be 84% (Kovacs, 2003). Additionally, numerous investigations have documented good concurrent validity for the CDI; the CDI is strongly correlated with other measures of depressive symptoms, as well as with measures of anxiety and self esteem (Kovacs, 2003). Separate norms were developed for older and younger children and for males and females (Barnhill, 2001). Developers of simi lar instruments suggest that standard scores of 60 and greater indicate that the individual is at risk for developing depression and standard scores of 70 and greater indicate clinically significant levels of depressive symptomatology. Use of the CDI to e by numerous research teams. The CDI has been used in investigations of depressive symptoms exhibited by typically developing youth and those with developmental disorders (Barnhill, 2001; Rudolph, Hammen, & Burge, 1994; Vickerstaff et al., 2007; Williamson, Craig, & Slinger, 2008). Use of the CDI allows researchers and clinicians to evaluate the magnitude of depressive symptoms in a quantitative manner. The CDI can also be used as a measure of change in symptoms of depression to indicate treatment effects (Barnhill, 2001; Sattler & Hoge, 2006).
54 reported and parent reported Total Scores, measures of overall depressive symptomatology, were used in the current analysis. Anal yses were run separately for each CDI Total Score and the results were interpreted accordingly. Predictor V ariables Biological and environmental risk factors Two primary indicators of biological and environmental risk for depression were assessed: particip Checklist (LEC), a standardized rating scale developed by the Nation al Center for Posttraumatic exposure to potentially traumatic events that are suspected to put individuals at risk for PTSD. For instance, the LEC assess exposure to natural disasters, physical and sexual abuse, accidents, illness or injury, and death of a loved one. the LEC. A composite score was generated by assigning one point to any item that the caregiver indicated the child having experienced personally. Gray, Litz, Hsu, and Lombardo (2004) evaluated the psychometric properties of the LEC and concluded that the t ool has adequate test retest reliability and good convergent reliability when compared with an established measure of trauma exposure, the Traumatic Life Events Questionnaire (TLEQ). The test retest correlation for the LEC was 0.82. When LEC items were com pared with corresponding items on the TLEQ, the average kappa coefficient was 0.55, indicating good convergent reliability. No kappa psychometric properties ma potentially traumatic events.
55 Family history of mood disorders was obtained through parent report. The primary caregiver was asked whether anyone who is biologically related to the participa nt has ever been diagnosed with a mood disorder, including Major Depressive Disorder, Dysthymic Disorder, Depressive Disorder Not Otherwise Specified, Bipolar Disorders, and Other Mood Disorders. Family History was reported as a dichotomous variable reflec Nonverbal cognitive ability standardized, norm referenced assessment tool: the Leiter International Performance Scale Revised (Lei ter R; Roid & Miller, 2002). The Leiter R is a nonverbal measure of intellectual ability. Brief administration of this instrument consists of four subtests that may be administered and responded to nonverbally, allowing for the valid assessment of childre n for whom traditional, verbally loaded intelligence tests are unsuitable. This includes students who are hearing impaired, have limited English proficiency, and have moderate to severe speech or language impairments. The tasks that comprise the Leiter R i nclude identifying a figure embedded in a larger picture, recognizing an object based on an array of its segmented parts, putting picture cards in sequential order, and selecting a picture that completes a sequence. According to the test developers, these subtests are indicators of Fluid Reasoning and Visual/Spatial abilities (Roid & Miller, 2002). The Leiter R yields a measure of nonverbal intellectual ability called the Brief IQ. The Brief IQ contributed to the analyses in the current investigation. The subtests of the Leiter R possess adequate internal consistency, as do the IQ and Composite scores. Internal consistency rating for the Brief IQ is 0.88. The Leiter R also demonstrates satisfactory test retest reliability. Test retest reliability correlatio ns for the Brief IQ
56 range from 0.88 to 0.96, with greater reliability found when assessing youth between the ages of 11 and 20 (Roid & Miller, 2002). Investigations have documented evidence of content validity; validity based on internal structure; and val idity based on correlations with other measures of nonverbal ability, verbal ability, and achievement (Braden & Athanasiou, 2005). The correlation between the Leiter R Brief IQ and the WISC III Full Scale IQ is 0.85. In addition, the correlations between t he Leiter R Brief IQ and commonly used achievement tests, such as the Woodcock Johnson Tests of Achievement Revised (WJ R) range from 0.69 to 0.82 (Roid & Miller, 2002). General language ability the Clinical Evaluation of Language Fundamentals 4 th Edition (CELF 4), a widely used standardized assessment of overall language skills (Semel, Wiig, & Secord, 2003). The CELF 4 yields the Receptive Language Index (RLI), a measure of spoken language com prehension, and the Expressive Language Index (ELI), a measure of verbal language production. The RLI consists of two to three subtests that require individuals to follow complex directions presented orally, select two words from a group that are related, and demonstrate understanding of sentences and paragraphs. The ELI consists of three subtests that require test takers to repeat sentences presented orally, create sentences to describe a picture, and explain why two words are related. m oment correlation coefficients were calculated across two test administrations to provide evidence for test retest reliability. The reliability estimates for the RLI are good, ranging from 0.84 to 0.93. The test retest reliability estimates for the ELI are very good, ranging from 0.87 to 0.94 (Semel, Wiig, & Secord, 2003). Coefficient alpha for the RLI and ELI are 0.89 and 0.93, respectively, indicating good internal consistency. The developers of
57 the CELF 4 also documented evidence for test validity, bas ed on content, internal structure, and response process. The validity of the CELF 4 is also supported by confirmatory factor analysis. Furthermore, the index scores yielded by the CELF 4 are strongly correlated with those of its predecessor, the CELF 3 (Se mel, Wiig, & Secord, 2003). Pragmatic language ability assessment tools: an observational rating scale and a standardized, norm referenced assessment of pragmatic knowledge. The Pragmatics Profile (PP) is a criterion referenced observational rating scale intended to evaluate use of pragmatic language skills based on the responses of a caregiver. The PP includes 52 items that describe pragmatics based on three domains: Rituals and Conversational Skills; Asking For, Giving, and Responding to Information; and Nonverbal Communication Skills. The PP also yields an indicator of overall pragmatic abilities, the Total Raw Score. The Total Raw Score is compared with a criterion score b The test developers reported that the PP has high reliability. The overall internal consistency reliability coefficient for the PP is 0.98 (Semel, Wiig, & Se cord, 2003). Based on the paucity of evidence based pragmatic assessment instruments, Russell and Grizzle (2008) reviewed 24 measurement tools designed to assess pragmatic language, including questionnaires, checklists, and tests. The authors evaluated the tools based on their ability to assess 17 core pragmatic areas, such as requests, speech characteristics and fluency, nonverbal communication, topic maintenance, conversational turn taking, and negotiations and instructions. No one instrument evaluated al l 17 competency areas. However, the PP was found to probe 13 total domains. Further, the PP was one of only four tools that assessed all of the six competency areas that were determined to be most critical by the authors. Based on this evidence of conten t
58 validity, the authors recommend the use of the PP in the assessment of pragmatic language skills (Russell & Grizzle, 2008). ability, a performance based indicator of p ragmatic knowledge was also administered. The Comprehensive Assessment of Spoken Language (CASL) is a standardized, norm referenced evaluation tool used in the assessment of overall expressive language skills (Carrow Woolfolk, 1999). The CASL is comprised of 15 subtests that evaluate the comprehension, expression, and retrieval skills of individuals between the ages of 3 and 21. Two subtests of the CASL, Pragmatic s that comprise the Pragmatic Judgment and Inference subtests require participants to generate appropriate conversational responses and use background knowledge to answer questions about a situation. Internal reliability estimates for the Pragmatic Judgme nt and Inference subtests range from 0.77 to 0.92. Test retest reliabilities range from 0.66 to 0.84. The CASL also demonstrates evidence of adequate content validity and construct validity, based on progression of scores based on age, intercorrelations am ong subtests, and factor analysis. In fact, the Inference and Pragmatic Judgment subtests have intercorrelations that range from 0.59 to 0.67, indicating a relationship that is strong enough to support their combined use but weak enough to suggest that the y are measuring slightly different abilities (Carrow Woolfolk, 1999). Use of these subtests as a complement to traditional caregiver rating skills in the evaluation of pragmatics is recommended by Reichow and colleagues (2008). A comparison indicated that, when both measures are administered to children with ASD, the Pragmatic Judgment and Inference subtests were significantly correlated with the Communication and Social domains of the Vineland
5 9 Adaptive Behavior Scales, a parent report of adaptive social co mmunication. Higher scores on the CASL subtests predicted higher Vineland scores. These results support the use of the Pragmatic Judgment and Inference subtests as a performance based measure of pragmatic language abilities (Reichow et al., 2008). Procedur e The primary caregiver of each potential participant was contacted via telephone by the investigator. At that time, the child was assessed to determine whether he or she met the inclusion criteria based on parent report. The investigator also determined w hether the child was ineligible for the study based on exclusion criteria. A study visit was then scheduled for participants who were deemed likely to meet eligibility standards. Participation required one study visit, which took place in a private examin ation room at the University of Florida Psychiatry Research Clinic. First, the investigator reviewed the informed consent form with the participant and his or her parent or caregiver. The caregiver read and signed the consent form. If possible, minor parti cipants provided assent to participate in the study. After their caregivers provided consent, participants were administered the Leiter R by a trained examiner. The Leiter R yields a Brief IQ score. Leiter R administration took approximately 30 minutes and occurred in a private examination room. Then, the caregiver was given three rating scales to complete: the CDI Parent, the PP, and the LEC. The caregiver was also asked to provide information regarding family history of mood disorders. The caregiver was a sked whether anyone who is biologically related to the participant had ever been diagnosed with a mood disorder, including Major Depressive Disorder, Dysthymic Disorder, Depressive Disorder Not Otherwise Specified, Bipolar Disorders, and Other Mood Disorde rs.
