Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Behaviour predictors of child development and parenting stress trajectories of children with autism 2006

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2006-198711.pdf
ubc_2006-198711.pdf [ 8.2MB ]
ubc_2006-198711.pdf
Metadata
JSON: 1.0054642.json
JSON-LD: 1.0054642+ld.json
RDF/XML (Pretty): 1.0054642.xml
RDF/JSON: 1.0054642+rdf.json
Turtle: 1.0054642+rdf-turtle.txt
N-Triples: 1.0054642+rdf-ntriples.txt
Citation
1.0054642.ris

Full Text

BEHAVIOUR PREDICTORS OF CHILD D E V E L O P M E N T A N D PARENTING STRESS TRAJECTORIES OF CHILDREN WITH A U T I S M by. K A R E N DOROTHY BOPP B.A. , The University of British Columbia, 1992 M.Sc., The University of British Columbia, 1995 A DISSERTATION SUBMITTED IN PARTIAL F U L F I L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Special Education) THE UNIVERSITY OF BRITISH C O L U M B I A April 2006 © Karen Dorothy Bopp, 2006 11 Abstract Little is known about the factors that predict change over time in children with autism and their parents. This research addressed the predictive relationship of six child behaviours - acting-out, sleep disturbances, eating difficulties, stereotypic behaviours, social unresponsiveness, and inattentiveness - on changes in child language, cognitive, and adaptive skills and on changes in parenting stress over 2 years. Participants were 70 young children with autism (mean age: 4;2) and their mothers. Child behaviour variables were constructed using a six step procedure. Structural equation modeling was used to explore the relationships between child behaviours and child and parenting stress trajectories from T l to T4. Results indicated that (a) high scores for acting-out behaviour at T l predicted a greater increase in the rate of change (ROC) of expressive vocabulary, expressive language, and social skills; (b) high scores for stereotypic behaviour at T l and reduced eating difficulties over 6 months predicted a greater increase in the ROC of social skills; (c) reduced stereotypic behaviour over 6 months predicted a greater increase in the ROC of cognitive skills; (d) reduced stereotypic behaviour over 1 year predicted a greater increase in the ROC of expressive language and daily living skills; (e) high scores for inattentive behaviour at T l predicted less increase in the ROC of receptive and expressive vocabulary, expressive language, and daily living skills; and (f) improvement in inattentive behaviours over 1 year predicted a greater increase in the ROC of expressive vocabulary. The results also revealed that: (a) reduced acting-out behaviour over 6 months predicted less increase in the ROC of maternal stress related to managing children's behaviour; (b) reduced sleep disturbances and stereotypic behaviour over 6 months predicted a greater increase in the ROC of overall maternal stress and specific stress related to parent competence; (c) reduced eating difficulties over 1 year predicted a greater increase in the ROC of maternal stress related to managing children's behaviour; and (d) reduced inattentive behaviour over 6 months predicted less increase in the ROC of maternal stress related to both parent competence and parent-child interactions. Interpretations of the findings, implications for treatment efficiency, and considerations for future research are discussed. IV T A B L E OF CONTENTS Abstract i i Table of Contents --iv List of Tables ••••• viii List of Figures • x Acknowledgements '. xi CHAPTER 1 •. 1 Introduction 1 The Effectiveness of Early Intervention for Children with Autism 2 Individual Child Differences and Links to Other Child Characteristics 4 Child and Family Characteristics and Parenting Stress 6 A Model of Family Stress 8 Summary 10 CHAPTER 2 12 Review of the Literature 12 Child Behaviours as Predictors of Other Child Characteristics 13 Acting-Out Behaviour: Links to Child Characteristics 13 Sleep Disturbances: Links to Child Characteristics 14 Eating Difficulties: Links to Child Characteristics 15 Stereotypic Behaviour: Links to Child Characteristics 16 Social Unresponsiveness: Links to Child Characteristics 18 Inattentiveness: Links to Child Characteristics 20 Summary of Child Behaviours as Predictors of Child Characteristics 21 V Child Behaviours as Predictors of Parenting Stress 28 Acting-Out Behaviour: Links to Parenting Stress 29 Sleep Disturbances: Links to Parenting Stress 31 Eating Difficulties: Links to Parenting Stress 32 Stereotypic Behaviour: Links to Parenting Stress 32 Social Unresponsiveness: Links to Parenting Stress 33 Inattentiveness: Links to Parenting Stress 33 Summary of Child Behaviours as Predictors of Parenting Stress 34 Negative Life Events and Coping Strategies: Influences on Parenting Stress 44 Negative Life Events 44 Coping Strategies , 44 Summary of the Research 45 Statement of the Problem and Overview of the Study • 46 Questions Related to Child Characteristics. 46 Questions Related to Parenting Stress 47 CHAPTER 3 49 Method 49 Participants 49 Measurement 50 Dependent Variables: Child Measures , 52 Dependent Variables: Parent Measures 55 Predictor Variables: Child Behaviour 57 Data Analysis 62 VI CHAPTER 4 '. •• 73 Results 73 Predictor Variable Creation: Results of Total Item Correlation Analysis 73 Descriptive Statistics 76 Descriptive statistics: Child behaviours (Tl to T3) predicting child measures 77 Descriptive statistics: Child behaviours (Tl to T3) predicting parenting stress 79 Descriptive statistics: Child measures (Tl to T4) 81 Descriptive statistics: Parenting stress (Tl to T4) 83 Unconditional Model: Changes in Child Measures and PSI-SF Over 2 Years 84 Results: S E M Analyses 86 Results for Question 1: 87 Results for Question 2: 90 Results for Question 3: 93 Results for Question 4:. 95 Results for Question 5: . 96 Results for Question 6: 98 Results for Question 7: 99 Summary of Results 101 CHAPTER 5 , 104 Discussion 104 Examining Change Over Time Using Multi-Wave Data 104 Child Behaviours as Predictors of Changes in Child Characteristics Over 2 Years... 106 Acting-Out Behaviour: Predictors of Child Characteristics 106 Vll Sleep Disturbances: Predictors of Child Characteristics 107 Eating Difficulties: Predictors of Child Characteristics 108 Stereotypic Behaviour: Predictors of Child Characteristics 109 Social Unresponsiveness: Predictors of Child Characteristics 110 Inattentiveness: Predictors of Child Characteristics 112 Child Behaviours as Predictors of Maternal Parenting Stress Over 2 Years 114 Acting-Out Behaviour: Predictors of Maternal Parenting Stress 115 Sleep Disturbances: Predictors of Maternal Parenting Stress 117 Eating Difficulties: Predictors of Maternal Parenting Stress 119 Stereotypic Behaviours: Predictors of Maternal Parenting Stress 120 Inattentiveness: Predictors of Maternal Parenting Stress 122 Limitations 123 Future Directions 125 REFERENCES 129 Appendix A: A l l Items Located for the Six Child Behaviour Variables 150 Appendix B: Item Overrepresentation in Child Behaviour Variables 167 Appendix C: Items Selected for the Six Child Behaviour Variables 187 Appendix D: Examples of LISREL Syntax and List of Restrictions. 198 VI11 LIST OF TABLES Table 2.1. Child behaviours linked to other child characteristics 22 Table 2.2. Child behaviours linked to parenting stress 35 Table 3.1. Measures collected at T l , T2, T3, and T4 for children and families 50 Table 4.1. Number of items per predictor variable and corresponding coefficient alpha.75 Table 4.2. Scores for child behaviour variables used to predict changes in child measures over 2 years (TV =70) 78 Table 4.3. Descriptive statistics for child behaviour variables used to predict changes in parenting stress over 2 years (N= 63) 80 Table 4.4. Raw scores of child measures from T l to T4 81 Table 4.5. Scores of Parenting Stress Index - Short Form (Tl to T4) 83 Table 4.6. Unconditional model of change in child measures from T l to T4 86 Table 4.7. Unconditional model of change in PSI-SF from T l to T4 87 Table 4.8. Path coefficients and (t-values) for T l child behaviours predicting rate of change (Tl to T4) of child measures 89 Table 4.9. Path coefficients and (t-values) related to difference scores (Tl - T2) for child behaviours predicting the rate of change (Tl to T4) of child measures .91 Table 4.10. Path coefficients and (t-values) for the rate of change (Tl to T3) of child behaviours predicting the rate of change (Tl to T4) in child measures 93 Table 4.11. Path coefficients and (t-values) for parent coping style and negative life events predicting the rate of change (Tl to T4) of parenting stress over 2 years 95 Table 4.12. Path coefficients and (t-values) for child behaviour variables at T l predicting the rate of change (Tl to T4) in PSI-SF scores 97 Table 4.13. Path coefficients and (t-values) related to difference scores (Tl - T2) for child behaviours predicting the rate of change (Tl to T4) of parenting stress 98 Table 4.14. Path coefficients and (t-values) for the rate of change in child behaviours (Tl to T3) predicting the rate of change in parenting stress (Tl to T4) 100 Table 4.15. Summary of predictive effects of child behaviours at T l , T l to T2, and ROC of T l to T3 on ROC of child characteristics and parenting stress from T l to T4 101 X LIST OF FIGURES Figure 3.1: Time 1 child behaviours as predictors of the slope of a child characteristic variable over 2 years 64 Figure 3.2: Change in child behaviour variable over first 6 months of intervention (T l - T2) as a predictor of the slope of a child characteristic variable over 2 years. 65 Figure 3.3: Change in a child behaviour variable over the first year of intervention (Tl to T3) as a predictor of the slope of a child characteristic variable over 2 years 66 Figure 3.4: Coping style (F-COPES) and negative life events (FILE) as predictors of the slope of parenting stress over 2 years 67 Figure 3.5: Time 1 child behaviour variables as predictors of the slope of parenting stress over 2 years, taking into account parent coping style and negative life events experienced 68 Figure 3.6: Change in child behaviour variables over the first 6 months of intervention (T1-T2) as predictors of the slope of parenting stress over 2 years, taking into account parent coping style and negative life events experienced ;..69 Figure 3.7: Change in a child behaviour variable over the first year of intervention (Tl to T3) as a predictor of the slope of parenting stress over 2 years, taking into account parent coping style and negative life events experienced 70 xi A C K N O W L E D G E M E N T S This dissertation is a culmination of my years of study, however, no words would have reached these pages without the inspiration I received from others. First and foremost, I would like to thank my advisor, Pat Mirenda. She is truly my mentor. Her knowledge, encouragement, compassion, and generous spirit have guided me to grow both as a researcher and a person. To Bruno Zumbo for the many hours he dedicated to helping me better understand the world of statistics. To Kim Schonert-Reichl for her encouragement and for demonstrating that the roles of mom and researcher are not mutually exclusive. To my colleague, Veronica Smith, whose enthusiasm is contagious and who was always there to lend an ear. To my parents, Angela and Edward Bopp, for their love, understanding, and limitless support throughout the long years of my education. To my sisters, Tricia McArthur and Linda Brothers, for believing in me. To my in-laws, Peter and Edith Matthews, who were always there to help. To the children and families who participated in this study, I thank-you for giving this research life. And finally, to my husband, Mark Matthews, and our two fabulous twin daughters, Riley and Hunter - you are my anchor and my sanctuary and without you my world would not be complete. 1 CHAPTER 1 Introduction Our understanding of autism, how to treat individuals with it, and how to support their families has come a long way since Leo Kanner first described 11 children as having inborn autistic disturbances of affective contact (Kanner, 1943). Currently, autistic disorder is one of five autism spectrum disorders (ASDs) that also include Rett syndrome, Childhood Disintegrative Disorder, Asperger's syndrome (AS), and Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS) (American Psychological Association, 2000). The symptoms of autism are evident within the first 3 years of life and include significant difficulties with social interaction; delayed or abnormal functioning in verbal and non-verbal communication; and unusual patterns of behaviour (e.g. restricted interests, repetitive activities, stereotyped movements, and/or unusual responses to sensory stimuli). It appears that the prevalence rate of ASD is on the increase. When epidemiological studies were first conducted, the rate of autism was estimated at approximately 4 to 5 cases per 10,000 births (Lotter, 1966). However, a recent study found that the prevalence rate was 58.7 per 10,000 preschool children for all pervasive developmental disorders, with autism at 22 per 10,000 and all other variants of autism at 36.7 per 10,000 (Chakrabarti & Fombonne, 2005). Today, many more children are diagnosed with autism than in the past and, as a result, more families are faced with the challenge of deciding how to best support their children's developmental progress. In both Canada and the United States, there has been much discussion, debate, and litigation regarding the effectiveness of various treatment 2 approaches, resulting in provinces such as British Columbia funding early intervention services for children with ASD. However, it is still premature to claim that one treatment approach is more effective than another (Prizant & Rubin, 1999). Instead, research must move forward to better understand the relationships among child and family characteristics, intervention approaches, and child change in order to aid families and governments to make informed decisions about how to maximize the effectiveness of scarce funding dollars. the Effectiveness of Early Intervention for Children with Autism Early intervention for children with autism has evolved substantially since the 1960s and 1970s, when researchers first demonstrated that these children are able to learn and develop through structured teaching using applied behaviour analysis (e.g. Ferster & Demyer, 1962; Jensen & Womack, 1967; Lovaas, 1974; Lovaas & Koegel, 1973; Lovaas, Koegel, Simmons, & Long, 1973). Today, there is mounting evidence that intensive early intervention that involves a range of behavioural and naturalistic approaches is effective for a substantial number of young children with autism (Woods & Wetherby, 2003). However, the outcomes for children receiving intervention continue to be variable. Some children display remarkable improvements, while others do not (Rogers, 1998). For example, Lovaas (1987) reported that, of 19 young children with autism who received up to 40 hours of intensive behavioural treatment over 2 years, 47% achieved normal intellectual and educational functioning, 40% were mildly retarded and assigned to special classes, and 10% were profoundly retarded and assigned to segregated settings at follow-up. A recent replication study in which 24 young children received an average of 38 hours per week of intensive behavioral intervention over 2 years reported similar 3. outcomes; 48% of the children achieved IQ scores in the normal range and were placed in regular, age-level classrooms, while the remaining children made less dramatic progress (Sallows & Graupner, 2005). Strain, Kohler, and Goldstien (1996) also reported that, of 51 children with autism who received 2 years of treatment in the L E A P preschool using incidental and naturalistic teaching techniques, 47% were placed in regular public school settings. Despite these and other positive outcomes, Volkmar and Pauls (2003) estimated that, even with early identification and intervention, only approximately 15% of individuals with autism can be reasonably self-sufficient as adults and another 15% to 20% can function well with periodic support. In a review of treatments for young children with autism, Schriebman (2000) noted this wide heterogeneity in the outcomes of intervention and remarked that these findings suggest that there is no "one size fits all" treatment for this population. The lack of 100% effectiveness of early intervention leads one to conclude that there are other variables affecting outcomes that have not yet been identified (Ingersoll, Schreibman, & Stahmer, 2001). Identifying these variables and understanding the process of development in autism and the factors that influence the differential outcomes observed in the literature are critical in order to improve treatment . efficiency. One approach used to identify variables that affect the outcomes of early intervention is to examine the relative effectiveness of one treatment compared to another — for example, traditional behavioural versus contemporary behaviour analytic interventions (e.g., "Lovaas therapy" versus incidental teaching). This approach, however, requires randomly assigned, matched-control samples, which is often difficult to achieve because many parents wish to choose - rather than be assigned — the method of intervention for their children (Prizant & Rubin, 1999). In addition, this approach focuses solely on treatment characteristics and ignores the individual differences among children with autism themselves. Kluth (2003) noted that autism is a complex spectrum disorder whereby no two individuals appear to have the same experiences. Thus, comparing one treatment approach to another without taking child differences into account-can be overly simplistic. An alternative approach that can be used to better understand differential outcomes is to examine the individual differences among children and how these differences affect development over time. Individual Child Differences and Links to Other Child Characteristics Pre-treatment IQ scores (Gabriels, Hi l l , Pierce, & Rogers, 2001; Harris & Handleman, 2000; Liss et al., 2001; Lovaas & Smith, 1998; Sallows & Graupner, 2005; Smith, Eikeseth, Klevstrand, & Lovaas, 1997), age at the start of intervention (Fenske, Zalenski, Krantz, & McClannahan, 1985; Harris & Handleman, 2000; Lovaas, 1987), and autism severity (DeMyer, 1973; Eaves & Ho, 1997; Liss et al., 2001) have all been found to be good predictors of cognitive, language, and adaptive behaviour outcomes of early intervention. However, these broad child characteristics are "unchangeable" and tell us little about which skills to target in order to improve treatment efficacy. Recently, researchers have taken a new direction and have examined specific "treatable" child skills that are associated with differential changes in child development. These skills include early language behaviours such as verbal imitation (Sallows & Graupner, 2005; Smith, Zaidman-Zait, & Mirenda, 2005; Stone & Yoder, 2001), synchronized behaviour with caregivers (Rollins & Snow, 1998; Siller & Sigman, 2002), early language and nonverbal skills (Szatmari, Bryson, Boyle, Streiner, & Duku, 2003), joint attention (Charman et al., 2003; Rollins & Snow, 1998; Travis, Sigman, & Ruskin, 2001; Sigman & McGovern, 2005; Smith et al., 2005), and early gestures (Brady, Marquis, Fleming, & McLean, 2004). A few longitudinal studies have examined the predictive effects of these skills on changes in child characteristics over time. For example, Rollins and Snow (1998) found that both joint attention and responsive parental interactions with children with autism at age 1;2 predicted the children's grammatical development at age 2;7. Stone and Yoder (2001) followed 35 children with autism and found that children who had stronger imitation skills at age 2 had better expressive language outcomes at age 4, over and above the effects of their initial language levels. Siller and Sigman (2002) examined 25 children with autism, 18 children with developmental disabilities, and 18 typically developing children from early childhood to the age of 16, over three time points. They found that children with autism who made significant gains in language and communication skills over 1-, 10-, and 16-year intervals had caregivers who showed higher levels of synchronized and undemanding utterances at baseline. Szatmari et al. (2003) found that early language and nonverbal skills measured at 4 to 6 years of age were important predictors of adaptive behaviour 2 years later in 47 children with autism and 21 children with Asperger's syndrome. Charman et al. (2003) found that children with autism and PDD-NOS who displayed increased responsiveness to joint attention and increased imitation at 20 months had increased levels of receptive (but not expressive) language at 42 months. Charman et al. (2005) found that the number of non-verbal communicative interactions observed at age 2 was a good predictor of language, communication, and social development at age 7 for 29 children with autism. Smith et al. (2005) found that 6 both gestures to initiate joint attention and verbal imitation skills predicted higher expressive vocabulary scores over 2 years in children with autism who received early intervention. Sallows and Graupner (2005) found that a combination of pre-treatment language age and verbal imitation, daily living, and socialization skills were highly predictive of early intervention outcomes. Sigman and McGovern (2005) found that functional play skills, responsiveness to others' bids for joint attention, and frequency of requesting behaviours in early childhood predicted language skills in adolescence for 48 children with autism followed over 16 years. Finally, in a study of children with developmental disabilities or PDD-NOS, Brady et al. (2004) found that children's communication rate, the extent to which they used proto-declarative pointing, and the amount of responsive parental interactions all predicted positive communicative development over 2 years. Despite these interesting findings, our understanding of which specific child skills are related to child change is in its infancy. It appears from recent research that treatments focused on promoting parent-child responsivity and on increasing prelinguistic communication skills (e.g., responding to and initiating gestures for joint attention, verbal imitation skills, and the rate of communicative acts) may improve the outcomes of children with autism over time. However, these may not be the only child behaviours involved in the differential changes in child development observed in children receiving early intervention. Child and Family Characteristics and Parenting Stress Intervention must be viewed within the context of the family and family outcomes, not just child development, must be examined. One family measure that has 7 received the most attention is parenting stress. Holroyd and Lazurus (1982) defined stress as occurring when "environmental and/or internal demands tax or exceed the individual's resources for managing them" (p. 22). Although most families experience stress resulting from discrete events or natural life changes over time (e.g., a parent's job loss or a child entering puberty), research has demonstrated that families of children with autism or other developmental disabilities experience higher levels of stress than families of typically developing children (Boyce & Behl, 1991). In fact, it seems that parents of children with autism endure more stress than parents of children with other disabilities (Dunn, Burbaine, Bowers, & Tantleff-Dunn, 2001). These parents tend to experience not only typical family stressors, but also a pile-up of stressors that can result from ongoing childcare demands (McCubbin & Patterson, 1982). In turn, these ongoing special care demands, along with multiple demands on family resources, can lead to significantly higher stress levels for parents and disruption of family relationships (Floyd & Gallagher, 1997; Smith, Oliver, & Innocenti, 2001). On the other hand, not all families with children with autism or other developmental disabilities experience elevated levels of stress compared to families raising children who are typically developing. In fact, some families report that having a child with a developmental disorder such as autism enhances family functioning and that the presence of a child with a disability does not always lead to decreased family well-being (Dyson, 1996; Floyd & Gallagher, 1997; VanRiper, 2001). The question then arises: Why do some families thrive while others become burdened by stress? In order to truly understand the impact that a child with autism has on parenting stress, one must examine the various processes that are related to stress within family systems. These include the impact of coping styles and negative life events 8 experienced and how both of these factors interact with specific characteristics observed in the child. A Model of FamUy Stress One model of stress and coping that examines the interactions within families systems is the Double A B C X model of family adaptation developed by McCubbin and Patterson (1983). In this model, " X , " the stress of raising a child with a disability, is influenced by " A , " the characteristics of the child and the pile-up of stressors and strains (e.g., life events); " B , " the parental perceptions of the child's disability and the parental coping style; and " C , " the family's external and internal resources and supports. These variables change over time in families, thus producing the "double" in the double A B C X model (Hodapp, Dykens, & Masino, 1997). This model is only one of many used to examine interactions within family systems and parenting stress; others include the Resiliency Model of Family Stress (McCubbin & McCubbin, 1991), the Family Caregiving Model (McDonald & Gregiore, 1997), and the parent-child interactive model of stress (Mash & Johnston, 1990). However, the Double A B C X model appears to provide a template upon which to describe the interplay of three key variables — (1) child characteristics, (2) negative life events, and (3) parent coping styles — that can influence parenting stress levels in families caring for children with autism. Child characteristics and parenting stress. Child characteristics are one of the factors that must be addressed when examining parenting stress. It goes without saying that all children with autism do not present with identical skill levels. Thus, it is more useful to examine how specific within-child characteristics make the impact of a specific disability more or less stressful for parents (Richdale, Francis, Gavidia-Payne, & Cotton, 9 2000) than to examine whether specific disabilities themselves (e.g. autism versus Down syndrome) are related to parenting stress. Several researchers have found that one of the primary sources of parenting stress is the presence of child behaviour problems (e.g., Cameron & Armstrong-Stassen, 1991; Floyd & Gallagher, 1997; McDonald & Gregoire, 1997; Richdale et al., 2000). For example, when comparing children with mental retardation to children with chronic illnesses, behaviour problems were found to be a better indicator of parenting stress than the type of disability (Floyd & Gallagher, 1997). Cameron and Armstrong-Stassen (1991) also found that child behaviour problems (e.g., harming oneself or others or interrupting the sleep of others) were significantly related to parenting stress, and that the more "extreme" the child's behaviour problem, the higher the level of stress. Other researchers have found that increased levels of behaviour problems also have indirect effects on parenting stress. Indirect effects occur when behavioural challenges limit family activities or school placements, are associated with sleep problems, or result in reduced extended family support — all of which, in turn, result in decreased coping and increased stress (Floyd & Gallagher, 1997; McDonald & Gregoire, 1997; Richdale et al., 2000). Coping style and parenting stress. Coping style is another factor that must be addressed when examining parenting stress. People cope with stress in different ways. Some people avoid or escape stressful situations entirely, while others use strategies such as emotional distancing; positive re-framing by redefining stressful events to make them . more manageable; positive or passive appraisal; accepting problem situations by minimizing their reactions; or a combination of these. The type of coping style used by parents of children with developmental disabilities has been found to correspond to 10 family outcomes. For example, Dunn et al. (2001) found that a coping style based on escape and avoidance was related to increased depression, isolation, and spousal relationship problems for parents of children with autism. They also found that parents who presented with higher levels of stress used a smaller range of coping styles and/or failed to use emotional distancing at all. Cameron and Armstrong-Stassen (1991) found that the use of positive re-framing resulted in decreased stress levels in mothers of adults with developmental delays. Thus, parent coping styles appear to have a significant impact on the stress levels experienced by parents of individuals with disabilities across the age range. In summary, the variables that are likely to affect the stress and coping process in families of children with autism appear to include child characteristics (e.g., behavioural difficulties), negative life events experienced, and internal family processes (e.g., coping style). It is these variables that practitioners who provide services to families of children with autism must take into consideration. Summary Past research has supported the effectiveness of early intervention for children with autism (Prizant & Rubin, 1999). However, little is known about the predictors of differential changes over time found for children and their parents. The present study was designed to expand our current understanding of the predictors of both child and family change by examining individual differences in the development of children with autism and their families over time. Specifically, the goal of this research was to determine the influences of six clusters of child problem behaviours - namely, acting-out behaviours, sleep disturbances, eating difficulties, stereotypic behaviours, social unresponsiveness, and inattentiveness -- on both child change and parenting stress in children with autism who received intensive behavioural intervention over a 2-year period. Chapter 2 reviews the existing literature on the relationship between child behaviours and both other child characteristics for children with autism and parenting stress for the families of these children. 12 CHAPTER 2 Review of the Literature The primary purpose of this investigation was to examine child behaviours as predictors of both changes in child characteristics and changes in parenting stress for children with autism receiving intensive behavioural intervention over 2 years. In this review, "child behaviours" refer to the following six behaviour clusters: acting-out behaviours, sleep disturbances, eating difficulties, stereotypic behaviours, social unresponsiveness, and inattentiveness. "Child characteristics" refer to children's language, cognitive, adaptive behaviour, and autism severity scores. Finally, the term "predict" refers to an influential, not a causal, relationship between two variables, in which the "predictor" variable (e.g., child behaviour) is measured prior to the outcome variable (e.g., child characteristic or parenting stress). Despite findings that imply that child behaviours may be related to other child characteristics and/or can impact parenting stress over time, very few studies have specifically, examined the sources of variability in child development and parental well- being longitudinally and within the context of early intervention (Hauser-Cram, Warfield, Erikson, Shonkoff, & Kraus, 2001). In addition, the majority of studies found in this area are descriptive or cross-sectional in nature; focus mainly on families and older children with developmental delays other than autism; and include children who did not receive early intervention services. This literature review will examine these issues in two sections. First, each of the six child behaviours used as predictors of child characteristics will be examined. Second, the same six child behaviours wil l be reviewed as predictor variables of parenting stress. These sections will then be followed by a cumulative 13 summary of the research and a statement of the current problem and overview of the study. Child Behaviours as Predictors of Other Child Characteristics Acting-Out Behaviour: Links to Child Characteristics In this review, acting-out behaviour refers to behaviours that can cause harm or. potential harm to a child and/or to the environment. Examples of acting-out behaviours include self-injury, tantrums, aggression towards others, and property destruction. Recent research findings have indicated that the presence of acting-out behaviours in children with-developmental disabilities appears to influence their cognitive development (Hauser-Cram et al., 2001). This may be because the inability to regulate one's behaviour during cognitively demanding tasks or situations places these children at a disadvantage over children who are able to do so (Bronson, 2000; Hauser-Cram et al., 2001). In addition, children with acting-out behaviours may be rejected by their peers and as a result are deprived of opportunities to practice and develop prosocial skills such as sharing and cooperating (Kaiser & Rasminsky, 2003). No studies have examined the impact of acting-out behaviours on the development of young children with autism receiving early intervention, and only one has examined this issue with children with other developmental disabilities. In this study, Hauser-Cram et al. (2001) examined acting-out behaviour as a predictor of child outcomes in a 10-year investigation of 183 young children with Down syndrome, motor impairments, or developmental delays of unknown etiology. They found that the children who were rated at 3 years of age by their teachers as having higher levels of acting-out behaviour demonstrated the least cognitive growth over the next 10 years. Thus, it 14 appears the acting-out behaviour was inversely related to cognitive growth over time. This research is the first to directly link acting-out behaviours to changes in other child characteristics in children with developmental disabilities. However, more research is needed to examine this issue for children with autism who display such behaviours. Sleep Disturbances: Links to Child Characteristics In this review, sleep disturbances refer to unusual sleep behaviours such as being often frightened by dreams; screaming during sleep and cannot be comforted; waking often and not falling back asleep; and not having a regular sleep schedule. Although the prevalence rate of sleep problems for children with autism has been estimated between 44-83% (Richdale, 1999), research examining the link between sleep disturbances and the development of other child characteristics in children with autism is in its infancy. In typically developing children, sleep problems appear to be associated with cognitive impairments; altered emotional states; inflexibility to change; reduced learning rates; and an increase in behaviours that interfere with learning, such as inattentiveness, aggression, hyperactivity, non-compliance, and irritability (Schreck, Mulick, & Smith, 2004). Wiggs and Stores (1996) assessed the sleep patterns, daytime behaviour, and challenging behaviour of 209 young children with severe learning disabilities (mean age: 10). They found that children who had severe sleep problems were more likely to be irritable, lethargic, hyperactive, and displayed more stereotypic behaviours and acting-out behaviours such as self-injury, aggression, screaming, temper tantrums, non-compliance, and impulsivity. Schreck, Mulick, et al. (2004) examined a database that included one- time parent reports of the sleep problems of 55 children with autism (mean age: 8;2) and found that sleep problems were related to various autistic symptoms and behaviours. 