.H..1.s 104:..‘d. .r :1. LE . m .., $.59... :1)? . (5.: . .1 ., J: a: THf‘Sis; ”,3 be Date Michigan State Unlverslty This is to certify that the thesis entitled Depression With and Without Comorbid Antisocial Involvement: Factors That Contribute to Comorbidity In An Adolescent Inpatient Sample presented by Allison Marie Schettini has been accepted towards fulfillment of the requirements for Masters Psychology degree in Angeli/LIN) jor professor August 11, 2000 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. } “(DATE DUE lr32 DATE DUE DATE DUE stein 4 2002 1.”! 6/01 c:/ClRC/DateDue.p65-p.15 Laofiwrz In‘ I - DEPRESSION WITH AND WITHOUT COMORBID ANTISOCIAL INVOLVEMENT: FACTORS THAT CONTRIBUTE TO COMORBIDITY IN AN ADOLESCENT INPATIENT SAMPLE By Allison Marie Schettini A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 2000 ABSTRACT DEPRESSION WITH AND WITHOUT COMORBID ANTISOCIAL INVOLVEMENT: FACTORS THAT CONTRIBUTE TO COMORBIDITY IN AN ADOLESCENT INPATIENT SAMPLE By Allison Marie Schettini Although depressed adolescents frequently are involved in antisocial activity, most are not comorbidly antisocial. This study assessed several factors proposed to discriminate between depressed adolescents with and without comorbid antisocial involvement: gender, ADHD, depressogenic concerns, reactance and school failure. Depressed adolescent inpatients (N=284), ages 1 1 to 17 (M=14.89 years), admitted over 36-months, completed the Functional Impairment Scale for Children and Adolescents — Self Report (FISCA-SR); Depressive Experiences Questionnaire -Adolescent Version (DEQ-A); Therapeutic Reactance Scale —Adolescent Version (TRS-A); and Reynold’s Adolescent Depression Scale (RADS). All participants reported severe impairment on the FISCA-SR Moods scale and reported clinically significant depression on the RADS and/or had a discharge diagnosis of depression. Comorbid patients were more Often boys, had an ADHD diagnosis, reported more school impairment and were more reactant. A discriminant function analysis indicated that regardless of gender, school impairment and reactance were the best discriminators of comorbid and non-comorbid patients. ACKNOWLEDGMENTS I would like to thank my advisor, Susan J. Frank, for her valuable contributions, support, and commitment throughout this endeavor. It was her interest and involvement that made this possible. I would also like to thank the other members of my committee, Alex von Eye and Hiram Fitzgerald, for their constructive feedback and statistical and conceptual assistance. I would like to thank Terry and Tara for their friendship, understanding, and laughter. To the patience, understanding, and support of my family who has been there with me through every step of the way. Nonny — thank you for your novenas and for lending me your angels who guide me through all my journeys. Melissa, Chrissy, and Vicky — you are my greatest and truest friends. Thank you for your listening ears and open hearts. To Mom and Dad - you have given me the courage to believe in myself and pursue my dreams. Your love, support, and determination to make me happy is incredible. You are both wonderful and you do more for me than you’ll ever know. To David, my greatest source of support and happiness, thank you for believing in me and sharing your life and love with me. TABLE OF CONTENTS List of Tables ................................................................................. vii List of Figures .............................................................................. viii Introduction ................................................................................. 1 Depression .......................................................................... l Antisocial Involvement ............................................................ 5 Evidence For Comorbidity of Depression and Externalizing Problems ...... 5 Family Factors ........................................................................ 7 Gender ................................................................................. 8 Attention Deficit Hyperactivity Disorder ....................................... 10 Academic Achievement ............................................................ 14 Depressogenic Concerns ............................................................ 16 Reactance ........................................................................... 21 Another Look At Gender ............................................................ 24 Summary and Hypotheses ......................................................... 26 Method ........................................................................................ 28 Participants ......................................................................... 28 Inclusion Criteria ............................................................ 28 Measures ........................................................................... 29 The Functional Impairment Scale for Children and Adolescents — Youth Self Report ....................................................... 29 Depressive Experiences Questionnaire, Adolescent Version ...... 35 Therapeutic Reactance Scale —— Adolescent Version .................. 36 Reynolds Adolescent Depression Scale ................................. 37 Discharge Diagnosis ....................................................... 38 Demographic Information ................................................ 38 Study Procedures ............................................................................ 39 Results .......................................................................................... 40 Preliminary Analyses ............................................................... 40 Selection of Participants for the Study ................................... 40 Characteristics of the Study Participants ................................ 43 Demographic Differences between Comorbid and Non-comorbid Groups of the Selected Patients ........................................... 44 Assessment of Sampling Biases ............................................ 45 Tests of the Hypotheses .............................................................. 49 Hypotheses l and 2: Gender and Comorbidity; ADHD and Comorbidity .................................................................. 49 Hypothesis 3: School Functioning and Comorbidity ................. 50 Hypothesis 4: Depressogenic Concerns and Comorbidity ........... 53 Hypothesis 5: Reactance and Comorbidity ..... I ........................ 54 Discriminant Function Analysis. .......................................... 55 Additional Analyses ................................................................... 56 Discussion ....................................................................................... 57 Research Predictions ................................................................... 57 Additional Considerations, Constraints, and Limitations ....................... 62 Implications for Treatment and Prevention ....................................... 64 Summary ........................................................................................ 66 Appendix ........................................................................................ 67 Table 1: Demographics and Depressogenic Concerns For Comorbid and Non-Comorbid Groups ................................................................ 68 Table 2: Descriptions Of 1 1 Diagnostic Categories ............................. 69 Table 3: Nonsignificant Differences Between Initially Eligible, Excluded and Initially Eligible, Included Groups on Demographics and Discharge Diagnoses ............................................................................................. 70 Table 4: Univariate Correlations Among the Discriminating Variables By Gender ............................................................................................. 71 Table 5: Results of the Discriminant Function Analysis for Boys and Girls Separately .............................................................................................. 71 List of References ............................................................................. 72 vi LIST OF TABLES Table 1 Criteria for Establishing Severe Levels of Impairment ................ 