60 While the caregiver was completing the rating scales, the child or adolescent participant engaged in four assessment tasks: (1) completing the CDI Self, (2) administration of two subtests of the CASL, (3) administration of the RLI and ELI of the CELF 4, and (4) administration of the ADOS. Completion of the study tasks took approximately three hours. The direct assessment tasks with each child took approximately two hours. Upon completion of the investigation, participants were debriefed and excused. Statistical Analyses As stated previously, the primary goal of the current investigation is to determine what portion of the variability in depressive symptomatology experienced by children and adolescents with ASD is accounted for by variations in pragma tic language ability, when variance accounted for by biological and environmental factors, general expressive language ability, and nonverbal Parent and CDI Self Total scores. Independ LEC ), (2) a dichotomous indicator of family history of mood disorders ( Fam ), dummy coded where 4 ( ELI ), an indicator of g eneral language ability, (4) Brief IQ, as measured by the Leiter R ( NVIQ ), (5) a measure of pragmatic language use in real world contexts (PP Total Raw Score ) CASL Inference and Pragmatic Judgment subtest standard scores (mean of 100 and standard setting. SPSS Version 16.0 was used for all statistical analyses.
61 Descriptive S tatistics In order to fully describe the sample in terms of the variables of interest, means, standard deviations, and ranges were calculated for all dependent and independent variables. Standard deviations indicate whether restriction of range occurred during the data collection phase. Correlational A nalyses A moment correlation matrix was constructed, indicating the zero order correlations between the variables. Correlational analyses also served as an indicator of multicollinearity between the independent variables Hierarchical Multiple R eg ression Hierarchical multiple regression analyses were used to predict depressive symptomatology based on the variables of interest: biological and environmental risk, general expressive language ability, nonverbal cognitive ability, and pragmatic language ability. Analyses were conducted at the ( p < .05) significance level. Four regression analyses were conducted: two which included CDI Parent Total Score as the outcome variable and two which included CDI Self Total Score as the outcome variable. Hierarch ical multiple regression has the capacity to address the research questions because it allows researchers to enter the predictor variables in the order of their choice, allowing for an evaluation of the relative contribution of each. The relative contribut ion of classes of variables scores on the assessment measures that is due to their chronological age was controlled. The next set of variables entered into the regression model was those that have been proposed by previous investigations: biological and environmental factors (Fam and LEC), general expressive language sk ill (ELI), and nonverbal cognitive ability (NVIQ). Because
62 previous research has not unequivocally determined the relative contributions of each predictor, these variables were entered simultaneously. Then, a second group of predictor variables was entere d into the model. For each outcome variable (CDI Parent and CDI Self), one regression analysis was conducted in which three indicators of pragmatic language ability were included in the third block. These indicators d Pragmatic Judgment subtest standard scores and PP Total Raw Scores. Then, both analyses were conducted again, this time with the PP Total Raw Score entered alone in the third block to determine whether this measure alone provided any additional predictiv e value to the model beyond the variables included in blocks one and two. Table 2 1 illustrates the variables included for each regression analysis.
63 Table 2 1 Variables included in four h ierarchical regression a nalyses Regression m odel Outcome variable Order of e ntry Predictor(s) 1 CDI Parent Total Score Block 1 Age in months Block 2 Life Events Checklist Family h istory CELF 4 ELI Leiter R Brief IQ Block 3 CASL: Inference CASL: Pragmatic j udgment PP Total Raw Score 2 CDI Parent Total Score Block 1 Age in months Block 2 Life Events Checklist Family h istory CELF 4 ELI Leiter R Brief IQ Block 3 PP Total Raw Score 3 CDI Self Total Score Block 1 Age in months Block 2 Life Events Checklist Family h istory CELF 4 ELI Leiter R Brief IQ Block 3 CASL: Inference CASL: Pragmatic j udgment PP Total Raw Score 4 CDI Self Total Score Block 1 Age in months Block 2 Life Events Checklist Family h istory CELF 4 ELI Leiter R Brief IQ Block 3 PP Total Raw Score Note 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken Language; PP = Pragmatics Profile.
64 CHAPTER 3 RESULTS The results obtained from these analyses are presented in three increments. First, descriptive statistics ar e presented for the demographic variables, outcome variables, and predictor variables. Then, results of the correlational analyses are presented, describing the relationships between the variables investigated in the study. Partial correlations are also pr esented, describing the unique relationships between the pragmatic language measures and depressive symptoms when other risk factors are ruled out. Finally, results of the hierarchical multiple regression analyses are presented, allowing for an evaluation of (1) the improvement in the model following the inclusion of the pragmatic language variables and (2) the amount of variability in depressive symptomatology that is accounted for by the predictor variables. Descriptive Statistics Demographic Variables Ta bles 3 1 through 3 3 describe the sample in terms of three demographic variables: gender, clinical diagnosis, and ADOS classification. Table 3 1 illustrates that more males participated in the study than did females. Specifically, 80% of the participants w ere male (n=40) and 20% were female (n=10). This proportion is similar to the relative prevalence of ASD found in the general population. Volkmar and Klin (2005) reported that ASD is approximately five times more prevalent among males than females. Further more, according to the CDC (2009), males are four to five times more likely to be diagnosed with ASD than are females. Table 3 2 illustrates the relative frequency with which each clinical diagnostic category was represented in the participant sample. The majority of participants had received clinical diagnoses of Asperger Syndrome (30%; n=15), followed by Autistic Disorder (28%; n=14) and PDD NOS (24%; n=12). The least common diagnostic category represented was ASD (14%; n=7). ASD is an
65 umbrella category t hat is not listed in the DSM IV. This diagnostic label is occasionally used more specific classification for their patients. Of note, caregivers of two participants did not diagnostic information exceeded the ADOS criteria for Autistic Disorder. Thus, despite missing clinical labels, the representativeness of the sample is believed to be intact. Table 3 3 illustrates the ADOS classification provided for each participant based on a sample of behavior that was observed and coded during their study session. According to these results, 56% of the sample met ADOS classi fication criteria for Autistic Disorder (n=28), 36% met criteria for ASD (n=18), and 8% did not meet criteria for any spectrum disorder (n=4). Thus, four participants did not obtain a classification of ASD based on their behavior during the ADOS assessment This trend can be explained by the parameters of the study. Because the study sought participants with cognitive abilities within the normal range who had sufficient language ability to respond to a standardized assessment instrument, many participants w ere high functioning and did not present with classical indicators of ASD, particularly during a brief behavior sample. Outcome Variables Table 3 4 displays descriptive statistics for two outcome variables: CDI Parent Total Score and CDI Self Total Score. The scores presented are T scores, which are normally distributed with a mean of 50 and a standard deviation of 10. Among study participants, the m ean CDI Parent score was 60.9 and the standard deviation was 11.1. These findings indicate that, according to parent report, the sample demonstrated a greater degree of depressive behaviors than were seen in the normative sample. Furthermore, the sample data was more highly concentrat ed in the above average range than was expected based on the normative data. Approximately 68% of the sample had CDI Parent scores between 49 and 71.
66 The mean CDI Self score for the sample was 54.1 and the standard deviation was 13.8. These scores are si milar to those found in the normative sample, indicating that children who participated in the study reported similar levels of depressive symptoms to children in the normative sample. Self report measures must be interpreted with caution, particularly whe n used with children with communication disabilities, such as ASD. It is likely that some children did not understand the test items. In addition, children are often inaccurate when providing self reports of behavior (Russell & Sofronoff, 2005). In order to further describe the depressive symptoms reported for the study sample, frequency counts were calculated for three descriptive categories: Within normal limits, At risk, and Clinically significant. According to the test developers, scores of 60 and abov e suggest that a risk range. Scores of 70 and above indicate that a which participants were rated in each of these catego ries are documented in Tables 3 5 and 3 6 Results indicate that parents reported clinically significant depression in 22% of the sample and at risk levels of depression in an additional 30% of participants. According to self reports, 10% of participants e xperienced depressive symptoms in the clinically significant range; an additional 16% reported that they may be at risk for depression. Predictor Variables Table 3 7 includes descriptive statistics for the following predictor variables: age in months, NVI Q, ELI, LEC, Fam, PP Total Raw Score, CASL: Inference Standard Score and CASL: Pragmatic Judgment Standard Score. The results obtained indicate that the mean ag e of the participants was 138.7 months, which corresponds to approximately 11 years, 6 months. The standard deviation is 36.9. Thus, the ages of study participants were widely distributed across a range from 85 to 215 months.