15 Children with autism who had increased sensitivity to stimuli in the sleeping environment, periods of screaming during the night, and/or fewer hours of sleep displayed difficulties in communication, increased stereotypic behaviour, and difficulties with social interactions. In addition, they found that the number of hours of sleep per night alone predicted the severity of autistic symptoms; children who had fewer hours of sleep had more severe symptoms, as measured by the Gilliam Autism Rating Scale (GARS: Gilliam, 1995). Hoffman et al., (2005) compared parent responses on the Children's Sleep Habits Questionnaire (CSHQ) from 80 parents of young children with autism (mean age: 8;2) with their children's GARS scores. They found that sleep disordered breathing was the best predictor of stereotypic behaviour, social interactions, and overall level of autism. In addition, they found that parasomnias such as sleep apnea, nightmares, and sleep walking were predictive of level of developmental disturbances and overall level of autism. These studies link sleep problems to child characteristics in the areas of communication, stereotypic behaviour, acting-out behaviour, social interactions, and autism symptomatology. Eating Difficulties: Links to Child Characteristics In this review, eating difficulties refer to unusual food intake behaviours such as eating a limited variety of foods, gagging, and/or having strong preferences for certain tastes and smells. It has long been noted that many children with autism present with unusual eating patterns and feeding difficulties (Archer & Szatmari, 1991; Cornish, 1998; Gray, 1994; Schreck, Williams, & Smith, 2004; Williams, Dalrymple, & Neal, 2000). Reports of the prevalence of eating difficulties in children with autism range from 42% (Archer & Szatmari, 1991) to 94% (De Meyer, 1979); in general, children with autism 16 exhibit more eating and mealtime problems than typically developing children (Schreck, Williams, et al., 2004). Such problems include eating a restricted range of foods, refusing foods, requiring specific presentations of foods, eating only specific textures, and using only specific utensils (Schreck, Williams, et al., 2004). Archer and Szatmari (1991) are the only authors to date who have examined the relationship between eating difficulties and other child characteristics in children with autism. They investigated children between 5 and 6 years of age who were typically developing, had autism, or were identified with specific eating problems. They found that 42% of the children with autism presented with eating and mealtime problems. For these children, there was a significant correlation between their total eating problem score and their total autism severity score, as measured by the Autism Behavior Checklist (ABC; Krug, Arick, & Almond, 1980). In other words, the children who had greater eating problems had more severe autism. Further analyses also revealed significant correlations between children's eating problems and their A B C scores on the subscales for relating, language, socialization, and self-help skills. These findings suggest that the eating problems present in many children with autism may be linked to other child characteristics. Stereotypic Behaviour: Links to Child Characteristics Stereotypic behaviours refer to repetitive behaviours that involve abnormal movements of the body such as rocking back and forth, flapping the arms, spinning around, squinting the eyes, and/or inappropriately touching items or people. They also include strong reactions to change and/or becoming involved in complicated rituals such as lining things up. Considerable research has focused on the functions of and 1 7 interventions related to stereotypic behaviours exhibited by children with autism (e.g., . Goldstein, 2002; Horner, Carr, Strain, Todd, & Reed, 2002; Polirstock, Rovert, Lawrence, & Serifino, 2003; Turner, 1999). However, only a few studies have examined the direct relationship between stereotypic behaviours and the development of other child characteristics. Epstein, Taubman, and Lovaas (1985) explored changes in stereotypic behaviours in six young children with autism who received intensive behavioural treatment over 2 years. They found that the children whose stereotypic behaviours changed most dramatically from "low-level" motor behaviours (e.g., rocking, spinning, or twirling) to "high-level" stereotypies (e.g., lining up objects or preoccupations) also had the best academic outcomes 2 years later, as measured by school placement. Dadds, Schwartz, Adams, and Rose (1988) examined stereotypic behaviours in 12 children with autism (mean age: 8;6) and found that those with the most stereotypic behaviours had lower levels of personal contact and lower language ages. Campbell et al. (1990) examined the stereotypic movements of 224 children with autism (mean age: 4;7) and found that children who had more stereotypic behaviours also had lower IQ scores and higher autism severity scores, as measured by the Clinical Global Impressions severity of illness item (CGI; National Institute of Mental Health, 1985). Venter, Lord, and Schopler (1992) followed 58 young children with high-functioning autism from preschool to school age over a period of 8 years. They found that, among other things, parental reports of restricted, repetitive behaviours in preschool predicted adaptive behavior and achievement scores 8 years later. Bodfish, Symons, Parker, and Lewis (2000) examined the rate of occurrence of repetitive behaviours in 32 adults with autism to see i f rate predicted overall autism severity. In their correlational analyses, they found that severity 18 scores for stereotypy, self-injury, compulsions, dyskinesia (i.e., repetitive, involuntary movements), and akathisia (i.e., repetitive, restless movements such as pacing) significantly predicted autism severity scores, as measured by the Autism Behavior Checklist. In other words, adults with a higher number of stereotypic behaviours had more severe autism. Finally, Gabriels, Cuccaro, Hi l l , Ivers, and Gbldsoh (2005) examined the relationship between repetitive behaviours and cognitive and adaptive . functioning levels in 14 children with autism (mean age: 10;7). They found that children with lower cognitive and communicative abilities had significantly more stereotypic behaviours when non-verbal IQ was not used as a covariate. Social Unresponsiveness: Links to Child Characteristics Another set of child behaviours that may be linked to other child characteristics is social unresponsiveness, which is defined as a decreased "capacity to engage in social exchanges in ways that 'invite' others to reciprocate" (Dunst, Trivette, Hamby, & Pollock, 1990, p. 208). Socially unresponsive behaviours include, for example, lack of smiling in response to others and/or lack of responsiveness to verbal or physical overtures. By definition, children with autism exhibit such socially unresponsive behaviours. For example, Sigman, Kasari, and Kwon, and Yirmiya (1992) found that young children with autism failed to respond when an adult displayed fear or distress, either verbally or through body language. Only a few studies, however, have examined the link between social unresponsiveness and other child characteristics. Dunst et al. (1990) examined children with mental retardation, physical impairments, or developmental delays (mean age: 2;4) and found that social responsiveness accounted for 34% of the variance in child progress over 1 year, as measured by mental age. 19 Dissanayake, Sigman, and Kasari (1996) followed young children with autism (mean age: 3;6) over 5 years and assessed their responsiveness to an adult's display of distress (i.e., an adult "hurting" her ankle and crying in pain). They found that, at each assessment time point, children's ability to respond to the adult was positively associated with cognitive level - that is, children who were more responsive had higher cognitive skills. Sigman and Ruskin (1999) examined children with autism, Down syndrome, or developmental delays between the ages of 10 and 13 and found that prelinguistic communication skills such as joint attention behaviours, the ability to engage in social exchanges, and social responsiveness to others were all predictors of peer engagement in children with autism over time. Finally, Beadle-Brown, and Murphy (2005) followed 91 individuals with intellectual disabilities, including autism, over 3 time points expanding 25 years (Tl - mean age: 8;11; T2 - mean age: 20;11; T3 - mean age: 34). They found that there was little change in social impairments over time as measured by the Schedule of Handicaps, Behaviors, and Skills. In addition, low levels of social impairment at T3, especially participants who were described as "aloof (i.e., indifferent or interacting only to obtain needs), more challenging behaviour at T3, and low IQ at T l predicted 81% of the variance in participants with poor or very poor outcomes (note: "outcome" was calculated from mean scores on the Schedule of Handicaps, Behaviors, and Skills including communication and social skills, work situation, residential placement, and a quality of life measure). From these findings, it appears that a lack of social responsiveness may be a predictor of delayed development of cognitive, daily living, and social skills for children with autism. 20 Inattentiveness: Links to Child Characteristics A final child behaviour that may be linked to other characteristics in children with autism is inattentiveness. In this review, inattentiveness refers to children's inability to make eye contact; maintain attention to others or activities; or attend to sudden changes in their environments, such as persons entering the room or loud noises. A few studies have linked the occurrence of inattentive behaviours with other child characteristics. Yarrow et al. (1983) found that visual attention was significantly correlated with developmental competence as measured by the Bayley Scales of Infant Development (BSID; Bayley, 1969) and the Mental Development Index (MDI) for typically developing 6-month-old infants. Edelson, Schubert, and Edelson (1998) found that, among other things, the presence of attention problems in 393 individuals with autism ranging in age from 4 to 41 years was predictive of their scores on the Test of Nonverbal Intelligence-Second Edition (TONI-2; Brown, Sherbenou, & Johnsen, 1997). In particular, they found that the participants who had a greater number of attention problems as measured by the Conner's Rating Scales (CRS; Goyette, Conners, & Ulrich, 1978) were also more likely to be untestable on the TONI-2. Hauser-Cram et al. (2001) found that mastery motivation (i.e., the ability to focus and persist in order to master a problem, skill, or task that is moderately challenging) was a predictor of change in both cognitive and adaptive behaviour skills in children with Down syndrome, motor impairments, or general developmental delays. Children with motor impairments or developmental delays who demonstrated higher levels of mastery motivation at the age of 3 also displayed more cognitive growth and better daily living skills over 10 years. 21 Summary of Child Behaviours as Predictors of Child Characteristics Table 2.1 summarizes the studies reviewed in this section regarding the relationship between child behaviours and other child characteristics. When reviewing this literature, three key issues emerge. First, the majority of studies found in this area are descriptive or correlational in nature (Archer & Szatmari, 1991; Bodfish et al., 2000; Campbell et al., 1990; Dadds et al., 1988; Dissanayake et al., 1996; Edelson et al., 1998; Schreck, Mulick, et al., 2004). Unfortunately, descriptive or correlational designs do not allow one to truly understand the sources of variability in child development over time. Thus, such designs do not help to identify which intervention strategies may be required early in treatment to promote positive changes in other child characteristics for young children with autism (Hauser-Cram et al., 2001). In addition, only half of the studies reviewed in Table 2.1 include children who were under the age of 6 (Archer & Szatmari, 1991; Campbell et al., 1990; Dissanayake et a l , 1996; Dunst et al., 1990; Epstein et a l , 1985; Hauser-Cram et a l , 2001; Sigman & Ruskin, 1999). Finally, the authors of only six of the investigations in Table 2.1 specified that participants were receiving or had received some type of early intervention services (Archer & Szatmari, 1991; Campbell et al., 1990; Dadds et al., 1988; Dunst et al., 1990; Epstein et al., 1985; Hauser-Cram et al., 2001). It is crucial that we begin to investigate how early intervention affects changes in child behaviours and how these changes, in turn, affect developmental trajectories in cognitive, language/communication, adaptive skill, and other areas. Knowing which child behaviours best influence positive child change might enable professionals to focus their interventions more precisely and thus improve the overall effectiveness of early intervention services. Table 2.1: Child behaviours linked to other child characteristics Study Child behaviours Participants Method used Results EI? Type? Hauser- Acting-out behaviour (e.g., 183 children: Longitudinal: Acting-out Yes Cram et al. apathy/withdrawal or Five time points: behaviour @ (2001) angry/defiant) 32.8% diagnosed with Down T3 predicted Community- syndrome and < 12 months of T l = entry into cognitive based EI Inattentiveness/mastery age @T1 early intervention growth programs motivation (i.e., ability to 39.3% diagnosed with motor T2 = 1 year later until age 3 focus and persist when impairment and < 24 months of T3 = Age 3 Inattentiveness/ problem-solving) age@Tl T4 = Age 5 mastery 27.9%o diagnosed with T5 = Age 10 motivation developmental delay and < 24 @T3 predicted months of age @T1 cognitive growth and daily living skills Schreck, Sleep problems (e.g., 55 children with autism Correlational Sleep problems Not Mulick, et increased sensitivity to between 5 and 12 years of age were related to specified al. (2004) stimuli in the sleeping (Mean age: 8;2) One time point communication environment/fewer hours of problems, School age sleep/periods of screaming stereotypic children during the night) behaviour, and attending social variety of interaction programs difficulties Table continues Table 2.1 (Continued) Study Child behaviours Participants Method used Results EI? Type? Archer & Eating difficulties (e.g., Szatmari eating a restricted range of (1991) foods/refusing foods/ requiring specific presentation or textures/using specific utensils) Epstein et Stereotypic behaviours (e.g., al. (1985) rocking, spinning, twirling, feeling/licking textures, vocal repetition, lining-up objects, repetitive actions with objects, etc.) 33 children with autism (Mean age = 5;3) 295 typically developing children (Mean age = 5; 8) 11 children with identified eating problems (Mean age = 6;2) 6 children (5 males and 1 female) with autism ranging in age from 5 to 9 Cross-sectional Correlational One time point Longitudinal Stereotypy measured monthly over 2 years Eating difficulties were related to autism severity (specifically relating, language, socialization, and self-help skills) Children who had the greatest improvement in their stereotypic behaviour also had the best achievement level scores Yes Preschool programs for children with autism Yes A l l had received treatment for more than 2 years in the U C L A Young Autism Project Table continues Table 2.1 (Continued) Study Child behaviours Participants Method used Results EI? Type? Dadds et al. Stereotypic behaviour (e.g., 12 children with autism (Mean Correlational Children who had Yes (1988) rocking, hand flapping, head age = 8;6) the most weaving, spinning, trapping One time point stereotypic Attended a objects, and twiddling small behaviours also therapy items/leaves/sand) had lower levels centre of personal contact and lower language ages Campbell Stereotypic behaviour 224 children with autism with a Correlational Children who had Yes et al. measured by the Children's mean age of 4;7 a greater number (1990) Psychiatric Rating Scale, the One time point of stereotypic A l l children Abnormal Voluntary behaviours also were Movement Scale, the had lower IQ inpatients at Abbreviated Dsykinesia scores and higher a hospital Rating Scale, and the Timed autism severity - Stereotypies Rating Scale scores Table continues Table 2.1 (Continued) Study Child behaviours Participants Method used Results EI? Type? Venter et Repetitive stereotypic 58 children with high- Longitudinal Children with N/A: all al. (1992) behaviours as measured by functioning autism (Mean age more repetitive children parental report on the = 14;7 at Time 1) Two time points behaviours in over 6 years Autism Diagnostic Interview . over 8 years preschool had at Time 1 (ADI) poorer adaptive behaviour and academic achievement at school age Bodfish et Stereotypic behaviours (e.g., 32 adults with autism Correlational Stereotypic N/A: adults al. (2000) compulsions, dyskinesia and behaviour was akathisia) 34 adults with mental One time point related to autism retardation severity Dunst et al. Social responsiveness (i.e., 47 children with mental Retrospective Social Yes (1990) engaging in social exchanges retardation, physical responsiveness in a way that invites others impairment, or developmental Two time points accounted for Weekly/bi- to participate) delay (Mean age = 2;4) for mental age: 34% of the weekly One at time of variance in child home-based study and a progress over 1 early second, 1 year year as measured intervention earlier by mental age Table continues Table 2.1 (Continued) Study Child behaviours Participants Method used Results EI? Type? Dissanayake Social responsiveness (i.e., et al. (1996) a child's responsiveness to an adult's display of distress) Sigman & Social responsiveness (i.e., Ruskin early nonverbal (1999) communication such as attention, social exchanges, and inviting others to reciprocate) 30 children with autism (Mean age = 3;6 at Time 1) 51 children with autism (Mean age @T1=3;11) 71 children with Down syndrome (Mean age @T1 = 2;7) 33 children with developmental delay(Mean age @T1=3;9) A l l children were between 10 and 13 years of age @ follow- up Correlational at each of three time points over 5 years Longitudinal T l and follow-up The children who were more responsive had higher cognitive skills at each time point Social responsiveness predicted peer engagement over time Not Specified Not Specified At Time 1, half of the children with autism were inpatients for 3-6 months on a psychiatric ward Table continues Table 2.1 (Continued) Study Child behaviours Participants Method used Results EI? Type? Edelson et Inattentiveness as measured 393 individuals with autism Correlational Participants who N/A: all al. (1998) by Conner's Rating Scales (Mean age = 12;0) One time point had a greater number of attention problems were also more likely to be untestable on the TONI-2 children were over 6 years of age N / A = Not applicable 28 Child Behaviours as Predictors of Parenting Stress Research has demonstrated that, in order to maximize the developmental gains of an individual child, one must consider that child as a part of the family system in which he or she lives (e.g., Floyd & Gallagher, 1997). As noted in Chapter 1, one dimension of the family system that has received considerable research attention is parenting stress. Many authors have noted that professionals must have a better understanding of parenting stress and how it impacts the caregiving experience, in order to enhance child outcomes (Boyce & Behl, 1991; McDonald & Gregoire, 1997). Thus, it is important to examine factors that may be related to changes in parenting stress over time; these include, for example, the impact of children's problem behaviour, family coping style, and negative life events, especially in families raising a child with a developmental disability (Weiss & Diamond, 2005). A variety of child behaviours have been linked to parenting stress in families with children with autism and other developmental disabilities. Stress can be defined as "the condition that results when person/environment transactions lead the individual to perceive a discrepancy between the demands of the situation and his/her resources or ability to cope" (Quine & Pahl, 1991, p. 57). Parenting stress is typically characterized by two components: Child-related stress, defined as "the parent's satisfaction.with and adaptation to the child's temperament and behavioural characteristics;" and parent-related stress, defined as "the parent's own emotional resources and adjustment to the parental role" (Hauser-Cram et al., 2001, p. 16). In the following sections, research on the relationship between child behaviours and parenting stress will be reviewed, along with a summary of the impact of coping strategies and life events on parenting stress. 29 Acting-Out Behaviour: Links to Parenting Stress It appears that acting-out behaviour is strongly related to parenting stress regardless of a child's disability. For example, in a comparison of families of children with mental retardation and those with chronic illnesses or emotional impairments, Floyd and Gallagher (1997) found that acting-out behaviours were more important than the nature of the child's disability in determining parenting stress. Baker et al. (2003) compared parents of children with developmental delays and parents of typically developing children (mean age: 3;0) and found that the children's acting-out behaviours predicted parenting stress levels but their developmental status did not. Quine and Pahl (1985) used a correlational design to examine the sources of stress in 200 families with children with mental handicaps. They found that the more acting-out behaviours (i.e., temper tantrums, destructiveness, spitting, biting, screaming, etc.) the children displayed, the more stress their mothers experienced. Similarly, Hodapp et al. (1997) found that children diagnosed with Prader-Willi syndrome (mean age: 10;3) who presented with high levels of acting-out behaviours also had parents who presented with higher levels of stress. Specifically, they found that children's behaviour problems accounted for 25% of the variance related to overall family stress. Cameron and Orr (1989) examined 84 families caring for children with developmental delays (ages 5 to 21) and divided the families into low, moderate, and high stress groups. They found that families in the high stress group had children who displayed more "irritating" behaviours (e.g., crying, whining, hanging on to parents) and aggressive behaviours (e.g., harming others, harming oneself, destroying property) than those in the other two groups. Cameron and Armstrong-Stassen (1991) found that acting- out behaviours such as self-injurious behaviour and aggression were significantly related 30 to parenting stress in 147 mothers of adult children with developmental delays; again the more "extreme" the behaviour problems, the higher the level of parenting stress. Finally, Ross and Blanc (1998) explored parenting stress among mothers of 92 young children (mean age: 4;6) with Attention Deficit Hyperactive Disorder (ADHD), Oppositional Defiant Disorder (ODD) and/or Conduct Disorder (CD). Some children had A D H D or ODD, while others had multiple diagnoses such as A D H D and ODD or all three disorders. They found higher levels of parenting stress in the mothers of multiply- diagnosed children and in those whose children had more acting-out behaviours, regardless of diagnosis. Finally, Hassell and McDonald (2005) found that acting-out behaviours, parental locus of control, and parenting satisfaction accounted for 59% of the variance of parenting stress in a correlational study which included 46 mothers of children with intellectual disabilities (mean age: 9;3). The majority of the research linking child behaviours and parenting stress has been descriptive or cross-sectional in nature. Hauser-Cram et al. (2001) were the first to examine the sources of variability in the stress experienced by parents of children with developmental disabilities in a longitudinal investigation. They found that children who displayed a higher number of acting-out behaviours upon exiting early intervention programs at age 3 had parents who experienced higher rates of child-related stress over 10 years. Their, findings also revealed differential predictors of stress outcomes between mothers and fathers. For mothers, a child's disability (i.e., Down syndrome, motor impairment, or developmental delay) was a significant predictor of changes in stress; while for fathers, a child's gender and the quantity of maternal-child interactions as measured by the Nursing Child Assessment Teaching Scale (NCATS; Barnard, 1978) were significant predictors of changes in stress over time. 31 Other researchers have also noted that maternal versus paternal stress appears to be predicted by different child behaviours. Konstantareas and Homatidis (1989) found that, while children's self-injurious behaviour was the best predictor of stress for both parents, children's inability to speak was a secondary stressor for fathers while children's preoccupations (e.g., inappropriately smelling, licking, and rubbing objects) was a secondary stressor for mothers. Hastings (2003) found that mothers of young children with autism had stress levels that were associated mainly with acting-out behaviours, while fathers displayed no such association. Weiss, Sullivan, and Diamond (2003) found that, for mothers of mostly adult children with developmental disabilities, acting-out behaviours accounted for a significant proportion of the variance in child-related stress; while for fathers, personal and social responsibility (i.e., a child's ability to function independently in relationships) was a better predictor of stress. Finally, Saloviita, Italinna, and Leinonen (2003) found that how mothers and fathers of children with intellectual disabilities defined their situations differentially predicted parenting stress. Mothers who defined their children's disabilities in terms of acting-out behaviours were more stressed, while fathers who defined their children's disabilities in terms of decreased social acceptance had higher levels of stress. Sleep Disturbances: Links to Parenting Stress Four studies have specifically examined the relationship between child sleep problems and parenting stress. Quine and Pahl (1985) analyzed the relative importance of several different child behaviours on parenting stress in 200 families of children with severe handicaps. They found a significant positive relationship between the number of night-time disturbances and parenting stress levels. They also found that parents who 32 reported not getting enough sleep had significantly higher levels of stress than those who were not sleep deprived. Similarly, Quine and Pahl (1991) found that higher scores on an index of sleep problems related significantly to higher stress scores in mothers of 162 children with severe learning difficulties. Finally, Richdale et al. (2000) found that night- waking was associated with increased stress in families of children with intellectual disabilities (mean age: 7;7). They also found that children with sleep onset and maintenance difficulties displayed more acting-out behaviours during the day, which in turn led to increases in parenting stress. Eating Difficulties: Links to Parenting Stress Only two studies to date have examined the influence of eating difficulties on parenting stress. Archer and Szatmari (1991) compared young children with autism and typically developing children both with and without eating problems and found a significant positive correlation between parenting stress and eating difficulties. In a qualitative study of 33 parents of children with autism (ages 6 to 12), Gray (1994) found that unusual child eating behaviours such as insisting food be presented on a specific plate appeared to cause considerable stress for many parents. Stereotypic Behaviour: Links to Parenting Stress A few studies have examined the link between stereotypic behaviour and parenting stress. Beckman (1983) found that the frequency of repetitive behaviours (e.g., head banging, finger flicking, etc.) displayed by infants with developmental delays (mean age: 1;10) was significantly associated with increased maternal stress. Stores, Fellows, and Buckley (1998) found that the amount of stereotypic behaviour as measured on the stereotypies subscale of the Aberrant Behaviour Checklist (ABC; Aman, Singh, Stewart, 33 & Field, 1985) was correlated with higher levels of maternal stress in 91 mothers of children with intellectual disabilities (mean age: 10;0). Stoddart (2003) found that child stereotypic behaviours accounted for a significant proportion of the variance in stress levels for 110 mothers of children with PDD (ages 6 to 18). Finally, Gabriels, Cuccaro, Hil l , Ivers, and Goldson (2005) found that parenting stress and child repetitive behaviours as measured by the Repetitive Behavior Scales-Revised (RBS-R; Bodfish, Symons, Parker, & Lewis, 2000) were highly positively correlated in 14 children with autism (mean age 10;7) and their parents. In general, these studies found that children who had higher levels of stereotypic behaviours had parents who were more stressed. Social Unresponsiveness: Links to Parenting Stress Researchers have also found a link between children's social responsiveness and parenting stress. Beckman (1983) found that how infants responded to various forms of social stimulation (e.g., via smiling, laughing, and responding to gestures from others) was significantly associated with maternal stress. Specifically, mothers who had infants who were less socially responsive were more stressed. Weiss et al. (2003) found that what they termed "personal and social responsibility" (i.e., children's ability to use interpersonal skills, socially respond to others, and function independently in relationships) accounted for the majority of variance (34%) in paternal child-related stress for fathers of children with developmental disabilities (ages 9 to 42). Inattentiveness: Links to Parenting Stress It appears that no studies have specifically examined inattentiveness and caregiver stress in families with children with autism. However, one study was found that linked inattentive behaviour to parenting stress in families of children with other developmental 34 disabilities. Podolski and Nigg (2001) found that, for 56 mothers of children with Attention Deficit Disorder (ADD) or A D H D (mean age: 10;2), child inattention (as assessed via the DSM-IV A D H D Inattention symptom list) was positively associated with stress related to the parenting role. In addition, they found that child inattentiveness had a unique association with maternal stress even when controlling for acting-out behaviours. Thus, inattentiveness alone may have a significant impact on parenting stress. Summary of Child Behaviours as Predictors of Parenting Stress Table 2.2 summarizes the studies linking child behaviours to parenting stress. The same three issues emerge in this body of work that were found in the research examining the links between child behaviours and other child characteristics. First, only two of the studies involved a longitudinal design (i.e., Baker et al., 2003: Hauser-Cram et al., 2001); thus, it is difficult to examine the sources of variability in parenting stress over time. Second, only six of the studies involved children on the autism spectrum (Archer & Szatmari, 1991; Gabriels et al., 2005; Gray, 1994; Hastings, 2003; Konstantareas & Homatidis, 1989; Stoddart, 2003) and only five included children with a mean age less than 6 years (Archer & Szatmari, 1991; Baker et al., 2003; Beckman, 1983; Hauser-Cram et al., 2001; Ross & Blanc, 