32 Table 2 Illustration of the F ISCA scoring procedure for two “severe” criteria from the Delinquency scale. ................................... 34 Table 3 Emotional Impairment as Indicated on the FISCA-SR (N =562). .. 40 Table 4 Relationship Between Parent and Adolescent Reports of Emotional Impairment ...................................................... 42 Table 5 Significant Differences Between Initially Eligible, Excluded and Initially Eligible, Selected Groups on Reactance, Depressogenic Concerns, Functional Impairment Scores for School, Delinquency, and Aggression, and Discharge Diagnoses ........................................................ 47 Table 6 Univariate Correlations Among the Discriminating Variables ....... 49 Table 7 Patient Frequency Broken Down By Comorbidity By Gender ......... 50 Table 8 Patient Frequency Broken Down by Comorbidity, Gender, and ADHD ............................................................................ 51 Table 9 Percent of Comorbid and Non-Comorbid Patients at Each Level of School Impairment ....................................................... 52 Table 10 Percent of Girls and Boys at Each Level of School Impairment ..... 53 Table 11 Patient Frequency Broken Down By Reactance and Gender and Comorbidity ........................................................................ 54 vii LIST OF FIGURES Figure 1: Inclusion Criteria and Decision Making Process ............... 30 viii INTRODUCTION Research indicates that depressed adolescents frequently are involved in antisocial activity (Compas, Connor, & Hinden, 1998; Compas, & Hammen, 1994; Domielly & Wilson, 1994; Messier & Ward, 1998). However, most depressed adolescents are not comorbidly antisocial. The factors distinguishing those who report depression and involvement in antisocial activity from those who report severe depression alone remain unclear. This study will consider several factors, suggested by the literature, that potentially discriminate between depressed adolescents with and without comorbid antisocial involvement. These factors are gender, ADHD, depressogenic concerns, reactance, and school failure. Depression Confusion and contradictions in research on childhood and adolescent depression are common because of the various uses of the word “depression.” Depression is conceptualized in three different ways — mood, syndrome, or disorder. More specifically, some researchers indicate “depression” by assessing the current feelings (of sadness, disphoria, etc.). Others identify a constellation of mood and other symptoms forming a depressive syndrome. Still others indicate “depression” by using diagnostic criteria established by well-known professional organizations (DSM-IV; American Psychiatric Association, 1994 or ICD-lO; World Health Organization, 1996). The three notions of depressive constructs have some commonalties; however, inherent to each is a unique set of assumptions and assessment procedures. The three constructs share a common set of symptoms reflecting negative affectivity but differ in their implications in duration and severity of the symptoms they include (Compas, Ey, & Grant, 1993). Mood measures assess depression, indicated as a symptom of sadness or unhappiness, and typically is evaluated using self-report rating scales. As a syndrome, depression is indicated by a constellation of symptoms (for example, lethargy, anhedonia, and insomnia/hypersomnia) occurring together as a recognizable and statistically coherent pattern (Achenbach, 1991). A depressive syndrome is often measured by self-report measures or by parent/teacher reports based on rating scales. A depressive disorder is diagnosed in the presence of certain established indicators, as seen in the DSM—IV (American Psychological Association, 1994); in particular, for Major Depressive Episode, five or more of a list of nine symptoms must have been present during the same 2-week period and represent a change from previous functioning in addition to evidence of impairment. Bearing in mind the definitional differences in “depression,” it is important to interpret research on childhood and adolescent depression with a certain degree of caution. Depression, as a syndrome or diagnosis, has not always been acknowledged in children and adolescents as a major mental health concern (Compas et al., 1998). Earlier psychoanalytic views held that depression could not occur during childhood because of inadequate development of the superego (Pozinski & Mokros, 1994), or else argued that if it did occur, it would be overshadowed by other factors (Pozinski & Mokros, 1994). It is only recently that these views have been challenged and that researchers have begun to show that children may show features of, or meet criteria for clinical depression (Cantwell & Carlson, 1983). In particular, many researchers advocate that children and adolescents can and do, manifest a constellation of affective, cognitive, and behavioral symptoms that reflect major depression as well as other forms of depressive disorders (Carlson & Cantwell, 1980; Hammen & Compas, 1994). Research on depression during these earlier years of development shows that the rate and symptoms of depression increase from childhood to adolescence (Leadbeater, Blatt, & Quinlan, 1995; Nolen-Hoeksema & Girgus, 1994). Despite differences in methods of assessment and case identification processes, many researchers report that the prevalence of depression increases during adolescence with 5% to 10% of teenagers manifesting a major depressive disorder at any point in time (Fleming & Offord, 1990) as compared to 2% to 3% of children (Angold & Rutter, 1992) and less than 1% of preschool children (Kashani & Carlson, 1987). At the high end, Peterson et a]. (1993) has found that one third of all adolescents meet criteria for clinical depression. Additionally, researchers have found that when self-report symptom scores are used, rather than diagnoses, approximately 10% to 30% (and sometimes even more) of adolescents exceed cutoffs for clinically significant levels (Albert & Beck, 1975; Roberts, Lewinsohn, & Selley, 1991) Even when diagnostic criteria are not met, subsyndromal depressive symptoms appear to be present frequently. For example, Cooper and Goodyer (1993) reported that 20.7% of their sample of 1 1 to 16 year old girls had significant symptoms but fell short of diagnostic criteria. Notably, there is evidence to suggest that the prevalence of depression in children and adolescence is increasing (e. g., Klerman et al., 1985) Early to middle adolescence is also considered the developmental period when girls begin to experience significantly more depression than boys (Angold & Rutter, 1992; Nolen-Hoeksema & Girgus, 1994; Peterson, Sarigiani, & Kennedy, 1991). Studies of preadolescent children provide mixed findings regarding gender differences. Some have found that there are higher rates of depression in boys in comparison with girls during childhood (Anderson, Williams, McGee, & Silva, 1987). In contrast, other researchers purport that there are no gender differences prior to age 1 1 (Angold & Rutter, 1992). Despite the contradictions found in samples of children, most researchers agree that gender differences, in the direction of higher prevalence rates for girls, are evident during adolescence (Cicchetti, Rogosch, & Toth, 1998). For example, Cohen et al. (1993) reported a prevalence of 7.6% in 14 to 16 year Old girls compared with 1.6% for boys of the same ages. Given the nature of the symptoms of depression, it is not surprising that depression (both syndromal and diagnostic) would be related to functional impairment. For example, Whitaker et al. (1990) report that ratings of functional impairment (as indicated by the Global Adjustment Scale) evidenced significant impairment in the majority of youth with major depression (85%) and dysthymic disorder (87%). Depression often disrupts multiple areas of functioning, such as school, family, and interpersonal relationships and when depressive mood and depressive symptoms lead to functional impairment, intervention is imperative. This disruption in functional impairment is partly recognized in the present criteria for diagnosing