Citation
Diagnostic Validity of DSM Symptoms and Criteria for Preschoolers

Material Information

Title:
Diagnostic Validity of DSM Symptoms and Criteria for Preschoolers An Exploration of Disruptive Behavior and Attention Deficit Hyperactivity Disorder Symptoms
Creator:
Chase, Rhea
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (80 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology
Clinical and Health Psychology
Committee Chair:
Eyberg, Sheila M.
Committee Members:
Marsiske, Michael
Johnson, James H.
Graber, Julia A.
Graduation Date:
8/9/2008

Subjects

Subjects / Keywords:
Adolescents ( jstor )
Age groups ( jstor )
Attention deficit and disruptive behavior disorders ( jstor )
Attention deficit hyperactivity disorder ( jstor )
Child psychology ( jstor )
Childhood ( jstor )
Factor analysis ( jstor )
Modeling ( jstor )
Parents ( jstor )
Symptomatology ( jstor )
Clinical and Health Psychology -- Dissertations, Academic -- UF
attention, conduct, defiant, deficit, diagnosis, disorder, dsm, hyperactivity, oppositional, preschool
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Psychology thesis, Ph.D.

Notes

Abstract:
Attention Deficit Hyperactivity Disorder (ADHD) and the disruptive behavior disorders of Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) represent the most common child psychological disorders diagnosed today, and are increasingly being diagnosed in preschool-aged children. Developmental theory suggests that the criteria outlined in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) may be inappropriate for younger children, but very little research has examined the validity of these diagnostic categories and assessment measures for this age group. This study therefore explored the validity of current diagnostic standards and assessment measures of ADHD, ODD, and CD specifically in preschool-aged children. The sample consisted of 207 children aged three to six years referred for treatment of symptoms of ADHD and disruptive behavior. Confirmatory factor analysis was used to compare three possible symptom models of these diagnoses: (a) a four-factor model of ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD currently implied by the DSM-IV; (b) a three-factor model that combined ADHD symptoms into one single factor, but included separate constructs of ODD and CD; and (c) a three-factor model allowing for separate ADHD-Inattentive and ADHD-Hyperactive factors but collapsed symptoms of ODD and CD into one disruptive behavior construct. Findings supported the four-factor model represented by the DSM-IV, which provided superior fit for the data as compared to both of the less differentiated models. The study also further explored current assessment strategies with this age group. We assessed and compared the validity of the fourth edition of the Diagnostic Interview Schedule for Children (DISC-IV), the Young Child Version of the Diagnostic Interview Schedule for Children (YC-DISC), and the fourth edition of the Early Childhood Inventory (ECI-4). All measures demonstrated similar correlations with other measures of child behavior, suggesting comparable validity across measures for this age group. Overall, results support the use of DSM-IV diagnostic categories with preschool-aged children and add to the limited research on the validity of current assessment methods for young children. These findings help promote accurate identification and subsequent treatment of psychopathology in preschoolers. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2008.
Local:
Adviser: Eyberg, Sheila M.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31
Statement of Responsibility:
by Rhea Chase.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Chase, Rhea. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2010
Classification:
LD1780 2008 ( lcc )

Downloads

This item has the following downloads:


Full Text





entire sample. These items were: (a)forcing sexual activity, (b)fire-setting, (c)breaking and

entering, (d)missing curfew, (e)running away, and (f)truancy. These six items were removed

from further analysis. The ECI-4 does not include breaking and entering, missing curfew,

running away, or truancy, and given the low frequency of these items when assessed in the

DISC-IV and DISC-YC, their exclusion seems warranted. Additionally, the ECI-4 assesses

stealing behaviors only, while the DISC-IV and DISC-YC adhere to DSM-IV criteria and

differentiate between stealing with and without confrontation. As can be seen in Table 3-2,

stealing with confrontation occurred at low frequencies in both the GIFT (5%) and SHAPE (2%)

samples. Further analysis therefore used a "stealing composite" symptom which included

stealing with and without confrontation.

Confirmatory Factor Analyses of DSM Symptoms

CFA was used to determine which of three models provided the best fit for the 35 DSM

symptoms of ADHD, ODD, and CD. The first model, referred to as the DSM-IV model, included

four separate factors representing ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD. The

second model, the combined ADHD model, collapsed the inattentive and hyperactive ADHD

symptoms and subsequently resulted in three factors of ADHD, ODD, and CD. The disruptive

behavior model also included three factors but combined ODD and CD symptoms, resulting in

the constructs of ADHD-Inattentive, ADHD-Hyperactive, and ODD/CD.

Fit indices for each of three models were compared and are presented in Table 3-4. All

three models resulted in a significant chi square; however, the ratio of the chi-square to the

degrees of freedom did not exceed the cutoff value of 2 for any of the three models. Therefore,

all three models provided an adequate fit for the data. However, as described above, the

conservative cutoff scores recommended for binary indicators and a small sample size are: CFI >

.95, TLI > .95, WRMR < 1.0, RMSEA < .06 (Yu & Muthen, 2002). Using these guidelines, the









category. The DSM-III rejected the psychoanalytic framework underlying its previous edition

and instead based diagnoses on empirically derived categories. The DSM-III also stipulated that

a "disorder" is characterized by functional impairment and subjective distress from the patient.

The multiaxial system was first introduced in the DSM-III, which required clinicians to assess

developmental disorders, personality disorders, medical conditions, and global functioning. The

DSM-III also provided information on clinical presentation of disorders, such as age of onset and

gender differences (APA, 1980). Inclusion of these factors demonstrated some acknowledgement

of the role of context and environment in the expression of psychopathology.

The number of specific childhood disorders significantly increased from the second to the

third edition of the DSM (Bemporad & Schwab, 1986). The DSM-II section of Behavior

Disorders of Childhood and Adolescence included short and somewhat vague descriptions of a

range of child problem behaviors, including "runaway reaction of childhood or adolescence" and

unsocializedd aggressive reaction of childhood or adolescence" (DSM-II; APA, 1968). The

DSM-III introduced the specific disruptive behavior disorders of Oppositional Defiant Disorder

(ODD) and Conduct Disorder (CD). ODD is characterized by a pattern of child noncompliance

and disrespect for adult authority, while CD is characterized by a repetitive and persistent pattern

of behavior in which the basic rights of others or major age-appropriate societal norms or rules

are violated (APA, 1980). The DSM-III also further developed the DSM-II diagnosis of

"hyperkinetic reaction of childhood" and introduced the diagnosis of Attention Deficit Disorder

(ADD), which was viewed primarily as a disorder of inattention and could be diagnosed with or

without hyperactivity. The refinement of childhood disorders continued with the revised version

of the DSM-III, which included age-appropriate modifications to several symptoms in an attempt

to increase their developmental sensitivity (DSM-III-R; APA, 1987). Importantly, the DSM-III-









symptoms entered into the factor analysis were assessed with different instruments. However, all

items mapped onto DSM-IV criteria. Indeed, the confirmation of the DSM-IV structure despite

the methodological difference may argue for the strength of the DSM-IV model. Also of note,

we found few differences in the convergent validity of the different measures, suggesting that

they may provide similarly accurate diagnostic information.

Drawing participants from different studies also introduced variability in the versions used

for study measures. Specifically, children from Project GIFT were assessed with an older version

of the CBCL (Achenbach; 1991; Achenbach, 1992) than children in Projects SHAPE and GANA

(Achenbach; 2000; Achenbach; 2001). Although these versions were highly similar, particularly

in the subscales of interest for this study, there were some differences across measures. Perhaps

most importantly, the 1992 version of the CBCL for 2 and 3-year-old children did not include an

Attention Problems subscale. Thus, 3 year-old children from Project GIFT were excluded from

the analyses involving this measure. However, this represented a relatively small group of

children relative to the entire sample, and it seems unlikely that their inclusion would have

significantly changed the results.

A relatively small sample size is another limitation of the current study. Our sample of 207

was enough to meet the recommendation of 5 cases per item when conducting factor analysis

(Floyd & Widaman, 1995). However, larger sample sizes are considered optimal for factor

analysis. The sample size also prevented cross-validation of either the DSM-IV or ECBI model

by age, race, or gender. Overall, previous research with older children and adolescents supports

the factors implied by the DSM-IV across age groups and gender (Lahey et al., 2008). However,

it will be important to confirm factor invariance in younger children.









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DIAGNOSTIC VALIDITY OF DSM SYMPTOMS AND CRITERIA FOR PRESCHOOLERS:
DISRUPTIVE BEHAVIOR AND ATTENTION DEFICIT HYPERACTIVITY DISORDER
SYMPTOMS

By

Rhea Marisa Chase

August 2008

Chair: Sheila M. Eyberg
Major: Psychology

Attention Deficit Hyperactivity Disorder (ADHD) and the disruptive behavior disorders of

Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) represent the most common

child psychological disorders diagnosed today, and are increasingly being diagnosed in

preschool-aged children. Developmental theory suggests that the criteria outlined in the fourth

edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) may be

inappropriate for younger children, but very little research has examined the validity of these

diagnostic categories and assessment measures for this age group. This study therefore explored

the validity of current diagnostic standards and assessment measures of ADHD, ODD, and CD

specifically in preschool-aged children. The sample consisted of 207 children aged three to six

years referred for treatment of symptoms of ADHD and disruptive behavior. Confirmatory factor

analysis was used to compare three possible symptom models of these diagnoses: (a) a four-

factor model of ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD currently implied by the

DSM-IV; (b) a three-factor model that combined ADHD symptoms into one single factor, but

included separate constructs of ODD and CD; and (c) a three-factor model allowing for separate

ADHD-Inattentive and ADHD-Hyperactive factors but collapsed symptoms of ODD and CD









R was the first version to use research from DSM field trials to clarify and validate diagnoses,

and results from these trials were used to inform and refine diagnostic criteria for several child

disorders, including the former ADD category now classified as Attention Deficit Hyperactivity

Disorder (ADHD; DSM-III-R; APA, 1987).

The fourth edition of the DSM (DSM-IV; APA, 1994) continued to develop childhood

diagnoses and expand subtypes of certain child disorders. DSM-IV criteria for ODD require that

children meet four out of eight possible symptoms, such as: arguing with adults, actively defying

or refusing to comply with adults' requests or rules, and deliberately annoying others. CD criteria

have changed somewhat since the disorder's first appearance in the DSM-III, although the core

feature, involving blatant and persistent disregard for others or societal rules, has remained the

same. Current criteria necessitate three or more behaviors relating to aggression, destruction of

property, deceitfulness or theft, and major violations of societal norms or rules. The text of the

DSM-IV states that a diagnosis of CD should reflect internal psychological dysfunction and not a

reaction to a negative environment. However, this qualification is not included in DSM-IV

diagnostic criteria, leading some to question the clinical utility of this text comment (Wakefield,

Pottick, & Kirk, 2002).

The DSM-IV also revised the diagnostic categories related to ADHD, which now consist

of three subtypes: predominantly inattentive, predominantly hyperactive, and combined type.

Children must meet six of nine possible inattentive symptoms, such as difficulty sustaining

attention and organizing tasks and activities, or six of nine hyperactivity/impulsivity symptoms,

such as fidgeting or squirming and interrupting or intruding on others. To warrant a diagnosis of

ODD or ADHD, symptoms must have persisted for at least 6 months; in the case of CD, three or

more criterion symptoms must have been present within the last twelve months, with at least one









The third model examined the predictive power of CD symptoms and child age in rates of

aggressive behavior. In this analysis, scores from the aggression scale of the CBCL were

regressed upon total number of CD symptoms, child age, and an interaction term between CD

symptoms and child age to test the hypothesis that age would moderate the relationship between

CD symptom count and severity of externalizing behavior. Total number of CD symptoms was

entered in the first block and acted as a significant predictor, F(1, 199) = 51.51, p < .001, R2

.206. Child age was entered into the model and did not add to the predictive power of the model,

R A = .002, nor was the interaction term between number of symptoms and age significant, R2 A

= .001. Thus, number of CD symptoms was a significant predictor of aggression scores on the

CBCL. However, the analysis failed to support the hypothesis that age would moderate that

relationship between symptom count and severity of aggressive behavior.

In sum, this study failed to find support for the hypothesis that age would significantly

moderate the relationship between symptom count and the severity of problem behaviors.

However, it should be noted that the sample represented a limited age range, and thus it may

have been difficult to find differences by age.









Further study in this area is crucial to promote effective assessment and subsequent treatment in

this age group.

Current Assessment Strategies

Along with further efforts to validate ADHD and the disruptive behavior disorders in

young children is the need to improve our assessment of these diagnoses in the preschool

population. Relatively few measures have been specifically developed for preschoolers; more

commonly, assessment methods developed for older children are applied to younger children.

One important direction therefore includes further study of the validity and reliability of

commonly used assessment instruments for preschoolers. The Diagnostic Interview Schedulefor

Children (DISC-IV; Shaffer et al., 2000) is a structured diagnostic interview of child

psychopathology based on DSM-IV criteria. The DISC-IV is a highly structured diagnostic

interview that is commonly used to assess psychopathology in children, particularly in research

studies. A parent version (DISC-IV-P) exists for parents or caregivers of 6- to 17-year-olds, and

a child self-report (DISC-IV-C) is used for 9-to 17-year-olds. Some research studies have used

the DISC-P with parents of children as young as 4 (Speltz, McClellan, DeKlyen, & Jones, 1999),

but most research on this instrument involves older children and adolescents. A recent

modification of the DISC-IV was designed to increase its developmental sensitivity for

preschoolers (Young Child Version of the Diagnostic Interview Schedule for Children; (YC-

DISC; Lucas, Fisher, & Luby, 2000). The YC-DISC differs from the DISC-IV by modifying the

language of symptoms to be more appropriate for young children. The interview also modified

the time frame for each symptom. In the DISC-IV, parents are asked whether a symptom has

been present for the past month and then for the past year. The YC-DISC asks only whether a

symptom has been present for the past three months. This newly developed instrument has yet to

be studied empirically to determine its validity in this age group.









LIST OF TABLES


Table page

2-1 Overview of relevant measures across datasets...........................................................35

3.1 P participant characteristics. ...................................................................... .....................45

3-2 Individual symptom percentages across the DISC-IV, YC-DISC, and ECI-4 ..................46

3-3 Percentage of diagnosis across assessment method.........................................................47

3-4 Absolute, relative, and component model fit for three DSM-IV CFA models..................48

3-5 Four-factor solution of D SM -IV sym ptom s ........................................... .....................49

3-6 C orrelations am ong D SM -IV factors ............................ .......................... .....................50

3-7 Pearson correlations between symptom counts and CBCL and ECBI scores ...................51

3-8 Absolute, relative, and component model fit for ECBI models......................................52

3-9 The CFA of the three-factor model of the ECBI ................... ........... ............... 53

3-10 O ne-factor C F A of the E C B I ............................ ...................................... .....................54

A-1 One-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR =. 10).............68

A-2 Two-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR= .08) ............69

A-3 Three-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR = .06) ..........70










Table 3-2. Individual symptom percentages across the DISC-IV, YC-DISC, and ECI-4
Symptom DISC-IV YC-DISC ECI-4
(GIFT; n = 99) (SHAPE; n= 53) (GANA; n = 55)
ADHD-Inattentive
Careless mistakes 43 67 24
Short attention span 55 87 35
Doesn't listen 93 93 46
Fails to finish tasks 70 91 51
Disorganized 53 76 40
Avoids concentrating 52 82 27
Loses items 44 65 36
Distractible 70 89 64
Forgetful 41 72 33
ADHD-Hyperactive
Fidgets 74 91 66
Leaves Seat 81 85 66
Runs/Climbs 90 91 76
Too Loud 63 67 55
Driven by motor 77 85 66
Excessive talking 71 78 56
Blurts out answers 48 59 31
Difficulty waiting turn 78 85 58
Interrupts 90 94 49
ODD
Loses temper 95 74 36
Argues 96 80 51
Defiant 95 90 75
Annoys others 76 63 56
Blames others 51 56 49
Easily annoyed 75 59 47
Angry 84 70 40
Spiteful 61 39 43
CD
Bullying 42 22 11
Fighting 27 24 16
Use of weapon 26 15 4
Cruel to people 24 9 9
Cruel to animals 27 15 5
Stole with confrontation 5 2 **
Forced sex* 0 0 2
Firesetting* 1 0 2
Destruction of property 50 50 10
Breaking and entering* 0 0 **
Lying 23 33 15
Stole without confrontation 29 31 **
Missed curfew* 0 0 **
Runaway* 0 0 **
Truancy* 0 0 **
Stealing composite*** 29 31 2
*Dropped from further analyses due to low frequency: **Not asked on ECI-4; ***Combination of stealing with and
without confrontation









CHAPTER 2
METHODS

Participants

Participants were 207 children between the ages of 3 and 6, inclusive, who received a

diagnosis of ADHD, ODD, CD, or a combination of ADHD and ODD/CD. To increase sample

size and to allow adequate power for the type of statistical analyses proposed, participants were

drawn from several studies, each of which is described below. All children were involved in

research on the effectiveness of Parent-Child Interaction Therapy (PCIT). Exclusion criteria

included immediate crisis requiring hospitalization or out-of-home placement.

Project GIFT

A subset of participants was drawn from a study of PCIT and the maintenance of treatment

gains (Project GIFT). One hundred boys and girls between the ages of 3 and 6 were enrolled in

the study after being referred to a university psychology outpatient clinic for treatment of

behavior problems; data were available from 99 of the original 100 participants. Families were

referred for treatment by pediatricians, teachers, other mental health providers, or were self-

referred. Inclusion criteria for Project GIFT stated that the child must meet DSM-IV diagnostic

criteria for the diagnosis of ODD and must live with at least one parent able to participate in

treatment. Children were excluded from the study if they scored below a standard score of 70 or

if their parent scored below a standard score of 75 on cognitive screening measures, if they had a

major sensory impairment such as blindness, or if they had been diagnosed with a pervasive

developmental disorder. Children who were taking psychotropic medication were required to be

stabilized on this medication for at least one month prior to entering the study.