1998). Finally, only four of the studies included children identified as receiving some type of early intervention. Thus, the relationship between changes in child behaviours in children receiving early intervention and changes in parent stress is unclear at the present time. Table 2.2: Child behaviours linked to parenting stress (EI = early intervention) Study Child behaviour Participants Method used Results EI? Type? Floyd & Acting-out behaviour 112 families of children Correlational Acting-out behaviours were more Not Gallagher (measured with the with mental retardation predictive of increased parenting specified (1997) Child Behavior (ages 6 to 18) One time stress than the nature of the child's Checklist, CBCL) point disability School aged 119 families of children children with chronic illness or enrolled in emotional impairment special (ages 6 to 18) education classes Baker et al. Acting-out behaviour 82 families of children Longitudinal Acting-out behaviours were Not (2003)' (measured with the with a developmental associated with higher parenting specified CBCL) delay Time 1 @ 3 stress but the presence of a 123 families of typical years developmental delay was not children Time 2 @ 4 (Mean age = 3;0) years Quine & Pahl Acting-out behaviour 200 families of children Correlational Acting-out behaviours were related Not (1985) (temper with mental handicaps to increased parenting stress specified tantrums/destructive- (no age provided) One time ness/spitting/biting/ point screaming, etc.) Table continues Of) Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Cameron & Orr(1989) Cameron & Armstrong- Stassen (1991) Ross & Blanc (1998) Acting-out behaviour (irritating and aggressive behaviours) Acting-out behaviour (harming self/other, interrupting the sleep of others) Acting-out behaviour (measured with the CBCL) 84 families of children with developmental delays (ages 5 to 21) 147 mothers of adult children with developmental delays 43 children with A D H D or ODD 39 children with A D H D and ODD 10 children with A D H D , ODD and CD (Mean age = 4;6) Correlational One time point Correlational One time point Cross- sectional Correlational One time point Acting-out behaviours and the number of handicapping conditions accounted for 50% of child-related stress for the parents Acting-out behaviours were related to increased parenting stress Mothers who coped by reframing and redefining reported significantly less stress Acting-out behaviours were related to increased parenting stress regardless of child diagnosis Not specified; school-aged children in a range of educational settings N/A: Adults Not specified Table continues ON Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Hodapp et al. (1997) Hauser-Cram etal. (2001) Acting-out behaviour (measured on the CBCL) Acting-out behaviour (measured with the C B C L and Classroom Problem Checklist) 42 children with Correlational Prader-Willi syndrome (Mean age = 10;4) One time point 60 children with Down Longitudinal: syndrome (age: < 12 T l = pre-EI months @ Tl ) T2 = 1 year 72 with motor later impairment (age < 24 T3 = age 3 months @ T l ) T4 = age 5 51 with developmental T5 = age 10 delay (age < 24 months @T1) Acting-out behaviours accounted for 25% of the variance in overall family stress Acting-out behaviours at age 3 years predicted child-related stress for parents over time N/A: school aged children Yes Enrolled in community- based EI programs until age 3 Table continues ~~0 Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? McDonald & Gregiore (1997) Johnston et al. (2003) Hodapp et al. (1998) Acting-out behaviour (measured with the C B C L Acting-out behaviour (as measured with the CBCL) Acting-out behaviour (as measured with the CBCL) 259 caregivers of children with atypical development or emotional disorder (ages 4 to 12) 75 families of children with Fragile X syndrome (Mean age = 10;9) 36 parents of children diagnosed with Smith- Magenis syndrome (Mean age = 8;4) Correlational One time point Correlational One time point Correlational One time point More acting-out behaviours, higher SES, and less use of positive coping strategies were all associated with increased caregiver stress Acting-out behaviours were associated with decreased family support which was associated with reduced coping strategies and an increase in caregiver stress More acting-out behaviours and less family support were associated with higher parental stress levels Acting-out behaviours were associated with increased parenting stress Not specified Not specified N/A: children were school-aged Table continues oo Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Konstantareas & Homatidis (1989) Hastings (2003) Soloviita, Italinna, & Leinonen (2003) Acting-out/self- injurious behaviour Stereotypic preoccupations (e.g., smelling, licking, and rubbing objects) Acting-out behaviour (as measured by the teacher report version of the Developmental Behavior Checklist) Acting-out behaviour 44 families of children with autism between ages 2 to 12 (Mean age = 6;10) 18 couples who were parents of children with autism (Mean age = i i ;8) Mothers (n= 116) and fathers (n = 120) of 236 children with intellectual disability (ages 1 to 10) Correlational One time point Correlational One time point Correlational One time point Table continues Self-injurious behaviour was the No best predictor of stress for both parents Children's preoccupations was a secondary stressor for mothers Children's inability to speak was a secondary stressor for fathers Acting-out behaviours were N/A: associated with higher maternal school-aged stress; no association between acting-out behaviours and paternal stress Mothers' definition of the child's Not disability as a burden or catastrophe specified was associated with difficulties controlling acting-out behaviour Fathers' definition of the child's disability as a burden or catastrophe was associated with his perceptions . regarding the child's social acceptance by others Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Quine & Pahl (1985) Quine & Pahl (1991) Richdale et al. (2000) Sleep disturbances (as measured by the number of night-time disturbances & parental report of "getting enough sleep") Sleep disturbances (as measured with an index of sleep problems) Sleep disturbances (i.e., sleep onset and maintenance and night-waking) 200 families of children with mental handicaps 166 mothers of children with severe learning difficulties 52 children with intellectual disabilities (Mean age = 7;7) 25 typically developing children (Mean age = 8;0) Correlational One time point Correlational One time point Cross- sectional Correlational One time point Sleep disturbances were associated with increased parenting stress Sleep disturbances were related to increased maternal stress scores Mothers whose coping strategy included positive acceptance of their child had lower stress scores Sleep disturbances were related to day-time behaviour problems which were associated with increased parenting stress Not specified Not specified N/A:school aged children Table continues Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Archer & Eating difficulties Szatmari (i.e., eating a (1991) restricted range of foods, refusing foods, requiring specific presentation or textures, and using specific utensils) Gray (1994) Eating difficulties (i.e., obsessive behaviour associated with food such as insisting to use the same plate) Beckman Stereotypic and (1983) repetitive behaviour (measured with the Carolina Record of Infant Behavior) 33 children with autism (Mean age = 5;3) 295 typically developing children (Mean age = 5; 8) 11 typical children with eating problems (Mean age = 6;2) 33 parents of children with autism (ages 6 to 12) 31 infants with developmental delays (Mean age = 1; 10) and their mothers Cross- sectional Correlational One time point Qualitative Correlational One time point Eating difficulties were related to increased parenting stress Eating difficulties appeared to cause much stress for parents Stereotypic repetitive behaviour was significantly associated with increased maternal stress Yes Preschool programs for children with autism N/A:school aged children Yes Minimum of 3 months of EI pre-study Table continues Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Stores, Fellows, & Buckley (1998) Stoddart (2003) Gabriels, Cuccaro, Hi l l , Ivers, & Goldson (2005) Beckman (1983) Stereotypic behaviour (as measured with the Stereotypies subscale of the Aberrant Behaviour Checklist) Stereotypic behaviours Stereotypic and repetitive behaviours (as measured by the Repetitive Behavior Scales-Revised) Social responsiveness (i.e., responses to social stimulation such as smiling, laughing, and gestures from others) 71 mothers of children with intellectual disabilities (Mean age = 10;0) 110 mothers of children with PDD (ages 6 to 18) 14 children with autism (Mean age = 10; 7) and their parents 31 infants (Mean age = 1;10) with developmental delays and their mothers Correlational One time point Correlational One time point Correlational One time point Correlational One time point Stereotypic behaviour was correlated with an increase in maternal stress Stereotypic behaviour and social phobia accounted for a 38% of the variance in maternal stress Stereotypic repetitive behaviours and parenting stress and were positively correlated Social responsiveness was significantly associated with increased maternal stress N/A:school aged N/A:school aged N/A: school aged Yes Minimum of 3 months of EI prior to the study Table continues 4̂ Table 2.2 (Continued) Study Child behaviour Participants Method used Results EI? Type? Weiss et al. (2003) Social responsiveness (i.e., ability to use interpersonal skills, socially respond to others, and function independently in relationships) 97 children with developmental disabilities (ages 9 to 42) and their parents Correlational One time point Social responsiveness accounted for 34% of the variance in child-related stress for fathers N/A:school aged and adults Podolski & Nigg (2001) Inattentiveness (as measured with the DSM-IV A D H D inattention symptom list) 56 mothers of children with Attention Deficit Disorder (ADD) or A D H D (Mean age = 10;2) Correlational One time point Inattentiveness was positively associated with stress relating to the parenting role N/A: school aged N / A = Not applicable 44 Negative Life Events and Coping Strategies: Influences on Parenting Stress Based on past research, it is unlikely that child behaviours alone can account for 100% of the variance in parenting stress over time. Research suggests that the relationship between child behaviours and parenting stress may change in the context of other variables such as negative life events and parent coping strategies. Negative Life Events Boyce and Behl (1991) found that families who experienced stressful life events (e.g., job loss, divorce, moving), reported more child-related stress but not more parent- related stress. They hypothesized that the increase in child-related stress was due to an interaction between child characteristics and life events. If Boyce and Behl (1991) are correct, one must account for life events that occur over a time period under investigation in order to truly understand the impact of child behaviours on parenting stress (Hauser- Cram et al., 2001). Coping Strategies Families' use of positive coping strategies (i.e., the actions they take to reduce stress) is another variable that must be taken into consideration when examining parenting stress. Hastings and Johnson (2001) found that lower parenting stress was related to, among other things, the use of reframing as a coping strategy (i.e., redefining events positively to make them more manageable) by 141 parents of young children with autism (mean age: 5;0). Cameron and Armstrong-Stassen (1991) found that mothers of adult children with developmental delays who used reframing reported significantly less stress than mothers who did not utilize this strategy. Quine and Pahl (1991) found that the 45 use of positive acceptance of the child as a coping strategy was related to reduced stress in mothers of children with learning difficulties. Finally, McDonald and Gregoire (1997) found an interaction between child external behaviour problems, family socio-economic status (SES), social support, and caregiver coping behaviours in 259 caregivers of children with atypical development or serious emotional disorders (ages 4 to 12). Specifically, they found that, when child external behaviour problems increased, relative and family support decreased, resulting in a reduction in parents' use of positive coping strategies and an increase in parenting stress. Summary of the Research The literature examining child behaviours as predictors of other child characteristics and parenting stress reveals four central themes that lead one to conclude that these areas of investigation are still in their infancy with regard to autism. First, very few studies have examined the influences of specific child behaviours such as acting-out behaviours, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness on other child characteristics and parenting stress. Second, many investigations have focused on children with developmental disabilities other than autism, and only a few studies have examined the relationships between child behaviours and other child characteristics in children under the age of 6. Third, the majority of studies conducted in this area have been descriptive or correlational in nature; thus, they provide only a restricted view of the relationship between child behaviours on either other child characteristics or parenting stress. Finally, only a few studies have included children with autism receiving early intervention services. It is imperative that we begin to better understand the relationship between child behaviours and changes in 46 other child characteristics or parenting stress associated with early intervention. This information will enable professionals to focus their interventions more precisely to promote healthy outcomes. Statement of the Problem and Overview of the Study The research reviewed in this Chapter justifies a need to examine the influence of child behaviours such as acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness on other child characteristics and parenting stress for young children with autism receiving early intervention over time. This investigation examined these issues in detail by identifying child behaviours at the start of early intervention and examining how changes in these behaviours over the first year of intervention affected changes in other child characteristics (i.e., language, adaptive skills, autism severity, and IQ) and in parenting stress over 2 years. Thus, this research was longitudinal in nature and addressed a number of specific questions, as detailed below. Questions Related to Child Characteristics 1. Does one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) at the onset of intervention predict child characteristics over 2 years? 2. Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social 47 unresponsiveness, and inattentiveness) over the first 6 months of intervention predict child characteristics over 2 years? 3. Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the 12 months of intervention predict child characteristics over 2 years? Based on the extant research literature, it is hypothesized that the predictive relationships between child behaviours at either T l , T l to T2, and/or T l to T3 and changes in other child characteristics from T l to T4 will be as follows: Acting-out behaviours will predict changes in cognitive skills; sleep disturbances will predict changes in communicative skills, social skills, and autism severity; eating difficulties will predict changes in language skills, social skills, daily living skills, and autism severity; stereotypic behaviours wil l predict changes in cognitive skills, communication skills, adaptive skills, and autism severity; social unresponsiveness will predict changes in cognitive skills, daily living skills, and social skills; and inattentiveness will predict changes in cognitive and adaptive skills. Questions Related to Parenting Stress 4. Is there a relationship between parent coping style, negative life events experienced at T4, and changes in parenting stress over 2 years? Based on the extant research literature, it is hypothesized that parents who exhibit more positive coping styles and/or who experience fewer negative life events will show evidence of greater reductions in parenting stress from T l to T4. 48 5. Does one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) at the onset of intervention predict a change in parenting stress scores over 2 years, when parent coping style and negative life events are taken into account? 6. Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the first 6 months of intervention predict a change in parenting stress scores over 2 years, when parent coping style and negative life events are taken into account? 7. Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the first 12 months of intervention predict a change in parenting stress scores over 2 years, when parent coping style and negative life events are taken into account? Based on the extant research literature, it is hypothesized at either T l , T l to T2, and/or T l to T3, acting-out behaviours, sleep disturbances, and/or stereotypic behaviour will have a strong predictive influence on changes in parenting stress from T l to T4. In addition, eating behaviours, social responsiveness, and inattentiveness at either T l , T l to T2, and/or T l to T3 will have a moderate predictive influence on changes in parenting stress from T l to T4. 49 CHAPTER 3 Method Participants For this study, data were accessed from a database established between 2001- 2004 for a project examining early intervention outcomes for 70 children with autism spectrum disorders and their parents in British Columbia (BC), Canada. Participants for that study were recruited from families receiving services from three Early Intensive Behavioural Intervention (EIBI) service providers (n = 39) and from families receiving direct early intervention funding (n = 31). Prior to entry into the study, 55 children had been diagnosed with autism and 15 had been diagnosed with pervasive developmental disorder-not otherwise specified (PDD-NOS). A l l diagnoses were done by experienced diagnosticians in the community not involved in the study. There were 58 males and 12 females (mean age: 4;2: range = 1;8 to 6;0). Ethnic backgrounds included European (n = 38), Asian (n• = 21), Hispanic (n = 4), Caribbean/African (n = 2), Middle Eastern (n = 1), and multiple (n = 4). At the beginning of the investigation, mothers had, on average, completed some college or university courses and were coded as semiskilled workers . (e.g., machine operator, grocery store clerk) (Hollingshead, 1962). Fathers had, oh average, some university training and were coded as skilled workers (e.g., department manager, administrative assistant etc.) (Hollingshead, 1962). At the outset, the families included 55 two-parent (married) families, three separated families, four divorced families, seven other arrangements (e.g., common-law), and one single parent. A l l of the children received approximately 15 to 20 hours per week of early intervention services year-round for 2 years. Thirty-nine of the children received clinic- 50 managed services and 31 received parent-managed services. Intervention for all children was based on the principles of applied behaviour analysis (ABA), but was, in general, considered eclectic. It consisted of individualized programming implemented by behavioural consultants and interventionists with input from speech-language pathologists, occupational therapists, and other professionals. Measurement Data were collected at baseline (i.e., prior to the initiation of early intervention, Tl) , and 6, 12, and 24 months later (T2-T4). Data collection occurred in each child's home or early intervention centre by registered psychologists, certified speech-language pathologists, and trained graduate students who acted as family interviewers. None of the assessors were involved in service provision to the children or their families. Two additional measures relating to parental coping style and life events experienced over 2 years were collected via mail-out, after each family was contacted by phone. A list of the measures collected at each time point can be found in Table 3.1. Table 3.1 Measures collected at T l , T2, T3, and T4 for children and families Assessment Instrument Purpose Administered by Childhood Autism Rating Scale (CARS) Mullen Scales of Early Learning Confirm diagnosis of ASD and establish a level of "severity" Psychologist IQ measure that provides subscores in gross Psychologist motor, visual perception, fine motor, receptive language, and expressive language Table continues 51 Table 3.1 (Continued) Assessment Instrument Purpose Administered by Wechsler Preschool and Primary Scales of Intelligence-Ill Vineland Adaptive Behavior Scales (VABS) Peabody Picture Vocabulary Scale-IIIA/B (PPVT) Expressive One-Word Picture Vocabulary Test (EOWPVT) Preschool Language Scale-3 (PLS-3) MacArthur Communicative Development Inventory (MCDI) Autism Behavior Checklist (ABC) Temperament and Atypical Behavior Scale (TABS) IQ measure for children who were above the Psychologist upper age range of the Mullen (i.e., age 69 months or above) Measurement of adaptive behaviour Measurement of single word receptive vocabulary Measurement of single word expressive vocabulary Measurement of global receptive and expressive language skills Measurement of parent report of single word vocabulary and pre-linguistic behaviours Assessment of specific behaviours related to autism Measurement of temperament and a variety of atypical behaviours Psychologist Speech-language pathologist Speech-language pathologist Speech-language pathologist Family interviewer Family interviewer Family interviewer Table continues 52 Table 3.1 (Continued) Assessment Instrument Purpose Administered by Sensory Profile Measurement of sensory abilities/deficits Parenting Stress Index-Short Measurement of parenting stress Form (PSI-SF) Measurement of life events and changes experienced by families over 2 years Family Inventory of Life Events and Changes (FILE) Family Crisis-Oriented Personal Evaluation Scales (F-COPES) Measurement of problem solving and behavioural strategies utilized by families Family interviewer Family interviewer Mail-out Mail-out Dependent Variables: Child Measures Child measures for the dependent variables were each child's raw scores1 on measures of adaptive behaviour, cognitive development, autism severity rating, language, and vocabulary skills. Measures relating to each child dependent variable existed in the database for T l , T2, T3, and T4. Adaptive behaviour. The Vineland Adaptive Behavior Scales (VABS; Sparrow, Balla, & Cicchetti, 1984) is a semi-structured interview administered by a psychologist that provides raw and standard scores for a total composite and four subscales: ' Raw scores were used for all variables except for cognitive development in which only standard scores were available. 53 communication, socialization, daily living, and motor skills. However, only the socialization, daily living, and communication subscale scores were used in this investigation because less than half of the participants (n = 33) obtained a raw score on the motor subscale at T4. The total composite score, which could not be calculated without the motor score, was also not used. Cognitive development. Cognitive development/IQ was measured with one of three tests implemented by a psychologist, depending on the child's age. The Mullen Scales of Early Learning (Mullen, 1995) was used with children under 5 years 9 months of age. It provides a standard score for a total early learning composite and five subscale scores: gross motor, visual reception, fine motor, receptive language, expressive language. The Wechsler Preschool and Primary Scale of Intelligence-Revised (WPPSI-R; Wechsler, 1989) was used for children between 5 years 9 months and 6 years of age; and the Wechsler Intelligence Scale for Children-Third Edition (WISC-III; Wechsler, 1991) was used for children over 6 years of age. Both of these measures provide a standard score for full scale intelligence. Autism severity rating. The autism severity rating was measured by the Childhood Autism Rating Scale (CARS; Schopler, Reichler, & Renner, 1988) and the Autism Behavior Checklist (Krug et al. 1980/1983). The CARS is a 15-item behaviour rating scale administered by a psychologist. It provides a total raw score and a raw score for each of 15 rating categories (i.e., relating to people; imitation; emotional response; body use; object use; adaptation to change; visual response; listening response; taste, smell, and touch response and use; fear or nervousness; verbal communication; nonverbal communication; activity level; level and consistency of intellectual response; and general 54 impressions). Children with total scores above 30 are categorized as having autism, with those scoring between 30 and 37 characterized as having "mild-moderate" autism and those scoring 38 to 60 characterized as having "severe" autism. Children with scores of 24-29 on the CARS were categorized as having pervasive developmental disorder-not- otherwise specified (PDD-NOS) (Perry et al., 2003). The A B C is a 57-item checklist of problem behaviours related to autism that is typically completed by the parent or primary caregiver. Each item is given a weighted score between 1 and 4, and children with total scores of over 68 are considered to have autism. Language and vocabulary skills. Language skills were measured with the Preschool Language Scale-3 (PLS-3; Zimmerman, Steiner, & Pond, 1992), and vocabulary skills were measured with the Peabody Picture Vocabulary Test-IIIA/B (PPVT-III; Dunn & Dunn, 1997), the Expressive One-Word Picture Vocabulary Test (EOWPVT; Brownell, 2000), and the MacArthur Communicative Development Inventories (Words and Gestures or Words and Sentences) (MCDI; Fenson, Dale, Reznick, Thai, Bates, Hartung, Pethick, & Reilly, 1993). The PLS-3 was administered by a speech-language pathologist (S-LP) and measures both receptive and expressive language abilities. It provides a raw and standard score for total language and two subscales, auditory comprehension and expressive communication. The PPVT-III was also administered by an S-LP and measures children's understanding of single-word vocabulary. It provides raw and standard scores for total words understood. The EOWPVT was administered by an S-LP and measures children's single word expressive vocabulary. It provides raw and standard scores of total words expressed. Finally, the MCDIs are parent report forms used to measure language and communication skills in 55 infants and young children. They come in two formats, Words and Gestures and Words and Sentences. At each time point, parents completed the appropriate form, depending on their child's estimated vocabulary size. Parents of children with an estimated vocabulary size of 50 words or less completed the Words and Gestures form and parents of children with an estimated vocabulary size over 50 words completed the Words and Sentences form. Over the four time points, some children switched from one form to the other; however, both forms provide a "total words said" score that was used in this investigation. Dependent Variables: Parent Measures Parent measures for the dependent variables were each parent's raw scores on measures of parenting stress, coping skills, and life events experienced. Measures relating to parenting stress existed in the database for T l , T2, T3, and T4. Measures for coping skills and life events were available for T4 only. Parenting stress. Parenting stress was measured with the Parenting Stress Index- Short Form (PSI-SF; Abidin, 1995). It was completed by parents and assessed their level of stress with regard to their relationship with their child with autism. Previous research has indicated that there are differential predictors of parenting stress for mothers versus fathers (see Chapter 2); therefore, only PSI-SFs completed by mothers were included in this study (n = 66). The PSI-SF consists of 36 key items derived from the PSI-Long Form. It provides a raw score for Total Stress which reflects overall stress experienced within the role of a parent; scores of 90 or above reflect clinically problematic levels of stress. The PSI-SF also provides three subscale scores. Parental Distress (PD) is a measure of how competent a parent feels about raising a specific child (i.e., the child with 56 autism). Parent-Child Dysfunctional Interaction (P-CDI) is a measure of the parent's perceptions that this child is not meeting her expectations and/or that her interactions with this child are not reinforcing. Difficult Child (DC) is a measure of child behavioural characteristics that make this child difficult to manage. The existing database included PSI-SF scores for Total Stress and all three subscales at all four time points. Coping style. Parent coping was measured with the Family Crisis-Oriented Personal Evaluation Scales (F-COPES; McCubbin, Olson, & Larson, 1981). F-COPES measures problem-solving and behavioural strategies utilized by families in difficult or problematic situations. It features 30 coping behaviour items and provides a raw score for total coping and five subscales: acquiring social support; reframing; seeking spiritual support; mobilizing the family to acquire and accept help; and passive appraisal. Only the total coping score was used for this investigation. A high total coping score on the F- COPES indicates that the parent is using more positive coping strategies. Life events. Life events were measured with the Family Inventory of Life Events and Changes (FILE; McCubbin, Patterson, & Wilson, 1983). FILE is a measure of life events and changes experienced by families. For the purposes of this investigation, families were asked to report on life events and changes that occurred over the past 24 months. This precedent was set in previous research by Gabriels et al. (2001) in an investigation that examined predictors of parenting stress outcomes over a 22-month period. In general, this measure assesses the accumulation of negative life events and changes experienced by a family over time. It provides a raw score on total life changes and nine subscales: intrafamily strains; marital strains; pregnancy and childbearing strains; finance and business strains; work-family transitions and strains; illness and 57 family "care" strains; losses; transitions "in and out;" and family legal violations. For the purposes of this study, only the raw score for total life changes was used. A high score on total life changes indicates that the parent has experienced more negative life events over the past 2 years. Predictor Variables: Child Behaviour The existing database did not include the six child behaviour variables that were utilized as independent variables/predictors in this study; thus, they were constructed in a six-step process. Identify relevant predictor variables. The first step was to identify relevant predictor variables. This process was accomplished by examining the current research literature for potential predictors of (a) child language/communication abilities, adaptive behaviour, IQ, and autism severity; and (b) parenting stress in families of young children with autism and other developmental disabilities. Six predictor variables were identified in the literature review (see Chapters 1 and 2), including acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness. Select individual items. The second step was to choose individual item indicators from the existing assessment measures at Time 1 in the data set. This process was completed by examining every item in the A B C , PSI-SF, Mullen, and V A B S , as well as in two additional measures: the Temperament and Aberrant Behavior Scale (TABS; Bagnato, Neisworth, Salvia, & Hunt, 1999) and the Sensory Profile (SP; Dunn, 1999). The TABS was completed by a parent/caregiver and consists of 55 yes/no response questions regarding each child's characteristic emotional style or disposition and 58 regulatory behaviour. The TABS is a reliable, valid, and norm-referenced assessment of dysfunctional behaviour used for children between 11 and 71 months of age (Bagnato et al., 1999). In general, it measures temperament, attention and activity, attachment and social behavior, neurobehavioral state, sleeping, play, vocal and oral behavior, senses and movement, and self-stimulatory behavior. The Sensory Profile was also completed by a parent/caregiver and consists of 125 judgment-based questions that report the frequency with which a child responds to various sensory experiences. It is used for children between the ages of 3 to 10 and provides a standardized method of reporting children's sensory and processing abilities. The choice of individual items was made in light of how each of the child behaviours was described in the early intervention and developmental disabilities literature (see Chapter 2). To eliminate any correlation effects between predictor variables and parenting stress outcomes, items from the PSI-SF were included only in the construction of the child behaviour variables used as predictors of changes in child measures. PSI-SF items were omitted from child behaviour variables used as predictors of parenting stress outcomes. A l l items chosen for each child behaviour variable are summarized in Appendix A. Convert items to dichotomous responses. The third step was to convert individual item indicators into dichotomous item responses (i.e., yes/no scores). As can be seen in Appendix A, some items were originally ordinal responses involving more than two scoring options. In cases where ordinal or Likert-type scales were used to score an individual item, "never," "no," strongly disagree," or "disagree" responses were coded as "no" (0) and all other responses were coded as "yes" (1). A second independent rater 59 coded a randomly selected 20% of the total items entered. Inter-rater reliability was 99.92%. Enter items into database. The fourth step was to enter each item identified as 1 or 0 (i.e., 1 = "yes'Ybehaviour was present and 0 = "no"/behaviour was not present) into an SPSS 13.0 database (Apache Software Foundation, 2000). Eliminate item over-representation. The fifth step was to eliminate over- representation of any single behaviour within a predictor variable. For example, three separate items related to "temper tantrums" were taken from three separate measures (the TABS, the A B C , and Sensory Profile). The question for this step was "Which item is the best representative of the behaviour of concern?" Answering this question involved several phases. First, items related to the same behaviour were grouped into subdomains on the basis of their definitions (see Appendix B). For example, the item "severe temper tantrums and/or frequent minor tantrums" from the A B C , item #108 "has temper tantrums" from the Sensory Profile, and item #26 "has wild temper tantrums" from the TABS were grouped into one subdomain. Second, a series of decision rules were applied to eliminate redundant items within each subdomain. Multiple items from the same measure were retained even when they referred to similar behaviors. For example, on the TABS, item #21, "upset by every little thing;" item #22, "often difficult to soothe when upset or crying," and item #52, "can't comfort self when upset" were all retained as separate items because they were considered separate items within the TABS. For items within a subdomain that originated from different measures, a Pearson correlation was calculated to determine whether or not the items were correlated. Cohen (1988) proposed that an r value of > 0.50, which 2 A l l items from the V A B S were reverse coded 60 indicates that at least 25% of the variance in one score can be accounted for by another, can be used to evaluate whether or not a correlation is of clinical significance when evaluating item over-representation. When r 2 was > .50, one item was eliminated, as follows: (a) items originally coded as dichotomous (e.g., those from the A B C and TABS) were chosen over items originally coded on a Likert-type scale (e.g. items from the PSI- SF, Sensory Profile, and VABS) ; (b) i f both items were originally coded dichotomously, the item that was most consistent with other chosen items in the variable was selected. For example, i f one subdomain in the acting-out variable had clinically correlated items from both the A B C and the Sensory Profile, according to the first rule of elimination, the A B C item was selected because it was originally coded dichotomously. Then, i f another subdomain in the same acting-out variable had correlated items from two measures that were both originally dichotomous (e.g., the A B C and TABS), the A B C item was chosen over the TABS to ensure that items were chosen from a consistent measure (e.g., the ABC) within each predictor variable. The third step for addressing item over-representation occurred only i f more than two items within a subdomain were clinically correlated at r = > 0.50. In this case, a Principal Components Analysis (PCA) was conducted to determine which item(s) to eliminate. P C A can be used to summarize the correlations among items to determine i f there is a single underlying variable that a given number of items have in common. Grimm and Yarnold (1995) noted that, given a collection of related items, P C A can identify a smaller set of items, called eigenvectors or factors. These eigenvectors can then be used to explain the majority of variation among the original set of items. P C A also provides correlation coefficients for each item within a factor. In this analysis, the item 61 with the highest correlation coefficient (i.e., the item that accounted for the greatest amount of variation within a factor) was retained. Calculate coefficient alphas and create composite scores. The final step in the process of creating each predictor variable was to examine the psychometric properties of the items within each predictor variable in order to compute a composite score (Zumbo, Gelin, & Hubley, 2002). First, however, missing data for individual item indicators needed to be addressed using the following rule: If the sum of the items at each time point was > 0.5, a missing datum for that time point was entered as 1. If the sum was < 0.5, a missing item was coded as 0. Once missing data were accounted for, the psychometric properties of the items within each behaviour variable were examined by using total item correlation analysis and calculating the coefficient alpha. Coefficient alpha is used to gauge the reliability of measurements (Cronbach, 2004). In the current investigation, it estimated the accuracy of the interrelated items within a variable by examining the consistency of scores from one item to the next and determining the average correlation of items within a variable (Cronbach, 2004). A coefficient alpha score of > .75 indicates that the items used to calculate the composite score for each predictor variable are internally consistent with little measurement error (Streiner & Norman, 1989). If the items did not hold together with an alpha of > .75, items were eliminated one at a time and the coefficient alpha was recalculated after each removal. If the omission of a single item increased the alpha to > .75, this item was removed permanently from the variable. If removal of any one item was insufficient to obtain the target alpha, items that increased the alpha the most were removed in descending order until an alpha of > .75 was obtained. Once the final items were chosen, the sum of those 62 items was calculated (i.e., 1 = the behaviour was present and 0 = the behavior was not present) to create a composite score for each of the predictor variables at each time point ( i .e . ,Tl ,T2, andT3). Data Analysis The data analysis procedures used to address each question in this research were similar. First, the predictor variables were constructed and then structural equation modeling (SEM) through the LISREL 8.72 Student Edition (Linear Structural RELations 8.7; Joreskog & Sorbom, 2005) computer program was used to answer each of the seven questions. Structural Equation Modeling (SEM). S E M was used to address each question in this investigation. S E M is a powerful method of data analysis that uses individual growth trajectories for building and testing models hypothesized by a researcher that are often predictive in nature. It is a technique that includes, among other things, path analysis and regression to explore relationships among changes in individual variables over time (Keith, 1993; Shumacker & Lomax, 2004). Individual growth trajectories offer a rich and flexible alternative to traditional methods for analyzing longitudinal data and take advantage of the multi-wave data present in this investigation. They also depict change as a continuous process, such that the amount of change between time periods for any given participant is a result of that participant's underlying growth trajectory (Francis, Fletcher, Stuebing, Davidson, & Thompson, 1991). By using SEM, this investigation was able to focus on describing the process of change in child behaviour predictors, child measures, and parenting stress over time using individual trajectories, instead of simply focusing on 63 the amount of change between two arbitrary points in time, such as T4 minus T l (Francis et al., 1991). Thus, individual variability was not ignored (Willett & Sayer, 1994). The software program LISREL 8.72 Student Edition (Joreskog & Sorbom, 2005) was used to examine the relationships among the raw scores3 of all child behaviour predictor variables, child measures, negative life events and parent coping style variables (when required), and the parenting stress variable. A model for each of the seven research questions was hypothesized and is presented below. 3 Raw scores were used for all variables except for IQ in which only standard scores were available. 64 Figure 3.1: Time 1 child behaviours as predictors of the slope of a child characteristic variable over 2 years Predictor @ Time 1 Figure 3.1 addresses Question #1 and examines whether predictor variables measured at baseline (Tl) explain a percentage of the individual difference variance in the Rate of Change (ROC) of the raw scores for each child characteristic variable (e.g., IQ, adaptive behaviour, and language skills) measured over 2 years (i.e. T l , T2, T3, and T4). 65 Figure 3.2: Change in child behaviour variable over first 6 months of intervention (T l - T2) as a predictor of the slope of a child characteristic variable over 2 years Predictor T2-T1 Difference Score e i A A A A Acting-Out T2-T1 .Sleep Disturbances T2-T1 Eating Difficulties T2-T1 Stereotypic Behaviours T2-T1 Social Un- responsive- ness T2-T1 In- (' '""^attentiveness T2-T1 Figure 3.2 addresses Question #2 and examines whether changes in predictor variables over the first 6 months of intervention (as measured by a T2 minus T l difference score) explain the individual difference variance in the ROC of the raw scores for each child characteristic variable measured over 2 years. 66 Figure 3.3: Change in a child behaviour variable over the first year of intervention (Tl to T3) as a predictor of the slope of a child characteristic variable over 2 years Child Figure 3.3 addresses Question #3 and examines whether the individual difference variance in ROC of the predictor variables over the first year of intervention (T l , T2, and T3) explains the individual difference variance in ROC in the raw scores of each child characteristic over 2 years. 67 Figure 3.4: Coping style (F-COPES) and negative life events (FILE) as predictors of the slope of parenting stress over 2 years Figure 3.4 addresses Question #4 and examines the relationship between life events and parent coping style and the individual difference yariance in the ROC of parenting stress over 2 years (i.e., T l , T2, T3, and T4). 68 Figure 3.5: Time 1 child behaviour variables as predictors of the slope of parenting stress over 2 years, taking into account parent coping style and negative life events experienced Figure 3.5 addresses Question #5 and examines whether predictor variables measured at baseline (Tl) explain a percentage of the individual difference variance in the ROC of parenting stress over 2 years in the context of negative life events and parent coping style. Figure 3.6: Change in child behaviour variables over the first 6 months of intervention (T1-T2) as predictors of the slope of parenting stress over 2 years, taking into account parent coping style and negative life events experienced Figure 3.6 addresses Question #6 and examines whether changes in predictor variables over the first 6 months of intervention (i.e., measured by a T2 minus T l difference score) explain the individual difference variance in the ROC of parenting 70 stress over 2 years in the context of parent coping style and negative life events experienced. Figure 3.7: Change in a child behaviour variable over the first year of intervention (Tl to T3) as a predictor of the slope of parenting stress over 2 years, taking into account parent coping style and negative life events experienced Finally, Figure 3.7 addresses Question #7 and examines whether the individual difference variance in ROC of the predictor variables over the first 12 months of intervention (T l , T2, and T3) explains the individual difference variance in the ROC of parenting stress over 2 years in the context of parent coping style and negative life events experienced. 71 SEM statistics reported. The guidelines for reporting S E M data dictate that one should examine goodness-of-fit indices (e.g. chi-square, GFI, CFI, NFI, RMSEA) for overall model fit before examining the individual path coefficients within the model (Keith, 1993; McDonald & Ho, 2002; Raykov, Tomer, & Nasselroade, 1991). If a model is found to be a "good fit," it accounts for the majority of the covariance (Chin, 1998). However, in this analysis, a decision was made not to examine or report fit indices for each structural model presented; rather, only individual path coefficients and their corresponding significance levels will be reported. This decision was based on three arguments found in the current literature with regard to the usefulness of fit statistics in general and the reliability of fit indices for small sample sizes in particular. First, fit indices have been found to be of little use when S E M is used to examine the individual structural paths between predictors and changes in outcome variables over time (Chin, 1998; Fan, Thompson, & Wang, 1999). The structural models presented in this research do not attempt to account for the majority of the covariance, as occurs in factor analysis. Instead, this investigation examined only a "piece" of the puzzle and attempted to identify some — but certainly not all — important variables that affect developmental changes over time in children with autism. Second, the general rules for determining cut-off values used by many researchers to examine fit (e.g., incremental fit indexes > .90) have been called into question recently (Marsh, Hau, & Wen, 2004; Tomarken & Waller, 2005; Yuan, 2005). For example, Tomarken and Waller (2005) argued that fit indices are negatively affected by factors such as sample size and model complexity. Yuan (2005) also concluded that cut-off values are questionable with regard to model fit/misfit when he demonstrated that most 72 fit indices change their distributions substantially when conditions such as sample size change. Finally, fit indices have been found to be unreliable for small sample sizes specifically (Curran, Bollen, Paxton, Kirby, & Chen, 2002; Yuan, 2005; Yuan, personal communication, September 27, 2005). Recent research has indicated that R M S E A (Root Mean Square Error of Approximation), one of the most commonly reported fit indices, is biased and unreliable when sample sizes are under 200 (Curren et al., 2002; Yuan, personal communication, September 27, 2005). Therefore, as the sample size in this investigation is considered small4 according to S E M standards (i.e., 70 participants over four testing occasions), fit indices will not be reported. Instead, only path coefficients and their corresponding t-values were examined. 4 It must be noted that in longitudinal research, small sample sizes are common, especially when examining special populations such as young children with autism. 73 CHAPTER 4 Results In this chapter, the results are presented in five sections. First, the results of the total item correlation analysis used to construct each child behaviour variable are summarized. Second, the descriptive statistics for all child behaviour variables from T l to T3, child measures from T l to T4, and parent stress measures from T l to T4 are provided. Third, the results of the unconditional model examining changes in each of the child measures and in the PSI-SF over 2 years are presented. Fourth, the results for Questions 1-3 that address the predictive effects of child behaviours on changes in other child characteristics over 2 years are summarized. Finally, the results for Questions 4-7 regarding the predictive effects of parent coping styles and negative life events and child behaviours on changes in parenting stress over 2 years are presented. Predictor Variable Creation: Results of Total Item Correlation Analysis Table 4.1 summarizes the total number of items chosen and the corresponding coefficient alpha found for each predictor variable. A coefficient alpha of > .75 indicated that the items used to calculate the composite score for each predictor variable were internally consistent with little measurement error (Streiner & Norman, 1989). Appendix C summarizes the complete list of items included in each child behaviour variable. Of note is the fact that the item content of the predictor variables for acting-out behaviour, stereotypic behaviour, and social unresponsiveness differed when these variables were used as predictors of child measures versus predictors of parenting stress measures. This is because individual items from the PSI-SF were included in the child 74 behaviour variables when they were used to predict changes in child measures over 2 years. However, PSI-SF items were omitted from the child behaviour variables when they were used to predict changes in parenting stress over 2 years, in order to eliminate any correlation effects. 75 Table 4.1 Number of items per predictor variable and corresponding coefficient alpha Child behaviour Total number of items Coefficient alpha Acting-out (as a predictor of 28 .825* child measures) Acting-out (as a predictor of 21 .833* parenting stress) Sleeping disturbances 4 .769* Eating difficulties 9 .817* Stereotypic behaviour (as 38 .750* predictor of child measures) Stereotypic behaviour (as a 37 .760* predictor of parenting stress) Social unresponsiveness (as a 22 .744** predictor of child measures) Social unresponsiveness (as a 0 Unable to obtain predictor of parenting stress) alpha of > .75 Inattentiveness 29 .755* *alpha = > .75 indicates that items are internally consistent with little measurement error ** alpha accepted at given value5 A coefficient alpha of .744 was felt to be sufficient because subsequent item omissions resulted in further reductions of alpha below the .75 target. 76 Overall, the acting-out behaviour variable, when used as a predictor of change in child measures, consisted of 28 items and had a coefficient alpha of .825. When this variable was used as a predictor of change in parenting stress, it consisted of 21 items and had a coefficient alpha of .833. The predictor variable sleeping disturbances consisted of only 4 items but had a coefficient alpha of .769. Eating difficulties consisted of nine items and had a coefficient alpha of .817. Stereotypic behaviour, when used as a predictor of change in child measures, consisted of 38 items and had a coefficient alpha of .75. When this variable was used as a predictor of change in parenting stress, it consisted of 37 items and had a coefficient alpha of .76. The social unresponsiveness variable, when used as a predictor of change in child measures, consisted of 22 items and had a coefficient alpha of .7446. However, when this variable was used as a predictor of change in parenting stress, it did not achieve a coefficient alpha greater than .75; thus, it was omitted as a child predictor variable from all parenting stress analyses. Finally, the inattentiveness variable consisted of 29 items and had a coefficient alpha of .755. In conclusion, all of the child behaviour variables with the exception of social unresponsiveness as a predictor of parenting stress, achieved the target alpha of > .75 and thus, were considered to be internally consistent with little measurement error. Descriptive Statistics Before providing descriptive statistics, it is important to address the issue of sample size. The sample size for all questions pertaining to predictors of change in child measures over 2 years consisted of the entire group of 70 children. However, the sample size for all questions pertaining to predictors of change in parenting stress measures over 6 A coefficient alpha of .744 was felt to be sufficient because subsequent item omissions resulted in further reductions of alpha below the .75 target. 77 2 years was reduced to 63 children, for two reasons. First, previous research suggests that the predictors of parenting stress differ for mothers and fathers (see Chapter 2); thus, the data for four participants whose fathers completed the PSI-SF were omitted. Second, according to the PSI-SF manual, a raw score of less than 10 on the Defensive Responding (DR) subscale indicates that a parent's responses on the instrument may reflect an attempt to present herself and her relationship with her child with autism in an overly positive light. Three parents with DR subscale scores below 10 during at least 2 of the 4 data collection points (i.e., at least 50% of the time) were considered to be defensive responders and were thus omitted from all parenting stress analyses. In combination, these two exclusion rules resulted in a sample size of 63 for the parenting stress analyses. • Descriptive statistics: Child behaviours (Tl to T3) predicting child measures. Composite scores were calculated for each child behaviour variable. Table 4.2 summarizes the mean, range, and standard deviation for each of the six child behaviours at baseline (Tl), 6 months (T2), and one year (T3) for all 70 children. 78 Table 4.2 Scores for child behaviour variables used to predict changes in child measures over 2 years (7V= 70) Child variable Time 1 Time 2 Time 3 Mean (range) SD Mean (range) SD Mean (range) SD Acting-out behaviours 14.37 • 5.37 11.94 5.91 11.61 5.94 (maximum score = 28) (4-27) (1-29) (0-24) Eating difficulties 5.74 2.73 5.70 2.87 5.77 2.91 (maximum score = 9) (0-9) (0-9) (0-9) Sleep disturbances 0.66 1.14 0.74 1.07 0.80 1.07 (maximum score = 4) (0-4) (0-3) (0-4) Stereotypic behaviour 20.17 5.46 17.51 6.93 16.86 6.94 maximum score = 38 (6-35) (2-36) (1-31) Social unresponsiveness 6.67 3.19 4.97 2.91 4.81 2.83 (maximum score = 22) (0-14) (0-12) (0-15) Inattentiveness 16.49 4.40 13.81 5.13 13.36 5.16 (maximum score= 29) (7-26) (2-24) (2-23) SD = standard deviation On average, the scores for acting-out behaviours, stereotypic behaviours, social unresponsiveness, and inattentive behaviours decreased over the first year of intervention (Tl to T3), whereas eating difficulties and sleeping disturbances displayed little average change over time. The relatively high mean scores for acting-out behaviours, stereotypic 79 behaviours, and inattentiveness appear to conform to the patterns that might be expected in children with autism (i.e., on average, the children displayed approximately 51% of all possible acting-out behaviours, 53% of stereotypic behaviours, and 57% of inattentive behaviours). However, the relatively low mean scores for social unresponsiveness were unexpected, given that this characteristic is typically associated with autism. On average, the children displayed only 29% of all possible socially unresponsive behaviours at T l , 23%atT2, and 22% at T3. Descriptive statistics: Child behaviours (Tl to T3) predicting parenting stress. Using the sample size of 63 and omitting all PSI-SF items, composite scores were re-calculated for each child behaviour variable, with the exception of social unresponsiveness which did not achieve an alpha of > .75. Table 4.3 summarizes the mean, range, and standard deviation for each of the five child behaviours at baseline (Tl), 6 months (T2), and one year (T3) for all 63 children. 80 Table 4.3 Descriptive statistics for child behaviour variables used to predict changes in parenting stress over 2 years (N = 63) Child behaviour variable Time 1 Time 2 Time 3 Mean (range) SD Mean (range) SD Mean (range) SD Acting-out behaviour 10.39 4.05 8.55 4.54 8.38 4.43 (maximum score = = 21) (1-20) (0-24) (0-18) Eating difficulties 5.59 2.76 5.68 2.88 5.75 2.90 (maximum score = = 9) (0-9) (0-9) (0-9) Sleep disturbances 0.73 1.18 0.74 1.05 0.84 1.11 (maximum score = = 4) (0-4) (0-3) (0-4) Stereotypic behaviour 19.65 5.21 17.16 6.84 16.59 6.81 (maximum score = = 37) (6-34) (2-35) (1-30) Inattentiveness 16.73 4.44 13.92 5.21 13.37 5.10 (maximum score =  29) (7-26) (2-24) (2-23) SD = standard deviation On average, acting-out behaviours, stereotypic behaviours, and inattentive behaviours decreased over the first year of intervention (Tl to T3), whereas eating difficulties and sleeping disturbances showed little change over time. 81 Descriptive statistics: Child measures (Tl to T4). Table 4.4 summarizes the raw score7 mean, range, and standard deviation for each of the 11 child measures at baseline (Tl), 6 months (T2), 1 year (T3), and 2 years (T4) for all 70 children. Table 4.4 Raw scores8 of child measures from T l to T4 Time 1 Time 2 Time 3 Time 4 Child Mean Mean Mean Mean measure (range) SD (range) SD (range) SD (range) SD A B C 67.90 23.68 49.68 23.93 47.47 26.40 47.80 24.85 (11-116) (0-99) (4-114) (4-110) CARS 36.14 5.95 35.43 6.81 34.74 7.18 34.34 7.90 (22-51) (18-49) (19-50) (17-51) IQ 52.39 12.12 54.83 12.90 58.51 15.45 61.76 18.91 (40-127) (48-112) (39-112) (40-117) V A B S 35.43 17.93 44.36 21.27 52.14 26.44 61.88 26.99 comm (4-91) (10-115) (10416) (12-114) Table continues 7 Raw scores were used for all child measures with the exception of IQ, for which only standard scores were available. 8 Raw scores were used for all child measures with the exception of IQ, for which only standard scores were available. 82 Table 4.4 (Continued) Time 1 Time 2 Time 3 Time 4 Child Mean Mean Mean Mean measure (range) SD (range) SD (range) SD (range) SD V A B S dls 34.83 14.80 42.33 20.35 51.36 22.79 60.61 25.02 (6-70) (8-117) (14- (17- 114) 136) V A B S soc 34.57 11.47 40.57 12.50 47.20 15.64 51.86 18.61 (13-72) (17-76) (17-84) (5-104) PPVT 10.64 18.71 20.80 23.05 28.80 27.67 39.20 30.04 (0-85) (0-87) (0-108) (0-125) EOWPVT 11.19 15.74 19.81 18.94 25.17 23.12 34.80. 26.91 (0-65) (0-75) (0-86) (0-96) PLS A C 13.39 9.78 20.01 11.46 23.91 13.02 29.19 13.37 (3-48) (3-47) (5-48) (8-48) PLS EC 14.38 8.10 18.39 8.89 21.57 10.27 26.05 . 12.43 (4-47) (6-44) (5-48) (7-48) PLS Total 27.77 17.45 38.41 19.95 45.48 22.94 55.24 ' 25.37 (8-95) (11-91) (10-96) (15-96) comm = communication; dls = daily living skills; soc = social skills On average, the mean scores on the A B C and CARS decreased over 2 years and the mean standard scores for IQ increased over 2 years. In addition, the mean raw scores 83 for the V A B S communication, social skills, and daily living skills subscales and for the PPVT, EOWPVT, PLS A C , PLS EC, and PLS Total all increased over 2 years. Descriptive statistics: Parenting stress (Tl to T4). Table 4.5 summarizes the mean, range, and standard deviation for each of the PSI-SF subscales and for the PSI-SF total score at baseline (Tl), 6 months (T2), 1 year (T3), and 2 years (T4) for 63 children. Table 4.5 Scores of parenting stress (Tl to T4) Time 1 Time 2 Time 3 Time 4 PSI-SF Mean Mean Mean Mean scores (range) SD (range) SD (range) SD (range). SD PD 33.13 8.87 30.19 8.38 29.52 10.08 26.46 8.73 (15-53) (15-51) (15-58) (12-46) PCDI 30.08 6.73 26.81 6.06 25.14 7.13 24.21 6.71 (18-50) (17-42) (15-46) (12-49) DC 38.17 7.39 34.23 8.39 34.00 8.27 32.90 9.19 (18-54) (14-52) (17-55) (16-53) Total 101.42 17.80 91.22 18.54 88.63 21.00 83.60 20.21 (59-147) (52-136) (53-149) (42-136) PD = Parental Distress subscale; PCDI = Parent-Child Dysfunctional Interaction subscale; DC = Difficult Child subscale 84 On average, there was a decrease in all three PSI-SF subscale scores and in the total score over 2 years. Overall, the mean PSI-SF total score at T l (101.42) decreased from above the cut-off value of > 90 that indicates a clinically significant level of parenting stress to below this cut-off value at T4 (83.60). Unconditional Model: Changes in Child Measures and PSI-SF Over 2 Years Before any analyses were conducted to examine the predictors of change in either child measures or parenting stress over 2 years, the unconditional model that examined change in each of these variables alone over 2 years was required. Table 4.6 presents the unconditional model for changes in child measures from T l to T4. Table 4.7 presents the unconditional model for changes in parenting stress as measured on the PSI-SF from T l to T4. Table 4.6 Unconditional model of change in child measures from T l to T4 V A B S V A B S V A B S EOW- MCDI PLS PLS PLS Value A B C C A R S IQ comm dls soc PPVT PVT total words AC EC total Intercept 59.60 35.66 52.62 35.72 34.17 34.27 12.27 12.97 155.54 16.22 14.63 31.63 Slope -6.91 -0.53 4.23 13.95 13.00 9.58 13.66 10.57 138.68 6.32 5.98 11.45 T-value -4.64* -1.42 5.23* 12.03* 14.42* 11.81* 11.63* 10.61* 11.39* 12.29* 12.04* 12.54* *t-value significant at > ±1.96 (p < 0.05) comm.= communication; dls = daily living skills; soc = social skills oo ON 86 Table 4.7 Unconditional model of change in PSI-SF from T l to T4 Value PSI total PSIPD PSI PCDI PSI DC Intercept 96.12 31.05 27.74 36.88 Slope -6.74 -2.53 -2.22 -2.01 T-value -6.14* -5.65* -6.87* -3.42* *t-value significant at > ±1.96 (p < 0.05) PD = parental distress; PCDI = parent-child dysfunctional interactions; DC = difficult child Overall, the results indicate that the children changed significantly over 2 years on all child measures, except for the CARS. In addition, the results demonstrated that parents changed significantly on all three subscales and on the total score of the PSI-SF over 2 years. Therefore, all measures except the CARS were included in subsequent conditional model analyses that examined the predictors of change for child measures and parenting stress over 2 years. Results: SEM Analyses Structural equation modeling (SEM) was used to examine whether one or more categories of child behaviour either as measured at the onset of intervention (i.e., Questions 1 and 5), measured with a difference score between baseline and 6 months (i.e., Questions 2 and 6), or measured by the rate of change from baseline to 1 year (i.e., Questions 3 and 7) predicted the individual difference variance in the rate of change of any child or parenting stress measure over 2 years. In this section, all results are 87 presented in tables instead of in S E M path diagrams for ease of reading. Each table includes the path coefficient (p) and corresponding t-value for the path. A path is defined from one of the six categories of child behaviours to the rate of change of each child measure over 2 years. A path coefficient with a t-value of > ±1.96 was considered significant at p < 0.05. Please note that analyses were not conducted on paths between child behaviour variables and child measures when items from the child measure were used in the creation of the predictor variable(s). Cases in which this situation occurred are coded as N /A (Not Analyzed) in each table. In addition, some of the conditional models required the use of mathematical restrictions such as allowing the error of a predictor at a specific time point to be greater than 0. Please refer to Appendix D for the examples of LISREL syntax for each question and a list of all restrictions used for each conditional model presented. Results for Question 1: Does one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) at the onset of intervention predict child characteristics over 2 years? Table 4.8 presents the results for this question. In this analysis the independent variable is the T l score for each child behaviour. 88 Table 4.8 Path coefficients and (t-values) for T l child behaviours predicting rate of change (Tl to T4) of child measures Child Acting-out Sleep Eating Stereotypic Social Inatten- measure behaviour disturbances difficulties behaviour unresp. tiveness A B C N / A -0.20 -1.03. N / A N / A - N / A (-0.16) (-1.87) IQ 0.12 -1.11 0.03 -0.11 -0.07 -0.01 (0.72) (-1.27) (0.09) (-0.56) (-0.21) (-0.03) V A B S comm 0.08 -0.43 -0.17 -0.10 -0.33 N / A (0.33) (-0.41) (-0:39) (-0.37) (-.75) V A B S dls 0.15 0.003 0.50 -0.12 -0.33 -0.66 (0.83) (0.004) (1.58) (-0.61) (-0.99) (-2.76)* V A B S soc 0.41 0.15 0.52 -0.40 N / A N / A (2.64)* (0.22) (1.88) (-2.39)* PPVT 0.51 -1.00 0.41 -0.10 0.15 -0.72 (2.14)* (-0.92) (0.94) (-0.40) (0.33) (-2.24)* EOWPVT 0.62 -0.73 0.16 -0.14 0.15 -0.62 (3.05)* (-0.81) (0.44) (-0.64) (0.39) (-2.28)* MCDI 5.74 15.98 0.65 -0.23 -0.16 -5.27 (2.37)* (1.44) (0.14) (-0.09) (-0.03) (1.63) Table continues 89 Table 4.8 (Continued): Child Acting-out Sleep Eating Stereotypic Social Inatten- measure behaviour disturbances difficulties behaviour unresp. tiveness PLS A C 0.03 0.30 0.07 -0.20 0.12 -0.14 (0.29) (0.61) (0.38) (-1.75) . (0.57) (-0.95) PLS EC 0.23 -0.22 0.05 -0.02 0.06 -0.37 - (2.18)* (-0.46) (0.27) ..(-0.17) (0.29) (-2.61)* PLS Total 0.24 0.30 . 