depression; however it is usually intuitively decided rather than concretely measured. This study recognizes and emphasizes the importance of functional impairment and therefore, concretely measures functional impairment in addition to either a diagnosis or a self-reported rating of depressive symptoms. Antisocial Involvement. Several different terms also have been used, Often interchangeably, to describe involvement in antisocial activity, such as conduct disorder, conduct problems, externalizing behavior, and delinquency. Strictly speaking, conduct disorder (CD) refers to a psychological disorder, as articulated in the DSM-IV (American Psychological Association, 1994) whereas delinquency refers to a legal term for behaviors that are against the law. Delinquent behaviors are part of the CD syndrome; however, some behaviors included in the syndrome (e. g., pushing and verbally acting-out) are not considered delinquent. Conduct problems refer to externalizing behaviors (illegal or not) and are synonymous with antisocial behaviors. Antisocial behavior includes both overt, confrontive behaviors (such as arguing, temper tantrums, and fighting) and covert, concealed behaviors (such as stealing, truancy, and fire setting); again these behaviors may or may not be illegal, but both are examples of externalizing behaviors as opposed to internalizing behaviors (such as anxiety and depression). The present paper will restrict use of the term delinquency to refer to behaviors that are illegal; conduct disorder to refer to a syndrome of problems defined as a disorder by DSM-IV; and the terms conduct problems, externalizing behaviors, and antisocial activity to refer to acting out behaviors in general. Evidence For Comorbidity of Depression and Externalizing Problems Despite its prevalence and profound importance in understanding childhood and adolescent problems, the implications of comorbidity have often been overlooked by focusing exclusively on one disorder and disregarding the impact of symptoms of other disorders. Yet, researchers continually find that comorbidity, in general, and comorbidity involving depression, in particular, is highly prevalent during childhood and adolescence (Angold, & Costello, 1993, Biederman, Newcorn, & Spirch, 1991; Brady & Kendall, 1992; Compas & Hammen, 1994). Depression is most often comorbid with other internalizing problems, particularly anxiety. In both clinical and epidemiological samples, anywhere between 30% to 75% of children and juveniles with major depression also report high rates of anxiety disorders, and endorse more anxiety items on anxiety/depression symptom checklists than children and juveniles not evidencing depression (Angold and Costello, 1993; Biederman, Farone, Mick, Moore, & Lelon, 1996; F inch, Lipovsky, & Casa, 1989). Nonetheless, evidence for a high degree of overlap between depression and externalizing problems exists as well. In general, research studies in the late 1980’s and early 1990’s identify high rates of conduct or oppositional defiant disorder, ranging from 21% to 83%, among juveniles with major depression (see Angold and Costello (1993) review). A somewhat later study of 16-year-olds in a general population sample also showed comorbidity between affective disorders (which include depression) and conduct disorders, and found that dimensionally scored measures of affective and conduct disorder symptoms were correlated in the region of .35 (Fergusson, Lynskey, & Horwood; 1996). Likewise, researchers studying 11 to 17 year olds have found a high rate of co-morbidity between depression and externalizing behaviors, with depression highly comorbid with substance abuse disorders (14% overlap) as well as disruptive behavior disorders (8% overlap; Compas et a1. 1998). How do we explain comorbidity? Achenbach (1990/1991) presents three general models for understanding why psychological problems or disorders co-occur, all of which may be applicable to the particular case of depression and externalizing behaviors. First, overlap may be due to chance. For example, if disorder B has a high prevalence rate in boys in general, than it is likely, among boys, that disorder B will overlap with disorder A, especially if disorder A also has a high prevalence rate in boys. In the strict sense of the term, this chance co-occurrence, to be referred to as a “prevalence” model of comorbidity, is not true comorbidity. A second model, to be referred to as a “common etiology” model, maintains that comorbidity results from a shared etiology. For example, the same variable, C, that predicts disorder A also predicts disorder B, accounting for the overlap between A and B. However, the relationship between A and B disappears after controlling for the common factor, C. Lastly, a third model posits that the presence of one disorder increases the risk for another disorder. For example, the associated risks or consequences of disorder A increase vulnerability for disorder B (or vice versa). This model will be referred to as the “linked risk” model. One or more of these models, the prevalence model, the common etiology model, and/or the linked risk model, can be used heuristically to provide a conceptual foundation for the study proposed here. Although this study is unable to rule in favor of one over another, each helps to provide a rationale for predicting that the variables identified here (gender, ADHD, academic failure, depressogenic concerns, and reactance) will be able to distinguish between depressed adolescents with and without comorbid antisocial involvement. Family Factors In attempting to explain why depression and conduct problems so frequently co- occur, researchers have focused a good deal on parenting variables; hypothesizing that both externalizing and internalizing problems have common etiological roots in family dysfunction. The data generally support hypotheses demonstrating a robust link between parenting practices and both externalizing problems a_nd_ internalizing problems (Burbach & Borduin, 1986; Capaldi, 1991; Conger et al., 1992, 1993; Conger, Ge, Elder, Lorenz, & Simons, 1994; Gelfand & Teti, 1990; Rutter, 1989). However, these results do not provide information on ways to distinguish between children at risk for internalizing problems and those at risk for externalizing problems. In this regard, studies by Ge, Best, Conger and Simons (1996) suggest that differences may be a matter of degree. In particular, these researchers have found that parents of 10th graders with elevated conduct problems were more hostile than parents of 10‘h grades with elevated depressive symptoms when these adolescents were in 7‘", 8th, and 9‘h grades. However, parents of 10th graders with both elevated depressive symptoms and conduct problems were most hostile and. last warm suggesting an interaction between the two types of disorders. The present study takes a different approach looking at factors that potentially are specific to depressed adolescents with and without comorbid conduct problems as opposed to those with depression alone. Ginger One difference between depressed adolescents with and without conduct problems, already suggested by the research literature, is that those who are comorbidly antisocial are more often boys than girls. This may have to do with the greater prevalence of conduct problems, per se, in boys as compared to girls. During adolescence, the incidence of depression, especially in clinical samples, typically is greater for girls than boys (Angold & Rutter, 1992; Compas et al., 1997; Nolen-Hoeksema & Girgus, 1994). However, in both younger and older children, externalizing behaviors in the form of delinquency, antisocial behavior, conduct disorder, and conduct problems are more consistently found in boys as compared to girls (Thorne & Luria, 1986; Biederman, et al., 1996). Accordingly, in samples of depressed adolescents, depressed boys are more likely to be antisocial than depressed girls (Loeber & Keenan, 1994; Loeber, F arrington, Stouthamer-Loeber, & Van Kammen, 1998) just as in samples of antisocial adolescents, antisocial girls are more likely to be depressed than antisocial boys. These gender-related patterns of comorbidity have been found with children and adolescents in both