67 The mean nonver bal IQ of participants was 91.5 with a standard deviation of 16.9, indicating that the majority of participan ts had nonverbal intellectual abilities between 75 and below average to average range. The mean ELI of the study participants was 81.3 with a standard deviation o f 26.6. This suggests that expressive language skills were highly variable across participants. On average, expressive language skills of study participants were lower than what would be expected based on the performance of age matched peers in the normati ve sample. Most participants were reported to have experienced very few traumatic life events, as indic ated by a mean LEC score of 1.5 In contrast, the majority of participants were reported to have a family history of mood disorder. Specifically, the me an of Family History ratings was 0.73. An examination of frequencies (Table 3 8 ) indicates that 72% of participants had a family history significant for mood disorders (n=36), 24% did not (n=12), and 4% opted not to respond (n=2). The PP Total Raw Score p rovides an unstandardized indicator of the pragmatic language skills that an individual has been observed to use in natural settings. Results presen ted in Table 3 7 indicate that the mean PP Total Raw Score amo ng study participants was 114.9 Furthermore, the CASL: Inference and CASL: Pragmatic Judgment subtest standard scores measure the pragmatic language knowledge that an individual possesses and can demonstrate in structured settings. The mean CASL: Inference and CASL: Prag matic Judgment scores were 78. 1 and 73.4 knowledge was lower than what would be expected based on their chronological ages. Standard
68 deviations for the CASL subtest scores were 22.9 (Inference) and 23.3 (Pra gmatic Judgment), suggesting that participants varied widely in their knowledge of pragmatic language rules. Correlational Analyses Table 3 9 displays correlations between the outcome and predictor variables included i n Parent Total Scores and CDI Self Total Scores were not significantly correlated with any predictor variable. Notabl y, CDI Self Total Scores were not significantly correlated with CDI Parent Total Scores ( r = .16, p > .05), indicating that an respondent. Parent Total Scores were significantly correlated with three predictor variables: Age in months ( r = .33, p <.05), LEC ( r = .48, p < .01) and PP Total Raw Score ( r = .29, p < .05). Older participants were reported to experience more depressive symptoms. A s expected, children who reportedly experienced more traumatic life events displayed more behaviors suggestive of depression. Surprisingly, no relationship was discerned between parent reported symptoms of depression and family history of mood disorders. A lso, in contrast to previous investigations suggesting a relationship between expressive language skill and emotional/behavioral disorders (e.g. Howlin et al., 2000; Prizant et al., 1999), no relationship was found between ELI and CDI Parent Total Score. M ost noteworthy is the significant, though weak to moderate, relationship between CDI Parent Total Scores and PP Total Raw Scores. Based on these results, participants with more well developed pragmatic language skills were less likely to experience symptom s of depression.
69 Among predictor variables, several significant relationships were identified. First, all of the standardized, norm referenced assessment measures were found to be significantly correlated. The NVIQ score was significantly and strongly rel ated to ELI ( r = .52, p < .01) and CASL: Inference ( r = .51, p < .01) and was significantly and moderately related to CASL: Pragmatic Judgment ( r = .45, p < .05). CASL: Inference and CASL: Pragmatic Judgment scores were significantly and very strongly corr elated ( r = .84, p scores were significantly and strongly correlated with both CASL: Inference ( r = .79, p < .01) and CASL: Pragmatic Judgment ( r = .81, p < .01) subtest standard scores. For exploratory purposes, Pears between the outcome variables and all possible indicators of pragmatic language. The findings are displayed in Table 3 10 Pragmatic language variables included the PP, CASL: Inference, CASL: Pragmatic Language and the ADOS Social Communication score. The ADOS Social Communication score was included based on its frequent use as a measurement of social language abilities among individuals with ASD. It was expected that the ADOS Social Communication score would be strongly related to other measure of pragmatic language. This prediction was partially supported. The ADOS Social Communication score was significantly associated with the CASL: Inference ( r = .37 p < .01) and CASL: Pragmatic Judgment ( r = .48 p < .01) subtest standard scores. However, it was not significantly correlated with r = .07 p > .05). Thus, it remains unclear whether the ADOS could be used as a stand alone measure of pragmatic language ability in individuals w ith ASD. Of note, the ADOS Social Communication score was not significantly correlated with parent or self reported symptoms of depression.
70 Partial Correlations Bivariate correlational analyses indicated a significant association between symptoms of depre ssion and pragmatic language use, as reported by parents. However, it is possible that relationships between predictors (e.g. PP Total Raw Score and ELI) could influence Product Moment correlations. In order to further investigate the relationshi p between pragmatics and depressive symptoms, partial correlations were conducted between measures of depression and measures of pragmatic language skills, while controlling for other suspected predictors (age, NVIQ, ELI, LEC, and Fam). Table 3 11 presents the partial correlations between CDI Parent Total Score and measures of pragmatic language, controlling for age, NVIQ, ELI, LEC, and Fam. In accord with the bivariate correlational analysis, the only predictor found to be significantly associated with dep ressive symptoms was the PP Total Raw Score ( r = .34, p < .05). Furthermore, removing the influence of the other predictors led to an increase in the strength of the relationship. This finding indicates that the variance in CDI Parent explained by PP Total Raw Score is stronger after controlling for other variab les. Table 3 12 displays the partial correlations between CDI Self Total Score and measures of pragmatic language, controlling for age, NVIQ, ELI, LEC, and Fam. No significant relationships were revealed. This finding is consistent with the outcome of the bivariate correlational analysis. Hierarchical Multiple Regression Four hierarchical multiple regression analyses were conducted to examine whether measures of pragmatic language predict a significant portion of the variance in depressive symptoms, beyon d the contributions of predictors supported by previous research (i.e. age, nonverbal IQ, expressive language skill, life stress, family history of mood disorder). For all regression analyses, age in months was entered in the first block. Predictors with established
71 research support were entered in the second block. Finally, measures of pragmatic language were entered in the third block. As depicted in Table 2 1 the analyses differed in regards to the outcome variables and the pragmatic language measures. The first two analyses investigate the utility of the set of independent variables in predicting parent reported depressive symptoms. The second set of analyses investigates the prediction of self reported depressive symptoms. Within each set, one regress ion analysis includes all three measures of pragmatic language in the final block and one includes only PP Total Raw Score in the final block. PP Total Raw Score was included as the sole measure of pragmatic language ability for these analyses based on res ults of correlational analyses indicating that this measure was the only predictor, besides LEC, which had a significant relationship with either outcome variable. Furthermore, the surprising lack of statistical association between PP Total Raw Score and t he CASL subtests suggests that these measures may not be assessing the same capacity. Of note, an analysis of multicollinearity was conducted for all regression analyses, including variance inflation factor (VIF) and tolerance. According to the criteria s et forth by Field (2005), no serious concerns were noted for any predictors. Further, all Durbin Watson statistics were between 1 and 3, indicating that the assumption of independent errors was met for all analyses. Visual inspection of scatterplots sugges ted no concerns with normal distribution, linearity, or homoscedasticity. Parent Report of Depression Regression 1 Table 3 13 displays the results of the hierarchical regression for CDI Parent, which includes all three indicators of pragmatic language. Acc ording to these findings, it can be concluded that the full model accounted for 42.8% of the variance in parent reported symptoms of depression. Furthermore, the analysis yielded an F ratio of 3.65 ( p < .01), indicating that the
72 full model significantly im proves prediction of CDI Parent, beyond using the mean as an Parent Total Score (R = .12, Adjusted R = .10, p < .05), predicting 12% of the variance in the ou tcome measure. The inclusion of the previously investigated set of predictors in the second block predicted an additional 22% of the variance, representing a significant improvement in prediction (R = .22, Adjusted R = .27, p < .05). The pragmatic measur es included in the third block did not significantly improve the predictive power of the model (R = .08, Adjusted R = .31, p > .05). Within the full model, two predictors were found to make significant contributions: LEC ( = .46, p < .01) and PP Total R aw Score ( = .29, p < .05). Unique contribution of Pragmatics Profile: Regression 2 Table 3 14 displays the results obtained from the multiple hierarchical regression for CDI Parent including PP Total Raw Score as the sole indicator of pragmatic language ability. The full model was responsible for 42% of the variance in parent reported depressive s ymptoms. An F ratio of 4.98 ( p < .001) indicates that the model represents a significant improvement in predictive power beyond the mean. Age in months accounted for 12% of the variance in CDI Parent when entered in the first block ( p < .05). The set of pr edictors entered in the second block predicted an additional 22% ( p < .05); however, LEC was the only significant contributor ( p < .01). Including PP Total Raw Score in the third block proved to add significant predictive power to the model, accounting for an additional 7.7% ( p < .05). Within the full model, the only significant predictors were LEC ( = .47, p < .01) and PP Total Raw Score ( = .29, p < .05) Self Report of D epression Regression 3 Table 3 15 presents the multiple hierarchical regression fo r CDI Self, which includes all three indicators of pragmatic language. These findings demonstrate that the full model did not
73 significantly predict variance in self reported symptoms of depression for children with ASD. This finding is further supported by an F ratio of .27 ( p > .05), indicating that the full model did not improve prediction of CDI in months was not a significant predictor of CDI Self Total Score (R = .00, Adjusted R = .03, p > .05). The inclusion of the previously investigated set of predictors in the second block was also not a significant improvement, predicting an additional 1% of the variance (R = .01, Adjusted R = .14, p > .05). The pragmatic measures included in th e third block did not significantly improve the predictive power of the model (R = .06, Adjusted R = .18, p > .05); however, this contribution was greater than any of the previously entered variable groups. Within the full model, no predictors were foun d to make significant contributions. Unique c ontribution of Pragmatics Profile: Regression 4 Table 3 16 displays the results of the multiple hierarchical regression for CDI Self including PP Total Raw Score as the sole indicator of pragmatic language abili ty. The full model did not significantly predict variance in self reported depressive symptoms ( p > .05). An F ratio of .16 ( p > .05) further confirms that the model was not an improvement in prediction beyond Self Total Score (R = .00, Adjusted R = .03, p > .05). The set of predictors entered in the second block was also not a significant improvement (R = .01, Adjusted R = .14, p > .05). When included in the third block, PP Total Raw Score did not significantly improve the predictive power of the model (R = .03, Adjusted R = .15, p > .05). Within the full m odel, no predictors were found to make significant contributions.