0.08 -0.20 0.07 -0.48 (1.22) (0.35) (0.22) (-1.00) (0.19) (-1.84) * ='t-value significant at > ±1.96 (p < 0.05) N/A = Not analyzed because items in the child measure were used to construct the predictor variable The results revealed that three child behaviours measured at T l predicted changes in other child characteristics from T l to T4 (i.e., over 2 years). First, a high score on acting-out behaviour at the onset of intervention was predictive of a greater increase in the rate of change in V A B S social, EOWPVT, MCDI, and PLS EC scores over 2 years. Second, a high score on stereotypic behaviour at the onset of intervention was predictive of less of an increase in the rate of change of V A B S social scores over 2 years. Finally, a high score on inattentive behaviour at the onset of intervention was predictive of less of an increase in the rate of change of V A B S social, V A B S daily living, PPVT, EOWPVT, and PLS EC scores over 2 years. The remaining child behaviours (i.e., sleep disturbances, 90 eating difficulties, and social unresponsiveness) did not significantly predict the rate of change of any child measure over 2 years. . The finding that children with more acting-out behaviours did better on social and language skills over time, and children with more inattentive behaviours did worse on these same skills required further investigation. The question was asked: "If these two behaviours predict opposing results on the same measures, what is the relationship between high acting-out children and high inattentive children?" To answer this question, children were coded as scoring either high on acting-out behavior at T l (i.e., scores at or above 18/28) or low on this variable (i.e., scores less than 18). Similarly, with regard to inattentiveness, they were coded as scoring either high (i.e., scores at or above 19/29) or low (i.e., scores below 19). Then, a Pearson product-moment correlation was calculated to determine i f children who scored high on acting-out behaviours also scored high on inattentiveness. The results revealed an insignificant correlation of r2= 0.96 (p = .428); thus, it appears that these were two separate groups of children. Results for Question 2: Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the first 6 months of intervention predict child characteristics over 2 years? Table 4.9 presents the results for this question. In this analysis, change in each of the child behaviours over the first 6 months of intervention was represented by a difference score between the two time points of interest (i.e., T2 minus T l ) for each predictor variable. Table 4.9 Path coefficients and (t-values) related to difference scores (T1-T2) for child behaviours predicting the rate of change (Tl to T4) of child measures Child Acting-out Sleep Eating Stereotypic Social Inatten- measure behaviour disturbances difficulties behaviour unresp. tiveness A B C N / A 1.61 0.17 N / A N / A . N / A (1.25) (0.25) IQ -0.10 0.69 0.39 -0.55 0.06 0.17 (-0.38) (0.85) (0.92) (-2.36)* (0.19) (0.56) V A B S comm 0.42 0.14 -0.19 -0.49 -0.01 N / A (1.17) (0.12) (-0.32) (-1.63) (-0.21) V A B S dls 0.33 -0.20 -0.58 -0.26 -0.24 0.23 (1.16) (-0.22) (-1.28) (-.1.11) (-0.66) (0.71) V A B S soc ' -0.24 0.87 -1.14 -0.40 N / A N / A (-0.97) (1.12) (-2.89)* (-1.93) PPVT 0.17 -0.11 -0.44 -0.05 -0.53 -0.09 (0.46) (-0.09) (-0.75) (-0.16) (-1.15) (-0.22) EOWPVT 0.08 -0.54 -0.12 -0.25 -0.21 0.10 (0.26) (-0.05) (-0.23) (-0.94) (-0.53) (0.26) MCDI 0.02 0.09 -0.001 -0.03 -0.07 0.05 (0.40 (0.79) • (-0.01) (-0.94) (-1.42) (1.08) Table continues 92 Table 4.9 (Continued) Child Acting-out Sleep Eating Stereotypic Social Inatten- measure behaviour disturbances difficulties behaviour unresp. tiveness PLS A C 0.23 -0.50 -0.08 -0.07 -0.22 -0.22 (1.39) (-0.98) (-0.31) (-0.49) (-1.08) (-1.20) PLS EC -.07 -0.27 -0.36 -0.09 -0.11 0.05 (0.42 (-0.52) (-1.37) (-0.63) (-0.56) (0.26) PLS Total 0.32 -0.98 -0.50 -0.09 -0.12 -0.22 (1.11) (-1.06) (-1.07) (-0.35) (-0.32) (-0.66) * = t-value significant at > ±1.96 (p < 0.05) N A = Not analyzed because items in the child measure were used to construct the predictor variable The results revealed that changes in two child behaviours from T l to T2 predicted changes in other child characteristics from T l to T4 (i.e., over 2 years). First, a reduction in eating difficulties over the first 6 months of intervention was predictive of a greater increase in the rate of change in V A B S social skills scores over 2 years. Second, a reduction in stereotypic behaviours over the first 6 months of intervention was predictive of a greater increase in the rate of change of IQ scores over 2 years. Change over the first 6 months in acting-out behaviours, sleep disturbances, social unresponsiveness, and inattentiveness did not predict the rate of change for any child measure over 2 years. 93 Results for Question 3: Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the first year of intervention predict child characteristics over 2 years? Table 4.10 presents the results for this question. In this analysis, change in each of the child behaviours over the first year of intervention was represented in the conditional SEM model through the creation of latent variables for each predictor yariable. The latent predictor variable was the slope or rate of change of each predictor variable from T1-T3. Table 4.10 Path coefficients and (t-values) for the rate of change (Tl to T3) of child behaviours predicting the rate of change (Tl to T4) in child measures Child Acting-out Sleep Eating Stereotypic Social Inatten- measure behaviour disturbances difficulties behaviour unresp. tiveness A B C N / A -0.75 1.20 N / A N / A N / A (-0.41) (1.31) IQ -0.69 11.63 -0.12 -0.25 -0.035 -0.97 (-1.14) (0.97) (-0.30) (-1.21) (-0.03) (-0.43) V A B S comm 0.005 -1.70 -0.15 -0.46 -0.26 N/A (0.006) (-0.75) (-0.27) (-1.85) (-0.19) Table continues 94 Table 4.10 (Continued) Child Acting-out Sleep Eating Stereotypic Social Inatten- measure behaviour disturbances difficulties behaviour unresp. tiveness V A B S dls -4.32 -21.82 0.16 -0.37 0.61 -0.34 (-1.52) (-0.96) (0.38) (-1.98)* (0.27) (-1.51) V A B S soc -1.74 -10.69 -0.27 -0.079 N / A N / A (-1.34) (-0.85) (-0.70) (-0.49) PPVT -6.79 -41.69 -0.16 -0.38 -0.70 -0.40 (-1.18) (-0.69) (-0:30) (-1.53) (-0.69) (-1.33) EOWPVT -4.91 -0.22 -0.19 -0.38 -0.32 -0.54 - (-1.07) (-0.12) (-0.41) (-1.72) (-0.64) (-2.26)* MCDI 5.56 -2.00 3.57 2.33 10.99 -0.28 (0.81) (-0.11) (0.61) (0.86) (1.59) (-0.09) PLS A C 0.11 -1.78 0.15 -0.075 .. -0.63 -0.19 (0.35) (-1.14) (0.62)' (-0.72) (-1.05) (-1.34) PLS EC -2.31 -28.62 -0.16 -0.24 -0.43 -0.21 (-1.05) (-0.29) (-0.62) (-2.10)* : (-0.89) (-1.57) PLS Total -0.05 -1.69 -0.15 -0.29 -0.57 -0.31 (-0.07) (-1.50) (-0.35) (-1.49) (-0.70) (-1.27) * = t-value significant at > ±1.96 (p < 0.05) N / A = Not analyzed because items in the child measure were used to construct the predictor 95 The results revealed that changes in two child behaviours from T l to T3 predicted changes in other child characteristics from T l to T4 (i.e., over 2 years). First, a reduction in stereotypic behaviour over the first year of intervention was predictive of a greater increase in the rate of change of V A B S daily living skills and PLS EC scores over 2 years. Second, a reduction in inattentive behaviours over the first year of intervention was predictive of a greater increase in the rate of change of EOWPVT scores over 2 years. Change over the first year in acting-out behaviours, sleep disturbances, eating difficulties, and social unresponsiveness did not predict the rate of change for any child measure over 2 years. Results for Question 4: Is there a relationship between scores for parent coping style (FCOPES), negative life events (FILE) at T4, and changes in parenting stress (PSI-SF) over 2 years? Table 4.11 presents the results for this question. Table 4.11: Path coefficients and (t-values) for parent coping style and negative life events predicting the rate of change (Tl to T4) of parenting stress over 2 years PSI-SF Scores FCOPES FILE PSI-SF Total ATU 0.36 (-2.20)* (2.50)* PSI-SF: Parental distress -0.01 0.10 (-0.34) (1.53) PSI-SF: Parent-child dysfunctional interaction -0.03 0.09 (-1.37) (1.76) Table continues 96 Table 4.11 (Continued) PSI-SF Scores FCOPES FILE PSI-SF: Difficult child -0.06 0.17 (-1.76) (2 .37)* * = t-value significant at > ±1.96 (p < 0.05) The results revealed that mothers with higher F-COPES scores (indicating that, in general, they utilized more positive problem-solving strategies in difficult situations) experienced greater reductions in total parenting stress from T l to T4 (i.e., over 2 years). In addition, mothers with higher scores on the FILE (indicating that they experienced more negative life events) experienced less of a reduction in total PSI-SF scores and in PSI Difficult Child subscale scores over 2 years. The Difficult Child subscale measures the extent to which the parent perceives the child's behaviour as difficult to manage. Results for Question 5: Does one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) at the onset of intervention predict a change in parenting stress scores over 2 years, when parent coping style and negative life events are taken into account? Table 4.12 presents the results for Question 5. In this analysis the independent variable is the T l score for parenting stress. 97 Table 4.12 Path coefficients and (t-values) for child behaviour variables at T l predicting the rate of change (Tl to T4) in PSI-SF scores Acting-out Eating Sleep Stereotypic Inattentive- PSI-SF scores behaviour difficulties disturbances behaviour ness PSI Total -0.28 -0.23 -1.15 0.06 . 0.23 (-0.83) (-0.56) (-1.20) (0.22) (0.82) PSI: Parental 0.05 -0.16 -1.03 -0.18 0.19 distress (0.31) (-0.84) (-2.35)* (-1.43) (1.46) PSI: Parent-child -0.20 -0.17 0.23 0.17 -0.05 dysf. interaction (-1.69) (-1.21) (0.69) (1.80) (-0.48) PSI: Difficult -0.10 0.11 -0.38 . 0.07 0.12 child (-0.54) (0.55) (-0.78) (0.50) (0.85) * = t-value significant at > ±1.96 (p < 0.05) Only one child behaviour variable at T l was found to be predictive of changes in parenting stress from T l to T4 (i.e., over 2 years). Children with higher scores on sleep disturbances at T l had parents who demonstrated greater reductions in stress related to their feelings of competency (i.e., the Parental Distress subscale). This finding was not in the expected direction. The remaining child behaviour variables at T l — acting-out behaviour, eating difficulties, stereotypic behaviour, and inattentiveness — did not significantly predict the rate of change of parenting stress over 2 years. 98 Results for Question 6: Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the first 6 months of intervention predict a change in parenting stress scores over 2 years, when parent coping style and negative life events are taken into account? Table 4.13 presents the results for this question. In these analyses, change in each of the child behaviours over the first 6 months of intervention was represented by difference scores between the two time points of interest (i.e., T2 minus T l ) for each predictor variable. Table 4.13 Path coefficients and (t-values) related to difference scores (T1-T2) for child behaviours predicting the rate of change (Tl to T4) of parenting stress Acting-out Sleep Eating Stereotypic Inatten- PSI-SF scores behaviour disturbances difficulties behaviour tiveness PSI Total -0.50 2.42 • 0.54 0.59 -0.59 (-1.31) (2.42)* (0.10) (2.05)* (-1.78) PSI: Parental distress -.0.17 1.37 -0.002 0.35 -0.40 (-1.02) (3.14)* (-0.008) (2.77)* (-2.71)* PSI: Parent-child 0.09 0.12 -0.47 0.17 -0.27 dysf. interaction (0.65) (0.35) (-0.25) (1.62) (-2.30)* PSI: Difficult child -0.41 0.88 0.10 0.06 0.54 (-2.03)* (1.68) (0.37) (0.39) (0.31) * = t-value significant at > ±1.96 (p < 0.05) 99 The results revealed that changes in four child behaviours from T l to T2 predicted changes in parenting stress from T l to T4 (i.e., over 2 years). Improvement in two child behaviour variables, acting-out and inattentiveness, had the opposite predictive effects on parenting stress than was expected. Children whose acting-out behaviours improved over the first 6 months had parents who were more stressed over 2 years regarding their ability to manage their children's behaviour (i.e., the Difficult Child subscale). In addition, children whose inattentive behaviours improved over the first 6 months had parents who were more stressed over 2 years regarding their feelings of competency and their interactions with their children (i.e., the Parental Distress and Parent-Child Dysfunctional Interaction subscales). On the other hand, children who had reduced stereotypic behaviours or sleep disturbances over the first 6 months had parents who were less stressed over 2 years (i.e., PSI-SF Total scores) as well as less stressed about how competent they felt about raising their children (i.e., the Parental Distress subscale). Results for Question 7: Does change in one or more categories of child behaviour (acting-out behaviour, sleep disturbances, eating difficulties, stereotypic behaviour, social unresponsiveness, and inattentiveness) over the first year of intervention predict a change in parenting stress scores over 2 years, when parent coping style and negative life events are taken into account? Table 4.14 presents the results for this question. In these analyses change in each of the child behaviours over the first year of intervention was represented in the S E M 100 model through the creation of latent variables for each predictor variable. The latent predictor variable was the slope or rate of change of each predictor variable from T1-T3. Table 4.14 Path coefficients and (t-values) for the rate of change in child behaviours (Tl to T3) predicting the rate of change in parenting stress (Tl to T4) Acting-out Sleep Eating Stereotypic Inatten- PSI-SF scores behaviour disturbances difficulties behaviour tiveness PSI Total 0.67 0.28 0.01 0.34 0.05 (0.59) (0.21) (0.62) (1.31) (0.21) PSI Parental distress -0.35 0.09 0.12 0.11 -0.09 (-1.13) (0.15) (0.49) (0.92) • (-0.77) PSI: Parent-child dysf. -0.34 -0.87 -0.03 0.49 -0.06 interaction (-1.52) (-1.84) (-0.17) (0.57) (-0.60) PSI: Difficult child 0.59 0.96 0.56 . 0.16 0.19 (1.20) (1.23) (1.99)* (1.30) (1.37) * = t-value significant at > ±1.96 (p < 0.05) The results revealed that changes in only one child behaviour from T l to T3 predicted changes in parenting stress from T l to T4 (i.e., over 2 years). Children whose eating difficulties reduced over the first year of intervention were found to have parents who experienced a greater decrease on the Difficult Child subscale over 2 years. 101 Summary of Results Table 4.15 presents a summary of the predictive effects of each child behaviour variable at T l , T l to T2, and from T l to T3 on both the rate of change (ROC) of child characteristics over 2 years and the ROC of parenting stress over 2 years. Table 4.15 Summary of predictive effects of child behaviours at T l , T l to T2, and T l to T3 on ROC of child characteristics and parenting stress from T l to T4. Child behaviour variable Child characteristic: ROC . Parenting stress measure: T l to T4 ROC T l to T4 Higher score on acting-out behaviours at T l Acting-out: T1-T2 - behaviour reduced Acting-out: T l to T3 - behaviour reduced Higher score on sleep disturbances at T l Sleep disturbances: T1-T2 behaviour reduced Greater increase in ROC of - expressive vocabulary - expressive language - social skills X . X X X X Greater increase in ROC of - difficult child X Greater decrease in ROC of - parental distress Greater decrease in ROC of -parental distress -total stress Table continues 102 Table 4.15 (Continued) Child behaviour variable Child characteristic: ROC Parenting stress measure: T l t o T 4 R O C T l t o T 4 Sleep disturbances: T l to T3 - behaviour reduced Higher score on eating difficulties at T l Eating difficulties: T1-T2 - behaviour reduced Eating difficulties: T l to T3 behaviour reduced Higher score on stereotypic behaviours at T l X X Greater increase in ROC of - social skills X Less of an increase in ROC of - social skills X X X Greater decrease in ROC of difficult child Stereotypic behaviour: T1-T2 Greater increase in ROC of Greater decrease in ROC of - behaviour reduced - cognitive skills -parental distress -total stress Stereotypic behaviour: Greater increase in ROC of T l to T3 - behaviour reduced - expressive language - daily living skills X Table continues 103 Table 4.15 (Continued) Child behaviour variable Higher score on social unresponsiveness at T l Social unresponsiveness: T l - T2 - behaviour reduced Social unresponsiveness: T l to T3 - behaviour reduced Higher score on inattentiveness at T l Inattentiveness: T1-T2 - behaviour reduced Inattentiveness: T l to T3 - behaviour reduced Child characteristic: ROC T l to T4 X X X Less of an increase in ROC of - receptive vocabulary - expressive vocabulary - expressive language - daily living skills X Grater increase in ROC of - expressive vocabulary Parenting stress measure: ROC T l to T4 N 7 A N / A N / A X Greater increase in ROC of - parental distress - parent-child dysfunctional interaction X ROC = rate of change; X = no predictive effect found; N / A = not analyzed 104 CHAPTER 5 Discussion This study examined a number of relationships between specific child problem behaviors and (a) child characteristics in the cognitive, language, social, and daily living skills domains; and (b) parenting stress in the mothers of young children with autism. There were two primary objectives. The first was to assess the predictive effects of acting-out behaviors, sleep disturbances, eating difficulties, stereotypic behaviours, social unresponsiveness, and inattentiveness on changes in child characteristics over 2 years for 70 young children with autism who received early intervention. The second was to assess the predictive effects of these same child behaviours on changes in parenting stress over 2 years for the mothers of 63 of the children. This chapter is presented in four sections. First, a discussion of the uniqueness of the statistical analysis used in this investigation is presented. Second, the findings for both objectives, organized according to each of the child problem behaviors investigated, are discussed. Third, the limitations of the study are described; and finally, directions for future research are proposed. Examining Change Over Time Using Multi-Wave Data The relationships between six child problem behaviors and changes in both child characteristics and parenting stress over 2 years were examined in two unique ways. First, child behaviours at the onset of intervention (i.e., at a single point in time, T l ) were examined as predictors of the rate of change in the developmental trajectories of other child characteristics and parenting stress over four time points (T l , T2, T3, and T4). This 105 is unique in that, historically, the majority of longitudinal studies in autism have sought to examine the extent to which child behaviours at one time point (Tl) predict child characteristics or parenting stress at a second time point (T2) alone. For example, Rollins and Snow (1998) found that joint attention and responsiveness to parental interactions in children with autism at age 1;2 (Tl) predicted grammatical development at age 2;7 (T2). Charman et al. (2005) found that nonverbal communicative interactions at age 2 (Tl) predicted language, communicative, and social skills at age 7 (T2). Baker et al. (2003) found that child acting-out behaviours at age 3 (Tl) predicted parenting stress levels 1 year later (T2). Other authors have also examined various child behaviors as predictors of change over two points in time (e.g., Brady et al., 2004; Charman et al., 2003; Sallows & Graupner, 2005; Sigman & McGovern, 2005; Sigman & Ruskin, 1999; Siller & Sigman, 2002; Stone & Yoder, 2001; Szatmari, 2003; Venter et al., 1992). This investigation moved beyond past research and used structural equation modeling (SEM) to explore the developmental trajectories over four time points rather than just two. The use of S E M allowed for the dependent variables, child characteristics and parenting stress, to be defined as the rate of change (ROC) of each child's continuous individual trajectory over 2 years (i.e., across T l , T2, T2, and T4) rather than simply as the mean T4 score. As a result, the variability for each child and parent over 2 years was not ignored (Willet & Sayer, 1994). In addition, this study moved beyond an exploration of T l predictors alone to examine the extent to which changes in each of six child behaviours over the first year (i.e., over the first six months, to T2; or over the first 12 months, to T3) predicted the differential developmental trajectories found in child characteristics and parenting stress over 2 years (i.e., from T l to T4). Thus, this investigation used both the mean difference 106 scores over 6 months (T2 minus T l ) and the individual rate of change over 1 year (i.e., T1-T3) to define the predictor variables. This examination of how changes in child behaviours over 1 year predicted differential developmental trajectories for children and their mothers over 2 years enables a deeper understanding of the inter-relationships between individual child behaviours and both child development and parenting stress over time. Child Behaviours as Predictors of Changes in Child Characteristics Over 2 Years Acting-Out Behaviour: Predictors of Child Characteristics The results revealed that only the T l measure of acting-out behaviours predicted differential developmental trajectories in other child characteristics over time. Specifically, young children with autism who displayed more acting-out behaviours at T l made more progress in expressive vocabulary, expressive language, and social skills over 2 years. This finding was not expected, as past researchers have theorized that acting-out behaviours may interfere during cognitively demanding tasks and/or may deprive children of opportunities to practice and develop social skills (Bronson, 2000; Hauser- Cram, 2001; Kaiser & Raminsky, 2003). - One possible explanation for the result is that acting-out behaviours in young children with autism may indicate their motivation to interact, albeit negatively, with their environment'. In the past, acting-out behaviours were viewed as maladaptive and nonfunctional, requiring simple elimination from a child's behavioral repertoire (Carr, Langdon, & Yarbrough, 1999). However, acting-out behaviors are now understood as serving one or more functions, at least some of which can be viewed as communicative (Durand & Merges, 2001). For example, acting-out behaviours may be used to gain 107 desired items such as drinks or toys (i.e., "I want X") (Sigafoos & Mirenda, 2002); to escape or avoid undesired activities, people, or situations (i.e., "I don't want X") ; and/or to gain attention or social interaction (Bopp, Brown, & Mirenda, 2004; Carr et al., 1999). Therefore, it may be that some (if not most) acting-out behaviours in young children with autism are an indication of underlying communicative and social skills, rather than simply being behaviours that can interfere with development. This may explain why children with more acting-out behaviors at T l made more progress in communication and social areas over time. Sleep Disturbances: Predictors of Child Characteristics Neither the number of behaviors related to sleep disturbances in young children with autism at T l nor the changes in these behaviors between either T1-T2 or T1-T3 predicted the developmental trajectories of other child characteristics over 2 years. However, this result may reflect problems with the items that comprised the predictor variable itself rather than a true lack of association between sleep disturbances and other child characteristics. This variable included only four items, the smallest number in all six predictor variables. Although the items in this variable achieved a coefficient alpha of 0.77 (indicating internal consistency and little measurement error), the variance between the four items may have been insufficient to predict the rate of change of other child measures over 2 years. In addition, recent studies suggest that two of the four items that constituted this variable, "crying or screaming during sleep" and "often frightened by dreams or having nightmares," are among the least frequently reported sleep problems in young children with autism (Wiggs & Stores, 2004; Williams, Sears, & Allard, 2003). Indeed, "crying or screaming during the night" was reported in only 11% of the children 108 in this study at T l , in 9% at T2, and in 6% at T3; and "often frightened by dreams..." was reported in only 16% at T l , in 14% at T2, and in 13% at T3. Future research is needed to examine how specific sleep disturbances (e.g., getting fewer hours of sleep, as per Schreck, Mulick et al., 2004) are predictive of changes in child characteristics over time. Eating Difficulties: Predictors of Child Characteristics Scores for eating difficulties at T l did not predict the rate of change in any child measures over 2 years; however, children whose eating behaviours improved between T1-T2 made more progress in social skills over this period of time. This finding is consistent with the work of Archer and Szatmari (1991), who found a significant correlation between eating behaviour and social skills at a single time point. The present study adds to the limited research on eating behaviors in autism by demonstrating that reductions in the frequency of these behaviours appears to be related to long-term gains in another area of child development. It is interesting to note that 7 of the 9 items (77.8%) that comprised the eating difficulties variable were behaviours related to narrow food preferences, such as "limits self to particular food textures/temperatures," "avoids certain tastes or food smells that are typically part of children's diets," and "seeks out certain tastes or smells." These behaviours can be considered stereotypic behaviours related to food, in that they reflect a child's insistence on sameness and restricted pattern of interest (Turner, 1999). Hence, the relationship between changes in eating behaviours and changes in other child characteristics will be discussed in more depth in the next section on stereotypic behaviour. 109 Stereotypic Behaviour: Predictors of Child Characteristics Both the level of stereotypic behaviour at T l and changes in stereotypic behaviour sometime between T1-T3 predicted differential developmental trajectories in other child characteristics over 2 years. Specifically, children who had more stereotypic behaviours at T l made less progress in social skills over 2 years (i.e., T1-T4). In addition, a decrease in stereotypic behaviours between T l and T2 predicted improvement in cognitive skills over 2 years, and a decrease in stereotypic behaviours from T l to T3 predicted improvement in expressive language and daily living skills over 2 years. Finally, as noted previously, reductions in primarily stereotypic eating behaviours between T1-T2 predicted improvement in social skills over 2 years. These findings are consistent with past correlational studies that have identified relationships between stereotypic behaviors and other child characteristics (e.g., autism severity, cognitive and/or language ability) at a single time point (e.g., Campbell et al., 1990; Dadds et al., 1988; Venter et al., 1992). The findings are also consistent with intervention studies that have identified an inverse relationship between stereotypic behaviour and the acquisition of communication and/or social skills (e.g., Koegel, Koegel, Hurley, & Frea, 1992; Lee & Odom, 1996). Stereotypic behaviors have also been found to compete against the acquisition of language and social skills (Lee & Odom, 1996; Polirstok et al., 2003); for example, children who demonstrate stereotypic behaviours often have fewer social learning opportunities (McConnell, 2002). Finally, the present research echoes the results reported by Epstein et al. (1985), who found that children with autism whose stereotypic behaviours improved the most also had the best academic achievement scores over 2 years. 110 In addition, this study was the first to demonstrate a predictive relationship between the three defining domains of autism (i.e., social interaction deficits; communication/language deficits; and repetitive, stereotypic behaviors). This is important because it provides us with potentially useful information about how to maximize the effectiveness of early intervention efforts. Past research suggests that, while interventions that focus on prelinguistic socio-communication skills such as joint attention appear to promote language and social development over time, these same interventions have little effect on changes in stereotypic behaviour (Charman & Swettenham, 2001; Charman et al., 2005; Eaves & Ho, 1996). However, the results of this study suggest that the opposite may be true — changes in stereotypic behaviour (including those related to restricted food preferences) appear to be related to changes in the developmental trajectories of social, language, daily living, and cognitive skills over time. This suggests that early intervention programs should include specific interventions designed to affect changes, in the frequency of stereotypic behaviors, since such changes appear to be related to changes in other domains as well. For example, interventions designed to increase play skills, teach alternative replacement behaviors, and use activity schedules or pictorial cues have all been found to reduce stereotypic behaviours over time (see Turner, 1999 for a review). Social Unresponsiveness: Predictors of Child Characteristics Neither scores reflecting the number of behaviors related to social unresponsiveness in young children with autism at T l nor changes in these scores from either T1-T2 or T l to T3 predicted the developmental trajectories of other child characteristics over 2 years. However, it is important to note that the 22 items in the I l l constructed social unresponsiveness variable obtained a coefficient alpha of 0.744, which was slightly less than the minimum level of significance (0.75) used to ensure internal consistency (Streiner & Norman, 1989). Efforts to strengthen the construct by removing one or more items were insufficient to achieve the target level, suggesting that there may have been an unacceptable degree of measurement error in the variable itself. In addition, the mean scores for social unresponsiveness at each time point were relatively low (e.g., the children displayed only 29% of all possible socially unresponsive behaviors at T l , 23% at T2, and 22% at T3). Given the fact that social unresponsiveness is a defining characteristic of autism, these scores suggest that the items used to construct the variable did not accurately reflect the social deficits that are typically seen in young children with autism. Taken together, these two concerns suggest that this predictor variable may have been insufficient to characterize the construct of social unresponsiveness accurately. On the other hand, it is possible that the predictor variable was reasonably accurate and valid despite the somewhat low coefficient alpha, and that children's low social unresponsiveness scores reflect a diagnostic problem rather than a measurement error. When the children in this study were diagnosed (i.e., prior to 2001), neither of the current "gold standard" diagnostic instruments for autism ~ the Autism Diagnostic Interview-Revised (ADI-R; Le Couteur, Lord, &, Rutter, 2003) and the Autism Diagnostic Observation Schedule (ADOS: Lord, Rutter, DiLavore, & Risi, 2002) — were commonly used in British Columbia, primarily because of a lack of trained diagnosticians. Rather, all of the children in this study were diagnosed by individual professionals or professional teams using the Childhood Autism Rating Scale (CARS; Schopler et al., 1988). In addition, an independent psychologist re-administered the CARS at each time point used in the study. 