general population samples and clinical pOpulations. The "linked risk” model, in addition to the “prevalence” model, may further account for the greater incidence of antisocial behavior among depressed boys as compared to depressed girls. Although the former model posits that, in general, the factors associated with or exacerbated by depression might increase the risk for externalizing problems, certain “linked risk” factors may be more importent for one gender as compared to the other. For example, depression has been linked to feelings of anger among both delinquent and non-delinquent boys and girls (Silver, 1996; Mirowsky & Ross, 1989). However, boys and girls cope with anger in different ways, such that “acting out” of angry feelings is more common in boys than girls. In particular, whereas adolescent boys’ anger is often accompanied by contempt for, and inflicting pain on others, anger in adolescent girls is more likely to coincide with surprise, shyness, shame, guilt, sadness, and self-directed hostility (Silver, 1996). Accordingly, anger accompanying depression in girls may only serve to intensify feelings of dysphoria whereas in boys anger paired with depression may direct negative affect towards others and away from the self. Gender differences in anger coping strategies may themselves be rooted in a differential emphasis on autonomy as opposed to connection: whereas girls place considerable value on connection with others, boys are more likely to place value on achieving autonomy and are less likely to value maintaining connections. (Blatt, Shaffer, Bars, & Quinlan, 1992; Gilligan & Attanucci, 1998). Accordingly, preserving the integrity of the self by blaming or inflicting pain on others is a more “palatable” defense for adolescent boys as compared to adolescent girls. Attention Deficit Hyperactivity Disorder Attention Deficit Hyperactivity Disorder (ADHD) is another factor that may increase the likelihood of comorbid antisocial behavior in the face of adolescent depression, with ADHD as an etiological factor in both types of difficulties. ADHD, as defined by DSM-IV (1994), is a disorder characterized by pervasive inattention and/or hyperactivity-impulsivity. In order to diagnose an individual with ADHD, symptoms of hyperactivity-impulsivity must persist for at least 6 months and to a degree that is maladaptive and inconsistent with developmental level. In addition, there must be clinically significant impairment in social, academic, or occupational functioning (American Psychiatric Association, 1994). ADHD by definition, has its onset during childhood, although its symptoms persist into adulthood (Compas et al., 1998). Recent research speaks to the importance of studying ADHD in children and adolescents; however, one of the foremost limitations in this research stems from failure to consider comorbidity. Only a limited number of studies of ADHD both include children with other comorbid disorders and directly examine the implications of comorbidity. More typically, some studies screen for comorbidity and focus exclusively on children with ADHD alone, which ultimately limits generalizeability (Foley, Carlton, & Howell, 1996). Alternatively, when children with comorbid conditions are included, the comorbid factors are not sufficiently described or considered. When it comes to comorbidity, ADHD has often been implicated in the onset of antisocial behavior, found in both retrospective and prospective studies (Farrington, Loeber, & Van Keammen, 1990b; Windle, 1993). This link is fairly easily understood; by definition, individuals with ADHD are hyperactive and impulsive. Hyperactivity often manifests itself in the form of fidgetiness, excessive running or climbing in situations where it is inappropriate, difficulty playing or engaging quietly in leisure activities, and a predilection to be “on the go” as if “driven by a motor” (American Psychological Association, 1994). Impulsivity as well refers to difficulty with behavior control and hyperactivity and impulsivity, together, suggest an inability to regulate behaviors according to rules; often an underlying mechanism in antisocial behavior (Barkely, 1990). Accordingly, although ADHD does not invariably lead to CD, children with ADHD or high levels of ADHD symptomatology are at high risk for conduct disorder, as well as substance abuse and criminality (Dalteg & Levander, 1998; Foley, et al., 1996; Mendelson, Johnson, & Stewart, 1971). This link has been found in retrospective assessments of attention deficit hyperactivity disorder in juvenile delinquent boys (Dalteg & Levander, 1998; Foley et al., 1996) and non-delinquents (Mendelson et al., 1971) ranging in ages from 12 to 16 years Less obvious, yet also significant, is the association between ADHD and depression. In fact, in both epidemiological and clinical studies, mood disorders and ADHD co-occur in 20% to 30% of children and adolescents diagnosed with one or the 11 other disorder as primary (Angold & Costello, 1993; F ergusson, et al., 1996; Loeber et al., 1998; Schmidt, Stark, Carlson, & Anthony, 1998; Wilson & Marcotte, 1996), with this pairing found in both longitudinal and cross-sectional work. Researchers initially wondered whether the ADHD-mood disorder overlap was truly indicative of comorbidity. However, to a large extent these questions have been resolved, with research suggesting that ADHD and depression are in fact separate disorders that often co-occur. The earlier argument was that both disorders include psychomotor disturbance and a diminished ability to concentrate so that the appearance of comorbidity was an artifact of overlapping symptoms rather than a true co-occurrence (Mannuzza, Fyer, & Klein, 1993). Milberger, Biederman, Farone, and Murphy (1995) used both a “subtraction” method and a “proportion” method to disentangle overlapping symptomatology. The subtraction method requires the researcher to re-diagnose subjects using criteria modified to remove overlapping symptoms; subjects must reach DSM-III-R threshold for diagnosis in spite of the removal of overlapping symptoms. In contrast, the proportion method lowers the threshold, so that the proportion of symptoms endorsed (minus the overlapping symptoms) is the same as the proportion required by DSM-III-R (including the overlapping symptoms). Milberger et a1. (1995) found that despite the adjustments for overlapping symptoms, children with ADHD continued to Show increased rates of depression, thereby supporting the phenomenon of comorbid depression and ADHD. Using a different approach, Biederman et a1. (1996) suggested that comorbidity of depression and ADHD was due to interviewer and assessment biases. Accordingly, they conducted a study that controlled for both overlapping symptoms as well as interviewer and assessment biases. Nonetheless, they found that children with major depression plus ADHD differed in several ways from those with either major depression alone or ADHD alone. The combined group had significantly higher rates of aggressive behavior and more marked irritability and oppositionalism than the pure major depression group; and higher rates of delinquency than the pure ADHD group. Therefore, after controlling for possible confounds, the link between depression, delinquency and ADHD remains. Obviously the comorbid occurrence of ADHD and antisocial behavior on the one hand, and ADHD and depression on the other hand, does not in and of itself provide sufficient evidence for ADHD as a common etiological factor in the co-occurrence of depression and antisocial involvement. However, longitudinal studies show that ADHD precedes both antisocial behavior and depression, which is a necessary (although, still not sufficient) condition for the “common etiology” model. Barkely (1990), for example, conducted an 8-year longitudinal study of psychosocial outcomes of 123 hyperactive children and 66 normal controls (initial ages 4 to 22 years). At the end of the study they found that 80% of the hyperactive children continued to qualify for an ADHD diagnosis, with 60% of the sample also qualifying for Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) at follow-up. These findings suggest that ADHD and hyperactivity often lead to behavioral problems, which