74 Table 3 1 Frequencies for demographic v ariables: Gender Frequency Percent Valid Male 40 80 Female 10 20 Total 50 100 Table 3 2 Frequencies for demographic variables: Clinical d iagnosis Frequency Percent Valid ASD 7 14 Autistic Disorder 14 28 Asperger Syndrome 15 30 PDD NOS 12 24 Missing 2 4 Total 50 100 Note. ASD = Autism Spectrum Disorder; PDD NOS = Pervasive Developmental Disorder Not Otherwise Specified. Table 3 3 Frequencies for demographic v ariables: Autism D iagnostic Observation Schedule (ADOS) c lassification. Frequency Percent Valid ASD 18 36 Autistic Disorder 28 56 Nonspectrum 4 8 Total 50 100 Note. ASD = Autism Spectrum Disorder. Table 3 4 Descriptive statistics for outcome v ariables Variable N M SD Range CDI Parent total score 50 60.9 11.1 40 84 CDI Self total score 42 54.1 13.8 38 100 Note. Table 3 5 Frequencies for outcome v Parent Frequency Percent Valid Clinically s ignificant 11 22 At r isk 15 30 Within normal l imits 24 48 Total 50 100
75 Table 3 6 Frequencies for o utcome v Depression Inventory Self Frequency Percent Valid Clinically s ignificant 5 10 At r isk 8 16 Within normal l imits 29 58 Missing 8 16 Total 50 100 Table 3 7 Descriptive statistics for p redictor v ariables Variable N M SD Range Age in months 50 138.7 36.9 85 215 Leiter R Brief IQ 50 91.5 17.0 70 126 CELF 4 ELI 50 81.3 26.6 45 134 Life Events Checklist 49 1.5 1.4 0 5 Family history of mood disorder 48 .7 .4 0 1 CASL: Inference 50 78.1 22.9 40 138 CASL: Pragmatic j udgment 50 73.4 23.3 40 122 PP Total Raw Score 49 114. 9 21.9 80 177 Note. Maximum score for PP Total Raw Score = 208. CELF 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken Language; PP = Pragmatics Profile. Table 3 8 Frequencies for outcome variable: Family history of mood d isorders Frequency Percent Valid Yes 36 72 No 12 24 Missing 2 4 Total 50 100
76 Table 3 9 Pearson p roduct m oment c orrelations between o utcome and p redictor v ariables 1 2 3 4 5 6 7 8 9 10 1. CDI P Total 1.00 2. CDI S Total .16 1.00 3. Age in months .33 .03 1.00 4. Leiter R Brief IQ .08 .05 .15 1.00 5. CELF 4 ELI .20 .01 .28 .52 ** 1.00 6. Life Events Checklist .48 ** .07 .24 .15 .11 1.00 7. Family h istory .17 .01 .40 ** .06 .0 8 .40 ** 1.00 8. CASL: Inference .08 .06 .02 .51 ** .79** .0 1 .11 1.00 9. CASL: Pragmatic j udgment .19 .08 .03 .45** .81 ** .07 .14 .84** 1.00 10. PP Total Raw Score .29* .12 .00 .10 .14 .13 .24 .23 .09 1.00 Note. p < .05; ** p 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken Language; PP = Pragmatics Profile.
77 Table 3 10 Pearson product moment correlations between outcome and pragmatic language v ariables 1 2 3 4 5 6 1. CDI P Total 1.00 2. CDI S Total .16 1.00 3. CASL: Inference .08 .0 6 1.00 4. CASL: Pragmatic j udgment .19 .08 .84** 1.00 5. PP Total Raw Score .29* .12 .23 .09 1.00 6. ADOS Social Communication .22 .20 .3 7 ** .48** .07 1.00 Note. p < .05; ** p < Communication = Autism Diagnostic Observation Schedule Social Communication Score; CASL = Comprehensive Assessment of Spoken Language; PP = Pragmatics Profile. Table 3 11 P artial c orrelations between CDI P and variables measuring p ragmatic language controlling for a ge, Leiter R Brief IQ, CELF 4 E LI, Life Events Checklist, and family h istory 1 2 3 4 1. CDI Parent Total 1.00 2. CASL: Inference .01 1.00 3. CASL: Pragmatic j udgment .13 .49** 1.00 4. PP Total Raw Score .34* .16 .09 1.00 Note. p < .05; ** p 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken Language; PP = Pragmatics Profile.
78 Table 3 12 Partial c orrelations between CDI S and variables measuring pragmatic language c ontrolling for a ge, Leiter R Brief IQ, CELF 4 E LI, Life Events Checklist, and family h istory 1 2 3 4 1. CDI Self Total 1.00 2. CASL: Inference .14 1.00 3. CASL: Pragmatic j udgment .19 .51** 1.00 4. PP Total Raw Score .13 .12 .14 1.00 Note p < .05; ** p < 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken La nguage; PP = Pragmatics Profile.
79 Table 3 13 Regression 1: Hierarchical r egression p Parent total s core (N=50) Predictor variables B SE R Step 1 Age Age in months 0.11 0.04 0.35* 0.12* Step 2 Previously reported predictors 0.22* Age in months 0.05 0.05 0.17 Leiter R Brief IQ 0.18 0.11 0.26 CELF 4 ELI 0.09 0.07 0.21 Life Events Checklist 3.84 1.08 0.49** Family h istory 2.29 3.67 0.09 Step 3 Pragmatic language ability 0.08 Age in months 0.07 0.05 0.22 Leiter R Brief IQ 0.16 0.10 0.24 CELF 4 ELI 0.04 0.12 0.10 Life Events Checklist 3.57 1.07 0.46** Family h istory 3.51 3.80 0.14 CASL: Inference 0.01 0.13 0.03 CASL: Pragmatic j udgment 0.07 0.13 0.14 PP Total Raw Score 0.15 0.07 0.29* Note Total Adjusted R = .31. F (8, 39) = 3.65. p < .05; ** p < Depression Inventory; CELF 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken La nguage; PP = Pragmatics Profile.
80 Table 3 14 Regression 2: Hierarchical r egression p Parent t otal s core (N=50) Predictor variables B SE R Step 1 Age 0.12* Age in months 0.11 0.04 0.35* Step 2 Previously reported predictors 0.22* Age in months 0.05 0.05 0.17 Leiter R Brief IQ 0.18 0.11 0.26 CELF 4 ELI 0.09 0.07 0.21 Life Events Checklist 3.84 1.08 0.49* Family h istory 2.29 3.67 0.09 Step 3 Pragmatic language ability 0.08* Age in months 0.06 0.05 0.19 Leiter R Brief IQ 0.17 0.10 0.25 CELF 4 ELI 0.10 0.07 0.24 Life Events Checklist 3.66 1.03 0.47* Family istory 4.17 3.58 0.17 PP Total Raw Score 0.15 0.06 0.29* Note Total Adjusted R = .34. F (6, 41) = 4.98. p < .05; ** p Depression Inventory; CELF 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken Language; PP = Pragmatics Profile.
81 Table 3 15 Regression 3: Hierarchical r egression p redicting Childre Self total s core (N=40) Predictor var iables B SE R Step 1 Age 0.00 Age in months 0.01 0.06 0.03 Step 2 Previously reported predictors 0.01 Age in months 0.00 0.08 0.01 Leiter R Brief IQ 0.03 0.18 0.04 CELF 4 ELI 0.00 0.11 0.00 Life Events Checklist 0.91 1.82 0.10 Family h istory 1.59 6.84 0.05 Step 3 Pragmatic language ability 0.06 Age in months 0.03 0.09 0.08 Leiter R Brief IQ 0.03 0.18 0.04 CELF 4 ELI 0.17 0.20 0.32 Life Events Checklist 1.17 1.90 0.12 Family h istory 2.57 7.42 0.08 CASL: Inference 0.10 0.22 0.16 CASL: Pragmatic j udgment 0.14 0.22 0.24 PP Total Raw Score 0.08 0.12 0.13 Note Total Adjusted R = .18. F (8, 31) = 0.27. p < .05; ** p Depression Inventory; CELF 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive As sessment of Spoken Language; PP = Pragmatics Profile.
82 Table 3 16 Regression 4: Hierarchical r egression p redicting Childre Self total s core (N=40) Predictor variables B SE R Step 1 Age 0.00 Age in months 0.01 0.06 0.03 Step 2 Previously reported predictors 0.01 Age in months 0.00 0.08 0.01 Leiter R Brief IQ 0.03 0.18 0.04 CELF 4 ELI 0.00 0.11 0.00 Life Events Checklist 0.91 1.82 0.10 Family h istory 1.59 6.84 0.05 Step 3 Pragmatic language ability 0.02 Age in months 0.00 0.08 0.01 Leiter R Brief IQ 0.03 0.18 0.04 CELF 4 ELI 0.01 0.11 0.03 Life Events Checklist 1.02 1.83 0.11 Family h istory 0.78 6.96 0.03 PP Total Raw Score 0.09 0.11 0.14 Note Total Adjusted R = .15. F (6, 33) = .16. p < .05; ** p < Depression Inventory; CELF 4 = Clinical Evaluation of Language Fundamentals Fourth Edition; ELI = Expressive Language Index; Leiter R = Leiter International Performance Scale Revised; CASL = Comprehensive Assessment of Spoken La nguage; PP = Pragmatics Profile.