112 Recently, Saemundsen, Magniisson, Smari, and Sigurdardottir (2003) examined both the CARS and the ADI-R with 54 children (ages 22 to 114 months) who were referred for possible autism. They found a significant correlation (r = 0.81; p < .001) between the ADI-R total score and the total score on the CARS. In addition, the observed agreement between the two instruments was 66.7% (i.e., moderate agreement) when all three domain scores on the ADI-R reached the threshold for autism and the cut-off score on the CARS was over 30 (i.e., defining autism). However, this agreement increased significantly to 94.4% when at least one domain score on the ADI-R reached the threshold for autism and the cut-off value on the CARS was over 30. Together, these results provide at least moderate support for the concurrent validity of the two measures. Similarly, Pilowsky, Yirmiya, Shulman, and Dover (1998) also examined the agreement between these two instruments in 70 participants who were suspected of having autism. They found an 85.7% agreement between the CARS and ADI-R diagnoses. From these findings, it appears that the diagnostic criteria used in the CARS are consistent with those in at least the ADI-R, suggesting that the children in this study who were diagnosed with the CARS did indeed have autism and that their low scores were reflective of problems with the constructed variable for social unresponsiveness rather than diagnostic inaccuracy. Inattentiveness: Predictors of Child Characteristics It is important to note at the outset that inattentive behaviours as defined in this research were not related to difficulties with either joint attention or gaze shifting. Rather, inattentiveness referred to the children's inability to make eye contact, maintain focused attention, and/or attend to sudden changes in their environments. Therefore, the 113 discussion of results related to this variable will focus only on inattentive behaviours as defined, without comparing those results to the large body of previous research on joint attention. Both the level of inattentiveness at T l and changes in inattentiveness between TI- TS predicted differential developmental trajectories in other child characteristics over 2 years. First, children with more inattentive behaviours at T l made significantly less progress over 2 years in receptive and expressive vocabulary, expressive language, and daily living skills. This is consistent with past research demonstrating a predictive relationship between the ability of children with developmental disabilities to focus and persist when trying to master a problem, task, or skill at T l (i.e., age 3) and increased cognitive and daily living skills at T2 (i.e., 10 years later) (Hauser-Cram et al., 2001). However, this is the first study to demonstrate this predictive relationship for young children with autism. Second, children whose in inattentive behaviors decreased from T l toT3 made significantly more progress in expressive vocabulary from T l to T4. This finding is intriguing. Given the previous finding that children with low levels of inattentiveness at T l improved in four areas of development over 2 years (i.e., expressive and receptive vocabulary, expressive communication, and daily living skills), one might anticipate that a reduction in inattentiveness from T l to T3 would predict improvements in all four of these areas. However, a positive rate of change in inattentiveness from T l to T3 predicted improvement in only one area, expressive vocabulary. In this regard, it is important to note that, on average, inattentive behaviours decreased by only 3.13 points from T1-T3; thus, it appears that the majority of children were still relatively inattentive at T3. Perhaps 114 larger reductions in inattentiveness were needed in order for changes in this behavior to have predictive value. Future research is needed in order to examine this issue. Child Behaviours as Predictors of Maternal Parenting Stress Over 2 Years The second objective of this study was to examine the same six child behaviours as predictors of differential changes in parenting stress for 63 mothers of children with autism over 2 years. However, the predictor variable social unresponsiveness was eliminated because the items used to define this variable for the parenting stress analyses did not hold together with a coefficient alpha of > .75, indicating a lack of internal consistency with an unacceptable amount of measurement error. Thus, only the remaining five behavior variables were examined as predictors. The purpose of these analyses was to better understand the inter-relationships between child problem behaviors and parenting stress over time. However, as described in Chapter 1, the Double A B C X model of family adaptation (McCubbin & Patterson, 1983) theorizes that the stress of raising a child with a disability (X) is influenced by child characteristics (A), parental coping style (B), and by external family supports and/or negative life events experienced (C). In addition, Hodapp et al. (1997) argued that relationships between A , B, C, and X change over time, as indicated by the "double" in the Double A B C X model. Thus, in order to better understand how specific child behaviours affected parenting stress over time, the influence of two other variables — parental coping style and negative life events experienced over 2 years — were first examined. The findings revealed that mothers of young children with autism who had more positive coping strategies and/or who experienced fewer negative life events had greater reductions in overall parenting stress over 2 years. These findings are consistent 115 with previous research with parents of children with other developmental disabilities (e.g., Cameron & Armstrong, 1991; Dunn et al., 2001; Hauser-Cram et al., 2001) and confirm the need for the inclusion of parental coping style and negative life events in any model examining the impact of specific child behaviors on parenting stress over time. Consequently, both parental coping style and negative life events were included in each of the structural equation models for all five child behaviours analyzed in this study. Acting-Out Behaviour: Predictors of Maternal Parenting Stress The level of child acting-out behaviours at T l did not predict the rate of change in maternal parenting stress over 2 years. This finding is not consistent with past research, which has demonstrated that acting-out behaviours are typically associated with higher levels of parenting stress for children with autism. However, the majority of past research has been correlational in nature (e.g., Hastings, 2003); has examined children with a variety of developmental disabilities other than autism (e.g., Ross & Blanc, 1998); and/or has examined the simple effects of mean levels of acting-out behaviours at one time point (Tl) on mean levels of parenting stress at another time point (T2) (e.g., Baker et. al., 2003). This study was unique in that S E M was used for the analysis, enabling an examination of the individual rate of change of each parent's stress scores over four time points spanning 2 years, rather than a simple examination of mean differences over two time points. Thus, the individual variability for these parents was not ignored — that is, change as reflected in their individual growth trajectories rather than in a simple group mean change score was examined (Francis, et al., 1991; Willet & Sayer, 1994). This finer-grained analysis may more accurately reflect the predictive influence of acting-out 116 behaviours at T l on the rate of change in parenting stress over time for mothers of young children with autism than the findings from previous research. In contrast, children whose acting-out behaviours decreased between T1-T2 had mothers who experienced more stress with regard to their ability to manage their children's behaviour over 2 years. This finding was also unexpected. One possible explanation is that the mothers of children with higher rates of acting-out behavior experienced increased stress related to implementing intervention programs aimed at ameliorating those behaviours. For example, many types of interventions that are used to replace acting-out behaviours with functional alternatives are typically implemented by parents in the home (see Lucyshyn, Dunlap, & Albin, 2002). The demands of implementing behavior intervention programs may have placed more stress on the mothers at the same time that they effectively reduced the children's acting-out behavior over time. This explanation is supported by anecdotal reports implying that home-based early intervention programs may increase parenting stress (e.g., Cattell-Gordon & Cattell- Gordon, 1998), although it is also challenged by research to the contrary (e.g., Hastings & Johnson, 2001). Additional research is required to examine the interactive relationships between child acting-out behaviours and the demands of parent-implemented behavioral interventions. An alternative explanation for this finding is that the positive changes observed in children's acting-out behaviours between T1-T2 (a 6-month period) were insufficient to have a significant impact on long-term (i.e., 2-year) reductions in maternal parenting stress. Acting-out behaviours in these children were not eliminated over this time period; they were simply reduced (from a mean of 10.4 to a mean of 8.4). As noted previously, the majority of research has demonstrated that acting-out behaviours are a strong and 117 consistent predictor of parenting stress, even when parental coping styles(e.g., McDonald & Gregoire, 1997) and negative life events experienced (e.g., Hauser-Cram et al., 2001) are taken into account. Thus, it is perhaps not surprising that, in this study, short-term, relatively minor reductions in acting-out behaviors were insufficient to affect decreases in parenting stress over 2 years. Finally, we know very little about the influences of specific acting-out behaviours in young children with autism on parenting stress over time. Perhaps, the "quality" of acting-out behavior has more influence on parenting stress than does the quantity of such behaviors. For example, relatively minor acting-out behaviours such as cries easily, is stubborn or uncooperative, or non-compliant (i.e., refusing to do what one is told to do) may have much less impact on parenting stress than more disruptive behaviours such as severe temper tantrums or hurting others by biting, kicking, and/or hitting. Both types of behaviors were included in the acting-out variable constructed for this study, which may have diluted the predictive value of the variable as a whole. Future investigations should examine the impact of specific acting-out behaviours on parenting stress over time, to clarify this issue. Sleep Disturbances: Predictors of Maternal Parenting Stress Both the level of sleep disturbances at T l and changes in sleep disturbances from T l to T2 predicted differential developmental trajectories in measures of maternal stress over 2 years. First, children who had more sleep disturbances at T l had mothers who were less stressed in terms of their feelings of competency as a parent (as reflected in scores on the PSI-SF subscale Parental Distress) from T l to T4. This finding was not expected. However, examining the impact of children's behaviour at a single time point 118 (Tl) on changes in parenting stress over 2 years may not reveal the whole the story between these two variables. This result may have been spurious - in that, the decrease observed in the subscale Parental Distress from T l to T4 was caused by a missing critical variable (Burns, 1997). In this case, it may not have been the high T l value, but perhaps was the decrease in this value over 6 months or 1 year (i.e., the difference score from T l - T2 or the rate of change from T1-T3) that predicted the decrease in maternal stress over time. Therefore, the next step in the analysis examined how actual changes in child behaviours sometime over one year predicted differential parenting stress measures over 2 years. In fact, this analysis did find that children whose sleep disturbances decreased from T l to T2 (i.e., in the first 6 months) had mothers who demonstrated less overall stress and were less stressed on the same subscale, Parental Distress, over 2 years. This finding is more revealing and indicates that it was most likely the decrease in stereotypic behaviours that predicted the reductions in maternal parental distress over time rather than the T l level alone. Recent research on sleep interventions appears to support the finding that decreases in child sleep disturbances can promote reductions in maternal stress. Although there is no way of knowing i f the children in this study received any intervention targeting sleep disturbances, the work of Wiggs and Stores (2001) supports the idea that sleep interventions can reduce parenting stress. They implemented a behavioural treatment program for sleep problems to the parents of 15 children (mean age: 8;2) with severe intellectual disabilities and severe sleep problems and compared them to 15 controls who received no treatment. The mothers of children in the sleep treatment group reported significantly less stress after 2 months than the mothers of children in the control 119 group. Therefore, it appears that helping mothers of children with autism to reduce their children's sleep disturbances can have a positively impact on parenting stress over time. A plausible explanation for the relationship between decreased sleep disturbances and reductions in maternal stress can be found by examining the individual items within this variable. A l l four items (i.e., "often frightened by dreams in the nighttime," "screams in sleep and can't be comforted," "wakes up often and doesn't fall back asleep," and "doesn't have.a regular sleep pattern") describe behaviours that disrupt both children's sleep and that of their mothers and other family members. Sleep deprivation has been found to be related to increased stress over time; for example, Kemp (2003) reported that sleep deprivation was one of the main contributors to parenting stress for parents caring for young children with severe eczema. Thus, when children sleep better, their mothers sleep better and may experience less stress as a result. Eating Difficulties: Predictors of Maternal Parenting Stress Neither scores reflecting the number of problem eating behaviors in young children with autism at T l nor changes in these scores from T1-T2 predicted differential maternal stress trajectories over 2 years. However, a decrease in eating difficulties from T1-T3 predicted a reduction in maternal stress related to the mothers' ability to manage their children's behaviour over 2 years (i.e., the PSI-SF Difficult Child subscale). This finding is consistent with past correlational research indicating that parents of children who have problems with eating also experience more stress (Archer & Szatmari, 1991; Gray, 1994). Gray (1994) noted that specific types of rigid eating behaviours such as "insisting that food be presented on a certain plate" appeared to cause the most stress for many parents of young children with autism. This type of behavior, which was prevalent 120 among the eating behaviors included in this study, can be considered to be stereotypic because of its repetitive, inflexible nature. Thus, the relationship between changes in eating behaviours and changes in maternal parenting stress will be discussed further in the next section on stereotypic behaviour. Stereotypic Behaviours: Predictors of Maternal Parenting Stress The level of stereotypic behaviours at T l did not predict differential maternal stress trajectories over 2 years. However, children whose stereotypic behaviours decreased from T1-T2 had mothers who experienced less overall stress over 2 years as well as less stress related to their feelings of competency in raising their children (as reflected in scores on the PSI-SF subscale Parental Distress). In addition, as previously noted, children whose stereotypic eating behaviours decreased between T1-T3 had mothers who reported less stress in terms of their ability to manage their children's behaviour over 2 years. These findings are consistent with past correlational research that has identified relationships between child stereotypic behaviours and parenting stress at one time point (e.g., Beckman, 1983; Gabriels et a l , 2005; Stoddart, 2003; Stores et al., 1998). The results are the first to suggest that decreasing stereotypic behaviours in young children with autism over a short period of time (i.e., 6 months or 1 year) can result in reduced maternal stress over a longer period of time (i.e., 2 years). One interpretation for this finding is that stereotypic behaviours act as a barrier to the development of children's adaptive communication, daily living, and socialization skills, which in turn results in increased parenting stress. The findings of this study provide some evidence to support the suggestion that stereotypic behaviours do indeed limit the development of adaptive skills (see the previous section entitled Stereotypic 121 Behaviours: Predictors of Child Characteristics). Other authors have reported results to support his contention. For example, Szatmari, Archer, Fisman, and Streiner (1994) provided additional evidence that fewer adaptive skills in children are related to increased parenting stress. They found that parents of 83 children with pervasive developmental disorder (mean age: 5;4) who reported fewer adaptive skills on the V A B S also experienced high levels of stress. Recently, Chadwick, Cuddy, Kusel, and Taylor (2005) conducted a 5-year follow-up study of 82 children with intellectual disability (mean age at T l : 8;0 and mean age at T2: 13;0). They, too, found a relationship between child adaptive behaviour and parenting stress. Specifically, they found that improvements in adaptive communication skills (as measured with the Vineland screener, a shortened version of the VABS) over 5 years were associated with reductions in parenting stress over the same period of time (r = 0.36: p = 0.02). In terms of the present study, these studies support the interpretation that as children's stereotypic behaviours improve, their adaptive skills improve and, as a result, their parents experience reduced stress over time. A second interpretation as to why reductions in stereotypic behaviours are related to reduced parenting stress concerns parents' level of comfort with their children in social situations. Many stereotypic behaviours - such as such as "repeatedly touching people or objects to the point of irritating others," "having strong reactions to changes in routine/environment," or "getting involved in complicated rituals such as lining things up" ~ can appear odd or disruptive to members of the general public. In an older study, O'Moore (1978) found that social contact activities such as going shopping or taking trips with the family produced higher levels of stress in mothers of children with autism. Related to this is a report that parents of children with autism spend relatively little time engaging in recreational and leisure activities outside of the home (Koegel, Schreibman, 122 Johnson, O'Neill, & Dunlap, 1984). Perhaps, social outings with children who display unusual stereotypic behaviours are more stressful for their parents (and are thus avoided) because these behaviours are not easily explained to or accepted by others. On the other hand, a recent study that specifically examined the relationship between stereotypic behaviours and parenting stress appears to provide evidence to the contrary. Tomanik, Harris, and Hawkins (2004) found that stereotypic behaviours had no correlational relationship with maternal stress for the mothers of 60 children with PDD (mean age: 5;0). However, the stereotypic behavior variable in this study was based on the stereotypy subscale of the Aberrant Behavior Checklist (Aman & Singh, 1986), which included only 7 stereotypic behaviours. Some of these were non-specific items such as "odd, bizarre behavior," while others were non-disruptive behaviors such as "meaningless, recurring body movements," "moves or rolls head back and forth repeatedly," or "repetitive hand, body, or head movements." Overall, many of these behaviours were less disruptive than the 38 stereotypic behaviours included in the present investigation. Perhaps, the different results reported by Tomanik et al. (2004) and the present study reflect differences between the specific stereotypic behaviors that were examined in each. Future research examining how reductions in specific types of stereotypic behaviours influence changes in parenting stress over time are needed, to clarify this issue. Inattentiveness: Predictors of Maternal Parenting Stress The level of inattentive behaviour at T l did not predict differential maternal stress trajectories over 2 years. However, children whose inattentive behaviours decreased from T1-T2 (i.e., over 6 months) had mothers who experienced more stress with regard to their 123 feelings of competency as a parent (i.e., scores on the PSI-SF Parental Distress subscale) and their feelings about interactions with their children (i.e., scores on the Parent-Child Dysfunctional Interaction subscale) over 2 years. This finding was not expected. One possible explanation is that, as the children became less inattentive, they may have sought more attention from their parents and placed more caretaking demands on their mothers which, in turn, resulted in increased maternal stress. In support of this suggestion, Keller (1999) examined the variances of stress in 62 mothers of elementary school age children with disabilities and found that caretaking demands were a significant predictor of parenting stress. However, this interpretation is only preliminary. Since this study is the first to examine the relationship between inattentiveness and changes in stress over time in mothers of children with autism, additional research is needed to clarify this result. Limitations This is the first autism study that has attempted to examine behavioural predictors of differential trajectories of changes in other child characteristics and maternal stress over time and, as is often case with "first" studies in particular, it has several limitations. First, relatively small sample sizes (N = 70 for child characteristics and N=63 for maternal stress) were used. Although small sample sizes are not uncommon in the field of developmental disabilities (e.g., Charman et al., 2005; Siller & Sigman, 2002; Sigman & McGovern, 2005; Stone & Yoder, 2001; Szatmari, 2003), future research is needed to examine the predictive relationships found here for a larger group of participants. Second, the timeline of 2 years for this investigation was relatively short for a longitudinal design. 124 Follow-up at 5 years and beyond for this group of children would be optimal, in order to ascertain the stability of the findings. Third, the data used in this investigation were originally collected for another purpose (i.e., an evaluation of the impact of early intervention over 2 years). This raises two issues. First, items included within each child behaviour variable were limited to only those items available in the existing measures in the data set. Thus, none of the predictor variables incorporated all possible individual manifestations of the behaviour constructs used as predictors. For example, the variable "sleep disturbances" was constructed from only four items and did not include behaviours such as getting fewer hours of sleep, which Schreck, Mulick et al. (2004) found to be related to other child characteristics. Second, even though considerable care was taken in the construction of the predictor variables to ensure psychometric validity, none of the variables were taken directly from existing valid and reliable standardized measures of child behavior. Thus, it may be that at least some of the unexpected (and difficult-to-interpret) results reflected problems with the constructed predictor variables themselves. In addition, as noted previously, the social unresponsiveness variable only obtained a coefficient alpha of 0.744 when used as a predictor of change in child measures and fell significantly short of the > .75 level of significance used to ensure internal consistency as a predictor of change in parenting stress. Even though social unresponsiveness was the only predictor variable that failed to achieve an adequate level of internal consistency, future research is needed to examine the predictive effects of all six of the child behaviours presented in this study using standardized, valid, and reliable measures. Fourth, it would be ideal to be able to interpret the predictive relationships between child behaviours and changes in other child characteristics and maternal stress in 125 the context of the early intervention these children received. Unfortunately, although the intervention was based on the principles of applied behavior analysis, it was quite eclectic in nature and varied considerably from child to child in terms of both focus and intensity. In addition, information about the extent to which interventions targeting the specific child behaviours examined in this research was not available. Thus, it was not possible to determine or speculate about the impact of specific interventions on changes in the child behaviour variables. Finally, one needs to be cautious when generalizing these findings beyond this group of 70 children. As with all "first studies," replication is needed to ascertain i f these findings persist with other young children with autism. That said, however, the 70 children in this study were reasonably representative of young children with autism in general, as evidenced by their overall developmental profiles and their autism rating scores at each of the four time points. Thus, while generalizability may be limited and must be approached with caution, it is likely that these findings are at least somewhat applicable to young children with autism other than those who participated. Future Directions The findings of this research provide a better understanding of some of the specific child behaviours that may predict the development of skills in young children with autism and/or the reduction of stress in their mothers. Identifying such predictive relationships may provide guidance for developing treatments that will result in better outcomes for young children with autism and their families (Koegel et al., 1992). For example, these data suggest that reductions in stereotypic behavior are related to improvements in several child development domains (i.e., cognitive development, 126 expressive language, social skills, and daily living skills) as well as in maternal stress. If future research supports this finding, empirically-supported early intervention approaches aimed at reducing or ameliorating children's stereotypic behaviour would appear to be in order. Similarly, the results also suggest that interventions designed to improve children's inattentive behaviours may result in improved expressive vocabulary skills; and interventions designed to improve children's sleep disturbances may result in reduced maternal stress. However, the data also suggest that treatments that improve inattentiveness and acting-out behaviours may result in more stress for mothers over time. This is a puzzling finding that requires additional research as it may indicate that these types of interventions need to include additional supports for parents. Nonetheless, future studies should consider the impact of specific child behaviours and the interventions related to them on differential trajectories of child development and parenting stress over time. In addition, more research is needed to explore i f the predictive relationships found in this investigation persist among other young children with autism receiving various types of early interventions. Only then will we be able to understand the influence of these child behaviours on the widely heterogeneous outcomes observed in young children with autism and their families (Schriebman, 2000). Future studies are also needed to explore the influence of changes in the specific manifestations of behaviours observed within each of the six child behaviour categories proposed in this investigation on changes in child characteristics and parenting stress over time. For example, stereotypic behaviours are manifested in a variety of ways, including motor stereotypies, rituals, compulsions, obsessions, sameness behaviours, echolalia, and self-injury (Bodfish et al., 2000; Militerni, et al., 2002). In fact, Bodfish et al. (2000) demonstrated that very few investigations examining the relationships between 127 stereotypic behaviours and either child or parent characteristics have operationally defined the specific behaviours involved. Indeed, even in this investigation, there was no way to determine whether changes in specific manifestations of stereotypic behaviours had a differential impact on changes in child characteristics and parenting stress over time. A standardized measure such as the Repetitive Behaviour Scale (Bodfish et al., 1999), which categorizes stereotypic behaviours into six distinct subgroups, may be better able to examine these relationships in the future. A similar argument can be made for the other child behaviours as well. For example, destructive acting-out behaviours such as "hurts others by kicking, biting, and hitting" may affect changes in child characteristics and parenting stress differently than minor behaviours such as "frequently irritable, touchy or fussy." Or, non-disruptive sleeping behaviours that such as "getting fewer hours of sleep per night" may differentially affect changes in child characteristics and parenting stress than behaviours . that affect the entire family, such as "crying or screaming during sleep." Therefore, future investigations should seek to define and examine subgroups of these six behaviours in order to determine their specific influences on the differential developmental trajectories observed in both young children with autism and their parents. Finally, it would be ideal for future studies to examine the predictive impact of changes in child behaviours on changes in other child characteristics and parenting stress using structural equation modeling (SEM) with a larger sample size and over a longer period of time. Implementing this type of S E M analysis with a larger population of young children with autism would allow for more variance within each measure. As a result, the covariance matrices that are produced may be better able to detect potential predictive relationships among the variables. In addition, more time may be needed for changes in 128 child behaviours to predict changes in other child characteristics and parenting stress. For example, a reduction in acting-out behaviours over 1 year may be insufficient to affect parenting stress over 2 years. In conclusion, the findings of this investigation open the door for future research examining the predictive relationships between specific child behaviours on changes in child characteristic and parenting stress over time. This type of research may assist service providers, families, and policy makers to make more informed decisions about how to maximize the effectiveness of their early intervention efforts. 129 REFERENCES Abidin, R.R. (1995). Parenting stress index (3rd ed.): Professional manual. Odessa, FL: Psychological Assessment Resources Inc. Achenbach T. & Edelbrock, C. (1983). Manual for the child behavior checklist and revised child behavior profile. Burlington: University of Vermont Department of Psychiatry Aman, M . G., Singh, N . N . , Stewart, A . W., & Field, C. J. (1985). The aberrant behavior checklist: A behavior rating scale for the assessment of treatment effects. American Journal of Mental Deficiency, 89, 485-491. American Psychological Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association. Archer, L. A. , & Szatmari, P. (1991). Eating and mealtime problems in young autistic children: Prevalence and correlates. EDRS, 25. Baker, B. L., Mclntyre, L. L. , Blacher, J., Crnic, K. , Edelbrock, C , & Low, C. (2003). Pre-school children with and without developmental delay: Behaviour problems and parenting stress over time. Journal of Intellectual Disability Research, 47, 217-230. Barnard, K. E. (1978). Nursing child assessment teaching scales. Seattle: University of Washington School of Nursing. Bayley, N . (1969). Bayley scales of infant development. New York: Psychological Corporation. Beckman, P. J. (1983). Influence of selected child characteristics on stress in families of handicapped infants. American Journal of Mental Deficiency, 88, 150-156. Bodfish, J. W., Symons, F. J., Parker, D. E., & Lewis, M . H . (1999). The repetitive behavior scales: A test manual. Morganton, NC: Western Carolina Center Research Reports. Bodfish, J. W., Symons, F. J., Parker, D. E., & Lewis, M . H . (2000). Varieties of repetitive behavior in autism: Comparisons to mental retardation. Journal of Autism and Developmental Disorders, 30, 237-243. Bopp, K.D. , Brown, K .E . , & Mirenda, P. (2004). Speech-language pathologists' role in the delivery of positive behavior support of individuals with developmental disabilities. American Journal of Speech-Language Pathology, 13, 5-19. Boyce, G. C , & Behl, D. (1991). Child characteristics, family demographics and family processes: Their effects on the stress experienced by families of children with disabilities. Counseling Psychology Quarterly, 4, 273-288. Brady, N . C , Marquis, J., Fleming, K. , & McLean, L. (2004). Prelinguistic predictors of language growth in children with developmental disabilities. Journal of Speech, Language, and Hearing Research, 47, 663-677. Bronson, M . B. (2000). Self-regulation in early education. New York: Guilford. Brown, L., Sherbenou, R. J., & Johnsen, S. K . (1997). Manual for the test of nonverbal intelligence (3rd ed.).Austin TX: PRO-ED. Brownell, R. (2000). Expressive One Word Picture Vocabulary Test. Novato, C A : Academic Therapy Publications. Bryant, F.B. & Yarnold, P.R. (1995). Principal-Components analysis and exploratory and confirmatory factor analysis. In L. G. Grimm & P.R. Yarnold (Eds.), Reading and understanding multivariate statistics. Washington, DC: American Psychological Association. 131 Burns, W.C. (1997). Spurious Correlations. Retrieved February 1, 2005, from http://www.burns.com/wcbspurcorl.htm Cameron, S. J., & Armstrong-Stassen, M . (1991). Stress, coping, and resources in mothers of adults with developmental disabilities. Counseling Psychology Quarterly, 4,301-310. Cameron, S. J., & Orr, R. R. (1989). Stress in families of school-aged children with delayed mental development. Canadian Journal of Rehabilitation, 2, 137-144. Campbell, M . , Locascio, J. J., Chorocco, M . C , Spencer, E. K. , Malone, R. P., Kafantaris, V. , et al. (1990). Stereotypies and tardive dyskinesia: Abnormal movements in autistic children. Psychopharmalogical Bulletin, 26, 260-266. Carr, E. G., Langdon, N . A. , & Yarbrough, S. C. (1999). Hypothesis-based intervention for severe problem behavior. In A. Repp & R. H . Horner (Eds.), Functional analysis of problem behavior (pp. 7-28). Belmont, C A : Wadsworth. CattelUGordon, D. & Cattell-Gordon, D. (1998). The development of an.effective applied behavior analysis program for a young child with autism: A parent's perspective. Infants and Young Children, 10, 79-85. Chadwick, O., Cuddy, M . , Kusel, Y . , & Taylor, E. (2005). Handicaps and the development of skills between childhood and early adolescence in young people with intellectual disabilities. Journal of Intellectual Disability Research, 49, 877- 888. Chakrabarti, S. & Fombonne, E. (2005). Pervasive developmental disorders in preschool children: Confirmation of high prevalence. American Journal of Psychiatry, 162, 1133-1141. Charman, T., Baron-Cohen, S., Swettenham, J., Baird, G., Drew, A. , & Cox, A . (2003). 