in turn may result in antisocial involvements. Importantly, the same study also pointed to a specific mechanism by which ADHD can result in depression, with the link between the two mediated by child functional impairment. ADHD at time one was predictive of school problems at time two. School failure has consistently been shown to relate to depressed mood in adolescence (Patterson, DeBaryshe, & Ramsey, 1989; Rossen, 1997), suggesting that 13 ADHD can also lead to depression in adolescents via its link to school (or other types of functional) impairment. Generally stated, ADHD compromises children's social competence and social role performance (Wilson & Marcotte, 1996). ADHD children have been shown to have difficulties processing and responding to social cues and intentions (Dodge & Shwartz, 1997; Matthys, Cuperus, Van Engeland, 1999; Sprouse, Hall, Webster, & Bolen, 1998). These difficulties, in turn, may lead to a trajectory of social role failure that increases feelings of inadequacy and subsequently results in depressed mood. Therefore, this research suggests that ADHD is directly related to antisocial activity as an etiological factor, and also indirectly related to depression via increased risk for social role failure, thereby supporting a linked risk model of comorbidity between depression and antisocial activity. Academic Achievement The discussion above implies a relationship between ADHD and. depression via school dysfunction, a specific example of functional impairment. However, the poor school functioning—depression link may be sufficiently critical, to warrant independent consideration. From the perspective of a common etiological factor model, poor academic achievement maybe associated with higher rates of depression as well as antisocial activity. Research offers support for this perspective. School problems, in the sense of academic difficulties, along with low measured intelligence, have been positively related to depression. Research conducted with three groups of clinic-referred youth (ages 8 to 27 years), one assessed before therapy and two afterward found that in all three groups, low levels of perceived academic competence were significantly correlated with children's Childhood Depression Inventory (CDI) scores (Weisz, Weiss, Wasserman, & Rintoul, 1987). Another study demonstrated a similar relationship between depressive symptoms and levels of competence in adolescents in both a “normal” subset of individuals as well as in a group of “severely depressed” individuals. School competence was found to be significantly negatively correlated with the adolescent's level of depressive symptoms (Rossen, 1997). Tulipan (1981) argues that it is possible that school failure may not only be a factor precipitating or resulting from depression but that it may be one of many factors that mask disturbing feelings more truly typical of depression. Research also indicates that involvement in antisocial activity, also is predictive of academic difficulties (as reviewed in Anderson & Hammen, 1993; Hammen & Rudolph, 1996; Kaslow & Racusin, 1990; Huizinga & Jakob-Chien, 1998). Some studies suggest that the child’s disobedience and undercontrolled behavior impede learning whereas others imply that antisocial children do not possess the skills necessary to excel in the classroom, such as attending to the teacher and remaining in their seats (Hops & Cobb, 1974). However, the interplay of academic achievement, depression, and involvement in antisocial activity may be quite complex and reciprocal, with low academic achievement leading to both depression and antisocial activity, and antisocial activity precipitating school failure. Regardless of the direction of the relationship, the interplay among the three types of difficulties has been consistently demonstrated (Paterson 1986; Patterson et al., 1989; Messier & Ward, 1998; Quay & Peterson, 1987). For example, Patterson and his colleagues (1984, 1986, 1989) have found that coercive and antisocial behaviors exhibited by children are likely to result in rejection by members of the normal peer group as well as academic failure, both of which increase risk for and can lead to depressive symptomatology and even greater involvement with deviant peer groups (Patterson, 1986; Patterson, et al., 1989; Patterson, Dishion, & Bank, 1984). Consequently, it is assumed that children following this developmental sequence are at high risk for engaging in chronic delinquent behavior. Other researchers have found that the nature of the relationship between depression and antisocial involvement changes when academic achievement is considered. Without taking into account academic achievement, Messier and Ward (1998) found that a sample of incarcerated juvenile delinquents scored higher on depression indices than normal populations, but lower than clinical samples (Messier & Ward, 1998). In fact, the results showed that slightly over one third of the sample had depression index scores in the range considered indicative of clinical depression as compared to 15% to 18% in normal populations. However, the nature of this relationship changes considerably when academic achievement and ability are taken into consideration. The findings Show that high ability youth who are involved in antisocial activity are M likely to be depressed than low ability youth who are involved in antisocial activity (Cornell, 1992; Messier & Ward, 1998). In sum, among depressed youth, school failure may be an outcome as well as a precipitating factor that increases depression and/or antisocial behavior, with the link between antisocial involvements and depression especially high among high ability youth. Depressogenic Concerns The prevalence of two different types of depressive concerns, referred to by Blatt and colleagues as interpersonal and self-critical concerns (Blatt, 1974) may also effect rates of comorbid depression and antisocial activity. Past research on comorbidity in depression has often overlooked these different dimensions of depression. Self-critical concerns are fueled by fears of failure and guilt and the need for achievement and autonomy. Self-critically preoccupied individuals want others’ approval, respect, and admiration but are afraid of being judged, criticized, or controlled (Blatt & Shichman, 1983). They try to avoid dependency at all costs and seek to dominate others. Presumably, depression is most likely to occur in these individuals in response to perceived achievement failure or lack of control over the environment (Robins & Block, 1988) In contrast, interpersonal concerns revolve around issues of dependency, helplessness, and feelings of loss and abandonment (Blatt, 1974; Blatt, D’Afflitti, & Quinlin, 1976; Blatt, et al., 1992). Interpersonally preoccupied individuals dread the potential loss of gratifying, protective relationships. Consequently, they have been described clinically as highly dependent, conformist, and bent on pleasing others. Depression in these individuals is most likely to occur in response to perceived loss or rejection in social relationships (Robins & Block, 1988). Researchers have found each type of depressogenic concerns to be linked to distinct personality characteristics suggested in clinical studies. (Blatt, 1974; Blatt, Hart, Quinlin, Leadbeater, & Auerbach, 1993). Evidence stemming mostly from research with adults suggests that self-critical individuals are more inclined to be involved in antisocial activity whereas interpersonally preoccupied individuals are more conformist and less likely to engage in antisocial activity. In contrast to interpersonally concerned individuals, self-critical individuals are more likely to be irritable, hostile or unlikable; and they often have discordant peer and family relationships (Blatt, 1991; Mongrain, 1993). Such 17 personality characteristics are likely to result in counteractive attempts to challenge authority by becoming involved in many forms of antisocial behavior (Loeber, 1990). Their alienation from others and their personal drive for success and autonomy help explain self-critical persons’ inclination for externalizing behaviors. Similar research (with adult samples) shows that, in contrast to self-criticals, interpersonally