83 CHAPTER 4 DISCUSSION Children with ASD are at increased risk for experiencing depression and other psychiatric comorbidities (Ghazuiddin et al., 2002; Sterling et al., 2008; Vickerstaff et al., 2007). Little is known about factors likely to contribute to the depressive symptomatology experienced by children with ASD. Previous research has postulated that expressive language skill, cognitive functioning, and environmental and biological r depression (Benner et al., 2002; Gallagher, 1996; Geurts & Embrechts, 2008). The current study utilized hierarchical multiple regression to assess whether pragmatic language ability, when considered along side the previously reported factors, can be of use in predicting depressive symptoms among children and adolescents with ASD. It was hypothesized that including measures of pragmatic language ability in the regression model would increase the usefulness o f the model in predicting depressive symptomatology. The investigators predicted that children with better pragmatic language ability would be found to exhibit fewer symptoms of depression. Key Findings Previous research has yielded inconsistent results re garding the presence of psychiatric comorbidities among children with ASD. For example, Vickerstaff et al. (2007) estimated that 29% of their sample of children with high functioning autism exhibited a clinical level of depressive symptoms; further, an add itional 24% demonstrated subclinical depressive symptoms. In contrast, Barnhill (2001) reported more modest rates of depression in a sample of adolescents with Asperger Syndrome. R esearch findings vary widely in terms of estimated prevalence rates of psych iatric disorders and risk factors for comorbid disorders. In addition, researchers utilize various assessment techniques to measure comorbidities. Thus, further study in this field is
84 needed to both confirm previous findings and provide preliminary evidenc e for new areas of inquiry. Two novel research questions were proposed by the investigators. In addition to addressing these questions, the current findings illuminated important information regarding risk for depression among individuals with ASD. The fo llowing findings address both primary research inquiries and general information that will be valuable in further understanding psychiatric comorbidities among children with ASD. General findings are presented first, followed by findings that address prima ry research questions. Do Parent and Self Reports Agree Regarding the Presence of Depressive Symptoms in Youth with ASD? Methods of assessing comorbidities in ASD have varied widely, including case studies, clinical interviews, and behavior rating scales completed by parents, teachers, and children (Sterling et al., 2008; Volkmar & Klin, 2005). Behavior rating scales are commonly used due to the availability of normative information and their efficiency in terms of time and cost. However, many researchers have questioned the reliability of behavior ratings across informants (Lopata et al., 2010; Russell and Sofronoff, 2005). In this study, bivariate correlational analysis indicates that there is no significant association between parent and self reported symptoms of depression in the sample. This disparity can be partially explained by validity concerns, including the possibility that children with communication impairments may have difficulty responding to self report items and that some children may not have been candid with their responses. Numerous investigations have documented inconsistency between child and parent ratings of psychological adjustment. For instance, Russell and Sofronoff (2005) assessed parent and self reported symptoms of anxiety amo ng children with ASD. Parents reported significantly more impairment than did children.
85 The authors concluded that children with ASD were likely to lack self awareness regarding their difficulties and, thus, parent reports were likely to be more accurate ( Russell & Sofronoff, 2005). Research comparing parent and child reported behavior ratings in the general population has indicated generally low correlations across respondents (Lopata et al., 2010). Future research may benefit from including a third sourc e of behavioral information, such as teachers or Is Depression More Prevalent among Children with ASD than among Typically Developing Children? Vickerstaff et al. (20 07) found that 24% of their sample of children with ASD were prevalence estimate is similar to findings reported by Leyfer and colleagues (2006) and Wing (1981, as cited in Ghazuiddin et al., 2002). However, rates of depression reported by other research teams have varied widely. For instance, a review by Lainhart (1999) reported that prevalence estimates ranged from 4.4% to 57.6% between 1967 and 1999. Furthermore, Solomon et al. (2004) collected pretest self reports of depression among a sample of children with ASD prior to implementing a social skills intervention program. No participant reported depressive symptoms above the average range at pretest (Solomon, Good lin Jones, & Anders, 2004). Among the sample of children with ASD included in the current investigation, parents reported higher ratings of depression symptomatology, relative to the symptoms reported by parents of children in the normative sample. The me an total score on the CDI Parent for study participants was higher than expected based on the normative information. Results from the National Comorbidity Study estimated the lifetime prevalence rate of Major Depression in adolescents to be 14%, with an ad ditional 11% reporting subclinical symptoms of depression
86 (Hammen & Rudolph, 2003). Results of the current study indicate that 22% of study participants were rated as displaying depressive symptoms in the clinically significant range by caregivers; an addi tional 30% were reported to display symptoms in the at risk range. In agreement with previous research, these findings suggest that children with ASD are at increased risk for depressive symptoms, as compared to typically developing children (Ghazuiddin et al., 2002; Leyfer et al., 2006; Vickerstaff et al., 2007). Are the Three Measures Believed to Represent Pragmatic Language Ability Significantly Associated? The assessment of pragmatic language skills is difficult to accomplish due to the lack of valid a nd reliable instrumentation that can be used across a wide range of age groups (Russell & Grizzle, 2008; Volden et al., 2009). Based on previous research findings, three instruments were chosen to measure pragmatic language ability among research participa nts: PP, CASL Inference, and CASL Pragmatic Judgment (Reichow et al., 2008; Russell & Grizzle, 2008). Significant intercorrelations among these tools would provide evidence that they indeed measured the same capacity. Surprisingly, an analysis of the bivari ate correlations among predictor variables revealed no significant association between the CASL subtests and the PP. Although the CASL subtests are highly correlated with each other, neither is significantly associated with the PP. Thus, it is possible tha t these tools in fact measured two different constructs. Furthermore, the strength of the relationships between the CASL subtests and the measure of expressive language (ELI) makes it unclear whether these tools are in fact measuring distinct domains. The highly significant correlations between these measures suggest that the CASL subtests may simply Alternatively, these three pragmatic language tools may be measuring two di fferent aspects of the same construct. While the CASL subtests are likely to measure knowledge of pragmatic
87 language rules, as demonstrated in a structured assessment setting, the PP Total Raw Score age skills in naturalistic environments. It is likely that children with ASD, especially those with nonverbal cognitive skills within the average range, are aware of appropriate social communication behaviors but fail to use them consistently in real world interactions, thus resulting in socioemotional difficulties. Interestingly, the ADOS Social Communication score, an index often used in clinical and research settings to determine whether a child meets social and language criteria for ASD, was foun d to be significantly correlated with CASL subtest scaled scores. This strong inverse relationship indicates that participants with a greater degree of social communicative impairment as estimated by the ADOS were more likely to obtain lower scores on a st andardized assessment of pragmatic language. Although exploratory in nature, this result is consistent with research findings stating that pragmatic language impairment is common among individuals with ASD (Tager Flusberg, 2004; Travis & Sigman, 1998). Alt ernatively, these findings could suggest that the CASL subtests may not be valuable measures of assessing relative pragmatic skills among individuals with ASD, as they may in fact be measuring an intrinsic feature of the disorder. In contrast, participants Communication scores were not significantly correlated with their PP scores. Thus, the PP may be a more sensitive measure of pragmatic language use among our ASD sample. What Predictor Cariables are Significantly Correlated with Depressive S ymptomatology? Past research has provided preliminary evidence that general language ability, cognitive experiencing depression (Benner et al., 2002; Gallagher, 1996; Geurts & Embrechts, 2008). However, inconsistencies were noted in the existing literature. For instance, although most research teams concluded that impaired language skills were related to more depressive
88 symptoms, Sterling et al. (2008) noted that bette r social communication abilities were predictive of increased depression among adults with ASD. In addition, although numerous research teams have speculated that better cognitive abilities may be related to greater risk for depression (Vickerstaff et al., 2007; Volkmar & Klin, 2005), Kim et al. (2000) found no relationship between Leiter R nonverbal IQ and mood disorder. Furthermore, support for risk factors is often theoretical, without consistent empirical support (Sterling et al., 2008). The investigato rs of the current study diverged from previous researchers by postulating that measures of pragmatic language may be associated with indicators of depression among children with ASD. No previous research was found which investigated the relationship betwee n pragmatic language ability and depression in children with ASD. Support for an empirical relationship between pragmatic language and depression could have far reaching implications for assessment and intervention for individuals with ASD. Surprisingly, t he results of the bivariate correlational analysis indicate that only two variables were significantly correlated with parent reported depressive symptoms: life stress and PP Total Raw Score. Although many investigators have postulated that language abilit ies may play a role in the etiology of depression among individuals with ASD (Beitchman, Brownlie, & Wilson, 1996; Howlin, Mawhood, and Rutter, 2000), the measure of general expressive language ability (ELI) was not significantly correlated with depressive symptomatology in the current study. This finding is consistent with Kim et al. (2 000), who concluded that better early language ability is not related to mood problems in later childhood. The relationship between depression and PP Total Raw Score is part icularly noteworthy as it supports the hypothesis that pragmatic language is associated with depression in children with ASD. In order to further investigate this relationship, partial correlations were conducted between measures of depression and measures
89 of pragmatic language skills, while controlling for other suspected predictors (age, NVIQ, ELI, LEC, and family history). A significant relationship between PP Total Raw Score and depression was found, despite removing the potential influence of other var iables. This finding provides further evidence that a significant relationship exists between parent reported pragmatic language use and parent reported depressive symptoms. Of note, no variables were significantly associated with self reported depressive symptoms. Research Question 1: To What Degree does the I nclusion of Pragmatic L anguag e Ability Increase the Predictive Power of the Previously Investigated F actors? Four hierarchical multiple regression analyses were utilized to answer Research Question 1 Two series of hierarchical multiple regression analyses were conducted : one predicting parent reported depressive symptoms and another predicting self reported depressive symptoms. Within each series, one regression analysis included all three measures o f pragmatic language skill and another included only PP Total Raw Score as the sole indicator of pragmatic language ability. In conducting each hierarchical multiple regression analysis, SPSS performed three separate regression analyses: one including Block one; another including Blocks one and two; months. Block two included the predictor variabl es supported by previous research findings: family history of mood disorders, traumatic life events, general expressive language skill, and nonverbal IQ. Block three included the novel predictor: pragmatic language ability. Entering the variables in this m anner allowed the researchers to evaluate the change in the overall predictive power of the model that occurred when pragmatic language was added to the existing set of variables. SPSS reported the squared multiple regression coefficient (R) for each blo ck of predictors entered into each regression analysis. In addition, the program determined whether the R for each model reached significance, thereby indicating how well the variables that had been
90 entered predicted depressive symptomatology. An evaluati on of the change in R that resulted from the inclusion of the pragmatic language variable(s) determined the degree to which the inclusion of pragmatic language improved the predictive power of the model. Of note, neither Regression 3 nor Regression 4 suc cessfully predicted self reported did not predict symptoms of depression. Furthermore, the addition of pragmatic language measures in Block three did not improve the predictive power of either model. Therefore, further discussion focuses solely on Regressions 1 and 2, which predicted parent reported depressive symptomatology. Regression 1 indicated that, when all proposed risk factors were included, the model was useful in the prediction of parent reported depressive symptoms. However, including the three measures of pragmatic language ability did not significantly improve the model. Although this finding does not support the hypothesis that pragmatic language skil l uniquely contributes to the prediction of depression among individuals with ASD, it suggests that this set of risk factors could be useful to consider when determining whether an individual is at risk for developing depression. The full model for Regres sion 2 was also useful in predicting depressive symptoms among the study sample. Furthermore, including the PP Total Raw Score in Block three improved the predictive power of the model by 7.7%. In addition, the change in F after adding Block three was sign ificant, indicating that the addition of pragmatic language made a meaningful contribution to associated with symptoms of depression, even after accounting for rel ated abilities, such as general expressive language and cognitive ability.