APPENDIX
STATISTICAL ANALYSES NOT INCLUDED IN MAIN TEXT

Table A-1. One-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR =.10)
ECBI item
12. Gets angry when doesn't get own way .89
8. Does not obey house rules on own .88
10. Acts defiant when told to do something .87
30. Is easily distracted .87
28. Constantly seeks attention .84
6. Slow in getting ready for bed .83
34. Has difficulty concentrating on one thing .82
32. Fails to finish task or projects .82
35. Is overactive or restless .80
14. Sasses adults .79
2. Dawdles or lingers at mealtime .79
20. Is careless with toys and other objects .79
16. Cries easily .78
4. Refuses to eat food presented .75
22. Lies .72
9. Refuses to obey until threatened with punishment .71
24. Verbally fights with friends own age .69
26. Physically fights with friends own age .67
18. Hits parents .63
36. Wets the bed .61
29. Interrupts .56
31. Has short attention span .46
13. Has temper tantrums .46
7. Refuses to go to bed on time .45
21. Steals -.44
5. Refuses to do chores when asked .42
15. Whines .40
17. Yells or screams .36
11. Argues with parents about rules .35
1. Dawdles in getting dressed .26
3. Has poor table manners .22
19. Destroys toys and other objects .12
23. Teases or provokes other children -.06
33. Has difficulty entertaining self alone .05
25. Verbally fights with sisters and brothers .01
27. Physically fights with sisters and brothers .01









following guidelines: CFI > .95, TLI > .95, RMSEA < .06, and WRMR < 1.0 (Yu & Muthen,

2002). The fit indices of each of the three models were held to these guidelines. Additionally,

adjusted 2 difference tests allowed direct comparison of the models. Specifically, the DSM-IV

model was compared to the combined ADHD model. The better-fitting of these two was then

compared to the disruptive behavior disorder model. A significant difference in the 2 suggested

that the more differentiated model fit the data better than the less differentiated (Sterba, Egger, &

Angold, 2007).









ODD/CD construct seems most relevant to preschoolers. However, further research examining

CD symptoms in preschoolers is necessary given the conflicting results surrounding the validity

of separate ODD and CD factors within child populations (Sterba, Egger, & Angold, 2007;

Burns et al., 1997). Indeed, further research specifically examining CD symptoms in clinical

preschool populations seems warranted. The disruptive behavior disorders and ADHD are

increasingly being diagnosed in preschool populations, highlighting the need for further

assessment of the validity of these disorders in this age group.

Issues in Preschool Diagnosis

Diagnosing psychopathology in preschool aged children raises important questions

regarding developmental considerations for this population. As discussed above, it was only

recently that the DSM began to consider differences between children and adults, and the current

diagnostic system reflected by the DSM-IV provides few modifications for very young children.

These may be important issues, as the rapid development characterizing early childhood makes it

difficult to assure that maladaptive symptoms are a sign of true psychopathology and not a

variation of normal child development. Additionally, the DSM provides little information on the

normative development of young children to guide diagnostic decisions in this age group

(Eyberg, Schuhmann, & Rey, 1998; Egger & Angold, 2004).

These issues in preschool diagnosis are particularly relevant in the diagnoses of ODD, CD,

and ADHD. Due to their cognitive and social development, and the struggle for independence

from their parents that commonly occurs in the second or third year, some level of aggression

and noncompliance is expected in preschool-aged children (Campbell, 2002). Thus, a certain

level of disruptive behavior symptoms may be present as part of normal child development.

Similarly, all preschool children are expected to display some level of inattention and

hyperactivity. A young child may have difficulty sitting still and sustaining attention, and









DSM four-factor model provided better fit for the data than either the combined ADHD three-

factor model or the disruptive behavior three-factor model. Chi square difference testing

confirmed these results (Muthen & Muthen, 2005). Specifically, when the DSM model was

compared to the combined ADHD model, the DSM model provided significantly better fit, X2 (2,

N= 207) = 14.02, p < .001). Similarly, when the DSM model was compared to the disruptive

behavior model, the DSM model provided significantly better fit, X2 (2, N= 207) = 11.09, p <

.001). Thus, results fail to find support for collapsing the ADHD symptoms across inattentive

and hyperactive domains or for combining ODD and CD symptoms in this clinical sample of

preschoolers. Table 3-5 shows that the loadings of the individual items onto their factors were

generally high; the path diagram is shown in Figure 1. The two ADHD factors were highly

correlated, as were the ODD and CD factors. Correlations between the ADHD and ODD, as well

as ADHD and CD, demonstrated a moderate relationship. Correlations among the six factors are

included in Table 3-6.

The CFA results provide support for the four-factor structure currently implied with the

DSM-IV. The modification indices provide further evaluation of model fit, as they provide the

expected reduction in chi square if a parameter were freely estimated. Notably, no modification

indices were above the minimum value often (Muthen & Muthen, 1998-2006).

Validity of the Three Assessment Methods

To provide some measure of validity of the three assessment measures included in this

study, ADHD, ODD, and CD symptom counts from the DISC-IV, DISC-YC, and ECI-4 were

correlated with the ECBI, CBCL Attention Problems and CBCL Aggressive Behavior subscales,

as well as the broadband Externalizing scale from the CBCL (See Table 3-7). Examination of the

correlation matrix revealed that symptom counts across different assessment methods correlated









below a standard score of 75 on cognitive screening measures, if they had a major sensory or

mental impairment such as blindness, or if they had been diagnosed with a pervasive

developmental disorder. One of the specific aims of Project SHAPE involved the investigation of

the effectiveness of psychosocial, rather than pharmacological, treatment of ADHD. Therefore,

children using psychotropic medications were excluded from this particular study.

Measures

Demographic and Background questionnaire: Parent questionnaires were used to obtain

descriptive information about the child and family including sex, age, race/ethnicity, occupation,

education level, and medical history.

Eyberg Child Behavior Inventory: (ECBI; Eyberg & Pincus, 1999). The ECBI is a 36-

item parent rating scale of disruptive behavior. The ECBI Intensity Scale measures the frequency

with which disruptive behavior occurs on a likert scale from 1 to 7, and the Problem Scale asks

whether the parent finds the behavior problematic for themselves on a yes-no scale. The Intensity

and Problem Scales of the ECBI yield internal consistency coefficients of .95 and .93 (Colvin,

Eyberg, & Adams, 1999), interrater (mother-father) reliability coefficients of .69 and .61

(Eisenstadt, McElreath, Eyberg, & McNeil, 1994), and test-retest reliability coefficients of .80

and .85 across 12 weeks and .75 and.75 across 10 months, respectively (Funderburk, Eyberg,

Rich, and Behar, 2003). Studies documenting convergent, discriminant, and discriminative

validity of the scales have been summarized in the test manual (Eyberg & Pincus, 1999). The

alpha coefficient for the ECBI in this study was .89.

Child Behavior Checklist: (CBCL; Achenbach, 1991; Achenbach, 1992; Achenbach &

Rescorla, 2000; Achenbach & Rescorla, 2001). The CBCL is a commonly used measure of child

behavior. The CBCL has different forms according to child age and recently underwent a

revision. Therefore, analysis involved four different versions of the CBCL. In Project GIFT,









and psychosocial and developmental risk factors. The authors concluded that, rather than

debating the superior assessment technique, efforts should instead be directed towards validating

diagnostic categories, because neither the categorical nor dimensional approach exhibited

particularly strong relations with the external validators (Jensen et al., 1996). Currently, the

"gold standard" of diagnosis involves a combination of categorical and dimensional approaches.

The Eyberg Child Behavior Inventory (ECBI) is a widely used parent rating scale

measuring disruptive behavior disorders in children ages 2 to 16 (Eyberg & Pincus, 1999). The

validity and reliability of this measure is well-established (Boggs, Eyberg, & Reynolds, 1990;

Funderburk, Eyberg, Rich, & Behar, 2003). The ECBI has also been correlated with diagnostic

interview ratings specifically in a preschool sample (Doctoroff & Arnold, 2004). The ECBI was

originally designed as a unidimensional measure of disruptive behavior (Eyberg & Robinson,

1983), and some studies confirm that a one-factor model provides the best fit for the structure of

the ECBI (Eyberg, 1992; Colvin, Eyberg, & Adams, 1999; Gross et al., 2007). However, other

researchers have found evidence for three separate factors within the ECBI representing

oppositional defiant behavior, conduct problems, and attention difficulties (Bums & Patterson,

1991; Bums & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). If a three factor structure is

valid, then component scores may be useful in screening for the specific disorders. However,

current data are mixed, and more research is warranted before these three factors are used as

separate diagnostic indicators. Notably, the two recent studies that examined the factor structure

of the ECBI specifically in preschool samples found conflicting results. In their predominantly

Caucasian sample of 2 to 6-year-old children, Weis, Lovejoy, & Lundahl (2005) found support

for the three-factor structure, whereas a more ethnically diverse sample of 2 to 4-year-olds

resulted in a one factor model (Gross et al., 2007). Both of these studies involved community,









confirmed in a diverse clinical sample of preschoolers. In light of the range of literature that

supports a three-factor model of the ECBI, we hypothesized that the three factors would provide

good fit in a diverse sample of clinic-referred preschoolers.

The final set of specific aims was related to the idea that children of different ages may

vary in the number of symptoms required to warrant a clinical diagnosis. Specifically, we

examined whether number of symptoms would be a significant predictor of severity on the

relevant subscale of the CBCL and whether age would moderate that relationship. We

hypothesized that number of ADHD symptoms would predict attention problem scores on the

CBCL and that age would moderate the relationship such that the positive relationship between

ADHD symptom count and attention problems would be stronger for older children. We also

hypothesized that number of ODD symptoms would predict scores on the externalizing subscale

of the CBCL and that age would moderate the relationship such that the relationship between

ODD symptom count and externalizing behavior would be stronger for older children. Finally,

we hypothesized that number of CD symptoms would predict scores on the aggressive behavior

subscale of the CBCL and that age would moderate the relationship such that the relationship

between CD symptom count and aggressive behavior would be stronger for older children.









The efficacy of these types of behavioral interventions, and specifically parent training programs,

is well-established for the treatment of ODD/CD in young children (Eyberg, Nelson, & Boggs,

2008). The results of our study do not support the lesser differentiation in the presentation of

psychopathology in preschoolers. However, treatment of externalizing disorders, specifically for

this population, may involve many of the same components. Due to their developmental level

and the significance of the parent-child relationship at this young age, a range of presenting

problems in preschoolers may be improved with comprehensive parent training programs that

apply behavioral strategies to a broad spectrum of child problem behaviors (Chase & Eyberg,

2008).

Validity of Assessment Measures

Results also provided validity for the DISC-IV, YC-DISC, and ECI-4 in a clinical

preschool sample. Regardless of the diagnostic approach, symptoms across diagnostic categories

correlated most highly with their conceptual counterpart on the Child Behavior Checklist

(CBCL). These correlations tended to be of similar magnitude for each assessment method.

Additionally, symptom counts across the three studies generally demonstrated similar

correlations with the ECBI intensity score. Some variability was seen in the correlations between

the ECBI and ODD symptom counts. Specifically, ODD symptoms as assessed by the DISC-YC

demonstrated the highest correlation with the ECBI. ODD symptoms from the DISC-IV resulted

in a small but significant correlation with the ECBI, while ODD symptoms from the ECI-4 were

not significantly correlated with the ECBI intensity score. However, this result is likely related to

the restricted range of ODD symptoms in Project GIFT and ECBI scores in Project GANA. As

noted previously, the inclusion criteria for Project GIFT necessitated higher rates of ODD

symptoms and the inclusion criteria for Project GANA resulted in higher ECBI scores. Thus, the

weaker relationship between DISC-IV ODD symptoms and the ECBI and the lack of relationship









most highly with their conceptual counterpart scale of the CBCL. Additionally, correlations

across the three different assessment measures tended to be of similar magnitude.

Confirmatory Factor Analysis of the Eyberg Child Behavior Inventory

CFA was used to confirm the three-factor solution of the ECBI (Burns & Patterson, 1991;

Burns & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). Although previous studies have all

reported factors reflecting oppositional behavior, conduct problems, and attention difficulties,

there are some differences in previous results regarding which items load onto the factors. The

current study will confirm the structure reported by Weis, Lovejoy, & Lundahl (2005) as theirs is

the only study that found support for a three-factor model in a sample of preschoolers, and they

sought to confirm the three-factor model previously supported in an older sample (Bums &

Patterson, 2000). The factor structure described in their study was highly similar to those

presented in other research supporting a three-factor model and required several ECBI items to

be dropped from the analysis, leaving 22 items in the model.

The observed indicators involved in this analysis are ordinal data. Due to the clinical

nature of the sample, ECBI items tended to be positively skewed. The MPlus program addresses

ordinal variables through analysis of the polychoric correlation matrix; the current analysis also

employed Weighted Least Squares with Mean and Variance Adjustment (WLSMV) to adjust for

non-normal data.

Results suggested that the three-factor solution was a relatively poor fit for the data. The

model resulted in a highly significant chi square and failed to meet the cut off scores proposed by

Yu and Muthen (2002) for any of the fit indices (see Table 3-8 for fit indices and Table 3-9 for

the factor structure). Examination of the modification indices suggested model improvement by

allowing several items to load onto all three factors. Additionally, results suggested that several

items from the oppositional behavior factor should be allowed to load onto the attention









Table A-2. Two-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR= .08)
ECBI item Factor One Factor Two
12. Gets angry when doesn't get own way .93 .12
8. Does not obey house rules on own .89 .05
10. Acts defiant when told to do something .89 .08
30. Is easily distracted .88 .03
6. Slow in getting ready for bed .85 .08
2. Dawdles or lingers at mealtime .84 -.16
32. Fails to finish task or projects .83 .02
14. Sasses adults .81 .10
16. Cries easily .81 -.09
34. Has difficulty concentrating on one thing .80 .06
4. Refuses to eat food presented .78 .11
35. Is overactive or restless .77 .08
20. Is careless with toys and other objects .76 .12
9. Refuses to obey until threatened with punishment .70 .03
22. Lies .68 .13
24. Verbally fights with friends own age .67 .07
18. Hits parents .65 .08
36. Wets the bed .64 -.09
26. Physically fights with friends own age .63 .14
29. Interrupts .47 .32
28. Constantly seeks attention .46 .32
7. Refuses to go to bed on time .40 .19
31. Has short attention span .39 .28
13. Has temper tantrums .39 .25
5. Refuses to do chores when asked .35 .25
15. Whines .34 .17
11. Argues with parents about rules .30 .16
17. Yells or screams .28 .25
1. Dawdles in getting dressed .24 .08
3. Has poor table manners .11 .38
33. Has difficulty entertaining self alone .01 .15
19. Destroys toys and other objects -.07 .68
23. Teases or provokes other children -.20 .65
25. Verbally fights with sisters and brothers -.18 .66
27. Physically fights with sisters and brothers -.17 .61
21. Steals -.61 .57









Diagnostic Interview Schedule for Children: (DISC-IV; Schaffer et al., 2000).The

DISC-IV is a structured diagnostic interview of child psychopathology based on Diagnostic and

Statistical Manual of Mental Disorders criteria (DSM-IV; American Psychiatric Association,

1994). Test-retest reliabilities have been reported at .54 for ODD, .54 for CD, and .79 for ADHD

on this instrument (Schaffer et al., 2000). The DISC-IV-P was administered to the child's

primary caregiver at pre-treatment by a trained graduate research assistant in Project GIFT.

Young-Child DISC. (Lucas, Fisher, & Luby, 2000). Project SHAPE involved the use of a

recent adaptation of the DISC interview for young children; the Young-Child DISC (YC-DISC).

The YC-DISC is highly similar to the DISC-IV in its format and administration. However, the

language has been modified to increase its sensitivity, and the time frame for each symptom has

been uniformly set at three months. The YC-DISC was administered to the child's primary

caregiver at pre-treatment by a trained graduate research assistant in Project SHAPE.

Early Childhood Inventory: (ECI-4; Gadow & Sprafkin, 2000). The ECI-4 is a parent

behavior rating scale that screens for DSM-IV emotional and behavioral disorders in children

between 3 and 6 years old. Parents are asked to rate the frequency of specific behaviors on a 4-

point scale from (0) never, (1) sometimes, (2) often, and (3) very often. Symptoms generally map

onto DSM-IV criteria; however, four symptoms of CD are not included: breaking and entering,

missing curfew, running away, and truancy. Additionally, the ECI-4 assesses for stealing

behaviors only, with no distinction between with and without confrontation. Thus, the CD

module on the ECI-4 includes 10 symptoms, compared to the 15 listed in the DSM-IV. The

wording of certain symptoms is altered to increase developmental sensitivity; for example, the

ECI-4 asks the parent to rate their preschoolers' preoccupation with sexual activity rather than

the DSM symptom of forcing sexual activity. A recent study reported internal consistency









Wolraich, M.L., Felice, M.E., & Drotar, D. (Eds). (1996). The classification of child and
adolescent mental diagnosis in primary care: Diagnosis and Statistical Manual for
Primary Care (DSM-PC) Child and Adolescent Version. Elk Gove, IL: American
Academy of Pediatrics.

World Health Organization (1993). The ICD-10 classification of mental and behavioral
disorders: Diagnostic criteriafor research, 10th edition. Geneva: World Health
Organization.

Yu, C., & Muthen, B. (2002). Evaluation of model fit indices for latent variable models with
categorical and continuous outcomes. Unpublished Dissertation available online at:
www.statmodel.com.









Factor Analysis of the Eyberg Child Behavior Inventory

Our study failed to support the three-factor structure of the ECBI reported in the literature

(Burns & Patterson, 1991; Burns & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). The

model with factors of oppositional behavior, inattention, and conduct problems yielded a highly

significant chi square and did not indicate good fit according to any of the fit indices. Most ECBI

items demonstrated significant loadings onto one factor of child behavior problems, consistent

with previous studies reporting a one-factor structure of the ECBI (Eyberg, 1992; Colvin,

Adams, & Eyberg, 1999; Gross et al, 2007). Several different reasons may explain the failure to

replicate the three-factor model. Notably, this is the first study to factor analyze the ECBI with a

clinical sample of preschoolers. The two studies that have examined the factor structure of the

ECBI specifically with preschoolers both used a community sample, and those two studies led to

conflicting results (Weis, Lovejoy, & Lundahl, 2005; Gross et al., 2007).

Certainly, the use of a clinical sample limits the generalizability of the current results.