preoccupied individuals have difficulty expressing opinions that oppose those of other individuals, find it difficult to express their anger and are overly conforming and hyperresponsible (Blatt & Maroudas, 1992; Fichman, Koestner, & Zuroff, 1996; Mongrain, 1993). Presumably, excessive concerns with maintaining close relationships diminish engagement in aggressive behaviors or delinquent acts which potentially, could push others away. Although the great majority of research on depressogenic concerns focuses on adult populations, studies with younger populations verify that personality characteristics associated with self-critical and interpersonal concerns in adults are similar in adolescent samples (Blatt, et al., 1993; Frank et al., 1997b). Accordingly, this research shows that during adolescence, self-critical concerns, but not interpersonal concerns, are linked to antisocial activity for boys and girls (Blatt, et al., 1993). For example, one study with an inpatient sample found that interpersonal concerns were positively associated with social introversion and negatively with family discord and deviance whereas self—critical preoccupations were associated with deviance, family discord, and general maladjustment/alienation (Frank et al., 1997b). These findings are consistent across various adolescent age groups (Blatt, 1991). However, the number of studies of depressogenic concerns during childhood and adolescents is still relatively small. 18 The attachment literature is also relevant to the discussion because of the theoretical correspondence between the two types of depressogenic concerns and the two types of insecure attachment styles, with self—critical concerns corresponding to dismissing attachment styles and interpersonal concerns to preoccupied attachment styles (Blatt & Maroudas, 1992). Both object relations theorists and attachment theorist argue that the self-critical/dismissing attachment styles and interpersonal/preoccupied attachment styles are rooted in earlier parent-child relationships and that the nature of these attachment bonds shape subsequent interpersonal relations as well as personality development and mental health (Kobak & Sceery, 1988; Main, 1981). In contrast, securely attached individuals typically have experienced responsive and sensitive parenting, allowing them to manifest positive emotions as well as heightened social interactions and competence (Kobak & Sceery, 1988). Childhood and adult outcomes are more problematic for those with either type of insecurity (George, Kaplan, & Main, 1985; Main, Kaplan, & Cassidy, 1985). Dismissing/self-critical attachments stem from parenting that is deprecating and rejecting of the child’s desires for love and attention (Blatt et al., 1992; Kobak, 1986; Main, 1981). This attachment organization and related self-cognitions and self-critical preoccupations are both associated with behaviors that seek compensation for feelings of failure and inadequacy. Individuals with dismissing attachments and self—critical concerns are both more distressed by signs of failure than rejection and are more focused on achieving positive results (Blatt & Maroudas, 1992). In attempting to achieve such results, coupled with a disregard for comiection with others, these individuals often ignore the effect that their actions may have on others (Blaney & Kutcher, (1991). These 19 individuals are uncomfortable in social interactions, are likely to appear irritable or hostile in these situations, and often have overtly conflicted and discordant relationships with family and peers (Kobak & Sceery, 1988; Mongrain, 1993). Preoccupied attachments, accompanied by excessive interpersonal concerns, are a result of early disruptions in caring relationships consisting of deprivation, inconsistency, or overindulgence that leads to an inordinate fear of loss of love, abandonment, and impoverishment. Both are linked to fears of rejection and excessive neediness to maintain close bonds with others (Blatt & Maroudas, 1992; Kobak & Sceery, 1988). Individuals characterized by these attachment styles and preoccupations are usually unable to openly express their feelings of anger or hostility and are more likely to conform to the wishes of others (Mongrain, 1993; Zuroff, Moskowitz, Wielgus, Powers, & Franko, 1983). Given this correspondence between attachment styles and depressogenic concerns, it is not surprising to find similar links between depressogenic concerns and antisocial activity on the one hand and attachment styles and antisocial activity on the other hand. Just as there is a link between self—critical preoccupations and antisocial activity, research suggests that individuals with dismissing attachments, are more likely to engage in antisocial activity than are individuals with preoccupied attachments or interpersonal concerns (Mongrain, 1993; Zuroff, et al., 1993). This relationship is seen with both children (Kobak & Sceery, 1988) and adolescents (Rosenstein & Horowitz, 1996). Additionally, Frank et al.’s (1997b) work supports the hypothesized link between depressogenic concerns and comorbidity of depression and antisocial involvement. Presumably, self-critical preoccupations are likely to lead to depressed mood, by nature of 20 the preoccupation itself, and to antisocial involvement as a result of an over-emphasis on independence and autonomy and a devaluing of connection to others. Reactance Psychological trait reactance is another factor that might help to account for depression comorbid with antisocial activity. Reactance originally was regarded as situationally induced (Brehm, 1966). When individuals were placed in situations that both emphasized the value of their personal freedom as well as threatened this freedom, they would, in turn, engage in behaviors directed at re-establishing their initial sense of freedom and control (Brehm, 1966; Brehm & Brehm, 1981). However, more recent research argues that reactance can be defined as a trait-like variable as well as a state-like variable (Dowd & Sanders, 1994). When reactance refers to stable aspects of personality, high reactant individuals are described as hypervigilant to threats to personal freedom, more likely to perceive a wide array of situations as potentially threatening their freedom, and generally are more oppositional than low reactant individuals (Beutler & Clarkin, 1990; Dowd & Wallbrown, 1993). Although the majority of studies have been situated in the laboratory rather than in ”real world" clinics, and almost exclusively have been done with adult populations, they nonetheless suggest that reactance is highly correlated with antisociality (Beutler, 1991; Dowd & Wallbrown, 1993; Graff & Luborsky, 1977; Weisz, Han, & Weiss, 1995). Pertinently, a recent study conducted with adolescent inpatients found that reactance correlated with anger, conduct problems, immaturity, authority problems, ego inflation, cynicism, social alienation, and proneness to, as well as actual use of, alcohol and drugs (F rank et al., 1998). Similarly, Frank, Schettini, and Lower (2000b) identified a strong 21 relationship between reactance and both undercontrolled aggression and delinquency in a sample of fourth through eighth graders. To some extent, reactance during adolescence appears to be normative. Research both in general population and clinic samples suggests a curvilinear relationship between age and trait reactance, with reactance during adolescence higher than during other periods of development (Frank et al., 1998; Tennen, Press, Rohrbaugh, & White, 1981), but with disturbed adolescents appearing unusually reactant, even more-so than normal college students (Frank et al., 1998). A recent study conducted with adolescents in a school setting suggested that reactance increases between 10 and 15 years of age (Frank, et al., 2000b), whereas a second investigator using an adult Australian sample found that levels of reactance decreased with age from 18 to 40 years. Somewhat heightened levels of reactance during adolescence may underlie a parallel increase in antisocial activity during the same developmental period. Some researchers suggest that reactance provides the context for adolescents to attain their developmental needs. For example, Dowd and Sieble (1990) suggest that reactance is instrumental in creating and maintaining autonomy, which is a critical developmental task for adolescents. In order to establish their independence, adolescents become resistant to advice and direction from others. Reactance can manifest itself in the form of argumentation, hostility, and oppositionality (Frank et al., 1998), which are often at the crux of adolescent antisocial activity. However, whereas "developmentally-normative reactant" adolescents may be oppositional and argumentative with their parents over relatively mundane and trivial matters, they are less-so regarding serious matters, and at worst, commit "minor” acts of delinquency, such as violating curfew or writing graffiti (Steinberg & Silverberg, 1986). 