132 Predicting language outcome in infants with autism and pervasive developmental disorder. International Journal of Language and Communication Disorders, 38, 265-285. Charman, T. & Swettenham, J. (2001). Repetitive behaviors and social-communicative impairments in autism: Implications for developmental theory and diagnosis. In J.A. Burack, T. Charman, N . Yirmiya, & P.R. Zelazo (Eds.), The development of autism: Perspectives from theory and research. Mahwah, New Jersey. Lawrence Erlbaum Associates. Charman, T;, Taylor, E., Drew, A. , Cockerill, H. , Brown, J., & Baird, G. (2005). Outcome at 7 years of children diagnosed with autism at age 2: Predictive validity of assessments conducted at 2 and 3 years of age and pattern of symptom change over time. Journal of Child Psychology and Psychiatry, 46, 500-513. National Institute of Mental Health (1985). Clinical global impressions. Psychopharmacology Bulletin, 21, 839-843. Cornish, E. (1998). A balanced approach towards healthy eating in autism. Journal of Human Nutrition and Dietetics, 11, 501-509. Cronbach, L. J. (2004). M y current thoughts on coefficient alpha and successor procedures. Educational and Psychological Measurement, 64, 391-418. Dadds, M . , Schwartz, S., Adams, T., & Rose, S. (1988). The effects of social context and verbal skill on stereotypic and task-involved behaviour of autistic children. Journal of Child Psychology and Psychiatry, 29, 669-676. Dawson, G., Toth, K. , Abbott, R., Osterling, J., Munson, J., Estes, A. , et al. (2004). Early social attention impairments in autism: Social orienting, joint attention, and attention to distress. Developmental Psychology, 40, 271-283. De Meyer, M . K . (1979). Parents and children with autism (Vol. Wiley): New York. DeMyer, M . K . (1973). Prognosis in autism: A follow-up study. Journal of Autism and Childhood Schizophrenia, 3, 199-246. Dissanayake, C , Sigman, M . , & Kasari, C. (1996). Long-term stability of individual differences in the emotional responsiveness of children with autism. Journal of Child Psychology and Psychiatry, 37, 461-467. Dunn, W. (1999). Sensory Profile: User's Manual. The Psychological Corporation. Dunn, L . M . & Dunn, L . M . (1997). Examiner's manual for the PPVT-III: Peabody picture vocabulary test - third edition: Form IIIA and form IIIB. Circle Pines: American Guidance Service. Dunn, M . E., Burbaine, T., Bowers, C. A. , & Tantleff-Dunn, S. (2001). Moderators in stress in parents with children with autism. Community Mental Health Journal, 37,39-51. Dunst, C. J., Trivette, C. M . , Hamby, D., & Pollock, B. (1990). Family systems correlates of the behavior of young children with handicaps. Journal of Early Intervention, 74,204-218. Durand, V. M . , & Merges, E. (2001). Functional communication training: A contemporary behavior analytic technique for problem behavior. Focus on Autism and Other Developmental Disabilities, 16, 110-119. Dyson, L. L. (1996). The experiences of families of children with learning disabilities: Parental stress, family functioning, and sibling self-concept. Journal of Learning Disabilities, 2°, 281-288. . Eaves. L.C. & Ho, H.H. (1996). Brief report: Stability and change in cognitive and behavioral characteristics of autism through childhood. Journal of Autism and 134 Developmental Disabilities, 26, 557-569. Eaves, L. C. & Ho, H . H. (1997). School placement and academic achievement in children with autistic spectrum disorders. Developmental and Physical Disabilities, 9, 277-292. Edelson, M . G. (1998). Factors predicting intelligence scores on the TONI in individuals with autism. Focus on Autism and Other Developmental Disabilities, 13, 17-27. Edelson, M . G., Schubert, D. T., & Edelson, S. M . (1998). Factors predicting intelligence scores on the TONI in individuals with autism. Focus on Autism and Other Developmental Disabilities, 13, 17-26. Epstein, L. J., Taubman, M . T., & Lovaas, O. I. (1985). Changes in self-stimulatory behaviors with treatment. Journal of Abnormal Psychology, 73, 281 -294. Feinfield, K . A . & Baker, B.L . (2004). Empirical support for a treatment program for families of young children with externalizing problems. Journal of Clinical Child and Adolescent Psychology, 33, 182-195. Fenske, E. C , Zalenski, S., & Krantz, P. J. (1985). Analysis and intervention in developmental disabilities. Early Intervention: Special Issue, 5, 49-58. Fenson, L., Dale, P.S., Reznick, S., Thai, D., Bates, E., Hartung, J.P. et al. (1994). MacArthur communicative development inventories: User's guide and technical manual. SanDiego: Singular Publishing Group. Ferster, C. B., & Demyer, M . K . (1962). A method for the experimental analysis of behavior of autistic children. American Journal of Orthopsychiatry, 32, 89-98. Field, T., Lasko, D., Mundy, P., Henteleff, T., Kabat, S., Talpins, S., & Dowling, M . (1997). Brief report: Autistic children's attentiveness and responsivity improve after touch therapy. Journal of Autism and Developmental Disorders, 27, 333- 135 338. Floyd, F. J., & Gallagher, E. M . (1997). Parental stress, care demands, and use of support services for school-aged children with disabilities and behavior problems. Family Relations, 46, 359-371. Francis, D. F., Fletcher, J. M . , Stuebing, K . K. , Davidson, K . C., & Thompson, N . M . (1991). Analysis of change: Modeling individual growth. Journal of Consulting and Clinical Psychology, 59, 27-37. Gabriels, R. L. , Cuccaro, M . L., Hi l l , D. E., Ivers, B. J., & Goldson, E. (2005). Repetitive behaviors in autism: Relationships with associated clinical features. Research in Developmental Disabilities, 26, 169-181. Gabriels, R. L. , Hi l l , D. E., Pierce, R. A. , & Rogers, S. J. (2001). Predictors of treatment outcome in young children with autism: A retrospective study. Autism: The International Journal of Research & Practice, 5, 407-429. Gilliam, J.E. (1995). Gilliam Autism Rating Scale (GARS). Austin, TX. PRO-ED. Goldstein, H. (2002). Communication intervention for children with autism: A review of treatment efficiency. Journal of Autism and Developmental Disorders, 32, 373- 396. Golstein, S., & Schwebach, A. J. (2004). The comorbidity of pervasive developmental disorder and attention deficit hyperactive disorder: Results of a retrospective chart review. Journal of Autism and Developmental Disorders, 34, 329-339. Goyette, C. H. , Connors, C. K. , & Ulrich, R. F. (1978). Normative data on revised conners parent and teacher rating scales. Journal of Abnormal Psychology, 6,221- 236. Gray, D. E. (1994). Coping with autism: Stresses and strategies. Sociology of Health and Illness, 16, 275-300. Harris, S. L., & Handleman, J. S. (2000). Age and IQ at intake as predictors of placement for young children with autism: A four- to six-year follow-up. Journal of Autism and Developmental Disorders, 30, 137-142. Hastings, R. P. (2002). Parental stress and behavior problems of children with developmental disability. Journal of Intellectual and Developmental Disability, 27, 149-160. Hastings, R. P. (2003). Child behavior problems and partner mental health as correlates of stress in mothers and fathers of children with autism. Journal of Intellectual Disability Research, 47', 231-237'. Hastings, R. P., & Johnson, E. (2001). Stress in U K families conducting intensive home- based behavioral intervention for their young child with autism. Journal of Autism and Developmental Disorders, 31, 327-336. Hassall, R., & McDonald, J. (2005). Parenting stress in mothers of children with an intellectual disability: The effects of parental cognitions in relation to child characteristics and family support. Journal of Intellectual Disability Research, 49, 405-418. Hauser-Cram, P., Warfield, M . E., Shonkoff, J. P., & Krauss, M . W. (2001). Children with disabilities: A longitudinal study of child development and parent well- being. Monographs of the Society for Research in Child Development, 66(3, Serial No. 266). Hodapp, R. M . , Dykens, E. M . , & Masino, L. L. (1997). Families of children with Prader- Will i syndrome: Stress-support and relations to child characteristics. Journal of Autism and Developmental Disorders, 27, 11-24. Hodapp, R. M . , Fidler, D. J., & Smith, A. C. M . (1998). Stress and coping in families of children with Smith-Magenis syndrome. Journal of Intellectual Disability Research, 42,331-340. Hoffman, C D . , Sweeney, D.P., Gilliam, J.E., Apodaca, D.D., Lopez-Wagner, M.C. , & Castillo, M . M . (2005). Sleep problems and symptomology in children with autism. Focus on Autism and Other Developmental Disabilities, 20, 194-200. Hollingshead, A . B . (1962). Socioeconomic status and mental illness. Sociology and Social Research, 46,387-396. Holroyd, K. A. , & Lazarus, R. S. (1982). Stress, coping, and somatic adaptation. In L. Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and clinical aspects. New York: The Free Press. Horner, R. H. , Carr, E. G., Strain, P. S., Todd, A. W., & Reed, H. K . (2002). Problem behavior interventions for young children with autism: A research synthesis. Journal of Autism and Developmental Disorders, 32, 423-446. Ingersoll, B., Schreibman, L. , & Stahmer, A. (2001). Brief report: Differential treatment outcomes for children with autistic spectrum disorder based on level of peer social avoidance. Journal of Autism and Developmental Disorders, 31, 343-349. Jensen, G. D., & Womack, M . G. (1967). Operant conditioning techniques applied in the treatment pf an autistic child. American Journal of Orthopsychiatry, 37, 30-34. Johnston, C , Hessl, D., Eliez, S., Erba, H. , Dyer-Friedman, J., Glaser, B. , et al. (2003). Factors associated with parenting stress in mothers of children with fragile x syndrome. Developmental and Behavioral Pediatrics, 24, 267-275. Joreskog, K . G . & Sorbom, D. (1989). LISREL 7: A guide to the program and applications (2 n d ed.). Chicago: SPSS Jung-Chang, T., Kennedy, C. H. , Koppekin, A. , & Caruso, M . (2002). Functional analysis of stereotypical ear covering in a child with autism. Journal of Applied Behavior Analysis, 35, 95-98. Kaiser, B., & Rasminsky, J. S. (2003). Challenging behavior in young children: Understanding, preventing, and responding effectively. Boston: Allyn and Bacon. Kanner, L. (1943). Autistic disturbances of affective contact. Nervous Child, 2, 217-250. Keith, T.Z. (1993). Latent variable structural equation models: LISREL in special education research. Remedial and Special Education Research, 14, 36-46. Keller, D.W. (1999). Dimensions of parenting stress of mothers and fathers of a school- aged child with a disability. Dissertation Abstracts International Section A: Humanities and Social Sciences, 60, 1347. Kemp, A.S. (2003). Cost of illness of atopic dermatitis in children: A societal perspective. Pharmacoeconomics, 21, 105-113. Kluth, P. (2003). Chapter 1: Defining autism. In P. Kluth (Ed.), You're going to love this kid! Teaching students with autism in the inclusive classroom (pp. 1-22). Baltimore: Paul H . Brookes. Koegel, R.L., Schreibman, L. , Johnson., J., O'Neill, R.E., & Dunlap, G. (1984). Collateral effects of parent training on families with autistic children. In R.F. Dangel & R A . Polster (Eds), Parent training: Foundations of research and practice (pp. 358-378). New York: Guilford. Koegel, R.L., Schriebman, L., Loos, L . M . , Dirlich-Wilhelm, H. , Dunlap, G., Robbins, F.R., & Plienis, A.J . (1992). Consistent stress profiles in mothers of children with autism. Journal of Autism and Developmental Disorders, 22, 205-216. Konstantareas, M . M . , & Homatidis, S. (1989). Assessing child symptom severity and 139 stress in parents of autistic children. Journal of Child Psychology and Psychiatry, 30, 459-470. Kraemer, H. C , Stice, E., Kazdin, A. , Offord, D., & Kupfer, D. (2001). How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. American Journal of Psychiatry, 158, 848-856. Krug, D.A., Arick, J.R., & Almond, P J . (1980/1983). Autism screening instrument for educational planning (2 n d ed.). Austin, TX: PRO-ED. Le Couteur, A. , Lord, C , & Rutter, M . (2003). Autism diagnostic interview-revised. Los Angeles, C A : Western Psychological Services. Lee, S. & Odom, S.L. (1996). The relationship between stereotypic behavior and peer social interaction for children with severe disabilities. Journal for the Association of Persons with Severe Handicaps, 21, 88-95. Liss, M . , Harel, B., Fein, D., Allen, D., Dunn, M . , Feistein, C , et al. (2001). Predictors and correlates of adaptive functioning in children with developmental disorders. Journal of Autism and Developmental Disorders, 31, 219-230.. Lord, C , Rutter, M . , DiLavore, P. C , & Risi, S. (2002). Autism diagnostic observation schedule. Los Angeles, CA: Western Psychological Services. Lotter, V . (1966). Epidemiology of autistic conditions in young children:. Prevalence. Social Psychiatry, 1, 124-137. Lovaas, O. I. (1974). A behavior modification approach to the treatment of autistic children. Journal of Autism and Childhood Schizophrenia, 4, 111-129. Lovaas, O. I. (1987). Behavioral treatment and normal educational and intellectual functioning in young autistic children. Journal of Consulting and Clinical Psychology, 55, 3-9. Lovaas, O. I., & Koegel, R. L. (1973). Behavior therapy with autistic children. National society for the study of education yearbook 72 Part 1, 230-258. Lovaas, O. I., Koegel, R. L., Simmons, J. Q., & Long, J. S. (1973). Some generalization and follow-up measures on autistic children behavior therapy. Journal of Applied Behavior Analysis, 6, 131-166. Lovaas, O. I., & Smith, T. (1998). Intensive and early behavioral intervention with autism: The U C L A young autism project. Infants and Young Children, 10, 67-78. Lucyshyn, J.M., Dunlap, G., & Albin, R.W. (Eds.) (2002). Families and positive behavior support: Addressing problem behavior in family contexts. Baltimore, Maryland: Paul H . Brooks McConnell, S.R. (2002). Interventions to facilitate social interaction for young children with autism: review of available research and recommendations for educational intervention and future research. Journal of Autism and Developmental Disorders, 52,351-372. McCubbin, H . I., & Patterson, J. M . (1982). Family adaptation to crisis. In H . I. McCubbin, A. E. Cauble & J. M . Patterson (Eds.), Family stress, coping and social support. Springfield, II: Charles C. Thomas. McCubbin, H. I., & Patterson, J. M . (1983). The family stress process: The double A B C X model of adjustment and adaptation. Marriage and Family Review, 6, 7-37. McCubbin, M . A. , & McCubbin, H. I. (1991). Family stress theory and assessment: The resiliency model of family stress, adjustment, and adaptation. In H . I. McCubbin & A. I. Thompson (Eds.), Family assessment inventories for research and practice (pp. 3-32). Madison: University of Wisconsin-Madison. McCubbin, H.I.,.Patterson, J. M . , & Wilson, L.R. (1996). FILE: Family inventory of life events and changes. In H . McCubbin, A. Thompson, & M ! McCubbin, Family assessment: Resiliency, coping, and adaptation: Inventories for research and practice (pp. 103-178). Madison, Wisconsin. University of Wisconsin Publishers. McCubbin, H.I., Olson, D.H. & Larsdn, A.S. (1996). F-COPES: Family crisis oriented personal evaluation scales. In H . McCubbin, A . Thompson, & M . McCubbin, Family assessment: Resiliency, coping, and adaptation: Inventories for research and practice (pp.455-507). Madison, Wisconsin. University of Wisconsin Publishers. McDonald, T. P., & Gregoire, T. K. (1997). Building a model of family caregiving for" . children with emotional disorders. Journal of Emotional and Behavioral Disorders, 5, 138-148. Militerni, R., Bravaccio, C , Falco, C , Fico, C , Palermo, M.T. (2002). Repetitive behaviors in autistic disorder. European Child and Adolescent Psychiatry, 11, 210-218. Mirenda, P. (2001). Parent satisfaction questionnaire, (adapted from Smith, T., Parent Satisfaction Questionnaire, University of Rochester). Vancouver, BC: University of British Columbia. Mirenda, P., & Smith, V . (2001). Preschool lifestyle inventory, (adapted from the Lifestyle Assessment in Horner, R., Newton, S., Ard, Jr., W, Olson, D., Sappington, G., Kerber, G., Maley, M . , Anderson, S., & Spoelstra. R. (1994). A residential outcomes system operations manual (pp. 2.1-2.51). Eugene, OR, Specialized Training program, University of Oregon). Vancouver, B C : University of British Columbia. 142 Mullen, E . M . (1995). Mullen scales of early learning. Circle Pines, M N : American Guidance Service Inc. Neisworth, J.T., Bagnato, S.T., Salvia, J., & Hunt, F . M . (1999). TABS manual for the temperament and atypical behavior scale: Early childhood indicators of developmental dysfunction. Baltimore: Paul H . Brookes. O'Moore, M . (1978). Living with autism. Irish Journal of Psychology, 4, 33-52. Perry, A. , Condillac, R.A., Freeman, N.L. , Dunn Geier, J . , & Belair, J. (2003, February). Multi-site study of the Childhood Autism Rating Scale (CARS) in four clinical groups of young children. Presentation at the Canadian Autism Intervention Research Network (CAIRN) Annual Meeting, Toronto, ON. Pilowsky, T., Yirmiya, N . , Shulman, C , & Dover, R. (1998). The autism diagnostic interview-revised and the childhood autism rating scale: Differences between diagnostic systems and comparison between genders. Journal of autism and Developmental Disorders, 28, 143-151. Podolski, C.-L., & Nigg, J. T. (2001). Parent stress and coping in relation to child ADHD severity and associated child disruptive behavior problems. Journal of Clinical Child Psychology, 30, 503-513. Polirstok, S. R., Dana, L. , Buono, S., Mongelli, V. , & Trubia, G. (2003). Improving functional communication skills in adolescents and young adults with severe autism using gentle teaching and positive approaches. Topics in Language Disorders, 23, 146-153. Prizant, B. M . , & Rubin, E. (1999). Contemporary issues in interventions for autism spectrum disorders: A commentary. Journal of the Association for Persons with Severe Handicaps, 24, 199-208. Quine, L., & Pahl, J. (1985). Examining the causes of stress in families with severely mentally handicapped children. British Journal of Social Work, 15, 501-517. Quine, L., & Pahl, J. (1991). Stress and coping in mothers caring for a child with severe learning difficulties: A test of Lazarus' transactional model of coping. Journal of Community and Applied Social Psychology, 1, 57-70. Richdale, A . L . (1999). Sleep problems in autism: Prevalence, cause, and intervention. • Developmental Medicine and Child Neurology, 41, 60-66. Richdale, A. , Francis, A. , Gavidia-Payne, S., & Cotton, S. (2000). Stress, behavior, and sleep problems in children with intellectual ability. Journal of Intellectual and Developmental Disability, 25, 147-161. Rogers, S. J. (1998). Empirically supported comprehensive treatments for young children with autism. Journal of Clinical Child Psychology, 27, 168-179. Rollins, P. R., & Snow, C. E. (1998). Shared attention and grammatical development in typical children and children with autism. Journal of Child Language, 25, 653- 673. Ross, C , N . , & Blanc, H. M . (1998). Parenting stress in mothers of young children with oppositional defiant disorder and other severe behavior problems. Child Study Journal, 25,93-111. Saemundsen, E., Magmisson, P., Smari, J. & Sigurdardottir, S. (2003). Autism diagnostic interview-revised and the childhood autism rating scale: Convergence and discrepancy in diagnosing autism. Journal of Autism and Developmental Disorders, 33,319-328. Sallows, G.O & Graupner, T.D. (2005). Intensive behavioral treatment for children with autism: Four-year outcome and predictors. Journal on Mental Retardation, 110, 144 ' 417-438. Saloviita, T., Italinna, M . , & Leinonen, E. (2003). Explaining the parental stress of fathers and mothers caring for a child with intellectual disability: A double A B C X model. Journal of Intellectual Disability Research, 47, 300-312. Schreck, K. A. , Mulick, J. A . , & Smith, A. F. (2004). Sleep problems, as possible predictors of intensified symptoms of autism. Research in Developmental Disabilities, 25, 57-66. Schreck, K. A. , Williams, K. , & Smith, A. F. (2004). A comparison of eating behaviors between children with and without autism. Journal of Autism and Developmental Disorders, 34, 433-438. Schreibman, L. (2000). Intensive behavioral/psychoeducational treatments for autism: Research needs and future directions. Journal of Autism and Developmental Disorders, 30, 373-378. Schopler, E., Reichler, R.J., & Rentier, B.R. (1988). The childhood autism rating scales (CARS). Los Angeles, C A : Western Psychological Services. Schumaker, R.E. & Lomax, R.G. (2004). A beginner's guide to structural equation modeling (2 n d ed.). Mahwah, New Jersey: Lawrence Erlbaum Associates. Sigman, M & McGovern, C. (2005). Improvement in cognitive and language skills from preschool to adolescence in autism. Journal of Autism and Developmental Disorders, 35, 15-23. Sigman, M . & Ruskin, E. (1999). Continuity and change in the social competence of children with autism, down syndrome, and developmental delays. Monographs of the Society for Research in Child Development, 64(\, Serial No. 256). Sigman, M . D., Kasari, C , Kwon, J.-h., & Yirmiya, N . (1992). Responses to the negative 145 emotions of others by autistic, mentally retarded, and normal children. Child Development, 63, 796-807. Sigafoos, J., & Mirenda, P. (2002). Strengthening communicative behaviors for gaining access to desired items and activities. In J. Reichle, D. Beukelman, & J. Light (Eds.). Exemplary practices for beginning communicators: Implications for AAC (pp. 123-156). Baltimore: Paul H . Brookes. Siller, M . , & Sigman, M . (2002). The behaviors of parents of children with autism predict the subsequent development of their children's communication. Journal of Autism and Developmental Disorders, 32,77-89. Smith, T., Eikeseth, S., Klevstrand, M . , & Lovaas, O. I. (1997). Intensive behavioral treatment for preschoolers with severe mental retardations and pervasive developmental disorder. American Journal of Mental Retardation, 102, 238-249. Smith, T. B., Oliver, M . N . I., & Innocenti, M . S. (2001). Parenting stress in families of children with disabilities. American Journal of Orthopsychiatry, 71, 257-260. Smith, V. , Zaidman-Zait, A. , & Mirenda, P. (2005). Development of Language in Autism: A Longitudinal Look at Vocabulary Development Using the Communication Development Inventory (CDI). Paper accepted for presentation at the Jean Piaget Society Annual Meeting, Vancouver, British Columbia. SoHyun, L. & Odom, S.L. (1996). The relationship between stereotypic behavior and peer social interaction for children with severe disabilities. Journal of the association for Persons with Severe Handicaps, 21, 88-95. Sparrow, S., Balla, D., & Cicchetti, D. (1984). Vineland adaptive behavior scales: Expanded manual. Circle Pines, M N : American Guidance service. Streiner, D.L. & Norman, G.R. (1989). Health Measurement Scales: A Practical Guide to 146 Their Development and Use. New York: Oxford University Press. Stoddart, K. P. (2003). Reported stress, personality, and mental health in parents of children with pervasive developmental disorders. Dissertation Abstracts International: Section A: Humanities and Social Sciences, 64, 1411. Stone, W. L. , & Yoder, P. (2001). Predicting spoken language level in children with autism spectrum disorders. Autism: The International Journal of Research & Practice, 5, 341-362. Stores, R., Stores, G., Fellows, B., & Buckley, S. (1998). Daytime behaviour problems and maternal stress in children with down's syndrome, their siblings, and non- intellectually disabled and other intellectually disabled peers. Journal of Intellectual Disability Research, 42,228-237. Strain, P. S., Kohler, F. W., & Goldstein, H . (1996). Learning experiences. A alternative program: Peer mediated interventions for young children with autism. In E. Hibbs & P. Jensen (Eds.), Psychosocial treatments for child and adolescent disorders: Empirically based strategies for clinical practice (pp. 573-586). Washington, DC: American Psychological Association. Szatmari, P., Archer, L. , Fisman, S., & Streiner, D.L. (1994). Parent and teacher agreement in the assessment of pervasive developmental disorders. Journal of Autism and Developmental Disorders, 24, 703-7-17.- Szatmari, P., Bryson, S. E., Boyle, M . H. , Streiner, D. L., & Duku, E. (2003). Predictors of outcome among high functioning children with autism and asperger syndrome. Journal of Child Psychology and Psychiatry, 44, 520-528. Tebruegge, M . , Nandini, V. , & Ritchie, J. (2004). Does routine child health surveillance contribute to the early detection of children with pervasive developmental 147 disorders? An epidemiological study in Kent, U.K. , BMC: Pediatrics. Tomanik, S., Harris, G.E., & Hawkins, J. (2004). The relationship between behaviours exhibited by children with autism and maternal stress. Journal of Intellectual and developmental Disability, 29, 16-26. Travis, L., Sigman, M . , & Ruskin, E. (2001). Links between social understanding and social behavior in verbally able children with autism. Journal of Autism and Developmental Disorders, 31, 119-130. Trivette, C M . & Dunst, C.J. (1988). Inventory of social support. In C. Dunst, C. Trivette, & A. Deal, Enabling and empowering families: Principles and guidelines for practice (ppl59-163). Cambridge, M A . Brookline Books. Turner, M . (1999). Annotation: Repetitive behavior in autism: A review of psychological research. Journal of Child Psychology and Psychiatry, 40, 839-849. VanRiper, M . (2001). Family variables associated with well-being in siblings of children with down syndrome. Journal of Family Nursing, 6, 267-287. Venter, A., Lord, C , & Schopler, E. (1992). A follow-up study of high functioning autistic children. Journal of Child Psychology and Psychiatry, 33, 489-507. Volkmar, F. R., & Pauls, D. (2003). Autism. The Lancet, 362, 1133-1141. Wechsler, D. (1989). The Wechsler preschool and primary scale of intelligence-revised: (WPPSI-R). San Antonio: The Psychological Corporation. Wechsler, D. (1991). The Wechsler intelligence scale for children -3rd ed: (WISC-III). San Antonio: The Psychological Corporation. Weiss, J. A. , Sullivan, A. , & Diamond, T. (2003). Parent stress and adaptive functioning of individuals with developmental disabilities. Journal of Developmental Disabilities, 10, 129-135. 148 Wiggs, L. & Stores, G. (1996). Severe sleep disturbance and daytime challenging behaviour in children with severe learning disabilities. Journal of Intellectual Disability Research, 40, 518-528. Wiggs, L. & Stores, G. (2001). Behavioural treatment for sleep problems in children with severe intellectual disabilities and daytime challenging behaviour: Effect on mothers and fathers. British Journal of Health Psychology, 6, 257-269. Wiggs, L. & Stores, G. (2004). Sleep patterns and sleep disorders in children with autistic spectrum disorders: Insights using parent report and actigraphy. Developmental Medicine and Child Neurology, 46, 372-380. Willett, J. B., & Sayer, A. G. (1994). Using covariance structure analysis to detect correlates and predictors of individual growth over time. Psychological Bulletin, 776,363-381. Williams, P. G., Dalrymple, N . J., & Neal, J. (2000). Eating habits of children with autism. Pediatric Nursing, 26, 259-264. Williams, P.G., Sears, L .L . , & Allard, A . (2003). Sleep problems in children with autism. Journal of Sleep Research, 13, 265-268. Woods, J. J., & Wetherby, A . M . (2003). Early identification of and intervention for infants and toddlers who are at risk for autism spectrum disorder. Speech, Language, and Hearing Services in Schools, 34, 180-193. Yarrow, L. J., McQuiston, S., MacTurk, R. PL, McCarthy, M . E., Klein, R. P., & Vietze, P. M . (1983). Assessment of mastery motivation during the first year of life: Contemporaneous and cross-age relationships. Developmental Psychology, 19, 159-171. Yuan, K. (2005). Fit indices versus test statistics. Multivariate Behavioral Research, 40, 149 115-148. Zimmerman, I.L., Streiner, V .G. , & Pond, R.E. (1992). PLS-3: Preschool Language Scale-3. San Antonio: The Psychological Corporation Harcourt, Brace Jovanovich, Inc. Zumbo, B. D., Gelin, M . N . , & Hubley, A . M . (2002). The construction and use of psychological tests and measures. In The psychology theme of the encyclopedia of life support systems (EOLSS).Oxford, U K : EOLSS Publishers. 150 Appendix A : A l l Items Located for the Six Child Behaviour Variables A l : Acting-out items Measure Item Coding A B C PSI-SF Severe temper tantrums and/or frequent minor tantrums Hurts self by banging head, biting hand etc. Hurts others by biting, kicking, hitting Is very destructive (toys and household items are soon broken) #25: M y child seem to cry or fuss more often than most children #26: M y child generally wakes up in a bad mood #27:1 feel that my child is moody and easily upset #28: my child does a few things which bother me a great deal yes/no yes/no yes/no yes/no strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree A l continues 151 A l (Continued) Measure Item Coding PSI-SF Sensory Profile #29: my child reacts very strongly when something happens that my child doesn't like. #30: my child gets upset over the smallest things #34: there are some things that my child does that really bother me a lot #35 my child turned out to be more of a problem than I had expected #36: M y child makes more demands on me then most children #105: displays excessive emotional outbursts when unsuccessful at a task strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree strongly agree/ agree/ not sure/ disagree/ strongly disagree always/ frequently/ occasionally/ seldom/ never #107: is stubborn or uncooperative always/ frequently/ occasionally/ seldom/ never #108: has temper tantrums always/ frequently/ occasionally/ seldom/ never #109: poor frustration tolerance always/ frequently/ occasionally/ seldom/ never A l continues 152 A l (Continued). Measure Item Coding Sensory Profile TABS #110: cries easily always/ frequently/ occasionally/ seldom/ never #21: upset by every little thing yes/ no #22: often difficult to soothe when yes/ no upset or crying #23: has wide swings in mood yes/ no #24: gets angry too easily yes/ no #25: too easily frustrated yes/ no #26: has wild temper tantrums yes/ no #27: frequently irritable, "touchy" or yes/ no fussy #30: controls adult's behavior, "is yes/ no the boss" #31: jealous too often yes/ no #35: almost always refuses to do yes/ no what is told #36: throws breaks things on yes/ no purpose #37: bites, hits, kicks others yes/ no #49: often cries too long yes/ no #52: can't comfort self when upset yes/ no 153 A2: Sleep disturbances Measure Item Coding Sensory #113: Has nightmares always/ frequently/ occasionally/ Profile seldom/ never TABS #50: often frightened by dreams or yes/ no the nighttime #51: screams in sleep and can't be yes/ no comforted #53: wakes up often and doesn't fall yes/ no back asleep #54: doesn't have a regular sleep yes/ no schedule #55: too often needs help to fall yes/ no asleep A3: Eating difficulties Measure Item Coding Sensory #54: gags easily with food textures or always/ frequently/ occasionally/ Profile food utensils in mouth seldom/ never #55: avoids certain tastes or food always/ frequently/ occasionally/ smells that are typically part of seldom/ never children's diet A3 continues 154 A3 (Continued) Measure Item Coding Sensory Profile #56: will only eat certain tastes always/ frequently/ occasionally/ seldom/ never #57: limits self to particular food always/ frequently/ occasionally/ textures/temperatures seldom/ never #58: picky eater, especially regarding always/ frequently/ occasionally/ food textures seldom/ never #60: shows strong preference for certain smells #61: shows strong preference for certain tastes #62: craves certain foods always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never #63: seeks out certain tastes or smells always/ frequently/ occasionally/ seldom/ never A 4 : Stereotypic behaviour Measure Item Coding A B C Whirls self for long periods of time yes/ no Does not use toys appropriately yes/ no (spins tires) Insists on keeping certain objects yes/ no with him/her Rocks self for long periods of time yes/ no Strong reactions to changes in yes/ no routine/environment Does a lot of lunging and darting yes/ no about, interrupting with spinning, toe walking, flapping, etc. Flaps hands yes/ no Walks on toes yes/ no Repeats phrases over and over yes/ no Twirls, spins and bangs objects a lot yes/ no Squints, frowns, or covers eyes when yes/ no in the presence of natural light Repeats sounds or words over and yes/ no over A 4 continues 156 A4 (Continued) Measure Item Coding A B C PSI Sensory Profile Echoes questions or statements made yes/ no by others Prefers to manipulate and be yes/ no occupied with inanimate things Wil l feel, smell, and/or taste objects yes/ no in the environment Gets involved in complicated rituals yes/ no such as lining things up, etc. Stares into space for long periods of yes/ no time #21: It takes a long time and it is strongly agree/ agree/ not sure/ very hard for my child to get used to disagree/ strongly disagree new things #8: enjoys strange noises/seeks to make noise for noises sake always/ frequently/ occasionally/ seldom/ never #15: Covers eyes or squints to protect eyes from light always/ frequently/ occasionally/ seldom/ never #16: looks intensely at objects/people always/ frequently/ occasionally/ (for example stares) seldom/ never A4 continues 157 A4 (Continued) Measure Item Coding Sensory Profile #24: seeks all kinds of movement activities and this interferes with daily routines (for example can't sit still, fidgets) #25: seeks out all kinds of movement activities (for example being whirled by adult, merry-go-rounds, playground equipment, moving toys) #26: Twirls/spins self frequently throughout the day (for example likes feeling dizzy) #27: rocks unconsciously (for example, while watching TV) #28: rocks in desk/chair/on floor always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never #40: touches people or objects to the always/ frequently/ occasionally/ point of irritating others seldom/ never A4 continues 158 A4 (Continued) Measure Item Coding Sensory #41: displays unusual need for always/ frequently/ occasionally/ Profile touching certain toys, surfaces, or seldom/ never textures (for example, constantly touching objects) #52: Walks on toes always/ frequently/ occasionally/ seldom/ never #59: Routinely smells nonfood always/ frequently/ occasionally/ objects seldom/ never #64: Chews or licks on nonfood always/ frequently/ occasionally/ objects seldom/ never #90: "on the go" always/ frequently/ occasionally/ seldom/ never #97: stares intensively at objects or always/ frequently/ occasionally/ people seldom/ never #121: Has difficulty tolerating always/ frequently/ occasionally/ changes in plans and expectations seldom/ never #122: Has difficulty tolerating always/ frequently/ occasionally/ changes in routines seldom/ never #124: Deliberately smells objects always/ frequently/ occasionally/ seldom/ never A4 continues 159 A4 (Continued) Measure Item Coding TABS- #1: Consistently upset by changes in schedule yes/ no #9: often stares into space yes/ no #13: makes strange throat noises yes/ no • #16: stares at lights yes/ no #17: overly interested in toy/object yes/ no #18: Flaps hands over and over yes/ no #19: shakes head over and over yes/ no A5: Social unresponsiveness Measure Item Coding A B C Has no social smile yes/ no Does not (or did not as a baby) yes/ no reach out when reached for Does not respond to own name yes/ no when called out among two others Not responsive to other peoples' yes/ no facial expressions feelings Is (or was as a baby) stiff and hard yes/ no to hold A5 continues 160 A5 (Continued) Measure Item Coding A B C Is flaccid (doesn't cling) when held yes/ no in arms Frequently has no visual reaction to yes/ no a "new" person PSI #15: M y child smiles at me much strongly agree/agree/not sure/ less than I expected disagree/ strongly disagree #17: When playing, my