However, it seems particularly important to replicate the three-factor structure in clinic-referred

children. The importance of using a clinical sample is highlighted by suggestions in the literature

that component scores from the three ECBI factors could be used for diagnostic purposes (Burns

& Patterson, 1991; Bums & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). Weis, Lovejoy,

& Lundahl (2005) found support for the three-factor model in a community sample of preschool-

aged children. However, the component scores from the individual factors were largely unable to

differentiate among the three different disorders in a clinical sample of children aged 4 to 6

years. Specifically, the Oppositional Defiant component did not distinguish children with

oppositional behavior from children with attention problems, and the Inattentive component did

not differentiate inattentive children from oppositional children. Indeed, the authors found that

the full ECBI intensity scale was more useful in differentiating these disorders than the









parents of children aged 4 through 6 completed the CBCL/4-18, which consists of 118 behavior-

problem items rated on a 3-point scale from (0) not true, to (2) very true or often true. The items

have been factor analyzed into two broadband scales of internalizing and externalizing behavior

problems, as well as narrow band scales assessing specific problem areas. Mean one-week test-

retest reliability has been reported at .93 for the Externalizing scale, .91 for the Aggressive

behaviors subscale, and .90 for the Attention Problems subscale (Achenbach, 1991). The

CBCL/2-3, administered to parents of 3-year-old children in Project GIFT, is similar in format to

the CBCL/4-18 but contains 99 items rated by the parent for frequency in the past 2 months on

the 3-point scale. Test-retest reliability of the CBCL/2-3 has been reported at .90 for the

externalizing scale and .91 for the Aggressive Behavior subscale over a three-week period (Koot

et al., 1997). Notably, the CBCL/2-3 did not include an Attention Problems subscale.

Both age forms of the CBCL underwent a recent revision; parents in Projects GANA and

SHAPE therefore completed the appropriate age form of the most recent version. The CBCL/6-

18 (Achenbach & Rescorla, 2001) is a 120-item parent rating scale designed to measure the

frequency of children's behavior or emotional problems on a 3-point scale from (0) not true, to

(2) very true or often true. One-week test-retest reliabilities have been reported as .92 for the

Externalizing Scale, .90 for the Aggressive behavior subscale, and .92 for the Attention Problems

subscale (Achenbach & Rescorla, 2001). The CBCL/1.5-5 (Achenbach & Rescorla, 2000) is

comprised of 99 items describing children's behavioral and emotional problems, which are rated

by parents on a 3-point scale from (0) not true, to (2) very true or often true. One week test-retest

reliabilities have been reported as .87 for the Externalizing scale, .87 for the Aggressive behavior

subscale, and .78 for the Attention Problems subscale (Achenbach & Rescorla, 2000).










Bemporad, J.R., & Schwab, M.E. (1986). The DSM-III and clinical child psychiatry. In T.
Million & G.L. Klerman (Eds.), Contemporary directions inpsychopathology: Toward
the DSM-IV(pp. 135-150). New York: Guilford Press.

Boggs, S.R., Eyberg, S., & Reynolds, L.A. (1990). Concurrent validity of the Eyberg Child
Behavior Inventory. Journal of Clinical Child Psychology, 19, 75-78.

Burns, G.L., Boe, B., Walsh, J.A., Sommers-Flangan, R., & Teegarden, L.A. (2001). A
confirmatory factor analysis on the DSM-IV ADHD and ODD symptoms: What is the
best model for the organization of these symptoms? Journal ofAbnormal Child
Psychology, 29, 339-349.

Burns, G.L., & Patterson, D.R. (1991). Factor structure of the Eyberg Child Behavior Inventory:
One-dimensional or multidimensional measure of disruptive behavior? Journal of
Clinical Child Psychology, 20, 439-444.

Burns, G.L., & Patterson, D.R. (2000). Factor structure of the Eyberg Child Behavior Inventory:
A parent rating scale of oppositional defiant behavior towards adults, inattentive
behavior, and conduct problem behavior. Journal of Clinical Child Psychology, 29, 569-
577.

Burns, G., Walsh, J., Patterson, D., Holte, C., Sommers-Flanagan, R., & Parker, C. (1997).
Internal validity of the disruptive behavior disorder symptoms: Implications form parent
ratings for a dimensional approach to symptom validity. Journal ofAbnormal Child
Psychology, 25, 307-319.

Campbell, S. (2002). Behavior problems in preschool children: Clinical and developmental
issues (2nd ed.). New York: Guilford Press.


Chase, R.M. & Eyberg, S.M. (2008). Clinical presentation and treatment outcome for children
with comorbid externalizing and internalizing symptoms. Journal of Anxiety Disorders,
22, 273-282.

Christiansen, H., Chen, W., Oades, R.D., Asherson, P., Taylor, E.A., Lasky-Su, J., Zhou, K.,
Banaschewski, T., Buschgens, C., Franke, B., Gabriels, I., Manor, I., Marco, R., Muller,
U.C., Mulligan, A., Psychogiou, L., Rommelse, N.N.J., Uebel, H., Buitelaar, J., Ebstein,
R.P., Eisenberg, J., Gill, M., Miranda, A., Mulas, F., Roeyers, H., Rothenberger, A.,
Sergeant, J.A., Sonuga-Barke, H.C., Steinhausen, M., Thompson, M., & Faraone, S.V.
(2008). Co-transmission of conduct problems with attention-deficit/hyperactivity
disorder: Familial evidence for a distinct disorder. Journal ofNeural Transmission, 115,
163-175.









differentiation between the subtypes of ADHD and between the two disruptive behavior

disorders, ODD and CD, has been more difficult to establish. We ran a series of three models: (a)

The four-factor model inherent in the DSM-IV: ADHD-Inattentive, ADHD-Hyperactive, ODD,

and CD; (b) A combined ADHD model that collapses symptoms across the inattentive and

hyperactive domains to create three factors of ADHD, ODD, and CD, and (c) A disruptive

behavior model that combines symptoms of ODD/CD into one factor, leaving three factors of:

ADHD-Inattentive, ADHD-Hyperactive, and ODD/CD.

The observed indicators involved in this analysis are dichotomous and consequently

violate assumptions of multivariate normality. Maximum likelihood estimation, commonly used

in factor analysis, is therefore inappropriate in this study because it uses dichotomous indicators.

Analyses were run using the Mplus program (version 4.2; Muthen & Muthen, 2006). The

program addresses dichotomous variables through analysis of the tetrachoric correlation matrix,

using Weighted Least Squares with Mean and Variance Adjustment (WLSMV) to adjust for non-

normal data.

To evaluate the three models, we examined the absolute fit (chi-square statistic), relative fit

indices (Comparative Fit Index, CFI; Tucker-Lewis Index, TLI), the Root Mean Square Error of

Approximation (RMSEA), which estimates discrepancy between model-implied and population

covariance matrices, and the Weighted Root Mean Squared Residual (WRMR). Traditional

guidelines for establishing good fit have been criticized as somewhat arbitrary (Marsh & Hau,

1996).Yu and Muthen (2002) evaluated fit indices with different types of data (i.e., continuous,

nonnormal continuous, and dichotomous) in small, moderate, and large sample sizes to

determine the cutoff values that would minimize both Type I and Type II errors. For

dichotomous indicators and N> 250, their results indicated that least error would occur using the









Nottelmann, E. D. & Jensen, P. S. (1995). Comorbidity of disorders in children and adolescents:
Developmental perspectives. In T. H. Ollendick & R. J. Prinz (Eds.), Advances in
Clinical Child Psychology: Vol. 17 (pp. 109-155). New York: Plenum Press.

Pillow, D.R., Pelham, W.E., Hoza, B., Molina, B.S.G., and Stultz, C.H. (1998). Confirmatory
factor analyses examining attention deficit hyperactivity disorder symptoms and other
childhood disruptive behaviors. Journal ofAbnormal Child Psychology, 26, 293-309.

Ross, C.N., Blanc, H.M., McNeil, C.B., Eyberg, S.M., & Hembree-Kigin, T.L. (1998). Parenting
stress in mothers of young children with oppositional defiant disorder and other severe
behavior problems. Child Study Journal, 28, 93-110.

Rushton, J.L., Fant, K.E1, & Clark, S.J. (2004). Use of practice guidelines in the primary care of
children with attention-deficit/hyperactivity disorder. Pediatrics, 114, E23-E28.

Schuhmann, E., Foote, R., Eyberg, S.M., Boggs, S., & Algina, J. (1998). Parent-child interaction
therapy: Interim report of a randomized trial with short-term maintenance. Journal of
Clinical Child Psychology, 27, 34-45.

Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone, M. E. (2000). NIMH
diagnostic interview schedule for children version IV (NIMH DISC-IV): Description,
differences from previous versions, and reliability of some common diagnoses. Journal
of the American Academy of Child & Adolescent Psychiatry, 39, 28-38.

Silk, J. S., Nath, S. R., Siegel, L. R., & Kendall, P.C. (2000). Conceptualizing mental disorders
in children: Where have we been and where are we going? Development and
Psychopathology, 12, 713-735.

Speltz, M., McClellan, J., DeKylen, M., & Jones, K. (1999). Preschool boys with Oppositional
Defiant Disorder: Clinical presentation and diagnostic change. Journal of the American
Academy of Child andAdolescent Psychiatry, 38, 838-845.

Spence, S., Rapee, R., McDonald, C., & Ingram, M. (2001). The structure of anxiety symptoms
among preschoolers. Behavior Research and Therapy, 39, 1293-1316.

Sprafkin, J., Volpe, R.J., Gadow, K.D., Nolan, E.E., & Kelly, K. (2002). A DSM-IV-referenced
screening instrument for preschool children: The early childhood inventory-4. Journal of
the American Academy of Child and Adolescent Psychiatry, 41, 604-612.

Sterba, S., Egger, H.L., & Angold, A. (2007). Diagnostic specificity and nonspecificity in the
dimensions of preschool psychopathology. Journal of Child Psychology andPsychiatry,
48, 1005-1013.

Task Force on Research Diagnostic Criteria: Infancy and Preschool. (2003). Research diagnostic
criteria for infants and preschool children: The process and empirical support. Journal of
the American Academy of Child and Adolescent Psychiatry, 43, 1504-1512.









individual component scores. They suggested that the failure of the component scores to

differentiate among the different disorders may reflect the lack of behavioral differentiation in

children's symptom expression (Weis, Lovejoy, & Lundahl, 2005). However, our findings from

the DSM-IV confirmatory factor analysis argue against a lesser differentiation of symptom

presentation in younger children. Results from the current study suggest that the ECBI

component scores fail to represent the distinct disorders in the same way that they are

conceptualized in the DSM-IV.

Notably, the ECBI was developed prior to the development of the DSM-III and was

developed as a unidimensional measure of child disruptive behavior. Although several ECBI

items appear to map onto DSM-IV criteria, some may not be clear indicators of DSM symptoms,

particularly in a clinical preschool sample. For example, the ECBI item "has short attention

span" appears very similar to the DSM-IV ADHD inattentive symptom, "often has difficulty

sustaining attention in tasks or play activities." However, expectations for a preschooler's

attention span are generally lower as compared to school-aged children and adolescents (Silk et

al., 2000). A "short attention span" in a 4 year-old may be viewed as problematic by the child's

parent or caregiver; however, it also may be less indicative of pervasive attention difficulties in a

younger child. As a child matures and expectations for the ability to attend increase, "short

attention span" may be more likely to reflect a significant problem with attention, and thus a

stronger indicator of the construct of ADHD-Inattentive Type. The wording of DSM-IV items

may be particularly important when assessing for these specific symptoms in preschoolers.

Indeed, when discussing developmental modifications of current DSM-IV criteria for

preschoolers, researchers have highlighted the importance of developmentally appropriate

modifications that maintain the clinical meaning of a particular symptom (Egger & Angold,









Cunningham, C., Abikoff, H.A., McCracken, J.T., Benedetto, V., Davies, M., McGough, J.J.,
Posner, K., Skrobala, A.M., Swanson, J.M., Wigal, T., Wigal, S.B., Ghuman, J.K., &
Chuang, S.Z. (2007). Factor structure of parent- and teacher-rated Attention-
Deficit/Hyperactivity Disorder symptoms in the preschoolers with Attention-
Deficit/Hyperactivity Disorder treatment study (PATS). Journal of Child andAdolescent
Psychopharmacology, 17, 621-634.

Harwood, M.D., O'Brien, K.A., Carter, C.G., & Eyberg, S.M. (under review). Mental health
services for preschool children in primary care: A survey of maternal attitudes and
beliefs. Journal ofPediatric Psychology.

Henggeler, S.W. & Borduin, C.M. (1990). Family therapy and beyond: A multisystemic
approach to treating the behavior problems of children and adolescents. Pacific Grove:
CA: Brookes and Cole.

Huffine, C.W. (2002). Conduct disorder should be eliminated from the diagnostic and statistical
manual of mental disorders. In L. Flaherty (ed). Adolescent Psychiatry: The Annals of the
American Society for Adolescent Psychiatry, Vol. 26, Hillsdale: NJ: The Analytic Press.

Jensen, P.S., Watanabe, H.K., Richters, J.E., Roper, M., Hibbs, E.D., Salzberg, A.D., & Liu, S.
(1996). Scales, diagnoses, and child psychopathology: II. Comparing the CBCL and the
DISC against external validators. Journal ofAbnormal Child Psychology, 24, 151-168.

Keenan, K. & Wakschlag, L. S. (2002). Can a valid diagnosis of disruptive behavior disorder be
made in preschool children? American Journal ofPsychiatry, 159, 351-358.

Keenan, K. & Wakschlag, L.S. (2000). More than the terrible twos: The nature and severity of
behavior problems in clinic-referred preschool children. Journal ofAbnormal Child
Psychology, 28, 33-46.

Koot, H. M, Van Den Oord, E. J. C. G., Verhulst, F. C., & Boomsma, D. I. (1997). Behavioral
and emotional problems in young preschoolers: Cross-cultural testing of the validity of
the child behavior checklist/2-3. Journal ofAbnormal Child Psychology, 25, 183-196.

Krueger, R. F., Caspi, A., Moffitt, T. E., & Silva, P. A. (1998). The structure and stability of
common mental disorders (DSM-III-R): A longitudinal-epidemiological study. Journal of
Abnormal Psychology, 107, 216-227.

Lahey, B.B., Rathouz, P.J., Van Hulle, C., Urbano, R.C., Krueger, R.F., Applegate, B., Garriock,
H.A., Chapman, D.A., & Waldman, I.D. (2008). Testing structural models of DSM-IV
symptoms of common forms of child and adolescent psychopathology. Journal of
Abnormal Child Psychology, 36, 187-206.

Lahey, B., Applegate, B., Waldman, I., Loft, J., Hankin, B., & Rick, J. (2004). The structure of
child and adolescent psychopathology: Generating new hypotheses. Journal ofAbnormal
Psychology, 113, 358-385.









(Burns, Boe, Walsh, Sommers-Flanagan, & Teegarden, 2001). When CD symptoms are allowed

to map onto their own latent construct, however, evidence supports its differentiation from ODD

(Lahey et al., 2008; Burns et al. 1997). Lahey and colleagues (2008) directly compared the

DSM-IV model, with separate ODD and CD factors, with a combined ODD/CD model inherent

in the International Classification of Diseases (ICD-10; World Health Organization, 1993) and

found support for separate ODD and CD factors.

Notably, the majority of research validating the DSM structure of psychopathology

involves school-aged children and adolescents. To date, only three studies have employed

confirmatory factor analysis to validate DSM-IV disorders in preschool populations. Two of

these studies focused on anxiety disorders and found support for separate constructs of

separation anxiety, social anxiety, obsessive compulsive disorder, specific phobias, and

generalized anxiety disorder (Eley et al., 2003; Spence, Rapee, McDonald, & Ingram, 2001). A

recent study by Sterba, Egger, & Angold (2007) sought to confirm DSM categories across both

internalizing and externalizing disorders in a sample of 307 2- to 5-year-olds recruited from a

primary care clinic. DSM-IV symptoms of ADHD, ODD, and CD, as well as social phobia,

separation anxiety, and generalized anxiety disorder were assessed with the Preschool Age

Psychiatric Assessment (PAPA). Separate confirmatory models were run for the internalizing

versus externalizing disorders. Within the disruptive behavior disorder models, a 4-factor model

of ADHD-Inattentive ADHD-Hyperactive, ODD, and CD was compared with two 3-factor

models, one of which collapsed inattentive and hyperactive symptoms into a single ADHD

factor, while the other combined ODD and CD symptoms. Results supported the three-factor

model that included separate ADHD-Inattentive and ADHD-Hyperactive factors but combined

ODD and CD symptoms into one factor. The authors concluded that an undifferentiated









coefficients of .91 for ADHD-Inattentive, .90 for ADHD-Hyperactive, .93 for ODD, and .87 for

CD. Study results also documented adequate convergent, discriminant, and discriminative

validity of the ECI-4, particularly for the ADHD and ODD scales (Sprafkin, Volpe, Gadow,

Nolan, & Kelly, 2002). According to the scoring guidelines in the ECI-4 manual, behaviors rated

as often or very often were included in the symptom count for this study to help ensure that the

behavior was truly reflective of a frequent and pervasive symptom (Gadow & Sprakfin 2000).

The ECI-4 was completed by the child's primary caregiver at the pre-treatment assessment in

Project GANA.

Procedures

Following the informed consent process, families in all studies completed a pre-treatment

assessment that included completion of the demographic questionnaire, ECBI, and CBCL. The

presence of ODD, CD, and ADHD symptoms was determined by the female primary caregiver's

responses on the DISC-IV interview in Project GIFT and the YC-DISC interview in Project

SHAPE. In Project GANA, the female primary caregiver completed the ECI-4 to assess child

symptoms of ODD, CD, and ADHD. Table 2-1 provides an overview of the assessment

measures across the three different studies.

To obtain access to data from all three datasets, the investigator first sought IRB approval

through the University of Florida Health Science Center. After receiving IRB approval, all data

were first de-identified and then transmitted through data-encrypted files.