22 In contrast, atypically high reactant adolescents appear to be more pervasively oppositional and more deviant than their “developmentally-normative reactant” counterparts, and in turn, are more involved in antisocial activity of a serious nature (Frank, et al., 1998; Frank, et al., 2000b). Although the link between reactance and antisocial involvement is fairly well documented, the relationship between reactance and depression remains ambiguous. Although many researchers argue that there is no relationship between reactance and depression, one study conducted with a large population of Australian adults, ages 17 to 40 years, found that self-reports of psychological reactance were positively related to depression (Hong & Giannakopoulos, 1994). A recent study helps to account for these mixed findings. Frank, et al. (2000b) found that children and early-adolescents (grades 4 to 8), reporting intrapsychic conflicts regarding separa'tion-individuation (also known as conflictual dependency) also reported greater reactance a_nd greater emotional impairment. In turn, more reactant children were more likely to be antisocial. However, whereas conflictual dependency was linked to both reactance and mood disturbance, reactance was not directly associated with emotional difficulty. Rather, this study suggests that depression is likely to be accompanied by antisocial behavior when conflicts over separation-individuation produce reactance as well as emotional dysfunction. An alternative, though not mutually exclusive argument, is that reactance and depression co-occur because both are rooted in self-criticism. Frank et al. (1997a) showed that self-criticism is a "by-product" of conflictual dependency. Because self- critical adolescents are excessively concerned with preserving their autonomy, it is reasonable to presume a relationship between self-criticism and reactance. Moreover, 23 because both reactance and self-critical preoccupations are both rooted in struggles with authority, they may independently result in greater antisociality. In short, various research findings argue that the relationship between depression, delinquency, reactance, and self-critical concerns all can be explained using a "common etiology" model, with conflictual dependency being the common source. Another Look At Gender The above discussion argues that there are several factors that potentially discriminate between depressed adolescents who are comorbidly antisocial and depressed adolescents who are not comorbidly antisocial. One such factor mentioned above is gender; however, gender may be both a discriminating factor as well as a moderating factor. Gender is linked to each of the other discriminating factors and in some circumstances gender may change the nature of the relationships between the different factors and the existence of comorbidity of depression and antisocial activity. Therefore, the factors that discriminate comorbid depression and antisocial involvement may be different when gender is considered. For example, ADHD is more common in boys than girls; hence, it follows that there is a greater rate of comorbid depression and antisocial behavior in boys, in part because ADHD is an etiological factor in antisociality and related to depression via increased risk for social role failure. In addition, researchers have found that boys are more likely than girls to have ADHD with aggression or conduct problems, as opposed to ADHD without aggression and conduct problems (Herrero, Hechtman, & Weiss, 1994), inferable from the higher prevalence of both disorders among boys. However, the possibility that girls, with ADHD, as compared to boys with ADHD respond to social- 24 role failure more often with depression rather than antisocial behavior to some extent balancing comorbid prevalence rates, is a question that needs to be considered as well. Similar relationships may also be found for the factors of school failure and depressogenic concerns. Research has found clear gender differences in depression with regards to its relationship to school failure. Depression is more often a result of school failure in girls than boys (Cole, Martin, Peeke, Seroczynski, & Fier, 1999; Ge, Lorenz, Conger, & Elder, 1994). Additionally, the research discussed above suggests that depressogenic concerns, particularly self-critical concerns are influential in predicting both depression and antisocial involvement; however, self-critical preoccupations are more common in boys than girls; and therefore, gender might moderate the proposed relationship. Therefore, gender may moderate the relationships between school failure as well as depressogenic concerns and comorbid depression and antisocial involvement. In contrast, the literature is still relatively unclear as to the relationship between gender and reactance. Research on gender differences for psychological reactance is mixed. Brehm and Brehm (1981) suggest that there is no reason to assume that reactance is gender-specific or that one gender exhibits higher levels of reactance than the other. Hong and Page (1989) and Hong (1993) supported this view. Both found no significant gender differences in psychological reactance for university student populations. In contrast, other researchers with American university graduates and undergraduates and early adolescents have found that that boys were significantly more reactant than girls (Courchaine, Loucka, & Dowd, 1995; Dowd, Wallbrown, Sanders & Yesenosky, 1994; Frank et al., 2000b; Joubert, 1990). Joubert (1990) reported that men scored higher than women on Hong’s reactance scale in a smaller sample of American university students. 25 Conceivably, if boys are more reactant, and reactance contributes to both depression and delinquency, it is not surprising that there is a higher prevalence of overlap between depression and delinquency in boys. Summary and Hypotheses To summarize, this study will examine factors contributing to antisocial involvement in a sample of adolescents with severe levels of emotional disturbance. In general, the research suggests that gender will discriminate between comorbid and non- comorbid groups, with boys being more characteristically comorbid. Moreover, ADHD, academic difficulties, self—critical preoccupations, and high levels of reactance are expected to be more characteristic of the comorbidly antisocial group than the depressed only group. An additional possibility is that gender will moderate whether or not ADHD, academic difficulties, self-critical preoccupations, and high reactance will discriminate between comorbidly antisocial and depressed-only adolescents. The current study will contribute to the understanding of comorbid. depression and antisocial activity in adolescents by overcoming several past limitations. The assessment of child and adolescent functioning has been limited by the use of unidimensional measures, which focus on symptomatology of one problem alone and disregarding the influence of comorbid symptomatology. This study will overcome this limitation by using a measurement devised to assess the multiple dimensions of child functioning. In addition, the majority of research described above has focused on adult populations, and most studies have been comprised of general population samples; thereby leaving both the adolescent years and the clinic-referred populations fairly unexplored. This is especially true for research on depressogenic concerns and reactance. Therefore, the 26 current study will overcome this limitation and extend the current research base by using an adolescent, inpatient sample. Additionally, much of the findings above stem from research with small sample sizes. The present study will use a large sample of inpatient adolescents. 