91 Across the models, LEC and PP were the only significant predictors of parent reported depression symptomatology. This finding conflicts with previous research indicating that famil y history of mood disorders, general expressive language, and cognitive ability are risk factors for depression among youth with ASD. However, the full model for Regression 2 is significant, indicating that, when considered together, these factors improve prediction of depressive symptomatology Research Question 2: For W hat Portion of the Variance in Depressive Symptomatology does Pragmatic L anguag e Ability Uniquely A ccount? In addition to the findings reported previously, the investigation determined the amount of variance in depressive symptoms uniquely associated with indicators of pragmatic language ability, while controlling for the variance accounted for by nonverbal cognitive ability, general expressive language ability, and biological and environmen tal risk factors. Research Question 2 was addressed by an examination of the standardized regression c oefficient and semi partial correlation for PP, CASL Inference, and CASL Pragmatic Judgment in Block three of Regressions 1 and 2. The standardized regres sion coefficients denote the change in depressive symptomatology expected based on a one unit change in each measure of pragmatic language. The squared semi partial correlation signifies the portion of variance in depressive symptomatology uniquely associa ted with each pragmatic language ability indicator. In Regression 1, the standardized regression coefficient for the PP Total Raw Score was statistically significant. Furthermore, the unique contribution of PP to the prediction of the parent reported depressive symptoms was moderate ( r = .26). This value indicates that PP scores uniquely accounted for 6.3% of the variance in parent reported depressive symptoms. Neither CASL subtest significantly predicted depressive symptomatology. This provides parti al support for the study hypothesis.
92 Results of Regression 2 provide further evidence for the unique contribution of the PP Total Raw Score in predicting parent reported symptoms of depression. In this analysis, the unique contribution of PP to the predict ion of depression was also moderate ( r = .27); PP scores uniquely accounted for 7.7% of the variance in parent reported depression. Implications The findings yielded by the current investigation have important implications for the assessment and treatment of children with ASD. First, the results of the current study provide evidence that youth with ASD have elevated levels of depressive symptomatology relative to typically developing children. Thus, clinicians conducting diagnostic evaluations for individua ls suspected of having ASD ought to include an assessment of socioemotional functioning in their test battery. Additionally, the study results may improve screening practices for comorbid psychiatric conditions among individuals with ASD, leading to earli er identification and intervention. Because of the communication and social difficulties inherent in ASD, most individuals are not referred for psychiatric evaluation due to perceived changes in mood. Rather, caregivers notice behavioral manifestations of depression, such as increased aggression, changes in sleep or appetite, increased self injury, or more pronounced repetitive behaviors. Consequently, referrals often do not occur until symptoms of depression are quite severe (Lainhart, 1999). Earlier inter vention for depression is likely to lead to improved treatment outcomes. Overall, it appears ression. This information is useful for clinicians, as it could assist in identifying children who are at greater risk for depression so that targeted prevention measures can be implemented. Increased
93 prevention efforts could reduce the prevalence of comor bid depression among individuals with ASD, leading to improved quality of life. Furthermore, the findings provide preliminary evidence that participation in pragmatic language therapy programs may reduce depressive symptoms among children with ASD. Experi mental treatment research utilizing single subject design will be necessary to confirm this hypothesis in the future. In her dissertation study, Deal (2009) implemented a pragmatic language therapy program with a group of children in a South Georgia public school. By comparing pretest and posttest reports of problematic behavior, Deal concluded that students who had participated in the pragmatic language therapy program exhibited a significant decrease in problematic behavior; the control group did not demo nstrate similar behavioral improvements (Deal, 2009). It is likely that similar outcomes would be noted among children with developmental disabilities, such as ASD. Continuing to document the positive effects of pragmatic language therapy on socioemotional development for children is likely to improve academic, social, and behavioral outcomes for children and families. The implications of the current investigation extend beyond the ASD population. In concordance with previous research findings (Benner et al ., 2002; Gallagher, 1999; Prizant et al., 1990; Redmond & Rice, 1998), the results of the current study provide evidence for an interaction between language ability and socioemotional functioning. Consequently, speech/language pathologists and psychologist s should work together to address emotional and behavior problems that co occur with language impairment. Currently, most speech/language pathologists do not receive training regarding the identification and treatment of socioemotional difficulties, such a s aggression and withdrawal (Brinton & Fujiki, 1993). In the future, speech/language therapists may benefit from increased awareness of signs of an underlying
94 emotional/behavioral issue. Additionally, psychologists and behavior specialists should be aware of red flags for communication difficulties when working with children with emotional and behavioral disturbances. Such interdisciplinary training is likely to allow professionals to identify and treat comorbid disorders earlier, leading to more positive t reatment effects. Limitations Results of the current study should be interpreted with caution due to the limited sample size. Wilson Van Voorhis and Morgan (2007) concluded that the general rule of thumb is to include at least 50 participants for a regress ion analysis. Harris (1985) recommended that, to conduct multiple regression analysis with optimal statistical power, a minimum of 10 participants should be included for every predictor variable. The current investigation included up to eight predictors, d epending on the regression model. Thus, it would have been optimal to include 80 participants. Participants were recruited for 14 months; however, due to the low incidence of ASD in the population and the time demands required for participation, only 50 pa rticipants were included in the final data set. In addition, some potential subjects were excluded from the final data analysis due to nonverbal IQ below the criterion level. R esearchers conducting comparable investigations have experienced similar recruit ment difficulties. For instance, Volden et al. (2009) conducted a multiple regression analysis predicting social outcomes for individuals with ASD, including 37 participants. Overall, research in the field of ASD would benefit from increased recruitment ef forts and collaboration among research teams to share data, allowing for larger sample sizes. The research findings presented here should also be considered in terms of possible threats to validity. When a strong association is found between two variables it is possible that an underlying confounding variable has inflated the relationship between the two. Because two of the primary measures used in the regression analysis were parent reports of behavior (i.e. PP and
95 CDI Parent), it is possible that both are influenced by parent response style. Parents who tend to emphasize problematic behaviors are more likely to under report pragmatic language skills and over report depressive symptoms. In contrast parents emphasizing positive aspects of development are likely to provide inverse ratings. Future studies can account for this influence by collecting data from multiple sources, such as teacher ratings, and from utilizing objective, observational instruments to assess pragmatic language skills. Additionally, although measurement tools were selected on the basis of psychometric value, their reliability and validity have not been established for individuals with developmental disabilities. In particular, the CDI includes items assessing social withdrawal, a cor e feature of ASD. Overlapping features of depression and ASD make it difficult to distinguish clinical depression from inherent features of autism. Clinical interviews may yield a more valid indicator of depression among individuals with ASD. The study res ults provide preliminary evidence for a significant and unique relationship between pragmatic language ability and depression in children with ASD. However, because this investigation was not experimental, causation cannot be determined. Thus, although it is possible that limitations in pragmatic language cause children with ASD to experience depression, it is possible that the pragmatic communication deficits noted in the sample are consequences of depression. In fact, depression is often manifested by lim ited social communication. Gravell and France (1991) noted that depressed patients often demonstrate reduced nonverbal communication and reduced initiations. In the future, longitudinal studies may be valuable in examining whether early pragmatic language impairment is related to depressive symptoms experienced later in life. In summary, although no research study is free from limitations, the current investigation provides compelling evidence that pragmatic language ability is uniquely related to risk for
96 depression among children with ASD. This finding builds on previous research conclusions by more precisely identifying which aspects of communication are associated with depression. Armed with this knowledge, future researchers and clinicians may develop pragmatic language treatment strategies that prevent and remediate socioemotional difficulties in children with ASD.
97 LIST OF REFERENCES American Psychiatric Association. (2000). Diagnostic and statistical manual of me ntal disorders (Revised 4 th ed.). Washington, DC: Author. Baker, L. & Cantwell, D. Factors associated with the development of psychiatric illness in children with early speech/language problems. Journal of Autism and Developmental Disorders, 17 (4), 499 51 0. Baltaxe, C. A. M., & Simmons, J. Q. (1988). Pragmatic deficits in emotionally disturbed children and adolescents. In Schiefelbusch, R. L. & Lloyd, L. L. (Eds.), Language perspectives: Acquisition, retardation and intervention (pp. 223 253). Austin, TX: Pro Ed. Barnhill, G. P. (2001). Social attributions and depression in adolescents with asperger syndrome. Focus on Autism and other Developmental Disabilities, 16 (1), 46 53. Baron Cohen, S. (1988). Social and pragmatic deficits in autism: Cognitive or affe ctive? Journal of Autism and Developmental Disorders, 18 (3), 379 402. Beitchman, J. H., Brownlie, E. B., & Wilson, B. (1996). Linguistic impairment and psychiatric disorder: Pathways to outcome. In Beitchman, J. H., Cohen, N. J., Konstantareas, N. M., & T annock, R. (Eds.). Language, learning, and behavior disorders: Developmental, biological, and clinical perspectives (pp.493 514). New York: Cambridge University Press. Bellini, S. (2006). The development of social anxiety in adolescents with autism spectru m disorders. Focus on Autism and other Developmental Disabilities, 21 (3), 138 145. Benner, G. J., Nelson, J. R., & Epstein, M. H. (2002). Language skills in children with EBD: a literature review. Journal of Emotional and Behavioral Disorders, 10 (1), 43 57. Billstedt, E., Gillberg, C., & Gillberg, C. (2005). Autism after adolescence: Population based 13 to 22 year follow up study of 120 individuals with autism diagnosed in childhood. Journal of autism and developmental disorders, 35 (3), 351 360. Brade n, J. P. & Athanasiou, M. S. (2005). A comparative review of nonverbal measures of intelligence. In Flanagan, D. P. & Harrison, P. L. (Eds.), Contemporary intellectual assessment: Theories, tests, and issues (pp. 557 578). New York: The Guilford Press. Bretherton, I., Fritz, J., Zahn Waxler, C., & Ridgeway, D. (1986). Learning to talk about emotions: A functionalist perspective. Child Development, 57 529 548. Brinton, B. & Fujiki, M. (1993). Clinical forum: Language and social skills in the school age p opulation: Language, social skills, and socioemotional behavior. Language, Speech, and Hearing Services in Schools, 24, 194 198. Carrow Woolfolk, E. (1999). Comprehensive Assessment of Spoken Language Manual Minneapolis, MN: NCS Pearson Inc.