Confirmatory Factor Analysis and Model Evaluation

This study examined the fit of the factor structure implied by the DSM-IV in a preschool

population. We therefore employed confirmatory factor analysis (CFA) to test the current

diagnostic system. A review of the literature provides support for the different disorders

represented by the DSM-IV, particularly in school-aged and adolescent samples. However,









Table 3-7. Pearson correlations between symptom counts and CBCL and ECBI scores
ADHD symptoms ODD symptoms CD symptoms


DISC- YC- ECI- DISC- YC- ECI- DISC- YC- ECI-
IV DISC 4 IV DISC 4 IV DISC 4
Attention Scale .49** .43** .52** .02 .09 .32* .08 .02 .26*

Aggression Scale .29** .20 .33* .37** .41** .49** .45** .46** .34*

Externalizing .31** .24 .43** .37** .41** .51** .41** .43** .35**
Scale
ECBI intensity .42** .47** .30* .27** .61** .19 .42** .51** .33*
score
*p< .05, **p<.01.









DIAGNOSTIC VALIDITY OF DSM SYMPTOMS AND CRITERIA FOR PRESCHOOLERS:
DISRUPTIVE BEHAVIOR AND ATTENTION DEFICIT HYPERACTIVITY DISORDER
SYMPTOMS

















By

RHEA MARISA CHASE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2008









between ECI-4 ODD symptoms most likely reflects a methodological weakness of our study

rather than a comment on the validity of either the DISC-IV or ECI-4.

Overall, this study provided evidence of convergent validity for the DISC-IV, YC-DISC,

and ECI-4 in a preschool sample. This is among the first studies to assess the validity of the

DISC-IV specifically in this age group and the validity of the newly developed YC-DISC.

Results suggest the two interviews are highly similar in their relationships with other measures of

child behavior. Additionally, symptom counts from the ECI-4, although lower than both the

DISC-IV and YC-DISC, demonstrated significant correlations with the appropriate subscale of

the CBCL. The ECI-4 is a relatively new measure, and this study provides important information

regarding its validity as a measure of DSM-IV disorders. Future research on preschool diagnosis

should address additional indicators of validity. A particularly important direction is the study of

predictive validity of assessment measures for this age group. Although ADHD and the

disruptive behavior disorders diagnosed in early childhood have been shown to persist into later

childhood and adolescence, symptoms remit in approximately half of children diagnosed in the

preschool years (Campbell, 2002). Instruments specifically designed for this age group may

relate to improved predictive validity. For example, the YC-DISC has altered the time duration

of each symptom and asks about the presence of a behavior over a three-month time span.

Longitudinal research will determine whether modifications such as these allow better

identification of those children who will continue to exhibit symptoms into later childhood and

adolescence. Additionally, modifying the language of certain symptoms to increase

developmental sensitivity may improve face validity with this population, as well as overall

consumer satisfaction.









difficulties factor. Overall, findings failed to provide strong evidence for a three-factor model,

and the modification indices were extensive and difficult to justify conceptually. Thus, an

exploratory factor analysis seemed the appropriate next step.

All 36 ECBI items were therefore subjected to an exploratory factor analysis (EFA).

Because other researchers found support for a three-factor model, we examined a one, two, and

three-factor EFA of the ECBI items. Results from all three of these analyses are included in the

Appendix. MPlus provides RMSR values for EFA analyses; significant decreases with increasing

number of factors suggest better fit of the more differentiated models. As can be seen in the

Appendix, the decrease in the RMSR from the one to two-factor model was small, and the

decrease from the RMSR from the two to three-factor was minimal. Perhaps more importantly,

the two and three-factor models did not conceptually match the factors of oppositional behavior,

attention difficulties, and conduct problems found in previous studies (Burs & Patterson, 1991;

Burns & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). Indeed, it was difficult to interpret

the two- and three-factor models.

In the two factor model, the first factor included most of the items and seemed to reflect a

more general disruptive behavior factor that included a range of oppositional and attention-

seeking behaviors. The second factor resembled the conduct problems factor found in previous

studies, reflecting overt aggression and provocation. This second factor was comprised of five

items: (a) destroys toys and other objects, (b) teases or provokes other children, (c) verbally

fights with sisters and brothers, (d) physically fights with sisters and brothers, and (e) steals.

Thus, the two-factor EFA seemed to reflect a general disruptive behavior construct with a

smaller factor of severe conduct problems and aggressive behaviors. The three-factor EFA was

exceedingly difficult to understand conceptually. The first factor again included the majority of









BIOGRAPHICAL SKETCH

Rhea Marisa Chase was born in Manchester, Connecticut on October 6, 1979. She

graduated magna cum laude with a B.A. in psychology in 2001 from Bates College in Lewiston,

Maine. She then worked in clinical research at Massachusetts General Hospital in Boston,

Massachusetts for two years. Her graduate work was done in the Department of Clinical and

Health Psychology at the University of Florida, and she completed her predoctoral internship at

Duke University Medical Center. Upon completion of her Ph.D. program, Rhea will begin a post

doctoral position through the Center for Anxiety and Related Disorders at Boston University.









Eyberg, S.M., Nelson, M.M., & Boggs, S.R. (2008). Evidence-based psychosocial treatments for
children and adolescents with disruptive behavior. Journal of Clinical Child &
Adolescent Psychology, 37, 215-237.

Eyberg, S. M., & Pincus, D. (1999). Eyberg Child Behavior Inventory and Sutter-Eyberg
Student Behavior Inventory: Professional manual. Odessa, FL: Psychological
Assessment Resources.

Eyberg, S.M., & Robinson, E.A. (1983). Conduct problem behavior: Standardization of a
behavior rating scale with adolescents. Journal of Clinical Child Psychology, 11, 130-
137.

Eyberg, S. M., Schuhmann, E. M., & Rey, J. (1998). Child and adolescent psychotherapy
research: Developmental issues. Journal ofAbnormal Child Psychology, 26, 71-82.

Floyd, F.J., & Widaman, K.F. (1995). Factor analysis in the development and refinement of
clinical assessment instruments. Psychological Assessment, 7, 286-299.

Funderburk, B.W., Eyberg, S.M., Rich, B.A., & Behar, L. (2003). Further psychometric
evaluation of the Eyberg and Behar rating scales for parents and teachers of preschoolers.
Early Education and Development, 14, 67-81.

Gadow, K.D. & Sprafkin, J. (2000). Early Childhood inventory -4 Screening Manual. Stony
Brook, NY: Checkmate Plus.

Glutting, J., Youngstrom, E., & Watkins, M. (2005). ADHD and college students: Exploratory
and confirmatory factor structures with student and parent data. Psychological
Assessment, 17, 44-55.

Greene, R. W., Biederman, J., Zerwas, S., Monuteaux, M., Goring, J. C., & Faraone, S. V.
(2002). Psychiatric comorbidity, family dysfunction, and social impairment in referred
youth with oppositional defiant disorder. American Journal ofPsychiatry, 159, 1214-
1224.

Greenhill, L., Kollins, S., Abikoff, H., McCracken, J., Riddle, M., Swanson, J., McGough, J.,
Wigal, S., Wigal, T., Vitiello, B., Skrobala, A., Posner, K., Ghuman, J., Cunningham, C.,
Davies, M., Chuang, S., & Cooper, T. (2006). Efficacy and safety of immediate-release
methylphenidate treatment for preschoolers with ADHD. Journal of the American
Academy of Child and Adolescent Psychiatry, 45, 1284-1293.

Gross, D., Fogg, L., Young, M., Ridge, A., Cowell, J., Sivan, A., & Richardson, R. (2007).
Reliability and validity of the Eyberg Child Behavior Inventory with African-American
and Latino parents of young children. Research in Nursing & Health, 30, 213-223.

Hardy, K.K., Kollins, S.K., Murray, D.W., Riddle, M.A., Greenhill, L.,










Lahey, B. B., Pelham, W.E., Stein, M A., Loney, J., Trapani, C., Nugent, K., Kipp, H., Schmidt,
E., Lee, S., Cale, M., Gold, E., Hartung, C.M., Willcutt, E., & Baumann, B. (1998).
Validity of DSM-IV attention-deficit/hyperactivity disorder for younger children. Journal
of the American Academy of Child and Adolescent Psychiatry, 37, 695-702.

Lavigne, J. V., Arend, R., Rosenbaum, D., Binns, H. J., Christoffel, K K., & Gibbons, R. D.
(1998). Psychiatric disorders with onset in the preschool years, part 1: stability of
diagnoses. Journal of the American Academy of Child andAdolescent Psychiatry, 37,
1246-1254.

Lilienfeld, S. O., Waldman, I. D., & Israel, A. C. (1994). A critical examination of the use of the
term and concept of comorbidity in psychopathology research. Clinical Psychology
Science andPractice, 1, 71-83.

Loeber, R., Burke, J.D., Lahey, B.B., Winters, A., & Zera, M. (2000). Oppositional defiant
disorder and conduct disorder: A review of the past 10 years, part I. Journal of the
American Academy of Child and Adolescent Psychiatry, 39, 1468-1484.

Luby, J.L., Heffelfinger, A.K., Mrakotsky, C., Hessler, M.J., Brown, K.M., & Hildebrand, T.
(2002). Preschool major depressive disorder: Preliminary validation for developmentally
modified DSM-IV criteria. Journal of the American Academy of Child and Adolescent
Psychiatry, 40, 928-937.

Lucas, C.P., Fisher, P., Luby, J. (2000) The Young ChildDISC. New York: Columbia DISC
Development Group.

Marsh, H.W. & Hau, K.H. (1996) Assessing goodness of fit: Is parsimony always desirable? The
Journal ofExperimental Education, 64, 364-390.

McClellan, J., & Speltz, M. (2003). Psychiatric diagnosis in preschool children. Journal of the
American Academy of Child and Adolescent Psychiatry, 42, 127-128.

Muthen, L. & Muthen B. (2005). Chi-square difference testing using the S-B scaled chi-square.
Note on Mplus website, www.statmodel.com.

Newcorn, J. H., Halperin, J. M., Jensen, P. S., Abikoff, H. B., Arnold, E., Cantwell, D. P.,
Conners, C. K., Elliott, G. R., Epstein, J. N., Greenhill, L. L., Hechtman, L., Hinshaw, S.
P., Hoza, B., Kraemer, H. C., Pelham, W. E., Severe, J. B., Swanson, J. M., Wells, K. G.,
Wigal, T., & Vitiello, B. (2001). Symptom profiles in children with ADHD: Effects of
comorbidity and gender. Journal of the American Academy of Child & Adolescent
Psychiatry, 40, 137-146.

Nolan, E. E., Gadow, K. D., & Sprafkin, J. (2001). Teacher reports ofDSM-IV ADHD, ODD,
and CD symptoms in schoolchildren. Journal of the American Academy of Child and
Adolescent Psychiatry, 40, 241-249.









however, inferential chi-square testing allowed direct comparison of the three models and found

support for the four-factor model. This is an important finding that adds to the limited research

on the validity of separate ODD and CD factors in younger children. Two previous studies found

similar results in large, nonclinical samples of school-aged children and adolescents (Lahey et

al., 2008; Burns et al., 1997). However, a recent study of DSM symptoms in 307 preschoolers

found support for collapsing ODD and CD symptoms into a single disruptive behavior factor

(Sterba, Egger, & Angold, 2007). Thus, the present study is the first to find support for the

differentiation of CD from ODD specifically in a preschool population. The clinical nature of the

sample may explain the discrepancy from the results by Sterba, Egger, and Angold (2007).

Given the extreme nature of many of the CD criteria, it may be difficult to establish a clear CD

factor in a community sample, particularly in younger children. Although more research is

needed, the present study adds to a small but growing body of literature that supports the validity

of CD in preschoolers (Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998; Keenan &

Wakschlag, 2000; Wilens et al., 2002).

Although results of this study support a separate CD factor, it is notable that several CD

symptoms were dropped from analyses due to low frequency: forcing sexual activity, fire-setting,

breaking and entering, missing curfew, running away, and truancy. This is hardly surprising in

light of preschoolers' developmental abilities and limited exposure to these types of situations.

However, the current DSM-IV provides no developmental modifications to current CD criteria.

Amidst claims that the diagnosis of CD is irrelevant to preschoolers, it seems important to

increase the developmental sensitivity of CD criteria for this age group. The Task Force on

Research Diagnostic Criteria for Infancy and Preschool (RDC-PF; 2003) has dropped breaking

and entering, breaking curfew, running away, and truancy; thus, it seems likely that the DSM-V









LIST OF FIGURES


Figure pe

3-1 Path diagram of four-factor solution of DSM-IV symptoms of ADHD-Inattentive,
AD H D -H yperactive, OD D and CD ...................................................... .....................55









The sample size may have affected the clarity of our results, particularly for the factor

analysis of the ECBI. Our findings resulted in an unclear factor structure, which was inconsistent

with some earlier research in this area (Burns & Patterson, 1991; Burns & Patterson, 2000; Weis,

Lovejoy, & Lundahl, 2005). However, it is notable that our results are more consistent with a

recent factor analytic study of the ECBI in a sample of 682 preschoolers (Gross et al., 2007).

Overall, results in this area of the literature are conflicted, highlighting the need for future studies

that examine the factor structure of the ECBI in larger samples.

Despite these limitations, findings provide valuable information on current diagnostic

standards and measures for the preschool population. The clinical nature of the sample is a

particular strength of the current study. Much of the literature examining issues of diagnostic

validity for this age group often involves community samples of non-referred children. Although

results from this study may not be generalized to non-clinical children, it seems particularly

important to focus on the children that exhibit the specific disorders. This may be particularly

important in examining diagnosis of CD in preschoolers, as our results found support for two

separate factors of ODD and CD in contrast to findings by Sterba, Egger, & Angold (2007).

Their sample of 307 preschoolers recruited from a primary care clinic may have been less

representative of children with externalizing disorders. The severity of the CD criteria may make

it difficult to establish a clear CD construct in non-clinical samples, particularly in younger

children, for whom certain behaviors are developmentally irrelevant to diagnosis. The

discrepancy of these results highlights the importance of future research that examines CD

symptoms in both clinical and nonclinical samples. Specifically, further study should involve a

broader range of clinical severity, perhaps including children from psychiatric inpatient units or

day treatment facilities.









CHAPTER 1
INTRODUCTION

The American Psychiatric Association (APA) introduced its first catalogue of mental

disorders in 1952, largely in response to the increased prevalence and awareness of mental

disturbance following World War II (Silk et al., 2000). The original version of the Diagnostic

and Statistical Manual of Mental Disorders (DSM-I; APA, 1952) made no distinction between

child and adult disorders. A child could be given any adult diagnosis, with little consideration to

any developmental differences in classification or expression of a disorder (Silk et al., 2000).

The second edition of the DSM (DSM-II; APA, 1968) introduced a section entitled "Behavior

Disorders of Childhood and Adolescence," which included such diagnoses as "hyperkinetic

reaction of childhood" and "overanxious reaction of childhood." DSM-II therefore began to

consider the differences in the expression of psychopathology in children and adolescents as

compared to adults. However, the DSM-II emphasized that these categories should be used only

when no other classification was possible, as this nascent childhood disturbance most likely

represented a developing adult disorder. The nosology of DSM-II reflected the widespread belief

that psychopathology in adulthood could always be traced back to childhood and adolescence,

and child mental illness could be conceptualized as a slowly emerging disorder that would be

fully realized in adulthood (Silk et al., 2000). Thus, although the DSM-II recognized that the

developing child may differ from the adult in the symptomatic expression of psychological

disturbance, its theoretical framework suggested that the disorder was always present in the same

form, but may not have fully revealed itself in a young child.

The third edition of the DSM (DSM-III; APA, 1980) underwent major revisions in

diagnostic method and overall approach to classifying psychopathology. For the first time, the

DSM-III included detailed descriptions of mental disorders and specific criteria for each









into one disruptive behavior construct. Findings supported the four-factor model represented by

the DSM-IV, which provided superior fit for the data as compared to both of the less

differentiated models. The study also further explored current assessment strategies with this age

group. We assessed and compared the validity of the fourth edition of the Diagnostic Interview

Schedule for Children (DISC-IV), the Young Child Version of the Diagnostic Interview

Schedule for Children (YC-DISC), and the fourth edition of the Early Childhood Inventory (ECI-

4). All measures demonstrated similar correlations with other measures of child behavior,

suggesting comparable validity across measures for this age group. Overall, results support the

use of DSM-IV diagnostic categories with preschool-aged children and add to the limited

research on the validity of current assessment methods for young children. These findings help

promote accurate identification and subsequent treatment of psychopathology in preschoolers.









2004). Differentiating components of the ECBI may be more relevant for older children, when

certain behaviors are inherently more problematic and more indicative of specific disorders.

Exploration of Age-Graded Criteria

Our study failed to find support for the use of age-graded criteria in assigning diagnoses

among 3- to 6-year-olds. Symptom counts of ADHD, ODD, and CD served as significant

predictors of behavior severity, as measured by the corresponding scale on the CBCL. However,

age failed to moderate that relationship. The failure to find support for our hypothesis is likely

related to the truncated age range of the sample. The study of age-related criteria would be more

appropriate across larger age ranges. In their comprehensive review of the presentation and

epidemiology of psychopathology in preschoolers, Egger and Angold (2006) systematically

evaluated the DSM qualification that behaviors included in ODD criteria must occur "often" in

order to represent a true symptom. ODD symptoms naturally occur at a higher frequency in

preschool populations as compared to school-aged children. Thus, applying the cutoff criteria

previously used for school-aged children (above the 90th percentile) for preschoolers would have

overestimated rates of ODD. This example highlights the importance of using established

developmental norms in determining the presence of a symptom (Egger & Angold, 2006), and

supports further study of the use of developmentally modified criteria for different age groups.

Study Limitations

In an effort to increase sample size and the demographic diversity of the study sample, this

study included participants from three different studies. Although all studies examined the

efficacy of Parent-Child Interaction Therapy (PCIT) in the treatment of disruptive behavior

disorders and/or ADHD in preschoolers, each study varied in its inclusion criteria. Consequently,

participants differed in diagnostic presentation and demographic characteristics, introducing

variability that may have affected the results. Notably, to provide the sample size necessary,









LIST OF REFERENCES


Abikoff, H. & Klein, R. G. (1992). Attention-deficit hyperactivity and conduct disorder:
Comorbidity and implications for treatment. Journal of Consulting and Clinical
Psychology, 60, 881-892.

Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 and the 1991 profile.
Burlington, VT: University of Vermont Department of Psychiatry.