27 METHOD Participants The sample includes 287 depressed adolescent inpatients (176 girls and 1 1 1 boys), ages 11 to 17 years (M: 14.89), admitted over a 36 month period to a private child and adolescent acute psychiatric hospital in the Midwest. Participants were mostly white (89.2%), with 67% living in a two-parent (natural, adopted, or step) home, 1% living in a therapeutic residence, and the remaining adolescents living in single-parent homes (natural, adopted, step, or relative). Information on household structure was missing for 4% of the cases. Mean annual income (available for 255 cases) was 3.67 (_S_D=l .80) as measured on an 8-point scale with l=<$8,000; 2=$8,000-$11,999; 3=$12,000-$l9,999; 4=$20,000 to $29,999; 5=$30,000-$44,999; 6=$45,000-$69,666; 7=$70,000-$100,000; and 8>$100,000. Approximately 6% of the households were earning less than $8,000 and 3% were earning $70,000 or more per year. Inclusion Criteria. All adolescents included in the study were moderately to severely depressed as indicated by two of the inclusion criteria. To be eligible for the study, adolescent patients had to report severe impairment associated with emotional difficulties, as indicated on the Functional Impairment Scale for Children and Adolescents — Youth Self Report version. In addition, they EM were required to (a) have a score of 77 or greater (indicating clinically significant depression) on the Reynolds Adolescent Depression Scale or (b) a discharge diagnosis, recorded in the discharge summary, of either Major Depressive Disorder, Dysthymic Disorder, or Depressive 28 Disorder-Not Otherwise Specified. The inclusion criteria and decision making process are illustrated in Figure 1. Measures The Functional Impairment Scalefor Children and Adolescents - Youth Self Report (FISCA-SR; Frank & Paul, 1995). The FISCA-SR will be used to assess functional impairment associated with emotional disturbance, school achievement, and antisociality (delinquent and aggressive behaviors). The FISCA-SR is a 172-item self- administered questionnaire with eight scales measuring separate dimensions of child and adolescent impairment. In addition to the scales used here assessing undercontrolled aggression, delinquency, problems related to feelings and moods, and school difficulties, other scales include alcohol and drug problems, problems at home, thinking problems, and self-harm. The content of the FISCA-SR is essentially the same as on a parent-report version of the same instrument (the F ISCA), which was developed before the self-report version. FISCA-SR items use a true-false, Likert or multiple-choice format. Impairment criteria for scoring the FISCA-SR are based on a modified and abbreviated version of the Child and Adolescent Functional Assessment Scale (CAF AS; Hodges, 1994; Hodges & Gust, 1995). An objective scoring procedure links item responses to specific criteria identifying mild (scored as 10), moderate (scored as 20) or severe (scored as 30) levels of impairment with each of the 8 domains measured by the FISCA-SR. A domain score of "0" (no impairment) indicates that an adolescent does not meet any criteria at any of the three levels within that domain of functioning. The number of criteria associated with each of the FISCA-SR domains ranges from 3 to 14. The final score for each impairment 29 Figure 1. Inclusion Criteria and Decision Making Process Adolescents Admitted N=708 YES (N=5 62) 1. FISCA—SR Severe Functional NO Between 12/19/95 and 12/9/98 ’ Not Included NO (N=146) Impairment on the 1—> Not Included Feelings & Moods N=221 N=341 2. RADS 3. Diagnosis OR 1. Major Depressive Raw Score277 Disorder NOS 2. Dysthymic Disorder 3. Depressive Disorder - NO YES: (Either 2 o_r 3) INCLUDED IN STUDY N=287 30 Not Included N=54 dimension reflects the highest level of impairment at which the child meets one or more criteria. Total impairment, ranging from 0 to 240, is computed by summing scores across domains. Table 1 shows criteria for severe impairment on the Feelings & Moods, Control of Aggression, and Delinquency scales. Table 2 illustrates scoring procedures for 2 (of 7) criteria indicating a severe level of impairment in school, along with related items on the FISCA-SR “keyed” to these criteria. Criteria are deemed “present”, scored as 1, or “not present”, scored as 0, depending on whether they meet the rules shown in column 4 of the table. To compute internal reliability, domain criteria are treated as items (scored as 0 or 1). Frank et al. (2000a) reported adequate internal reliability for all but the Home scale (9g = .26). Alpha coefficients for the seven other scales (with 3 to 14 criteria, each) ranged from .61 to .88 with a mean of .79. Alpha levels for the scales used in this study are: Feelings and Moods (a measure of emotional dysfunction) (g=.72, 14 items), Undercontrolled Aggression, (g=.77, 13 items), Delinquency (g=.66, 8 items), and School (pg-=69, 14 items). The parent-report FISCA was found to discriminate inpatients from outpatients (Paul, 1996), predict hospital recidivism (Frank, et al., 2000a), and correlate meaningfully with daily staff ratings of child and adolescent behaviors on an inpatient unit (Van Egeren, Frank, & Paul, 1999). A recent study using agreement between parent and self-report forms of the FISCA, showed that a hypothesized three- factor model of child functional impairment (based on extant literature) that distinguishes “overt” from “covert” forms of externalizing behaviors and both of these from 31 Table 1 Criteria For Establishing Severe Levels of Impairment Feelings & Moods 1. Mood problems are accompanied by suicidal intent 2. Extreme emotional dysregulation; unusual or very intense expression of emotions that others see as odd or strange 3. Depression or anxiety are associated with academic incapacitation (school absences, poor grades, performance deficits, etc.) 4. Depression or anxiety are associated with social isolation or withdrawal Control of Aggression 1. Child has no age-appropriate friends because behavior is chronically hostile, belligerent, or exploitative 2. Frequently very cruel to animals 3. Frequent and/or serious physical aggression; threatened or used a weapon against others, attacked and/or seriously hurt others; or behavior is so dangerous or out of control that child has been removed from home or school 4. Sexually abused, molested, or assaulted someone of the same or opposite sex Criteria For Establishing Severe Levels of Impairment cont. Delinquency: 1. Intentionally and severely damaged property outside the home 2. Set fires with malicious intent 3. Severely delinquent or criminal behavior involving confrontation or harm to a victim or severe law violation (e.g. auto theft, robbery, mugging, purse snatching, dealing or carrying drugs, threatening with a weapon, break-ins, physical assault, murder, sexual assault) 4. Delinquent or criminal behavior resulted in incarceration or confinement in a jail or Detention center 33 0:008:80 80:: H0 0802 A0 .H00:8 H0 80 80:8 0: 0:88 H 80:: 880000 %8 00803 28050 30:8 00 0008088 803 H 90 888:0 >8 80:0 803800 00 80:80:»: 00 880000 >8 00,005 :0 00:00 83050 H00:0m H0 808% 0000 w8t0w 803 00 0.30000 8800 82 H A: .808: 0008 :0 m 828000: 038 00. 00: H 00 200:8 00. 0m 00 0830: 00 H00:8 0008va 50% mmD880 082 .vm 88: cm as: sea a 88: em 9 : e ”H0 >80 H E 808: CH 00 m A0 808: v 00 H 800300: A: 00>08 90 00880800 3800000 0000 Q24 200:8 00 80:00 03:88 8 0 .S pnmm 0008on 50% 030: 8:808 m 88H 0:: mars: 8088 808 >03 .mm 00:80: >088: 088:0 8% oz a 8008830 8008830 05 8 03088088; 0:0 8 030088088: \erH 00 80> 80810 0003 :0» 28 00:88 0:: 8:808 m 88H 0:: wcgfl .me 31003008 EEO 000>0m >0vH w8000m 880: 80>0H0~H 80:00:00 60: .0H08 H00:0m 0:: 80¢ 0:80:00 200058,. 030, 00H 0800008 .0888 ;~ .. e: .....r....... .2. L ..:r . . 2...... . ..... £71.. 37...... Titti: . 3.. .7 . 1. m S7 5- if... ya c. .. .mxms... A. .5395: :_ r. u 2...“ I. 43.1. . 3:25. .......I i. E... i... ... t. . «u .3 . .. : ...... i .2». . 5;: a: : 7.