98 Carter, A. S. Davis, N. O., Klin, A., & Volkmar, F. (2005). Social development in Autism. In Volkmar, F.R., Paul, R., Klin, A., & Cohen, D. (Eds.). Handbook of Autism and Pervasive Developmental Disorders (pp. 312 334). Hoboken, NJ: John Wiley & Sons, Inc. Centers for Disease Control and Prevention (2009). Prevalence of Autism Spectrum Disorders Autism and Developmental Disabilities Monitoring Network, United States, 2006. Morbidity and Mortality Weekly Report Surveillance Summaries 58 (SS10), 1 20. Chaban, P. (1996 ). Understanding language dysfunction from a developmental perspective: An overview of pragmatic theories. In Beitchman, J. H., Cohen, N. J., Konstantareas, N. M., & Tannock, R. (Eds.). Language, learning, and behavior disorders: Developmental, biological, and clinical perspectives (pp.23 37). New York: Cambridge University Press. Cohen, N. J. (1996). Unsuspected language impairments in psychiatrically disturbed children: Developmental issues and associated conditions. In Beitchman, J. H., Cohen, N. J., Kon stantareas, N. M., & Tannock, R. (Eds.). Language, learning, and behavior disorders: Developmental, biological, and clinical perspectives (pp.105 127). New York: Cambridge University Press. Cooper, P.J. & Goodyer, I. (1993). A community study of depression in adolescent girls: I: Estimates of symptom and syndrome prevalence. British Journal of Psychiatry, 163 369 374. Dale, P. S. (1996). Language and emotion: A developmental perspective. In Beitchman, J. H., Cohen, N. J., Konstantareas, N. M., & Tannock, R (Eds.). Language, learning, and behavior disorders: Developmental, biological, and clinical perspectives (pp.5 22). New York: Cambridge University Press. Deal, S. (2009). The relationship between pragmatic language skills and behavior. Dissertation Abstr act International, 70 (12). (UMI No. AAT 3380348). Retrieved March 26, 2010 from Dissertations and Theses database. DeThorne, L. S. & Watkins, R. V. (2006). Language abilities and nonverbal IQ in children with language impairment: Inconsistency across meas ures. Clinical Linguistics & Phonetics, 20 (9), 641 658. Dodge, K. A. & Frame, C. L. (1982). Social cognitive biases and deficits in aggressive boys. Child Development, 53 620 635. Donahue, M. and Cole, D (1994). Links between language and emotional behav ioral disorders. Education and Treatment of Children, 17 3, 244 255. Downs, A. & Smith, T. (2004). Emotional understanding, cooperation, and social behavior in high functioning children with autism. Journal of Autism and Developmental Disorders, 34 (6), 62 5 635. Eaves, L. C. & Ho, H. H. (2008). Young adult outcome of autism spectrum disorders. Journal of Autism and Developmental Disorders, 38 739 747.
99 Elder, L., Caterino, L., Chao, J., Shaknai, D., & De Simone, G. (2006). The efficacy of social skills trea tment for children with Asperger syndrome. Education and Treatment of Children, 29 (4), 635 663. Elliott, C. D. (2005). The Differential Ability Scales. In Flanagan, D. P. & Harrison, P. L. (Eds.), Contemporary intellectual assessment: Theories, tests, and issues (pp. 402 424). New York: The Guilford Press. Elliott, C. D. (2007). Differential Ability Scales Second Edition: Introductory and Technical Handbook San Antonio, TX: Harcourt Assessment, Inc. Evans, J.R., van Velsor, P., & Schumacher, J. E. (2002 ). Addressing adolescent depression: A role for school counselors. Professional School Counseling, 5 211 219. Fombonne, R. (2005). The changing epidemiology of autism. Journal of Applied Research in Intellectual Disabilities, 18 281 294. Fujiki, M. Brint on, B., Morgan, M., & Hart, C. (1999). Withdrawn and sociable behavior of children with language impairment. Language, Speech, and Hearing Services in Schools, 30 183 195. Gallagher, T. M. (1993). Language skill and the development of social competence in school age children. Language, Speech, and Hearing Services in Schools, 24 199 205. Gallagher, T. M. (1996). Social interactional approaches to child language intervention. In Beitchman, J. H., Cohen, N. J., Konstantareas, N. M., & Tannock, R. (Eds.). La nguage, learning, and behavior disorders: Developmental, biological, and clinical perspectives (pp.418 435). New York: Cambridge University Press. problems. Topi cs in Language Disorders, 19 (2), 1 15. Geurts, H. M., & Embrechts, M. (2008). Language profiles in ASD, SLI, and ADHD. Journal of Autism and Developmental Disorders, 38 1931 1943. Ghazuiddin, M., Ghazuiddin, N., & Greden, J. (2002). Depression in persons with autism: Implications for research and clinical care. Journal of Autism and Developmental Disorders, 32 (4), 299 306. Gravell, R. & France, J. (1991). Speech and communication problems in psychiatry San Diego, CA: Singular Publishing Group, Inc. Gray, M. J., Litz, B. T., Hsu, J. L., & Lombardo, T. W. (2004). Psychometric properties of the life events checklist. Assessment, 11 (4), 330 341. Hammen, C. & Rudolph, K. D. (2003). In Mash, E. J., & Barkley, R. A. (Eds.), Child psychopathology: Second edi tion (pp. 233 278). New York: The Guilford Press.
100 Harris, R. J. (1985). A primer of multivariate statistics: Second edition. New York: Academic Press. Howlin, P. (2005). Outcomes in Autism Spectrum Disorder. In Volkmar, F.R., Paul, R., Klin, A., & Cohen, D. (Eds.). Handbook of Autism and Pervasive Developmental Disorders (pp. 201 222). Hoboken, NJ: John Wiley & Sons, Inc. Howlin, P., Mawhood, L., & Rutter, M. (2000). Autism and developmental receptive language disorder A follow up comparison in early adu lt life. II: Social, behavioral, and psychiatric outcomes. Journal of Child Psychology and Psychiatry, 41 (5), 561 578. Hyter, Y. D., Rogers Adkinson, D. L., Self, T. L., Simmons, B. F., & Jantz, J. (2001). Pragmatic language intervention for children with language and emotional/behavioral disorders. Communication Disorders Quarterly, 23 (1), 4 16. Johnston, E. B., Weinrich, B. D., & Glaser, A. J. (1991). A sourcebook of pragmatic activities Revised: Theory and intervention for language therapy (PK 6). Tucs on, AZ: Communication Skill Builders, Inc. & Gray, D. B. (Eds.). intervention (pp.3 10). Baltimore, MD: Paul H. Brookes Publishing Co. Kanner, L. (1943). Autistic disturbances of affective contact. Nervous Child, 2, 217 250. Kanner, L. (1973). Childhood Psychosis: Initial Studies and New Insights New York: Winston/Wiley. Kelly, A. B., Garnett, M. S., Atwood, T., & Peterson, C. (2008). Autism spectrum symptomatology in children: The impact of family and peer relationships. Journal of Abnormal Child Psychology, 36 1069 1081. Kim, J., Szatmari, P., Bryson, S., Steiner, D., & Wilson, F. (2000). The preval ence of anxiety and mood problems among children with autism and asperger syndrome. Autism, 4 (2), 117 132. Tonawanda, NY: Multi health Systems. Lainhart, J. (1999). Ps ychiatric problems in individuals with autism, their parents and siblings. International Review of Psychiatry, 11 278 298. Larson, V. L. & McKinley, N. L. (1987). Communication assessment and intervention strategies for adolescents Eau Claire, WI: Thinki ng Publications.