Achenbach, T. M. (1992). Manual for the Child Behavior Checklist/2-3 and 1992 profile.
Burlington, VT: University of Vermont Department of Psychiatry.

Achenbach, T.M. (1995). Diagnosis, assessment, and comorbidity in psychosocial treatment
research. Journal ofAbnormal Child Psychology, 23, 45-65.

Achenbach, T. M., & Rescorla, L. A. (2000). Manualfor ASEBA Preschool Forms & Profiles.
Burlington, VT: University of Vermont, Research Center for Children, Youth, &
Families.

Achenbach, T. M., & Rescorla, L. A. (2001). Manual for ASEBA School-Age Forms & Profiles.
Burlington, VT: University of Vermont, Research Center for Children, Youth, &
Families.

American Psychiatric Association. (1952). Diagnostic and Statistical Manual of Mental
Disorders (1st ed.). Washington, D.C.: Author.

American Psychiatric Association. (1968). Diagnostic and Statistical Manual of Mental
Disorders (2nd ed.). Washington, D.C.: Author.

American Psychiatric Association. (1980). Diagnostic and Statistical Manual of Mental
Disorders (3rd ed.). Washington, D.C.: Author.

American Psychiatric Association. (1987). Diagnostic and Statistical Manual of Mental
Disorders (3rd ed., Rev.). Washington, D.C.: Author.

American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders (4th
ed.). Washington, D.C.: Author.

Baker, B., & Heller, T. (1996). Preschool children with externalizing behaviors: Experience of
fathers and mothers. Journal ofAbnormal Child Psychology, 24, 513-532.

Barkley, R.A. (2006). Attention-deficit hyperactivity disorder: A handbook for diagnosis and
treatment (3rd ed.). New York: Guilford Press.

Bauermeister, J. (1992). Factor analyses of teacher ratings of attention-deficit/hyperactivity and
oppositional defiant symptoms in children aged four through thirteen years. Journal of
Clinical Child Psychology, 21, 27-34.









Table A-3. Three-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR= .06)
ECBI item Factor One Factor Two Factor Three
12. Gets angry when doesn't get own way .91 .05 .01
26. Physically fights with friends own age .84 -.31 -.23
28. Constantly seeks attention .83 .04 .05
24. Verbally fights with friends own age .81 -.20 -.15
30. Is easily distracted .78 .03 .17
14. Sasses adults .76 .06 .09
34. Has difficulty concentrating on one thing .74 -.07 .15
20. Is careless with toys and other objects .74 .14 .11
10. Acts defiant when told to do something .72 .12 .26
32. Fails to finish task or projects .72 .03 .20
22. Lies .72 -.19 .02
8. Does not obey house rules on own .71 .08 .28
35. Is overactive or restless .68 -.06 .21
18. Hits parents .67 .01 -.03
36. Wets the bed .66 .02 -.04
6. Slow in getting ready for bed .66 .13 .29
2. Dawdles or lingers at mealtime .61 .23 .33
16. Cries easily .60 .15 .31
4. Refuses to eat food presented .60 .15 .26
9. Refuses to obey until threatened with .45 .09 .42
punishment
29. Interrupts .34 -.22 .33
31. Has short attention span .28 -.20 .29
23. Teases or provokes other children .00 -.70 -.05
19. Destroys toys and other objects .02 -.65 .12
25. Verbally fights with sisters and brothers -.06 -.64 .06
27. Physically fights with sisters and brothers -.01 -.62 .00
21. Steals -.59 -.46 .17
3. Has poor table manners -.19 -.16 .63
5. Refuses to do chores when asked .09 -.07 .52
7. Refuses to go to bed on time .14 -.02 .50
1. Dawdles in getting dressed -.05 .10 .49
15. Whines .09 -.02 .46
11. Argues with parents about rules .08 -.01 .43
17. Yells or screams .14 -.14 .34
13. Has temper tantrums .26 -.16 .31
33. Has difficulty entertaining self alone -.11 -.05 .24









Cole, D., Truglio, R., & Peeke, L. (1997). Relation between symptoms of anxiety and depression
in children: A multitrait-multimethod-multigroup assessment. Journal of Consulting and
Clinical Psychology, 65, 110-119.

Colvin, A., Eyberg, S.M., & Adams, C. (1999). Standardization of the Eyberg Child Behavior
Inventory with chronically ill children. Gainesville: University of Florida, Child Study
Laboratory. Available on-line at http://www.pcit.org

Craighead, W.E. (1991). Cognitive factors and classification issues in adolescent depression.
Journal of Youth & Adolescence, 20, 311-326.

Doctoroff, G.L., & Arnold, D.H. (2004). Parent-rated externalizing behavior in preschoolers:
The predictive utility of structured interviews, teacher reports, and classroom
observations. Journal of Clinical Child and Adolescent Psychology, 33, 813-818.

dosReis, S., Zito, S.M., Safer, D.J., Soeken, K.L., Mitchell, J.W., & Ellwood, L.C. (2003).
Parental perceptions and satisfaction with stimulant medication for attention-deficit
hyperactivity disorder. Journal ofDevelopmental & Behavioral Pediatrics, 24, 155-162.

DuPaul, G.J., McGoey, K.E., Eckert, T., & VanBrakle, J. (2001). Preschool children with
attention-deficit/hyperactivity disorder: Impairments in behavioral, social, and school
functioning. Journal of the American Academy of Child and Adolescent Psychiatry, 40,
508-515.

Egger, H.L. & Angold, A. (2006). Common emotional and behavioral disorders in preschool
children: Presentation, nosology, and epidemiology. Journal of Child Psychology and
Psychiatry, 47, 313-337.

Egger, H.L. & Angold, A. (2004). The preschool age psychiatric assessment (PAPA): A
structured parent interview for diagnosing psychiatric disorders in preschool children. In
R. Delcarmen-Wiggens and A. Carter (Eds.) A Handbook of Infant and Toddler Mental
Health Assessment. New York, NY: Oxford University Press.

Eisenstadt, T.H., McElreath, L.S., Eyberg, S.M., & McNeil, C.B. (1994). Interparent agreement
on the Eyberg Child Behavior Inventory. Child & Family Behavior Therapy, 16, 21-28.

Eley, T.C., Bolton, D., O'Connor, T.G., Perrin, S., Smith, P., & Plomin, R. (2003). A twin study
of anxiety-related behaviors in pre-school children. Journal of Child Psychology and
Psychiatry, 44, 945-960.

Eyberg, S.M. (1992). Parent and teacher behavior inventories for the assessment of conduct
problem behaviors in children. In L. VandeCreek, S. Knapp, & T.L. Jackson (Eds).,
Innovations in Clinical Practice: A Source Book (Vol. 12, pp.377-382). Sarasota, FL:
Professional Resource Exchange.









Table 3-10. One-factor CFA of the ECBI
ECBI item
10. Acts defiant when told to do something .89
8. Does not obey house rules on own .89
13. Has temper tantrums .89
32. Fails to finish task or projects .89
6. Slow in getting ready for bed .83
30. Is easily distracted .83
34. Has difficulty concentrating on one thing .83
36. Wets the bed .83
2. Dawdles or lingers at mealtime .80
22. Lies .79
4. Refuses to eat food presented .77
18. Hits parents .76
26. Physically fights with friends own age .75
9. Refuses to obey until threatened with punishment .74
28. Constantly seeks attention .73
24. Verbally fights with friends own age .72
20. Is careless with toys and other objects .61
31. Has short attention span .53
7. Refuses to go to bed on time .51
5. Refuses to do chores when asked .45
14. Sasses adults .44
16. Cries easily .41
12. Gets angry when doesn't get own way .36
35. Is overactive or restless .06
29. Interrupts .01











Fails to Disorganized Avoids
finish concentrating

.86 .73 .80


ages 111adeLLCHlVe
Bullying


aves Seat .81 .35 .79
.8 .6 Fighting
.63

imbs.88
Use of
.77 weapon
bs: 82 .88 AD H D 77
>o Loud .82
Hyperactive CD .68
Cruel to
.90 g\ people
/\~ / .\46 Cul
.68 /Cruel to
animals
ilkative .71
Y 67 /.54 .74
.73
.73 f Destroys
urts out property
------ / {1 \ < .7
Lying
fficulty .72 ODD
citing

terrupts Stealing




.67 .84 .89 .72 .71 .68 .74 .65
/ ,/ _. \ ^ \\ \ .
Loses Argues Defiant Annoys Blames Easily Angry Spiteful
temper others others annoyed


Figure 3-1. Path diagram of four-factor solution of DSM-IV symptoms of ADHD-Inattentive,
ADHD-Hyperactive, ODD, and CD.









Project GANA

Another subset of participants was drawn from a study of the efficacy of PCIT with

Mexican American families (Project GANA). Participants were 58 Mexican American families

of 3- to 7-year-old children with clinically significant behavior problems seen for treatment at a

community mental health clinic; the three 7-year-old children were dropped from the current

study, leaving data from 55 participants. Families were referred from schools/teachers, hospitals

or other clinics, social workers, or were self-referred. Families were eligible if they met the

following criteria: (1) Parent identified the child as a Mexican American child between the ages

of 3 and 7, (2) Child received a score above the clinical cutpoint on the Intensity Scale of the

Eyberg Child Behavior Inventory (ECBI; Eyberg & Pincus, 1999), and (3) Neither parent nor

child were participating simultaneously in any other psychosocial treatment targeting the child's

behavior problems.

Project SHAPE

The final group of participants was drawn from an ongoing study on the efficacy of PCIT

in treating children with ADHD and comorbid ADHD + ODD (Project SHAPE). Data from 53

boys and girls between the ages of 4 and 6 were available for the current study. Families were

referred for treatment by pediatricians, teachers, other mental health providers, or were self-

referred. Inclusion criteria for Project SHAPE stated that the child must meet DSM-IV diagnostic

criteria for the diagnosis of ADHD, either hyperactive/impulsive or combined type. Due to the

need to establish the presence of ADHD symptoms across two or more settings (APA; 1994),

children were required to attend preschool or daycare for some part of each week. They were

also required to live with a female primary caregiver able to participate in treatment. Along with

a diagnosis of ADHD, some of the children additionally met criteria for ODD. Children were

excluded from the study if they scored below a standard score of 70 or if their parent scored









ACKNOWLEDGMENTS

I thank my mentor and chair, Dr. Sheila Eyberg, for her unending enthusiasm and constant

encouragement. I am very appreciative to the members of my doctoral committee for their

guidance, and would like to thank Dr. Michael Marsiske for his statistical knowledge and the

amazing skill and patience with which he teaches others. I would also like to thank Dr.

Chongming Yang for introducing me to the MPlus program. I thank Drs. Kristen McCabe and

Christopher Lucas, as well as members of the Child Study Laboratory at the University of

Florida, for their contributions to the current research. This dissertation is the culmination of five

rewarding but challenging years. I am eternally grateful to the many friends and family members

who have supported me along the way. I especially acknowledge my father, Stephen Chase,

whose love and support has been critical to any success I have achieved in graduate school.









Table 2-1. Overview of relevant measures across datasets
Project GIFT Project SHAPE Project GANA
(N= 99) (N= 53) (N= 55)
Measure of DSM-IV Symptoms
DISC-IV-P X
YC-DISC X
ECI-4 X
CBCL
CBCL/2-3 and CBCL/4-18 X
CBCL/1.5-5 and CBCL/6-18 X X
ECBI X X X
Demographic Questionnaire X X X
Note. All measures are from the pre-treatment assessment.









and child age, to serve as a preliminary examination of age-graded criteria for the disruptive

behavior disorders and ADHD.

The first aim of this study was to examine the validity of the DSM constructs of ODD, CD,

and ADHD in a clinical preschool sample. Specifically, we focused on the differentiation

between the different subtypes of ADHD, as well as separate factors of ODD and CD. The

majority of previous research supports two separate factors relating to the inattentive and

hyperactive domains within ADHD, even within a preschool sample. Additionally, a recent study

found support for collapsing ODD and CD symptoms into one single construct of disruptive

behavior. We therefore hypothesized that a three-factor model, with constructs representing

ADHD-Inattentive, ADHD-Hyperactive, and ODD/CD would provide the best fit in a sample of

clinic-referred preschoolers.

The second specific aim further examined current assessment techniques with a preschool

sample. We compared the validity of three measures commonly used to assign DSM-IV

symptoms and diagnoses. Specifically, symptom counts from the Diagnostic Interview Schedule

for Children (DISC-IV; Schaffer et al., 2000), the Young-Child version of the DISC (YC-DISC;

Lucas, Fisher, & Luby, 2000), and the Early Childhood Inventory (ECI-4; Gadow & Sprafkin,

2000) were correlated with other measures of child disruptive behavior, including the Attention

Problems and Aggressive Behavior subscales from the Child Behavior Checklist (CBCL) and the

ECBI. We hypothesized that the three assessment methods would demonstrate similar

correlations with other measures of child behavior, suggesting similar validity across diagnostic

interview type. To examine further a commonly used dimensional measure of disruptive

behavior in preschoolers, we conducted a confirmatory factor analysis of the ECBI to determine

whether a three factor model which maps onto the three externalizing behavior diagnoses can be









Thomas, J. & Guskin, K. (2001). Disruptive behavior in young children: What does it mean?
Journal of the American Academy of Child andAdolescent Psychiatry, 40, 44-51.

Verduin, T.L., Abikoff, H., & Kurtz, S.M.S. (2008). Evidence-based treatment of attention
deficit/hyperactivity disorder in a preschool-age child: A case study. Journal of Clinical
Child & Adolescent Psychology, 37, 477-485.

Vitiello, B., Abikoff, H.B., Chuang, S.Z., Kollins, S.H., McCracken, J.T., Riddle, M.A.,
Swanson, J.M., Wigal, T., McGough, J.J., Ghuman, J.K., Wigal, S.B., Skrobala, A.M.,
Davies, M., Posner, K., Cunningham, C., & Greenhill, L.L. (2007). Effectiveness of
methylphenidate in the 10-month continuation phase of the preschoolers with ADHD
treatment study (PATS). Journal of Child and Adolescent Psychopharmacology, 17, 593-
603.

Wakefield, J.C., Pottick, K. J., & Kirk, S.A. (2002). Should the DSM-IV diagnostic criteria for
conduct disorder consider social context? American Journal ofPsychiatry, 159, 380-386.

Wakschalg, L. S., Leventhal, M. D. & Thomas, M. D. Disruptive behavior disorders & ADHD in
preschool children: Characterizing heterotypic continuities for a developmentally-
informed nosology for DSM V. In D. Reiger, M. First & W. Narrow (Eds). Age and
gender considerations in psychiatric diagnosis: A research agenda for DSM-V.
Washington, D.C.: American Psychiatric Publishing Inc. (in press).

Weis, R., Lovejoy, M.C., & Lundahl, B.W. (2005). Factor structure and discriminative validity
of the Eyberg Child Behavior Inventory with young children. Journal of
Psychopathology and Behavioral Assessment, 27, 269-278.

Wigal, T. Greenhill, L., Chuang, S., McCough, J., Vitiello, B., Skrobala, A., Swanson, J., Wigal,
S., Abikoff, H., Kollins, S., McCracken, J., Riddle, M., Posner, K., Ghuman, J., Davies,
M., Thorp, B., & Stehli, A. (2006). Safety and tolerability of methylphenidate in
preschool children with ADHD. Journal of the American Academy of Child and
Adolescent Psychiatry, 45, 1294-1303.

Wilens, T., Biederman, J., Spencer, T., & Monuteaux, M. (2003). Psychiatric diagnosis in
preschool children: Reply. Journal of the American Academy of Child and Adolescent
Psychiatry, 42, 128-129.

Wilens, T. E., Biederman, J., Brown, S., Tanguay, S., Monuteaux, M.C., Blake, C., & Spencer,
T. J. (2002). Psychiatric comorbidity and functioning in clinically referred preschool
children and school-age youths with ADHD. Journal of the American Academy of Child
and Adolescent Psychiatry, 41, 262-268.

Willoughby, M., Curran, P., Costello, E., & Angold, A. (2000). Implications of early versus late
onset of attention-deficit/hyperactivity disorder symptoms. Journal of the American
Academy of Child and Adolescent Psychiatry, 39, 1512-1519.










TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ..............................................................................................................4

LIST OF TABLES .....................................................................7

LIST OF FIGURES .................................. .. ..... ..... ................. .8

ABSTRAC T ...........................................................................................

CHAPTER

1 INTRODUCTION ............... ............................ ................................. 11

ADHD and the Disruptive Behavior Disorders...................................................................14
Validity of the D SM -IV N osology ...................................................................... 15
Issues in P reschool D iagnosis........................................................................ ...................18
Current A ssessm ent Strategies .................................................. .................. ............... 20
Developmental Modifications to Current DSM Standards.........................................23
R e search Q u e stio n s.................................................................................................... .. 2 4

2 M E T H O D S .............. ...............................................................................................................2 7

P articip an ts .........................................................................2 7
P roje ct G IF T ......................................................................................................2 7
P roje ct G A N A ......................................................................................................2 8
P roje ct S H A P E .....................................................................................................2 8
M e a su re s ................... ...................2...................9..........
Procedures .................................... ... .. .................. ...................... ......... 32
Confirmatory Factor Analysis and Model Evaluation............................ 32

3 R E SU L T S .............. ... ................................................................36

P participant C characteristics ...............................................................36
Sym ptom and D diagnostic Patterns ..........................................................................................37
Confirmatory Factor Analyses of DSM Symptoms .......................................... 38
Validity of the Three Assessment Methods ..... ............................................... 39
Confirmatory Factor Analysis of the Eyberg Child Behavior Inventory ............................. 40
E exploration of A ge-G raded C criteria ................................................................................. 42

4 D IS C U S S IO N ........................................................................................................5 6

Validity of the DSM-IV for Preschoolers ................................ .................................56
Validity of Assessment Measures............................................. 60
Factor Analysis of the Eyberg Child Behavior Inventory ...................................................62
E exploration of A ge-G raded C riteria ................................................................................. 64









Table 3-3. Percentage of diagnosis across assessment method
* DISC-IV YC-DISC ECI
S (GIFT; N= (SHAPE; N= (GANA; N=
99) 53) 55)
* ADHD- 1 5 4
Inattentive

* ADHD- 24 11 32
Hyperactive

* ADHD- 46 76 22
Combined

* ODD 91 79 56

* CD 42 29 9









Table 3-8. Absolute, relative, and component model fit for ECBI models
Model 2 DF P CFI TLI RMSEA WRMR

Three-factor CFA 805.24 53 < .0001 .75 .86 .27 2.28

One-factor CFA 356.84 66 <.0001 .92 .97 .15 1.2


































2008 Rhea Marisa Chase









present in the previous six months. As with all DSM-IV categories, the symptoms must cause

significant impairment in everyday functioning to indicate a true disorder (DSM-IV; APA,

1994).