child does strongly agree/ agree/ not sure/ not often giggle or laugh disagree/ strongly disagree #19: M y child doesn't seem to smile strongly agree/agree/not sure/ as much as most children disagree/ strongly disagree Sensory Profile #6: Appears to not hear what you always/ frequently/ occasionally/ say (for example, does not "tune-in" seldom/ never to what you say, appears to ignore you) #7: Does not respond when name is always/ frequently/ occasionally/ called but you know child is hearing seldom/ never OK #94: Is overly affectionate with always/ frequently/ occasionally/ others seldom/ never A5 continues 161 A5 (Continued) Measure Item Coding Sensory Profile #95: Doesn't perceive body language or facial expressions (for example, unable to interpret) #116: Doesn't express emotions TABS going on #5: Acts like others are not there #8: Seems to be in "own world" #38: Rarely smiles, giggles, or laughs at funny things #42: Doesn't react to own name #43: Doesn't care when others are hurt V A B S #3: Smiles in response to presence Communication of caregiver #4: Smiles in response to presence of familiar person other than caregiver always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never #2: Emotions don't match what is yes/ no yes/ no yes/ no yes/ no yes/ no yes/ no yes/ sometimes/ no yes/ sometimes/ no A5 continues A5 (Continued) Measure Item Coding V A B S #5: Raises arms when caregiver yes/ sometimes/ no Communication says, "Come here" or "up" V A B S #2: Responds to voice of caregiver yes/ sometimes/ no Socialization or another person #5: expresses two or more yes/ sometimes/ no recognizable emotions such as, pleasure, sadness, fear, or distress #6: shows anticipation of being yes/ sometimes/ no picked up by caregiver #7: shows affection towards familiar yes/ sometimes/ no people #9: reaches for familiar person yes/ sometimes/ no #15: laughs or smiles appropriately yes/ sometimes/ no in response to positive statements A5 continues I 163 A5 (Continued) Measure Item Coding CARS III 4: Severely abnormal emotional yes/ no Emotional response: responses are seldom Response appropriate to situation; once the child gets in a certain mood, it is very difficult to change the mood. Conversely, the child may show wildly different emotions when nothing has changed. A6: Inattentiveness Measure Item Coding A B C Frequently does not attend to yes/ no social/environmental stimuli Seems not to hear so hearing loss yes/ no is suspected Sometimes shows no startle yes/ no response to a load noise (may have thought child was deaf) Actively avoids eye contact yes/ no "Looks through" people yes/ no A6 continues 164 A6 (Continued) Measure Item Coding A B C Sensory Profile Frequently unaware of yes/ no surroundings and may be oblivious to dangerous situations #3: Has trouble completing tasks always/ frequently/ occasionally/ when the radio is on #4: Is distracted or has trouble functioning i f there is a lot of noise seldom/ never around seldom/ never always/ frequently/ occasionally/ #5: Can't work with background noise (for example, fan refrigerator) #43: Doesn't seem to notice when someone touches arm or back (for example, unaware) #48: Has difficulty paying attention #49: Looks away from tasks to -notice all actions in the room always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never always/ frequently/ occasionally/ seldom/ never A6 continues 165 A6 (Continued) Measure Item Coding Sensory Profile #50: Seems oblivious within an always/ frequently/ occasionally/ TABS active environment (for example, unaware of activity) #96: Avoids eye contact seldom/ never always/ frequently/ occasionally/ seldom/ never #99: Doesn't notice when people always/ frequently/ occasionally/ come into the room #123: jumps from one activity to another so that it interferes with toy #32: mostly on the go, "in high gear" #33: doesn't sit still seldom/ never always/ frequently/ occasionally/ seldom/ never C play #3: Seems to look through or past yes/ no people #4: Resists looking you in the eye yes/ no #10: "Tunes out," loses contact yes/ no with what is going on #28: can't wait at all for food or yes/ no yes/ no yes/ no A6 continues 166 A6 (Continued) Measure Item Coding TABS #34: too "grabby," impulsive #39: Doesn't pay attention to sights and sounds #40: Doesn't seem to watch moving objects V A B S #10: Listens attentively to Communication instructions #17: Listens to a story for at least five minutes V A B S #1: Looks at face of caregiver Socialization PLS A C minutes Mullen: Scale 4 #8: attends to words and Receptive Language movements yes/ no yes/ no yes/ no yes/ sometimes/ no yes/ sometimes/ no yes/ sometimes/ no #9: Maintains attention for two yes/ no yes/ no B l : Acting-out Appendix B: Item Overrepresentation in Child Behaviour Variables Measure Item description Correlation PCA Item chosen to be included in variable? A B C TABS Sensory Profile Severe temper tantrums and/or frequent minor tantrums #26: has wild temper tantrums #108: has temper tantrums Only the A B C item and TABS #26 correlate at r 2>0.50 where r 2 = .574* p = .000 N / A Yes No Yes B l continues O N ^ 1 BI (Continued) Measure Item description Correlation xl = .642* p = .000 PCA Item chosen to be included in variable? A B C TABS Hurts others by biting, kicking, hitting #37: bites, hits, kicks others N/A Yes No A B C TABS Is very destructive (toys and household items are soon broken) #36: throws breaks things on purpose xl = .484* p = .000 N/A Yes No BI continues B l (Continued) Measure Item description Correlation PCA Item chosen to be included in variable? PSI-SF #25: M y child seems to cry or fuss Only PSI #25 N/A No more often than most children and TABS #49 Sensory Profile #110: cries easily correlate at Yes TABS #21: upset by every little thing r 2>0.50 Yes #22: often difficult to soothe when where Yes upset or crying r 2 = .473* #49: often cries too long p = .000 Yes #52: can't comfort self when upset Yes B l continues BI (Continued) Measure Item description Correlation PCA Item chosen to be included in variable? PSI-SF Sensory Profile #27:1 feel that my child is moody and easily upset #29: my child reacts very strongly when something happens that my child doesn't like. #30: my child gets upset over the smallest things #105: displays excessive emotional outbursts when unsuccessful at a task #109: poor frustration tolerance Only PSI #27 and TABS #24 correlate atr 2>0.50 where r 2 = .523* p = .000 N/A No Yes Yes Yes Yes BI continues BI (Continued) Measure TABS Item description #24: gets angry too easily #25: too easily frustrated Correlation see above PCA see above Item chosen to be included in variable? Yes Yes PSI-SF TABS #36: M y child makes more demands None of the on me then most children items across #30:controls adult's behavior, "is the measures boss" correlate #29: Demands attention continually at r 2 > 0.50 N / A Yes Yes Yes BI continues BI (Continued) Measure Item description Correlation PCA Item chosen to be included in variable? PSI-SF Sensory Profile TABS #26: M y child generally wakes up in a None of the bad mood items across #107: is stubborn or uncooperative #23: has wide swings in mood #27: frequently irritable, "touchy" or fussy #31: jealous too often #35: almost always refuses to do what is told measures correlate atr z>0.50 N/A Yes Yes Yes Yes Yes Yes N / A = not applicable B2: Sleep disturbances Measure Item description Correlation PCA Item Chosen to be included in variable? Sensory Profile #113: Has nightmares TABS #50: often frightened by dreams or the nighttime #51: screams in sleep and can't be comforted None of the N / A Yes items across Yes measures correlate Yes atr 2>0.50 N / A = not applicable B3: Stereotypic behaviour Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? A B C Rocks self for long periods of time Sensory Profile #27: rocks unconsciously (for example, while watching TV) None of the items across measures correlate atr 2>0.50 N / A Yes Yes A B C TABS Flaps hands #18: Flaps hands over and over r = .680* p = .000 N / A Yes No B3 continues 4^ B3 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? A B C Sensory Profile Whirls self for long periods of time #26: Twirls/spins self frequently throughout the day (for example likes feeling dizzy) None of the items across measures correlate atr 2>0.50 N / A Yes Yes A B C Sensory Profile Squints, frowns, or covers eyes when None of the in the presence of natural light items across #15: Covers eyes or squints to protect measures eyes from light correlate N / A Yes Yes atr'>0.50 B3 continues B3 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? A B C TABS TABS B'tiif Stares into space for long periods of time #9: often stares into space Sensory Profile #8: enjoys strange noises/seeks to make noise for noises sake #13: makes strange throat noises r = .528* p = .000 None of the items across measures correlate N / A N / A Yes No Yes Yes atr'>0.50 B3 continues B3 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? A B C Sensory Profile Walks on toes #52: Walks on toes r z = .480* p = .000 N/A. Yes No A B C Sensory Profile Sensory Profile Wil l feel, smell, and/or taste objects in None of the the environment items across N / A #40: touches people or objects to the point of irritating others #41: displays unusual need for touching certain toys, surfaces, or textures (for example, constantly touching objects) measures correlate atr'>0.50 Yes Yes Yes B3 continues B3 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? Sensory Profile Sensory Profile Sensory Profile #59: Routinely smells nonfood objects #64: Chews or licks on nonfood objects #124: Deliberately smells objects see above see above Yes Yes Yes Sensory Profile TABS #90: "on the go" #32: mostly on the go, "in high gear" None of the items across measures correlate N / A Yes Yes atr 2>0.50 B3 continues B3 (Continued) Measure Item Description - Possible Correlation PCA Item chosen to be Overrepresentation included in variable? A B C Strong reactions to changes in A B C and A B C = .854 Yes routine/environment sensory profile SP#121 = Sensory Profile #121: Has difficulty tolerating #121 correlate .803 No changes in plans and expectations atr 2=.552* TABS #1 = TABS #1: Consistently upset by changes in schedule p = .000 and A B C and TABS #1 correlate at r 2 = .498* p = .000 .766 No N / A = not applicable B4: Social unresponsiveness Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? A B C PSI-SF TABS Has no social smile #15: M y child smiles at me much less than I expected #17: When playing, my child does not often giggle or laugh #19: M y child doesn't seem to smile as much as most children #38: Rarely smiles, giggles, or laughs at funny things Only PSI # 17 and TABS #38 correlate at r2 = .456 p = .000 N/A Yes Yes No Yes Yes B4 continues B4 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? V A B S - Comm. #3: Smiles in response to presence of see above caregiver #4: Smiles in response to presence of familiar person other than caregiver V A B S - Soc #15: laughs or smiles appropriately in response to positive statements see above Yes Yes Yes A B C Does not respond to own name when called out among two others None of the items across measures correlate N / A Yes atr 2>0.50 B4 continues B4 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item chosen to be included in variable? Sensory Profile T A B S #6: Appears to not hear what you say (for example, does not "tune-in" to what you say, appears to ignore you) #7: Does not respond when name is called but you know child is hearing O K #42: Doesn't react to own name V A B S - Comm #5: Raises arms when caregiver says, "Come here" or "up" see above see above Yes Yes Yes Yes N / A = not applicable B5: Inattentiveness Measure A B C TABS Sensory Profile Item Description - Possible Overrepresentation Actively avoids eye contact #4: Resists looking you in the eye #96: Avoids eye contact A B C TABS "Looks through" people #3: Seems to look through or past people Correlation PCA Item Chosen to be included in variable? None of the items across measures correlate atr 2>0.50 N / A Yes Yes Yes r 2 = .522* p = .000 'N/A Yes No B5 continues B5 (Continued) Measure Item Description - Possible Overrepresentation Correlation PCA Item Chosen to be included in variable? A B C Frequently does not attend to social/environmental stimuli Seems not to hear so hearing loss is suspected Sometimes shows no startle response to a load noise (may have thought child was deaf) Frequently unaware of surroundings and may be oblivious to dangerous situations None of the items across measures correlate atr 2>0.50 N/A Yes Yes Yes Yes B5 continues B5 (Continued) Measure Item Description - Possible Correlation PCA Item chosen to be Overrepresentation included in variable? Sensory profile #43: Doesn't seem to notice when Yes someone touches arm or back (for example, unaware) Sensory profile #50: Seems oblivious within an active see above see above Yes continued environment (for example, unaware of activity) #99: Doesn't notice when people Yes come into the room B5 continues B5 (Continued) Measure TABS Item Description - Possible Overrepresentation Correlation #5: Acts like others are not there #8: Seems to be in "own world" #10: "Tunes out," loses contact with what is going on PCA Item chosen to be included in variable? Yes Yes Yes Sensory profile TABS #123: jumps from one activity to None of the another so that it interferes with play items across #28: can't wait at all for food or toy measures #33: doesn't sit still correlate #34: too "grabby," impulsive at r 2 > 0.50 N/A Yes Yes Yes Yes N / A = not applicable Appendix C: Items Selected for the Six Child Behaviour Variables C1: Acting-out Measure Item description Number A B C Hurts self by banging head, biting hand etc 1 Severe temper tantrums and/or frequent minor 2 tantrums Hurts others by biting, kicking, hitting 3 Is very destructive (toys and household items are 4 soon broken) *PSI-SF #29: my child reacts very strongly when 5 something happens that my child doesn't like. #30: my child gets upset over the smallest things 6 #36: M y child makes more demands on me than 7 most children #26: M y child generally wakes up in a bad mood 8 #28: my child does a few things which bother me 9 a great deal #34: there are some things that my child does that 10 really bother me a lot #35 my child turned out to be more of a problem 11 than I had expected C l continues CI (Continued). Measure Item description Number Sensory #108: has temper tantrums 12 profile #110: cries easily 13 #105: displays excessive emotional outbursts 14 when unsuccessful at a task #109: poor frustration tolerance 15 #107: is stubborn or uncooperative 16 TABS #21: upset by every little thing 17 #22: often difficult to soothe when upset or 18 crying #49: often cries'too long 19 #52: can't comfort self when upset 20 #24: gets angry too easily • 21 #25: too easily frustrated 22 #30:controls adult's behavior, "is the boss" 23 #29: Demands attention continually 24 #35: almost always refuses to do what is told 25 #23: has wide swings in mood 26 #27: frequently irritable, "touchy" or fussy 27 #31: jealous too often 28 PSI-SF items were omitted when used as predictors of parenting stress 189 C2: Sleep disturbances Measure Item description Number TABS #50: often frightened by dreams or the nighttime 1 #51: screams in sleep and can't be comforted 2 #53: wakes up often and doesn't fall back asleep 3 #54: doesn't have a regular sleep schedule 4 C3: Eating difficulties Measure Item description Number Sensory #54: gags easily with food textures or food 1 profile utensils in mouth #57: limits self to particular food 2 textures/temperatures #58: picky eater, especially regarding food 3 textures #55: avoids certain tastes or food smells that are 4 typically part of children's diet #56: will only eat certain tastes . 5 #60: shows strong preference for certain smells 6 C3 continues 190 C3 (Continued) Measure Item description Sensory #61: shows strong preference for certain tastes profile #62: craves certain foods continued #63: seeks out certain tastes or smells C4: Stereotypic behaviour Measure A B C Item description Rocks self for long periods of time Flaps hands Whirls self for long periods of time Twirls, spins and bangs objects a lot Squints, frowns, or covers eyes when in the presence of natural light Stares into space for long periods of time Walks on toes Wil l feel, smell, and/or taste objects in the environment Echoes questions or statements made by others Repeats sounds or words over and over Number Number 1 2 3 4 5 6 7 8 9 10 C4 continues / 191 C4 (Continued) Measure A B C continued *PSI-SF Sensory profile Item description Number Does a lot of lunging and darting about, 11 interrupting with spinning, toe walking, flapping, etc. Prefers to manipulate and be occupied with 12 inanimate things Does not use toys appropriately (spins tires) 13 Insists on keeping certain objects with him/her 14 Gets involved in complicated rituals such as 15 lining things up, etc. Strong reactions to changes in 16 routine/environment #21: It takes a long time and it is very hard for 17 my child to get used to new things #27: rocks unconsciously (for example, while 18 watching TV) #28: rocks in desk/chair/on floor 19 #26: Twirls/spins self frequently throughout the 20 day (for example likes feeling dizzy) #15: Covers eyes or squints to protect eyes from 21 light C4 continues 192 C4 (Continued) Measure Item description Number Sensory #16: looks intensely at objects/people (for 22 profile example stares) continued #97: stares intensively at objects or people 23 #8: enjoys strange noises/seeks to make noise for 24 noises sake #40: touches people or objects to the point of 25 irritating others #41: displays unusual need for touching certain 26 toys, surfaces, or textures (for example, constantly touching objects) #59: Routinely smells nonfood objects 27 #64: Chews or licks on nonfood objects 28 #124: Deliberately smells objects 29 #24: seeks all kinds of movement activities and 30 this interferes with daily routines (for example can't sit still, fidgets) #25: seeks out all kinds of movement activities 31 (for example being whirled by adult, merry-go- rounds, playground equipment, moving toys) #90: "on the go" 32 C4 continues 193 C4 (Continued) Measure Item description Number Sensory profile continued TABS #122: Has difficulty tolerating changes in routines #16: stares at lights #13: makes strange throat noises #19: shakes head over and over #32: mostly on the go, "in high gear" #17: overly interested in toy/object 33 34 35 36 37 38 *PSI-SF items were omitted when used as predictors of parenting stress C5: Social unresponsiveness Measure Item description Number A B C Has no social smile Does not (or did not as a baby) reach out when reached for Does not respond to own name when called out among two others Not responsive to other peoples' facial expressions feelings Is (or was as a baby) stiff and hard to hold 1 2 C5 continues 194 C5 (Continued) Measure Item description A B C Is flaccid (doesn't cling) when held in arms continued Frequently has no visual reaction to a "new" person PSI-SF #15: M y child smiles at me much less than I expected #19: M y child doesn't seem to smile as much as most children Sensory #6: Appears to not hear what you say (for profile example, does not "tune-in" to what you say, appears to ignore you) Sensory #7: Does not respond when name is called but profile you know child is hearing O K continued #116: Doesn't express emotions TABS #2: Emotions don't match what is going on #38: Rarely smiles, giggles, or laughs at funny things #42: Doesn't react to own name #43: Doesn't care when others are hurt V A B S -soc #5: expresses two or more recognizable emotions such as, pleasure, sadness, fear, or distress Number 6 7 10 11 12 13 14 15 16 17 C5 continues 195 C5 (Continued) Measure Item description Number V A B S -soc #6: shows anticipation of being picked up by 18 continued caregiver #7: shows affection towards familiar people 19 #9: reaches for familiar person 20 #15: laughs or smiles appropriately in response to 21 positive statements #2: Responds to voice of caregiver or another 22 person soc = socialization C6: Inattentiveness Measure Item Description Number A B C Actively avoids eye contact 1 "Looks through" people 2 Frequently does not attend to 3 social/environmental stimuli Seems not to hear so hearing loss is suspected 4 Sometimes shows no startle response to a load 5 noise (may have thought child was deaf) C6 continues C6 (Continued) Measure Item Description Number A B C continued Mullen Sensory profile Sensory profile continued Frequently unaware of surroundings and may be 6 oblivious to dangerous situations #8: attends to words and movements 7 #96: Avoids eye contact 8 #43: Doesn't seem to notice when someone 9 touches arm or back (for example, unaware) #50: Seems oblivious within an active 10 environment (for example, unaware of activity) #99: Doesn't notice when people come into the 11 room #3: Has trouble completing tasks when the radio 12 is on #4: Is distracted or has trouble functioning i f 13 there is a lot of noise around #5: Can't work with background noise (for 14 example, fan refrigerator) #48: Has difficulty paying attention 15 #49: Looks away from tasks to notice all actions 16 in the room C6 continues 197 C6 (Continued) Measure Item Description Number #123: jumps from one activity to another so that it 17 interferes with play TABS #4: Resists looking you in the eye 18 #10: "Tunes out," loses contact with what is 19 going on #5: Acts like others are not there 20 #8: Seems to be in "own world" 21 #39: Doesn't pay attention to sights and sounds 22 #40: Doesn't seem to watch moving objects 23 #28: can't wait at all for food or toy 24 #33: doesn't sit still 25 #34: too "grabby," impulsive 26 VABS-soc #1: Looks at face of caregiver 27 VABS-comm #10: Listens attentively to instructions 28 #17: Listens to a story for at least five minutes 29 soc = socialization; comm. = communication 198 Appendix D: Examples of LISREL Syntax and List of Restrictions Example of LISREL syntax for time one predictors of child measures from Tl to T4 Conditional model spss-Data from file IQ_Regardless.sav Observed variables: outcl outc2 outc3 outc4 st ia su ao e s sample size = 70 Latent variables: int_cept slope Relationships: outcl = l*int_cept outc2 = l*int_cept outc3 = l*int_cept outc4 = 1 *int_cept outcl = 0*slope outc2 = 0.5*slope outc3 = 1.0*slope outc4 = 2.0*slope int_cept=const slope=const slope = st ia su ao e s int_cept = st ia su ao e s Method = maximum likelihood path diagram 199 D l : Syntax restrictions for T l predictors of child measures (Tl -T4) Child measure A B C V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total Restriction None required None required None required Let error of outc4 be greater than 0 None required None required Let error of outc4 be greater than 0 None required None required None required Let error of outc4 be greater than 0 200 Example of LISREL syntax for time one predictors ofparenting stress from Tl to T4 Conditional model spss-Data from file PSI_med_pred_Tl_no_dads_no_DR.sav Observed variables: outcl outc2 outc3 outc4 fcopes life st ia ao e s sample size = 63 Latent variables: int_cept slope Relationships: outcl = l*int_cept outc2 = l*int_cept outc3 = l*int_cept outc4 = 1 *int_cept outcl = 0*slope outc2 = 0.5*slope outc3 = 1.0*slope outc4 = 2.0*slope int_cept=const slope=const slope = fcopes life st ia ao e s int_cept = fcopes life st ia ao e s Method = maximum likelihood path diagram 201 D2: Syntax restrictions for T l predictors of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total PSI-SF parental distress PSI-SF parent-child dysfunctional interaction None required Let error of slope greater than 0 Let the Error of slope be greater than 0 Let the errors correlate PSI-SF difficult child None required Example of LISREL syntax for difference score (T2 minus Tl) predictors of child measures from Tl to T4 Conditional model spss-Data from file IQ_Diff_T2_Tl.sav Observed variables: outcl outc2 outc3 outc4 st ia su ao e s sample size = 70 Latent variables: intcept slope Relationships: outcl = l*int_cept outc2 = l*int_cept outc3 = l*int_cept outc4 = l*int_cept outcl = 0*slope outc2 = 0.5*slope outc3 = 1.0*slope outc4 = 2.0*slope intcept^const slope=const slope = st ia su ao e s int_cept = st ia su ao e s Method = maximum likelihood path diagram 203 D3: Syntax restrictions for difference score (T2 minus T l ) predictors of child measures (Tl -T4) Child measure Restriction A B C IQ V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total None required None required None required Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate None required None required Let error of outc4 be greater than 0 None required. Let the error of outc4 greater than 0 None required Let the error of outc4 be greater than 0 204 Example of syntax for difference score (T2 minus Tl) predictors ofparenting stress from Tl to T4 Conditional model spss-Data from file PSI_T2_diff_med_nodadsDR.sav Observed variables: outcl outc2 outc3 outc4 fcopes life st ia ao e s sample size = 63 Latent variables: int_cept slope Relationships: outcl = l*int_cept outc2 = l*int_cept outc3 = l*intcept outc4 = l*int_cept outcl = 0*slope outc2 = 0.5*slope outc3 = 1.0*slope outc4 = 2.0*slope int_cept=const slope=const slope = fcopes life st iaao e s int_cept = fcopes life st ia ao e s Method = maximum likelihood path diagram 205 D4: Syntax restrictions for difference score (T2 minus T l ) predictors of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total None required PSI-SF parental distress Let error of slope greater than 0 PSI-SF parent-child dysfunctional Let error of slope greater than 0 interaction PSI-SF difficult child None required 206 Example of LISREL syntax for rate of change (Tl - T3) predictors of child measures from Tl to T4 Conditional model spss-Data from file VABS_dls_T3_AO.sav Observed variables: outcl outc2 outc3 outc4 predl pred2 pred3 sample size = 70 Latent variables: int_cept slope int_pred slope_pred Relationships: outcl = l*in t_cept outc2 = l*in t_cept outc3 = l*in t_cept outc4 = l*int_cept outcl = 0*slope outc2 = 0.5*slope outc3 = 1.0*slope outc4 = 2.0*slope predl = l*int_pred pred2 = l*int_pred pred3 = l*int_pred predl = 0*slope_pred pred2 = 0.5* slope jpred pred3 = 1.0*slope_pred int_cept=const slope=const int_pred=const slope_pred=const slope = int_pred slope_pred int_cept = int_pred slope_pred Method = rnaximum likelihood path diagram 208 D5: Syntax restrictions for rate of change (Tl - T3) of acting-out predicting child measures (Tl -T4) "* Child measure Restriction A B C IQ V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total Not applicable Let error of int_cept be greater than 0 Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate None required None required None required Let the error of outc4 be greater than 0 Let the Error of outc4 greater than 0 None required None required Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate Let the Error of outc4 greater than 0 209 D6: Syntax restrictions for rate of change (Tl - T3) of sleep disturbances predicting child measures (Tl -T4) Child measure Restriction A B C IQ V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total None required Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate None required None required None required None required Let the Errors of the int_pred and slope_pred correlate Let the Errors of the intcept and slope correlate None required Let the Error of outc4 greater than zero None required Let the Error of pred3 be greater than 0 Let the Error of outc4 be greater than 0 210 D7: Syntax restrictions for rate of change (Tl - T3) of eating difficulties predicting child measures ( T l - T 4 ) Child measure Restriction A B C IQ V A B S communication VABS daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total None required None required Let the error of pred3 be greater than 0 Let the Error of the predl greater than zero Let the error of the pred3 greater than zero Let the Error of outc4 be greater than 0 Let the Error of the pred3 greater than 0 Let the Error of pred3 greater than 0 Let the error of pred3 be greater, than 0 Let the Error of outc4 be greater than 0 Let the Error of pred3 greater than zero Let the Errors pred3 greater than 0 Let the Errors of outc4 greater than 0 None required Let the Errors of pred3 greater than 0 Let the Errors of outc4 greater than 0 211 D8: Syntax restrictions for rate of change (Tl - T3) of stereotypic behaviour predicting child measures (Tl -T4) Child measure Restriction A B C IQ V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total Not applicable Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate Let the Error of the predl greater than 0 Let the Error of predl greater than 0 Let the Error of outc4 be greater than 0 Let the Error of the predl greater than 0 Let the Error of the predl greater than 0 Let the Error of outc4 be greater than 0 Let the Error of the predl greater than 0 Let the Error of the predl greater than 0 Let the Error of the predl greater than 0 Let the Error of the predl greater than 0 Let Error of outc4 be greater than 0 Let the Error of the predl greater than 0 212 D9: Syntax restrictions for rate of change (Tl - T3) of social unresponsiveness predicting child measures (Tl -T4) Child measure Restriction A B C IQ V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total N / A None required Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate Let the Errors of the int_pred and slope_pred correlate Let the Errors of the int_cept and slope correlate None required None required Let error of pred3 be greater than 0 Let the Error of outc4 be greater than 0 Let the Error of the pred3 greater than 0 Let the error of outc4 be greater than 0 None required Let the Error of the outc4 greater than 0 213 D10: Syntax restrictions for rate of change (Tl - T 3 ) of inattentiveness predicting child measures (Tl -T4) Child measure Restriction A B C IQ V A B S communication V A B S daily living skills V A B S socialization PPVT EOWPVT MCDI total words said PLSAC PLSEC PLS total Not applicable None required None required Let the Error of the predl greater than 0 Let the Error of the outc4 greater than 0 None required None required Let the Error of the predl greater than 0 Let the error of outc4 be greater than 0 Let the Error of the predl greater than 0 None required None required Let the Error of the outc4 greater than 0 214 Example of syntax for rate of change (Tl - T3) predictors ofparenting stress from Tl to T4 Conditional model spss-Data from file PSI_T3_E_no_dadsDR.sav Observed variables: outcl outc2 outc3 outc4 predl pred2 pred3 fcopes life sample size = 70 Latent variables: int_cept slope int_pred slope_pred lfcopes llife Relationships: outcl = l*int_cept outc2 = l*int_cept outc3 = l*int_cept outc4 = l*int_cept outcl = 0*slope outc2 = 0.5*slope outc3 = 1.0*slope outc4 = 2.0*slope predl = l*int_pred pred2 = l*int_pred pred3 = l*int_pred predl = 0*slope_pred pred2 = 0.5*slope_pred pred3 = 1.0*slope_pred fcopes = l*lfcopes life = l*llife int_cept=const slope=const . int_pred=const slope_pred=const slope = lfcopes llife int_pred slope_pred int_cept = lfcopes llife int_pred slope_pred Set the Error variance of fcopes to 0 Set the Error variance of life to 0 Method = maximum likelihood path diagram options: AD=OFF iT=500 end of problem 216 D l 1: Syntax restrictions for rate of change (Tl - T3) of acting-out predicting of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total PSI-SF parental distress PSI-SF parent-child dysfunctional interaction PSI-SF difficult child Set the Error variance of fcopes to 0 Set the Error variance of life to 0 Let the error of predl be greater than 0 Let error of slope be greater than 0 . Let the error of predl be greater than 0 Let error of slope be greater than 0 Let the error of predl be greater than 0 Let error of slope be greater than 0 None required D12: Syntax restrictions for rate of change (Tl - T3) of sleep disturbances predicting of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total Let the error of pred3 be greater than 0 Let error of slope be greater than 0 PSI-SF parental distress None required PSI-SF parent-child dysfunctional Let error of pred3 be greater than 0 interaction Let error of slope be greater than 0 PSI-SF difficult child Let error of slope be greater than 0 217 D13: Syntax restrictions for rate of change (Tl - T3) of sleep disturbances predicting of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total Set the Error variance of fcopes to 0 Set the Error variance of life to 0 PSI-SF parental distress Let the error of pred3 be greater than 0 Let error of slope be greater than 0 PSI-SF parent-child dysfunctional Let the error of predl be greater than 0 interaction Let error of slope be greater than 0 PSI-SF difficult child Let error of predl be greater than 0 D14: Syntax restrictions for rate of change (Tl - T3) of stereotypic behaviour predicting of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total Let error of predl be greater than 0 PSI-SF parental distress Let the error of predl be greater than 0 Let error of slope be greater than 0 PSI-SF parent-child dysfunctional Let the error of predl be greater than 0 interaction Let error of slope be greater than 0 PSI-SF difficult child Let the error or predl be greater than 0 218 D15: Syntax restrictions for rate of change (Tl - T3) of inattentiveness predicting of parenting stress (Tl -T4) Parenting stress measure Restriction PSI-SF total PSI-SF parental distress PSI-SF parent-child dysfunctional interaction PSI-SF difficult child Let error of predl be greater than 0 Let the error of predl be greater than 0 Let error of slope be greater than 0 Let error of slope be greater than 0 None required

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 25 0
China 21 0
Germany 3 6
Japan 3 0
France 2 0
Republic of Korea 1 0
City Views Downloads
Beijing 21 0
Washington 16 0
Unknown 7 6
Piscataway 6 0
Tokyo 3 0
Sunnyvale 1 0
Ashburn 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}

Share

Share to:

Comment

Related Items