101 Leyfer, O. T., Folstein, S. E., Bacalman, S., Davis, N. O., Dinh, E., Morgan, J., Tager Flusberg, H., & Lainhart, J. E. (2006) Comorbid psychiatric disorders in children with autism: Interview development and rates of disorders. Journal o f Autism and Developmental Disorders, 36 849 861. Lopata, C., Toomey, J., Fox, J., Volker, M., Chow, S., Thomeer, M., Lee, G., Rodgers, J., McDonald, C., & Smerbeck, A. (2010). Anxiety and depression in children with HFASDs: Symptom levels and source diff erences. Journal of Abnormal Child Psychology, 38 765 776. Lord, C., Rutter, M., DiLavore, P., & Risi, S. (2008). Autism Diagnostic Observation Schedule Manual Los Angeles, CA: Western Psychological Services. Lord, C., Shulman, C., & DiLavore, P. (2004). Regression and word loss in autistic spectrum disorders. Journal of Child Psychology and Psychiatry 45, 234 252. Mack, A. E. & Warr Leeper, G. A. (1992). Language abilities in boys with chronic behavior disorders. Language, Speech, and Hearing Services in Schools, 23 214 223. Matson, J. L. & Nebel Schwalm, M. S. (2007). Comorbid psychopathology with autism spectrum disorder in children: An overview. Research in Developmental Disabilities, 28 341 352. McCarthy, J., Downes, E. J., & Sherman, C. A. (2008) Looking back at adolescent depression: A qualitative study. Journal of Mental Health Counseling, 30 (1), 49 68. McDonough, K. M. (1989). Analysis of the expressive language characteristics of emotionally handicapped students in social interactions. Behavi oral Disorders, 14 127 149. Montague, M., Enders, C., Dietz, S., Dixon, J., & Cavendish, W. M. (2008). A longitudinal study of depressive symptomology and self concept in adolescents. The Journal of Special Education, 42 (2), 67 78. Mupawose, A., Katijah, K., & Naran, V. (2007). Depression and pragmatics in adult patients with mild traumatic brain injury: An exploratory study. Journal of Psychology in Africa, 17 111 114. Nelson, J. R., Benner, G. J., & Cheney, D. (2005). An investigation of the language s kills of students with emotional disturbance served in public school settings. The Journal of Special Education, 39 (2), 97 105. Nolan, S. A., Flynn, C., & Garber, J. (2003). Prospective relations between rejection and depression in young adolescents. Journ al of Personality and Social Psychology, 85 (4), 745 755. Orsmond, G. I., Krauss, M. W., & Seltzer, M. M. (2004). Peer relationships and social and recreational activities among adolescents and adults with autism. Journal of Autism and Developmental Disorde rs, 34 (3), 245 256.
102 Paul, R., Orlovski, S. M., Marcinko, H. C., & Volkmar, F. (2009) Conversational behaviors in youth with high functioning ASD and Asperger syndrome. Journal of Autism and Developmental Disorders, 39 (1), 115 125. Price, J. & Dodge, K. T. J. & Ladd, G. (Eds.) Peer Relationships in Child Development (pp. 341 370). New York: John Wiley. Prizant, B. M., Audet, L. R., Burke, G. M., Hummel, L. J., Theadore, G., & Mahe r, S. R. (1990). Communication disorders and emotional/behavioral disorders in children and adolescents. Journal of Speech and Hearing Disorders, 55 179 192. Redmond, S. M. & Rice, M. L. (1998). The socioemotional behaviors of children with SLI: Social ad aptation or social deviance? Journal of Speech, Language, and Hearing Research, 41 688 700. Reichow, B., Salamack, S., Paul, R., Volkmar, F. R., & Klin, A. (2008). Pragmatic assessment in autism spectrum disorders: A comparison of a standard measure with parent report. Communication Disorders Quarterly, 29 (3), 169 176. Reid, R., Gonzalez, J. E., Nordness, P. D., Trout, A., & Epstein, M. H. (2004). A meta analysis of the academic status of students with emotional/behavioral disturbance. The Journal of Speci al Education, 38 (3), 130 143. role of depressive symptoms, social competence, and gender. Infant and Child Development, 15 89 107. and social interactions. In Kaiser, A. P. & Gray, D. B. (Eds.). communication: Research foundations for intervention (pp.139 158). Baltimore, MD: Paul H. Brookes Publish ing Co. Rice, M. Sell, M. & Hadley, P. (1991). Social interactions of speech and language impaired children. Journal of Speech and Hearing Research 34 1299 1308. Roid, G. H., & Miller, L. J. (2002). Leiter International Performance Scale Revised: Examine rs Manual Wood Dale, IL: Stoelting, Co. Rudolph, K. D., Hammen, C., & Burge, D. (1994). Interpersonal functioning and depressive symptoms in childhood: Addressing the issues of specificity and comorbidity. Journal of Abnormal Child Psychology, 22 (3), 355 372. Russell, E. & Sofronoff, K. (2005). Anxiety and social worries in children with Asperger syndrome. Australian and New Zealand Journal of Psychiatry, 39 633 638.
103 Russell, R. L. & Grizzle, K. L. (2008). Assessing child and adolescent pragmatic language competencies: Toward evidence based assessments. Clinical Child and Family Psychology Review, 11 59 73. Rutter, M. (1970). Autistic children: Infancy to adulthood. Seminars in Psychiatry, 2 435 450. Rutter, M. (1978). Diagnosis and definition. In Autis m: A Reappraisal of Concepts and Treatment (Eds. Rutter, M. and Schopler, E.). Plenum, New York. rejection experiences: Impact on social and emotional adjustment. Social Development, 12 (4), 530 550. Sattler, J. M. & Hoge, R. D. (2006). Assessment of Children: Behavioral, social, and clinical foundations (5 th ed.). San Diego, CA: Jerome M. Sattler, Publisher, Inc. Seltzer, M. M., Krauss, M. W., Shattuck, P. T., Orsm ond, G., Swe, A., & Lord, C. (2003). The symptoms of autism spectrum disorders in adolescence and adulthood. Journal of autism and developmental disorders, 33 (6), 565 581. Seltzer, M. M., Shattuck, P., Abbeduto, L. & Greenberg, J. S. (2004). Trajectory of development in adolescents and adults with autism. Mental retardation and developmental disabilities research reviews, 10 234 247. Semel, E., Wiig, E. H., & Secord, W. A. (2003). Clinical Evaluation of Language Fundamentals nual. San Antonio, TX: Harcourt Assessment, Inc. Solomon, M., Goodlin Jones, B.L., & Anders, T. F. (2004). A social adjustment enhancement developmental disorder NOS. Journal of A utism and Developmental Disorders, 34 (6), 649 668. Sterling, L., Dawson, G., Estes, A., & Greenson, J. (2008). Characteristics associated with presence of depressive symptoms in adults with autism spectrum disorder. Journal of Autism and Developmental Disorders, 38, 1011 1018 Stevenson, J. (1996). Developmental changes in the mechanisms linking language disabilities and behavior disorders. In Beitchman, J. H., Cohen, N. J., Konstantareas, N. M., & Tannock, R. (Eds.). Language, learning, and behavior disorders: Developmental, biological, and clinical perspectives (pp.78 100). New York: Cambridge University Press. Tager Flusberg, H. (1989). A psycholinguistic perspective on language development in the autistic child. In Dawson, G. (Ed.). Autism: Nature, Diagnosis, and Treatment (pp.92 118). New York: Guilford Press. Tager Flusberg, H. (2004). Strategies for conducting research on language in autism. Journal of Autism and Developmental Disorders, 34 (1), 75 80.
104 Tager Flusb erg, H., Baron Cohen, S. & Cohen, D. (1993). An introduction to the debate. In Tager Flusberg, H., Baron Cohen, S. & Cohen, D. (Eds.). Understanding other minds (pp. 3 9). New York: Oxford University Press Inc. Tager Flusberg, H., Paul, R., & Lord, C. (200 5). Language and communication in Autism. In Volkmar, F.R., Paul, R., Klin, A., & Cohen, D. (Eds.). Handbook of Autism and Pervasive Developmental Disorders (pp. 335 364). Hoboken, NJ: John Wiley & Sons, Inc. Tannock, R. & Schachar, R. (1996). Executive dy sfunction as an underlying mechanism of behavior and language problems in attention deficit hyperactivity disorder. In Beitchman, J. H., Cohen, N. J., Konstantareas, N. M., & Tannock, R. (Eds.). Language, learning, and behavior disorders: Developmental, bi ological, and clinical perspectives (pp.128 155). New York: Cambridge University Press. Travis, L. L. & Sigman, M. (1998). Social deficits and interpersonal relationships in autism. Mental Retardation and Developmental Disabilities, 4 65 72. Turner, M. (1999). Annotation: Repetitive behavior in autism: A review of psychological research. Journal of Child Psychology and Psychiatry, 40 (6), 839 849. Vickerstaff, S., Heriot, S., Wong, M., Lopes, A., & Dossetor, D. (2007). Intellectual ability, self perceived social competence, and depressive symptomology in children with high functioning autism spectrum disorders. Journal of Autism and Developmental Disorders, 37 1647 1664. Volden, J., Coolican, J., Garon, N., White, J., & Bryson, S. (2009). Brief report: Pr agmatic language in autism spectrum disorder: Relationships to measures of ability and disability. Journal of Autism and Developmental Disorders, 39 388 393. Volkmar, F. R. & Klin, A. (2005). Issues in the classification of Autism and related conditions. In Volkmar, F.R., Paul, R., Klin, A., & Cohen, D. (Eds.). Handbook of Autism and Pervasive Developmental Disorders (pp. 5 41). Hoboken, NJ: John Wiley & Sons, Inc. Vygotsky, L. S. (1986). Thought and Language (A. Kozulin, Trans.). Cambridge, MA: The Massac husetts Institute of Technology. (Original work published 1962). Williamson, S., Craig, J. & Slinger, R. (2008). Exploring the relationship between measures of self esteem and psychological adjustment among adolescents with Asperger Syndrome. Autism, 12 (4 ), 391 402. Wilson Van Voorhis, C. R. & Morgan, B. L. (2007). Understanding power and rule of thumb for determining sample sizes. Tutorials in Quantitative Methods for Psychology, 3 (2), 43 50.
105 BIOGRAPHICAL SKETCH Emily Ann Wray was born in Flint, M ichigan. She and her twin sister, Elizabeth Buckley, were raised in North Carolina and graduated from Piedmont High School in Monroe, North Carolina Emily attended the University of Fl orida and obtained her Bachelor of Science degree in April of 2005. She entered the school p sychology program at the University of Florida in August of 2005 and obtained her Master of Arts in Education degree in December of 2008. She will earn her Doctor of Philosophy in School Psychology in August of 2011. Emily completed a clinical internship at the Carolina Institute for Developmental Disabilities (CIDD) at the University of North Carolina. She will continue her training at the CIDD for her postdoctoral fellowship year. Emily and her husband, Jason, have settled in Chapel Hill, North Carolina.