ADHD and the Disruptive Behavior Disorders

ADHD and the disruptive behavior disorders, ODD and CD, are the most common

childhood disorders diagnosed today. ODD and CD are estimated to occur in 2 to 16% of school-

age children, depending on the population sampled and the method of assessment (Loeber et al.,

2000). ODD prevalence tends to be highest in younger samples whereas the more severe

behaviors inherent in the diagnosis of CD become more prevalent at older ages. Additionally, the

CD diagnosis supersedes that of ODD if symptoms of both disorders are present; thus, the two

disorders are viewed on a spectrum of severity. ADHD is diagnosed in 1 to 5% of school-aged

children (APA, 1994) and 2% of preschool-aged children (Lavigne et al., 1998), although

community studies based on teacher report have included estimates as high as 16% (Nolan,

Gadow, & Sprafkin, 2001). ODD/CD and ADHD are linked with a myriad of detrimental effects

on a variety of psychosocial outcomes and are associated with higher levels of parenting stress

(Baker & Heller, 1996; DuPaul et al., 2001), greater social impairment (DuPaul et al., 2001;

Greene et al., 2002; Wilens et al., 2002), and increased risk for emotional problems (Thomas &

Guskin, 2001) when compared to controls. The disruptive behavior disorders and ADHD are

generally recognized as common childhood disorders that warrant early identification and

treatment (Abikoff & Klein, 1992; Ross, et al., 1998).

Complicating the clinical presentation of these common childhood disorders is the high

rate of overlap among them. As noted above, ODD and CD reflect similar patterns of

negativistic, defiant, and hostile behaviors, and CD is conceptualized as a more severe form of

ODD. In fact, some researchers have proposed the elimination of CD from the DSM, as









"externalizing" and "internalizing" disorders may be more relevant in diagnosing young children

(Nottelmann & Jensen, 1995).

The differentiation of DSM-IV disorders in child and adolescent samples has become an

increasing area of interest within the literature. Studies of this nature generally involve subjecting

DSM symptoms to confirmatory factor analysis to determine whether the underlying factor

structure implied by the DSM-IV provides the best fit for the data or whether a more

parsimonious model is more appropriate (Sterba, Egger, & Angold, 2007). In summarizing the

current literature on DSM syndrome differentiation, Lahey and colleagues (2004) conclude that

many areas of psychopathology are well-represented by the constructs implied by DSM-IV

criteria. However, results are more ambiguous in differentiating between Oppositional Defiant

Disorder (ODD) and Conduct Disorder (CD) and between the two types of ADHD (Lahey et al.,

2004).

Several factor analytic studies support separate inattentive and hyperactive-impulsive

factors within the constellation of ADHD symptoms (Lahey et al., 2008; Glutting, Youngstrom,

& Watkins, 2005; Willoughby, Curran, Costello, & Angold, 2000). However, some research

suggests a unidimensional ADHD factor, particularly for younger children. One study found that

a unifactorial model of ADHD was most appropriate for 4 -5 year-old children, whereas a

bifactorial model was the best fit for 6 7 year-old children (Bauermeister, 1992).

Differentiation between ODD and CD has also been difficult to establish, primarily

because of the low occurrence of CD symptoms in both child and adolescent samples. Some

studies have combined ODD and CD symptoms and focused on its differentiation from ADHD

(Pillow, Pelham, Hoza, Molina, & Stultz, 1998), while others simply exclude CD symptoms

because of their low frequency and focus only on the differentiation between ODD and ADHD









Table 3.1. Participant characteristics.
Total Sample GIFT SHAPE GANA


Age (years)

Gender

Ethnicity


Caucasian

African American

Hispanic

Asian

Native American


M 4.48
SD = 1.01
71% male


56%

7%

25%

0.5%

0.5%


M= 4.38
SD= 1.09
69% male


76%


M= 4.85
SD = 0.75
74% male


76%

14%

1%

0%


Biracial/Other


Family yearly
income
CBCL Externalizing
T Score

CBCL Internalizing
T score


ECBI Intensity Raw
score


10%
M= 36,278.83
SD = 29,779.31
M= 70.57
SD = 8.95


M= 61.91
SD= 9.91


M 169.56
SD 28.41


12%
M= 34,308.52
SD = 25,302.84
M= 72.82
SD = 6.91


M= 62.54
SD= 9.58


M 168.34
SD 25.06


M= 54,251.54
SD = 39,681.04
M= 69.00
SD = 9.05


M= 59.83
SD = 9.95


M 154.08
SD 29.68


2%
M= 22,582.25
SD = 13,299.15
M= 67.49
SD = 11.07


M= 62.67
SD = 9.63


M 182.58
SD 26.95


M= 4.33
SD = 1.31
80% male


7%

0%

91%

0%









CHAPTER 4
DISCUSSION

The present study provided important information regarding the validity of DSM-IV

symptoms of ODD, CD, and ADHD in a relatively diverse, clinical preschool sample. Findings

also provide support for the validity of current assessment measures for preschoolers. Results can

be used to inform future diagnostic methodology and assessment techniques in this population.

Ultimately, findings from the current study can support effective assessment and treatment of

psychopathology in preschoolers.

Validity of the DSM-IV for Preschoolers

Results suggest that the four constructs of ODD, CD, ADHD-Inattentive and ADHD-

Hyperactive implied by the DSM-IV are valid in younger children. Even when using the

relatively conservative standards outlined by Yu and Muthen (2002), indices suggested good fit.

The factor loadings of each symptom onto its corresponding underlying construct were generally

high. None fell beneath the commonly used level of .4 (Floyd & Widaman, 1995) and only one

symptom had a factor loading below .6. Results generally support the growing body of research

supporting the structure of the DSM-IV in older children (Lahey et al., 2004; Lahey et al., 2008)

and adds to the limited body of research examining the validity of DSM-IV nosology in

preschoolers.

The four-factor DSM-IV model provided the best fit in this sample of clinic-referred

preschoolers when compared to a three-factor model that combined the inattentive and

hyperactive domains into a single ADHD factor, with separate constructs of ODD and CD. The

DSM-IV model also proved superior when compared to a three-factor model of ADHD-

Inattentive, ADHD-Hyperactive, and a disruptive behavior factor that collapsed ODD and CD

symptoms. The two alternate models resulted in adequate fit according to certain fit indices;









the ECBI items, representing oppositional and attention-seeking behaviors. The only significant

loadings onto the second factor demonstrated a negative relationship with the underlying

construct. The third factor included five items seeming to reflect a lack of respect for rules, such

as having poor table manners and refusing to do chores when asked. None of the items

demonstrated particularly strong loadings onto the third factor.

The one-factor model was then subjected to CFA to examine the fit of the model. Items

with factor loadings below .4 on the EFA were dropped from the analysis. This included many

items that failed to demonstrate significant loadings onto any factors in previous studies, such as

whining and having poor table manners. The CFA also excluded several items that have

previously been included on the "conduct problems" factor in previous studies, as these items

seemed to be reflective of a more severe behavior than most other ECBI items. Thus, 25 items

were subjected to a one-factor CFA. All measures of model fit improved and the results

suggested good fit according to the TLI. The CFI value was marginal according to the guidelines

proposed by Yu and Muthen (2002), although comparable to the value reported in previous

studies (Weis, Lovejoy, & Lundahl, 2005), and the WRMR was also marginal. However, the

RMSEA continued to suggest poor fit (see Table 3-8). The factor structure of the one-factor CFA

is presented in Table 3-10.

Exploration of Age-Graded Criteria

As a preliminary examination of the possibility of age-graded criteria, a series of

hierarchical regressions were conducted in which scores on the corresponding scale of the CBCL

were regressed upon total number of symptoms, child age, and an interaction between those two

variables. To test the hypothesized moderator effects of age, we evaluated the significance of the

change in R2 after the addition of the interaction term.









The overall MANOVA revealed a significant main effect for study source, F(6, 378) =

10.94, p < .01, 2 = .148. To further understand the model, univariate ANOVAs were conducted

for each dependent variable. The main effect of study source was significant for Externalizing

score, F(2, 190) = 7.32, p < .01, r2= .071, and ECBI intensity score, F(2,190) = 13.65, p < .001,

r2= .126. However, the three groups did not differ in Internalizing scores, F(2, 190) = 1.32, p =

.271. Post hoc tests with Bonferroni correction were conducted to decompose further the

significant main effect of study source. Children enrolled in GIFT demonstrated higher

Externalizing scores than GANA participants (p < .01), but GIFT and SHAPE participants did

not differ in their Externalizing scores (p = .06). Additionally, GANA participants demonstrated

higher ECBI scores than both GIFT (p < .01) and SHAPE (p < .001) participants. GIFT

participants also exhibited higher ECBI scores as compared to SHAPE participants (p < .05).

Table 3-1 includes the means and standard deviations for the three studies on each of the

three dependent variables. Overall, the results suggest some differences in the clinical

presentation of participants across different studies. Children enrolled in GIFT demonstrated

higher rates of externalizing behaviors than children in GANA. However, GANA participants

demonstrated higher total ECBI scores than both the GANA and SHAPE participants. Thus, it

seems that GANA children cannot uniformly be characterized as less severe in their clinical

presentation and rates of overall disruptive behavior.

Symptom and Diagnostic Patterns

Participants were drawn from three different studies, and DSM symptoms were assessed

with a different assessment measure in each study. Symptom frequencies and rates of diagnosis

are therefore presented by assessment and study source (See Tables 3-2 and 3-3). Notably, six of

the fifteen DSM symptoms for CD occurred between zero and two percent of families in the









Table 3-5. Four-factor solution of DSM-IV symptoms
ADHD-I ADHD-H ODD CD
Symptom
ADHD-I
Careless mistakes .74
Short attention span .90
Doesn't listen .89
Fails to finish tasks .86
Disorganized .73
Avoids concentrating .80
Loses items .74
Distractible .90
Forgetful .76
ADHD-H
Fidgets .81
Leaves Seat .88
Runs/Climbs .88
Too Loud .82
Driven by motor .90
Excessive talking .68
Blurts out answers .67
Difficulty waiting turn .69
Interrupts .72
ODD
Loses temper .67
Argues .84
Defiant .89
Annoys others .72
Blames others .71
Easily annoyed .68
Angry .74
Spiteful .65
CD
Bullying .79
Fighting .63
Use of weapon .77
Cruel to people .68
Cruel to animals .46
Destruction of property .71
Lying .73
Stealing composite .71









Table 3-6. Correlations among DSM-IV factors
ADHD- ADHD- ODD CD
Inattentive Hyperactive


ADHD-I 1.00

ADHD-H .81 1.00

ODD .49 .54 1.00

CD .35 .39 .74 1.00









field is beginning to consider the possibility of different criteria for various developmental levels.

Although the DSM-IV advises its users to consider what is "developmentally appropriate" for a

child of a certain age, this is a vague recommendation, and it is unclear how much credence

clinicians give to this warning. A more formal age-graded system based on developmental norms

would concretize the importance of the child's developmental level in assigning diagnoses. This

concept is included in the Diagnostic and Statistical Manual for Primary Care (DSM-PC;

Wolraich, Felice, & Drotar, 1996), developed to facilitate the identification of psychological

disorders in the primary care setting. The framework of the DSM-PC reflects the idea that

psychopathology occurs along a spectrum: the clinician can classify problematic child behaviors

as developmental variations, problems, and disorders, with each category denoting an increasing

level of problem severity. The category of "developmental variation" allows the clinician to

recognize that a child behavior viewed as problematic to the caregiver may reflect a normative

variation of child development. "Problems" include those behaviors that may be distressing or

causing some impairment in the child's overall functioning but do not warrant a formal DSM-IV

diagnoses. The final category of "disorders" reflects those categorized in the DSM-IV. The

DSM-PC also includes concrete examples of developmental variations, problems, and disorders

across different age groups. Thus, the DSM-PC recognizes that psychopathology occurs along a

continuum, and that the level at which a behavior is reflective of psychopathology is influenced

by child age (Wolraich, Felice, & Drotar, 1996).

Research Questions

This study was designed to examine the validity of DSM-IV symptoms for ODD, CD, and

ADHD in preschoolers and explore the validity of current diagnostic assessment measures for

this age group. The study also examined relations between DSM symptoms, symptom severity,









will exclude these symptoms in the assessment of preschoolers. Further research is needed on

potential modifications to CD criteria to increase developmental sensitivity while also

maintaining the clinical essence of the symptoms (Egger & Angold, 2006).

Results also confirm the validity of two separate factors of inattention and hyperactivity

within the ADHD domain, specifically in a preschool sample. A substantial body of research

supports this distinction in older children and adolescents (Glutting, Youngstrom, & Watkins,

2005; Willoughby, Curran, Costello, & Angold, 2000). Some research has suggested that a single

ADHD factor is most appropriate for younger children (Bauermeister, 1992). Yet results from

our study replicate recent findings by Sterba, Egger, and Angold (2007) and add to the validity of

the DSM-IV model of ADHD in young children. Nevertheless, further study is needed. Given

the necessity of ADHD symptoms in the school setting for diagnosis, it will be particularly

important to validate the two domains of inattention and hyperactivity with teacher ratings. Data

from the Preschoolers with Attention Deficit/Hyperactivity Disorder Treatment Study (PATS)

found conflicting results between parent and teacher ratings of ADHD symptoms. Specifically,

the two-factor model of inattention and hyperactivity provided adequate fit when applied to

parent responses on a behavior rating scale of ADHD symptoms, but neither a one-factor nor

two-factor model provided adequate fit for teacher ratings (Hardy et al., 2007). The authors note

that the use of a behavior rating scale may not have provided the most accurate representation of

DSM symptoms; thus, the failure of the study to find support for the two-factor model may

reflect a methodological artifact. However, it also may be that the structure of ADHD symptoms

is different when applied to classroom behavior (Hardy et al., 2007).

This study focused on the validity of specific diagnostic categories in preschoolers. The

validation of psychopathology for preschoolers is an important step. However, more research is



































In loving memory of my mother, Paula Joyce Chase. Her strength is my inspiration, her wisdom
a source of guidance, and her love a constant support.









The first model examined the predictive power of ADHD symptoms and child age in

inattention scores. Specifically, scores from the attention scale of the CBCL were regressed upon

total number of ADHD symptoms, child age, and an interaction term between ADHD symptoms

and child age to test the hypothesis that age would moderate the relationship between ADHD

symptom count and severity of inattention. Total number of ADHD symptoms was entered in the

first block and acted as a significant predictor, F(1, 169) = 81.28, p < .001, R2 = .325. Child age

was entered into the model and did not add to the predictive power of the model, R2 A = .001, nor

was the interaction term between number of symptoms and age significant, R2 A = .001. Thus,

number of ADHD symptoms was a significant predictor of attention scores on the CBCL.

However, the current analysis failed to support the hypothesis that age would moderate that

relationship between symptom count and severity of inattention.

The second model examined the predictive power of ODD symptoms and child age in

externalizing behavior scores. In this analysis, scores from the externalizing scale of the CBCL

were regressed upon total number of ODD symptoms, child age, and an interaction term between

ODD symptoms and child age to test the hypothesis that age would moderate the relationship

between ODD symptom count and severity of externalizing behavior. Total number of ODD

symptoms was entered in the first block and acted as a significant predictor, F(1, 198) = 63.04, p

< .001, R2 = .241. Child age was entered into the model and did not add to the predictive power

of the model, R2 A = .011, nor was the interaction term between number of symptoms and age

significant, R2 A = .001. Thus, number of ODD symptoms was a significant predictor of

externalizing scores on the CBCL. However, the analysis failed to support the hypothesis that

age would moderate the relationship between symptom count and severity of externalizing

behavior.









Table 3-9. The CFA of the three-factor model of the ECBI
ECBI item Factor One Factor Two Factor Three
12. Gets angry when doesn't get own way .94
10. Acts defiant when told to do something .93
8. Does not obey house rules on own .90
14. Sasses adults .82
9. Refuses to obey until threatened with .76
punishment
13. Has temper tantrums .50
11. Argues with parents about rules .41
5. Refuses to do chores when asked .47
7. Refuses to go to bed on time .45
17. Yells or screams .36

30. Is easily distracted .96
34. Has difficulty concentrating on one thing .90
32. Fails to finish task or projects .88
31. Has short attention span .56

24. Verbally fights with friends own age .87
22. Lies .85
26. Physically fights with friends own age .84
25. Verbally fights with sisters and brothers .55
27. Physically fights with sisters and brothers .53
19. Destroys toys and other objects .21
23. Teases or provokes other children .15
21. Steals -.47









Future Directions

This study adds to the limited literature on the validity of current diagnostic standards and

assessment methods for preschoolers. Overall, results suggest that DSM-IV categories and

common assessment measures are appropriate for this age group. Future research should

continue to examine the validity and reliability of preschool diagnosis. An important area of

research includes the predictive validity of preschool diagnosis. Although a growing body of

research supports DSM disorders in preschoolers, it is unknown which preschoolers will

continue to demonstrate symptoms into childhood and adolescence. Longitudinal research is

needed on long-term outcomes of children diagnosed with disruptive behavior disorders in early

childhood. Further research in this area will increase our understanding of the factors that

distinguish developmental variations from true psychopathology and promote the continuing

evolution of the DSM.









rather than clinical, samples. Future research examining the underlying factor structure of the

ECBI therefore should focus on clinic-referred preschool populations.

Developmental Modifications to Current DSM Standards

The field is increasing its focus on the expression of psychological disorders in

preschoolers. Along with recognition that certain behaviors listed in the DSM are, in fact, found

in normally developing children, comes the realization that a more appropriate approach to

diagnosis in children may involve developmentally modified criteria. Some researchers argue

that criteria should be less stringent for younger children because it would be unrealistic for

young children to demonstrate stable patterns of problem behaviors (Luby et al., 2002); others

argue for more stringent criteria, given that fleeting problem behaviors can characterize normal

development (Wakschlag, Leventhal, & Thomas, in press). Limited research has systematically

explored developmental modifications to DSM criteria, yet this research is increasingly

recognized as an important future direction to improve the developmental sensitivity of current

diagnostic tools (Wakschlag, Leventhal, & Thomas, in press; Egger & Angold, 2004; Eyberg,

Schuhmann, & Rey, 1998).

One mechanism that may increase developmental sensitivity involves the use of age-

graded criteria. According to the DSM-IV, a six-year-old and a fourteen-year-old must

demonstrate the same number of symptoms to warrant a diagnosis of ADHD. This standard

seems somewhat counterintuitive given that the six-year-old will naturally be more active and

less attentive than the fourteen-year-old. The young child is therefore more likely to receive the

diagnosis, even if the adolescent deviates more severely from age-appropriate behaviors (Silk et

al., 2000). The concept of age-graded criteria is particularly relevant in diagnosing preschool

ADHD; the high rates of ADHD-like behaviors in typically developing children has led to the

recommendation of longer symptom duration for very young children (Barkley, 2006). Thus, the









diagnostic criteria include a wide range of heterogeneous behaviors that generally reflect deviant

and defiant behaviors; Huffine (2002) argues that the diagnosis of CD has little clinical utility

and does not relate to any specific treatment program beyond those typically used to target

symptoms of ODD. Additionally, high rates of comorbid ODD/CD and ADHD are well

documented, particularly in school-aged samples, with estimates of co-occurrence ranging from

30 to 62% (Newcorn et al., 2001; Wilens et al., 2002). These findings call into question the

uniqueness of these disorders and have led to the proposal of an ODD/CD subtype of ADHD

(Nottelmann & Jensen, 1995; Christiansen et al., 2008).

Validity of the DSM-IV Nosology

Comorbidity in child psychopathology is common, with estimates as high as 75%

(Nottelmann & Jensen, 1995). One explanation for the high rates of comorbidity relates to the

idea that comorbid disorders could result from conceptual flaws in our current diagnostic system.

That is, comorbidity may reflect a methodological artifact: a failure of the DSM to represent

properly the constructs underlying psychological disorders. The high rates of comorbidity might

suggest that a more parsimonious nosology would be more accurate than that currently implied

by the DSM (Lilienfeld, Waldman, & Israel, 1994; Krueger, Caspi, Moffitt, & Silva, 1998). This

argument is particularly salient when discussing child psychological disorders. It has been

proposed that due to their developmental level, children vulnerable to psychological problems

are more likely to demonstrate a range of psychological symptoms that cut across different

diagnostic categories (Nottelmann & Jensen, 1995). Indeed, the differentiation of child

psychological disorders is often difficult to establish; research involving the internalizing

disorders suggests that anxiety and depressive symptoms are frequently undifferentiated in

young children (Cole, Truglio, & Peeke, 1997). It has been argued that specific psychopathology

dimensions evolve with development (Craighead, 1991) and that a more general framework of









The Early Childhood Inventory, fourth edition (ECI-4) is a parent behavior rating scale

that screens for DSM-IV emotional and behavioral disorders and is specifically designed for

children between 3 and 6 years old (Gadow & Sprafkin, 2000). Parents are asked to rate the

frequency of specific behaviors that map onto DSM criteria for common child diagnoses.

However, certain symptoms deemed developmentally irrelevant are excluded, such as the

truancy item from the CD module. Additionally, the wording of certain items is altered to

increase their developmental sensitivity. The ECI-4 may be scored according to either the

categorical or dimensional approach.

Structured and semi-structured interviews based on DSM criteria represent the categorical

approach to diagnosis in which a disorder is rated as either present or absent. Behavior rating

scales, in which a child's behavior is rated along a continuum and then compared to a normative

sample, represent the dimensional approach. Both approaches include their own strengths and

weaknesses. Categorical approaches have been criticized for poor specificity; some research

suggests high false-positive rates across diagnostic categories. Categorical approaches have also

been criticized for failing to account for overlapping symptoms, subsequently inflating rates of

comorbidity (Achenbach, 1995). Dimensional scales are not without their own weaknesses, and

have been criticized for poor sensitivity and for failing to provide important information

regarding onset, duration, and severity of symptoms. Although the relative superiority of these

two approaches has been a popular area of debate, research does not clearly support one method

over the other. Jensen and colleagues (1996) directly compared these two approaches and found

little difference between them. Both the dimensional approach (i.e., Child Behavior Checklist;

CBCL) and the categorical approach (i.e., DISC, Version 2.1) demonstrated low to moderate

correlations with a variety of external validators, such as school impairment, service utilization,









Study Lim itations........ ......... ...................................... ........ .. ... .....64
F utu re D direction s .................................................................67

APPENDIX: STATISTICAL ANALYSES NOT INCLUDED IN MAIN TEXT........................68

L IS T O F R E F E R E N C E S .................................................................................... .....................7 1

B IO G R A PH IC A L SK E T C H .............................................................................. .....................80















































6









Table 3-4. Absolute, relative, and component model fit for three DSM-IV CFA models
Model 2 DF p CFI TLI RMSEA WRMR

DSM-IV 170.66 82 < .001 .95 .96 .06 1.0
(4 factor)
Combined ADHD 177.05 88 < .001 .93 .95 .07 1.1
(3 factor)
Disruptive behavior 174.90 89 < .001 .93 .95 .07 1.1
(3 factor)









needed to demonstrate the clinical utility of these specific disorders in young children. As noted

by Huffine (2002), treatment recommendations for CD are virtually identical to those for ODD,

leading some to question the utility of two separate disorders. This argument seems particularly

relevant to preschoolers because the more intensive, multimodal treatment programs sometimes

used to treat CD, such as Multisystemic Therapy (MST; Henggeler & Borduin, 1990) would be

inappropriate for preschoolers. A similar argument may be relevant when discussing the clinical

diagnosis of ADHD. Treatment recommendations for ADHD in school-aged children and

adolescents frequently involve a medication trial and more individualized behavioral

interventions, such as organizational skills training. Results from the Preschoolers with Attention

Deficit/Hyperactivity Disorder study (PATS) suggest that stimulant medications can be effective

in this age group (Greenhill et al., 2006; Vitiello et al., 2007). However, the efficacy of stimulant

medication tends to be less (Greenhill et al., 2006) and the side effects more severe when

compared to school aged children and adolescents (Wigal et al., 2006). Notably, no medication is

currently approved by the Food and Drug Administration (FDA) for the treatment of ADHD in

preschool aged children. A recent study examined maternal perceptions of mental health services

for their preschool-aged children. The majority of mothers surveyed expressed interest in

receiving assistance in managing their child's behavior problems; however, most also indicated

that they would not be interested in pursuing medication (Harwood, O'Brien, Carter, & Eyberg,

under review). Due to the controversy surrounding medication use in this age group, and parental

concerns regarding the safety of medication for young children, first line treatment

recommendations for preschoolers with ADHD often involve behavioral interventions such as

parent training and contingency management, the same treatments used for ODD/CD in young

children (dosReis, et al., 2003; Rushton, Fant, & Clark, 2004; Verduin, Abikoff, & Kurtz, 2008).









generally demonstrate certain ADHD behaviors as part of their natural development.

Consequently, the boundaries between normative and pathological may be particularly

challenging to establish when assessing for ODD, CD, or ADHD in young children. The

diagnosis of CD in preschoolers is particularly controversial, and the validity of this diagnosis in

very young children has been hotly debated (McClellan & Speltz, 2003; Wilens et al., 2003).

One study followed a sample of 92 preschoolers diagnosed with ODD over two years and found

very low rates of the CD diagnosis over time (Speltz, McClellan, DeKylen, & Jones, 1999). The

authors argued that most CD criteria are irrelevant to preschoolers due to their developmental

level and typical environment. In contrast, however, several studies have found support for the

diagnosis in clinical populations, with estimates ranging from 22 to 40% (Schuhmann, Foote,

Eyberg, Boggs, & Algina, 1998; Keenan & Wakschlag, 2000; Wilens et al., 2002).

Research does support the validity of these disorders in very early childhood in regards to

distress and interference in daily functioning. Preschoolers diagnosed with disruptive behavior

disorders demonstrate clinically significant levels of aggression, noncompliance, and destructive

behavior that lead to impairments in social, academic, and family functioning (Keenan &

Wakschlag, 2000; Keenan & Wakschlag, 2002). Similarly, preschoolers diagnosed with ADHD

demonstrate significant social and academic impairment relative to controls (Lahey et al., 1998),

and ADHD preschoolers display similar impairments in functioning when compared to their

school-aged counterparts (Wilens et al., 2002). Therefore, evidence supports the validity of

ODD/CD and ADHD in preschoolers, but questions remain regarding the applicability of the

current diagnostic standards to younger populations. Research is needed to determine whether

younger children are accurately classified according to the categories included in the DSM-IV.









CHAPTER 3
RESULTS

Participant Characteristics

The 207 families who participated in the study included 147 boys (71%) and 60 girls

(29%) with a mean age of 4.48 years (SD = 1.01). Fifty six percent of children were Caucasian

(by parent report); 25% Hispanic, 7% African-American, 0.5% Asian, 0.5% Native American,

6% biracial, and 4% identified as "Other." Mean yearly income for the entire sample was $36,

278.83 (SD= $29,779.31); however, income varied significantly across the three studies.

Families enrolled in Project SHAPE reported significantly higher income than families in both

GIFT (t(142) = 3.68, p < .001) and GANA (t(105) = 5.60, p < .001). Additionally, families in

Project GIFT reported higher income than families in Project GANA (t(145) = 3.18, p < .001).

Full information on participants is presented in Table 3.1.

The clinical presentation of these three groups was also compared. Specifically, a between

factor MANOVA was conducted with three dependent variables: CBCL externalizing score,

CBCL internalizing score, and the ECBI intensity score. The GIFT sample was significantly

larger than the other two groups. Multicollinearity among the three dependent variables (CBCL

Externalizing score, CBCL Internalizing score, and the ECBI intensity score) was examined and

was within acceptable limits.

The Box-M test for the homogeneity of variance-covariance matrices produced a

significant result, F(12, 78260.34) = 5.33, p < .01. However, Levene's test for the Externalizing

score was the only significant result, F(2, 190) = 6.32, p < .01. Due to this violation of the

assumption of homogeneity of variance, the relatively conservative Pillai's trace was used in the

following MANOVA.




Full Text

PAGE 1

1 DIAGNOSTIC VALIDITY OF DSM SYMPTOMS AND CRITERIA FOR PRESCHOOLERS: DISRUPTIVE BEHAVIOR AND ATTENTION DEFICIT HYPERACTIVITY DISORDER SYMPTOMS By RHEA MARISA CHASE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Rhea Marisa Chase

PAGE 3

3 In loving memory of my mother, Paula Joyce Chas e. Her strength is my inspiration, her wisdom a source of guidance, and her love a constant support.

PAGE 4

4 ACKNOWLEDGMENTS I thank m y mentor and chair, Dr. Sheila Eybe rg, for her unending enthusiasm and constant encouragement. I am very appreciative to th e members of my docto ral committee for their guidance, and would like to thank Dr. Michael Marsiske for his statis tical knowledge and the amazing skill and patience with which he teach es others. I would also like to thank Dr. Chongming Yang for introducing me to the MPlus program. I thank Drs. Kristen McCabe and Christopher Lucas, as well as members of the Child Study Laboratory at the University of Florida, for their contributions to the current rese arch. This dissertation is the culmination of five rewarding but challenging years. I am eternally gr ateful to the many friends and family members who have supported me along the way. I especia lly acknowledge my father, Stephen Chase, whose love and support has been critical to a ny success I have achieved in graduate school.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT.....................................................................................................................................9 CHAP TER 1 INTRODUCTION..................................................................................................................11 ADHD and the Disruptive Behavior Disorders ...................................................................... 14 Validity of the DSM-IV Nosology.........................................................................................15 Issues in Preschool Diagnosis................................................................................................. 18 Current Assessment Strategies............................................................................................... 20 Developmental Modifications to Current DSM Standards .....................................................23 Research Questions............................................................................................................. ....24 2 METHODS.............................................................................................................................27 Participants.............................................................................................................................27 Project GIFT....................................................................................................................27 Project GANA.................................................................................................................28 Project SHAPE................................................................................................................28 Measures.................................................................................................................................29 Procedures..................................................................................................................... ..........32 Confirmatory Factor Analys is and Model Evaluation ............................................................32 3 RESULTS...............................................................................................................................36 Participant Characteristics.................................................................................................... ..36 Symptom and Diagnostic Patterns.......................................................................................... 37 Confirmatory Factor Analyses of DSM Symptoms................................................................ 38 Validity of the Three Assessment Methods............................................................................ 39 Confirmatory Factor Analysis of th e Eyberg Child Behavior Inventory ............................... 40 Exploration of Age-Graded Criteria....................................................................................... 42 4 DISCUSSION.........................................................................................................................56 Validity of the DSM-IV for Preschoolers............................................................................... 56 Validity of Assessment Measures........................................................................................... 60 Factor Analysis of the Eyberg Child Behavior Inventory ......................................................62 Exploration of Age-Graded Criteria....................................................................................... 64

PAGE 6

6 Study Limitations.............................................................................................................. ......64 Future Directions....................................................................................................................67 APPENDIX: STATISTICAL ANALYSES NOT INCLUDED IN MAIN TEXT ........................ 68 LIST OF REFERENCES...............................................................................................................71 BIOGRAPHICAL SKETCH.........................................................................................................80

PAGE 7

7 LIST OF TABLES Table page 2-1 Overview of relevant measures across datasets.................................................................35 3.1 Participant characteristics................................................................................................ ..45 3-2 Individual symptom percentages acr oss the DIS C-IV, YC-DISC, and ECI-4..................46 3-3 Percentage of diagnosis across assessm ent method........................................................... 47 3-4 Absolute, relative, and component m odel fit for three DSM-IV CFA models .................. 48 3-5 Four-factor solution of DSM-IV symptoms...................................................................... 49 3-6 Correlations among DSM-IV factors................................................................................. 50 3-7 Pearson correlations between symptom counts and CBCL and ECBI scores...................51 3-8 Absolute, relative, and component model fit for ECBI m odels......................................... 52 3-9 The CFA of the three-factor model of the ECBI............................................................... 53 3-10 One-factor CFA of the ECBI............................................................................................. 54 A-1 One-factor EFA of the Eyberg Chil d Behavior Inventory(E CBI) (RMSR = .10)............. 68 A-2 Two-factor EFA of the Eyberg Child Behavior Inventory(E CBI) (RMSR = .08)............ 69 A-3 Three-factor EFA of the Eyberg Chil d Behavior Inventory(E CBI) (RMSR = .06).......... 70

PAGE 8

8 LIST OF FIGURES Figure page 3-1 Path diagram of four-factor solution of DSM-IV sym ptoms of ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD.................................................................................. 55

PAGE 9

9 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIAGNOSTIC VALIDITY OF DSM SYMPTOMS AND CRITERIA FOR PRESCHOOLERS: DISRUPTIVE BEHAVIOR AND ATTENTION DEFICIT HYPERACTIVITY DISORDER SYMPTOMS By Rhea Marisa Chase August 2008 Chair: Sheila M. Eyberg Major: Psychology Attention Deficit Hyperactivity Disorder (ADHD) and the disruptive behavior disorders of Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) represent the most common child psychological disorders diagnosed today, and are increasingly being diagnosed in preschool-aged children. Developmental theory s uggests that the criteria outlined in the fourth edition of the Diagnostic and Statistical Ma nual of Mental Disorders (DSM-IV) may be inappropriate for younger children, but very little research has examined the validity of these diagnostic categories and assessment measures fo r this age group. This study therefore explored the validity of current diagnos tic standards and assessment m easures of ADHD, ODD, and CD specifically in preschool-aged children. The sample consisted of 207 childr en aged three to six years referred for treatment of symptoms of AD HD and disruptive behavior. Confirmatory factor analysis was used to compare three possible sy mptom models of these diagnoses: (a) a fourfactor model of ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD currently implied by the DSM-IV; (b) a three-factor model that combin ed ADHD symptoms into one single factor, but included separate constructs of ODD and CD; and (c) a three-factor model allowing for separate ADHD-Inattentive and ADHD-Hyperactive factors but collapsed symptoms of ODD and CD

PAGE 10

10 into one disruptive behavior construct. Finding s supported the four-factor model represented by the DSM-IV, which provided superior fit for the data as compared to both of the less differentiated models. The study also further explored current assess ment strategies with this age group. We assessed and compared the validity of the fourth edition of the Diagnostic Interview Schedule for Children (DISC-IV), the Young Ch ild Version of the Diagnostic Interview Schedule for Children (YC-DISC), and the fourth edition of the Early Childhood Inventory (ECI4). All measures demonstrated similar correla tions with other measures of child behavior, suggesting comparable validity across measures for this age group. Overal l, results support the use of DSM-IV diagnostic categories with pr eschool-aged children and add to the limited research on the validity of current assessmen t methods for young children. These findings help promote accurate identification and subsequent tr eatment of psychopathology in preschoolers.

PAGE 11

11 CHAPTER 1 INTRODUCTION The Am erican Psychiatric Association (APA) introduced its first catalogue of mental disorders in 1952, largely in response to the in creased prevalence and awareness of mental disturbance following World War II (Silk et al., 2000). The origin al version of the Diagnostic and Statistical Manual of Mental Disorders (D SM-I; APA, 1952) made no distinction between child and adult disorders. A child could be given any adult diagnosis, with little consideration to any developmental differences in classification or expression of a diso rder (Silk et al., 2000). The second edition of the DSM (DSM-II; APA, 1968) introduced a section entitled Behavior Disorders of Childhood and Adolescence, which included such diagnoses as hyperkinetic reaction of childhood and overa nxious reaction of childhood. DSM-II therefore began to consider the differences in the expression of psychopathology in childr en and adolescents as compared to adults. However, the DSM-II emphasi zed that these categorie s should be used only when no other classification was possible, as this nascent childhood disturbance most likely represented a developing adult disorder. The nos ology of DSM-II reflected the widespread belief that psychopathology in adulthood could always be traced back to childhood and adolescence, and child mental illness could be conceptualized as a slowly emerging di sorder that would be fully realized in adulthood (Silk et al., 2000). Thus, although the DSM-II recognized that the developing child may differ from the adult in the symptomatic expression of psychological disturbance, its theoretical framew ork suggested that the disorder was always present in the same form, but may not have fully re vealed itself in a young child. The third edition of the DS M (DSM-III; APA, 1980) underwent major revisions in diagnostic method and overall approach to clas sifying psychopathology. For the first time, the DSM-III included detailed descriptions of ment al disorders and specific criteria for each

PAGE 12

12 category. The DSM-III rejected the psychoanaly tic framework underlying its previous edition and instead based diagnoses on empirically derived categories. The DSM-III also stipulated that a disorder is characterized by functional impairment and subject ive distress from the patient. The multiaxial system was first introduced in the DSM-III, which required clinicians to assess developmental disorders, personality disorders, medical conditions, and global functioning. The DSM-III also provided information on clinical presen tation of disorders, such as age of onset and gender differences (APA, 1980). Inclusion of thes e factors demonstrated some acknowledgement of the role of context and environmen t in the expression of psychopathology. The number of specific childhood disorders signi ficantly increased from the second to the third edition of the DSM (Bemporad & Schw ab, 1986). The DSM-II section of Behavior Disorders of Childhood and Adolescence included s hort and somewhat vague descriptions of a range of child problem behaviors, including runa way reaction of childhood or adolescence and unsocialized aggressive reaction of chil dhood or adolescence (DSM-II; APA, 1968). The DSM-III introduced the specific disruptive behavior disorders of Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD). ODD is charact erized by a pattern of child noncompliance and disrespect for adult authority, while CD is ch aracterized by a repetitive and persistent pattern of behavior in which the basic ri ghts of others or major age-appr opriate societal norms or rules are violated (APA, 1980). The DSM-III also further developed the DSM-II diagnosis of hyperkinetic reaction of childhood and introduced the diagnosis of Atten tion Deficit Disorder (ADD), which was viewed primarily as a disorder of inattention and could be diagnosed with or without hyperactivity. The refineme nt of childhood disorders conti nued with the revised version of the DSM-III, which included age-appropriate modi fications to several symptoms in an attempt to increase their developmental sensitivity (DSM-III-R; APA, 1987). Im portantly, the DSM-III-

PAGE 13

13 R was the first version to use research from DS M field trials to clarify and validate diagnoses, and results from these trials were used to inform and refine diagnostic cr iteria for several child disorders, including the former ADD category now classified as Attention Deficit Hyperactivity Disorder (ADHD; DSM-III-R; APA, 1987). The fourth edition of the DSM (DSM-IV; AP A, 1994) continued to develop childhood diagnoses and expand subtypes of certain child disorders. DSM-IV criteria for ODD require that children meet four out of eight possible symptoms, such as: arguing with adults, actively defying or refusing to comply with adults requests or ru les, and deliberately a nnoying others. CD criteria have changed somewhat since the disorders fi rst appearance in the DS M-III, although the core feature, involving blatant and persistent disregard for others or so cietal rules, has remained the same. Current criteria necessitate three or more behaviors relating to aggression, destruction of property, deceitfulness or theft, and major violations of societal no rms or rules. The text of the DSM-IV states that a diagnosis of CD should re flect internal psychologi cal dysfunction and not a reaction to a negative environment. However, this qualification is not included in DSM-IV diagnostic criteria, leading some to question the clinical utility of this text comment (Wakefield, Pottick, & Kirk, 2002). The DSM-IV also revised the diagnostic cat egories related to ADHD, which now consist of three subtypes: predominantly inattentive, predominantly hyperactive, and combined type. Children must meet six of nine possible inatte ntive symptoms, such as difficulty sustaining attention and organizing tasks a nd activities, or six of nine hype ractivity/impulsivity symptoms, such as fidgeting or squirming and interrupting or intruding on others. To warrant a diagnosis of ODD or ADHD, symptoms must have persisted for at least 6 months; in th e case of CD, three or more criterion symptoms must have been present w ithin the last twelve m onths, with at least one

PAGE 14

14 present in the previous six mont hs. As with all DSM-IV categor ies, the symptoms must cause significant impairment in everyday functioning to indicate a true di sorder (DSM-IV; APA, 1994). ADHD and the Disruptive Behavio r Disorders ADHD and the disruptive behavior disord ers, ODD and CD, are the most common childhood disorders diagnosed today. ODD and CD ar e estimated to occur in 2 to 16% of schoolage children, depending on the population sampled and the method of assessment (Loeber et al., 2000). ODD prevalence tends to be highest in younger samples whereas the more severe behaviors inherent in the diagnos is of CD become more prevalen t at older ages. Additionally, the CD diagnosis supersedes that of ODD if symptoms of both disorders are present; thus, the two disorders are viewed on a spectrum of severity. ADHD is diagnosed in 1 to 5% of school-aged children (APA, 1994) and 2% of preschool-a ged children (Lavigne et al., 1998), although community studies based on teacher report have included estimates as high as 16% (Nolan, Gadow, & Sprafkin, 2001). ODD/CD and ADHD are linke d with a myriad of detrimental effects on a variety of psychosocial outcom es and are associated with hi gher levels of parenting stress (Baker & Heller, 1996; DuPaul et al., 2001), gr eater social impairment (DuPaul et al., 2001; Greene et al., 2002; Wilens et al., 2002), and increased risk for emotional problems (Thomas & Guskin, 2001) when compared to controls. Th e disruptive behavior disorders and ADHD are generally recognized as common childhood disord ers that warrant early identification and treatment (Abikoff & Klein, 1992; Ross, et al., 1998). Complicating the clinical presentation of these common childhood disorders is the high rate of overlap among them. As noted above ODD and CD reflect similar patterns of negativistic, defiant, and hostile be haviors, and CD is conceptualiz ed as a more severe form of ODD. In fact, some researchers have proposed the elimination of CD from the DSM, as

PAGE 15

15 diagnostic criteria include a wide range of heter ogeneous behaviors that ge nerally reflect deviant and defiant behaviors; Huffine ( 2002) argues that the diagnosis of CD has little clinical utility and does not relate to any speci fic treatment program beyond those typically used to target symptoms of ODD. Additionally, high rates of comorbid ODD/CD and ADHD are well documented, particularly in school-aged samples, with estimates of co-occurrence ranging from 30 to 62% (Newcorn et al., 2001; Wilens et al., 2002). These findings call into question the uniqueness of these disorders and have led to the proposal of an ODD/CD subtype of ADHD (Nottelmann & Jensen, 1995; Christiansen et al., 2008). Validity of the DSM-IV Nosology Com orbidity in child psychopathology is common, with estimates as high as 75% (Nottelmann & Jensen, 1995). One explanation for the high rates of comorbidity relates to the idea that comorbid disorders could result from con ceptual flaws in our current diagnostic system. That is, comorbidity may reflect a methodological artifact: a failure of the DSM to represent properly the constructs underlyi ng psychological disorders. The hi gh rates of comorbidity might suggest that a more parsimonious nosology would be more accurate than that currently implied by the DSM (Lilienfeld, Waldman, & Israel, 1994; Krueger, Caspi, Moffitt, & Silva, 1998). This argument is particularly salient when discussi ng child psychological disorders. It has been proposed that due to their deve lopmental level, children vulne rable to psychological problems are more likely to demonstrate a range of ps ychological symptoms that cut across different diagnostic categories (Nottelmann & Jensen, 1995). Indeed, the differentiation of child psychological disorders is often difficult to establish; resear ch involving the internalizing disorders suggests that anxiety and depressive symptoms are fr equently undifferentiated in young children (Cole, Truglio, & Peeke, 1997). It has been argued that specific psychopathology dimensions evolve with development (Craighead 1991) and that a more general framework of

PAGE 16

16 externalizing and internalizi ng disorders may be more rele vant in diagnosing young children (Nottelmann & Jensen, 1995). The differentiation of DSM-IV disorders in child and adolescent samples has become an increasing area of interest within the literature. St udies of this nature ge nerally involve subjecting DSM symptoms to confirmatory factor analys is to determine whether the underlying factor structure implied by the DSM-IV provides the best fit for th e data or whether a more parsimonious model is more appropriate (Sterb a, Egger, & Angold, 2007). In summarizing the current literature on DSM syndrome differentiati on, Lahey and colleagues (2004) conclude that many areas of psychopathology ar e well-represented by the c onstructs implied by DSM-IV criteria. However, results are more ambiguous in differentiating between Oppositional Defiant Disorder (ODD) and Conduct Disord er (CD) and between the two t ypes of ADHD (Lahey et al., 2004). Several factor analytic studies support separate inattentive a nd hyperactive-impulsive factors within the constellation of ADHD symp toms (Lahey et al., 2008; Glutting, Youngstrom, & Watkins, 2005; Willoughby, Curran, Costello, & Angold, 2000). However, some research suggests a unidimensional ADHD factor, particular ly for younger children. One study found that a unifactorial model of ADHD wa s most appropriate for 4 -5 year-old children, whereas a bifactorial model was the best fit for 6 7 year-old children (Bauermeister, 1992). Differentiation between ODD and CD has also been difficult to establish, primarily because of the low occurrence of CD symptoms in both child and adolescent samples. Some studies have combined ODD and CD symptoms and focused on its differentiation from ADHD (Pillow, Pelham, Hoza, Molina, & Stultz, 1998), while others simply exclude CD symptoms because of their low frequency and focus only on the differentiation between ODD and ADHD

PAGE 17

17 (Burns, Boe, Walsh, Sommers-Flanagan, & Teegar den, 2001). When CD symptoms are allowed to map onto their own latent construct, however evidence supports its di fferentiation from ODD (Lahey et al., 2008; Burns et al. 1997). Lahe y and colleagues (2008) directly compared the DSM-IV model, with separate ODD and CD factor s, with a combined ODD/CD model inherent in the International Cl assification of Diseases (ICD-10; World Health Organization, 1993) and found support for separate ODD and CD factors. Notably, the majority of research vali dating the DSM structure of psychopathology involves school-aged children and adolescents. To date, only three studies have employed confirmatory factor analysis to validate DSM-IV disorders in preschool populations. Two of these studies focused on anxiety disorders a nd found support for separa te constructs of separation anxiety, social anxiety, obsessive compulsive di sorder, specific phobias, and generalized anxiety disorder (Eley et al., 2003; Spence, Rapee, McDonald, & Ingram, 2001). A recent study by Sterba, Egger, & Angold (2007) sought to confirm DSM categories across both internalizing and externalizing di sorders in a sample of 307 2to 5-year-olds recruited from a primary care clinic. DSM-IV symptoms of AD HD, ODD, and CD, as well as social phobia, separation anxiety, and generali zed anxiety disorder were assessed with the Preschool Age Psychiatric Assessment (PAPA). Separate confirma tory models were run for the internalizing versus externalizing disorders. Within the disruptive behavior disorder models, a 4-factor model of ADHD-Inattentive ADHD-Hyperactive, ODD, and CD was compared with two 3-factor models, one of which collapsed inattentive and hyperactive symptoms into a single ADHD factor, while the other combined ODD and CD symptoms. Results support ed the three-factor model that included separate ADHD-Inattentiv e and ADHD-Hyperactive factors but combined ODD and CD symptoms into one factor. The authors concluded that an undifferentiated

PAGE 18

18 ODD/CD construct seems most relevant to presch oolers. However, further research examining CD symptoms in preschoolers is necessary given the conflicti ng results surroundi ng the validity of separate ODD and CD factor s within child populations (S terba, Egger, & Angold, 2007; Burns et al., 1997). Indeed, further research speci fically examining CD symptoms in clinical preschool populations seems warranted. The disruptive behavior disorders and ADHD are increasingly being diagnosed in preschool populations, highlighting the need for further assessment of the validity of these disorders in this age group. Issues in Preschool Diagnosis Diagnosing psychopathology in preschool aged children raises im portant questions regarding developmental considerations for this population. As disc ussed above, it was only recently that the DSM began to consider differen ces between children and adults, and the current diagnostic system reflected by the DSM-IV prov ides few modifications for very young children. These may be important issues, as the rapid de velopment characterizing early childhood makes it difficult to assure that maladaptive symptoms are a sign of true psychopathology and not a variation of normal child development. Additiona lly, the DSM provides little information on the normative development of young children to guide diagnostic decisions in this age group (Eyberg, Schuhmann, & Rey, 1998; Egger & Angold, 2004). These issues in preschool diagnosis are partic ularly relevant in the diagnoses of ODD, CD, and ADHD. Due to their cognitive and social de velopment, and the struggle for independence from their parents that commonly occurs in the second or third year, some level of aggression and noncompliance is expected in preschool-a ged children (Campbell, 2002). Thus, a certain level of disruptive behavior symptoms may be present as part of nor mal child development. Similarly, all preschool children are expected to display some level of inattention and hyperactivity. A young child may have difficulty s itting still and sustaining attention, and

PAGE 19

19 generally demonstrate certain ADHD behaviors as part of their natural development. Consequently, the boundaries between normative and pathological may be particularly challenging to establish when assessing for ODD, CD, or ADHD in young children. The diagnosis of CD in preschoolers is particularly controversial, and th e validity of this diagnosis in very young children has been hotly debated (McC lellan & Speltz, 2003; Wilens et al., 2003). One study followed a sample of 92 preschooler s diagnosed with ODD over two years and found very low rates of the CD diagnosis over time (Speltz, McClellan, DeKylen, & Jones, 1999). The authors argued that most CD criteria are irreleva nt to preschoolers due to their developmental level and typical environment. In contrast, however, several st udies have found support for the diagnosis in clinical populati ons, with estimates ranging from 22 to 40% (Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998; Keenan & Wakschlag, 2000; Wilens et al., 2002). Research does support th e validity of these disorders in ve ry early childhood in regards to distress and interference in daily functioning. Pr eschoolers diagnosed with disruptive behavior disorders demonstrate clinically significant levels of aggressi on, noncompliance, and destructive behavior that lead to impairments in soci al, academic, and family functioning (Keenan & Wakschlag, 2000; Keenan & Wakschlag, 2002). Si milarly, preschoolers diagnosed with ADHD demonstrate significant social and academic impair ment relative to contro ls (Lahey et al., 1998), and ADHD preschoolers display similar impairment s in functioning when compared to their school-aged counterparts (Wilens et al., 2002). Therefore, evid ence supports the validity of ODD/CD and ADHD in preschoolers, but questions remain regarding the applicability of the current diagnostic standards to younger populations Research is needed to determine whether younger children are accurately classified accordi ng to the categories included in the DSM-IV.

PAGE 20

20 Further study in this area is crucial to promote effective assessment and subsequent treatment in this age group. Current Assessment Strategies Along with further efforts to validate ADHD and the disruptive behavior disorders in young children is the need to im prove our assess ment of these diagnos es in the preschool population. Relatively few measures have been specifically deve loped for preschoolers; more commonly, assessment methods developed for ol der children are applied to younger children. One important direction therefore includes furt her study of the validity and reliability of commonly used assessment instruments for preschoolers. The Diagnostic Interview Schedule for Children (DISC-IV; Shaffer et al., 2000) is a structured diagnostic interview of child psychopathology based on DSM-IV criteria. The DISC-IV is a highly structured diagnostic interview that is commonly used to assess psychopathology in child ren, particularly in research studies. A parent version (DISC-IV-P) exists for parents or caregivers of 6to 17-year-olds, and a child self-report (DISC -IV-C) is used for 9-to 17-year-olds. Some research studies have used the DISC-P with parents of children as young as 4 (Speltz, McClellan, DeKlyen, & Jones, 1999), but most research on this instrument involve s older children and adolescents. A recent modification of the DISC-IV was designed to increase its developmental sensitivity for preschoolers (Young Child Version of the Dia gnostic Interview Schedule for Children; (YCDISC; Lucas, Fisher, & Luby, 2000). The YC-DISC differs from the DISC-IV by modifying the language of symptoms to be more appropriate for young children. The interview also modified the time frame for each symptom. In the DISC -IV, parents are asked whether a symptom has been present for the past month and then for the past year. The YC-DISC asks only whether a symptom has been present for the past three months This newly developed instrument has yet to be studied empirically to determine its validity in this age group.

PAGE 21

21 The Early Childhood Inventory, fourth edition (ECI-4) is a parent behavior rating scale that screens for DSM-IV emotional and behavioral disorders and is spec ifically designed for children between 3 and 6 years old (Gadow & Sp rafkin, 2000). Parents are asked to rate the frequency of specific behaviors that map onto DSM criteria for common child diagnoses. However, certain symptoms deemed developmen tally irrelevant are excluded, such as the truancy item from the CD module. Additionally, the wording of certain items is altered to increase their developmental sensitivity. The ECI-4 may be scored according to either the categorical or dimens ional approach. Structured and semi-structured interviews ba sed on DSM criteria represent the categorical approach to diagnosis in which a disorder is rated as either pr esent or absent. Behavior rating scales, in which a childs behavi or is rated along a continuum a nd then compared to a normative sample, represent the dimensional approach. Both approaches include their own strengths and weaknesses. Categorical approach es have been criticized for p oor specificity; some research suggests high false-positive rates across diagnostic categories. Cate gorical approaches have also been criticized for failing to account for overlappi ng symptoms, subsequently inflating rates of comorbidity (Achenbach, 1995). Dimensional scales are not without their own weaknesses, and have been criticized for poor sensitivity a nd for failing to provide important information regarding onset, duration, and severity of symp toms. Although the relative superiority of these two approaches has been a popular area of debate research does not clearly support one method over the other. Jensen and colleagues (1996) dire ctly compared these two approaches and found little difference between them. Both the dimens ional approach (i.e., Child Behavior Checklist; CBCL) and the categorical appro ach (i.e., DISC, Version 2.1) demonstrated low to moderate correlations with a variety of external validators, such as schoo l impairment, service utilization,

PAGE 22

22 and psychosocial and developmental risk factor s. The authors conclude d that, rather than debating the superior assessment te chnique, efforts should instead be directed towards validating diagnostic categories, because neither the cate gorical nor dimensional approach exhibited particularly strong relations with the external validators (Jensen et al., 1996). Currently, the gold standard of diagnosis involves a combina tion of categorical and dimensional approaches. The Eyberg Child Behavior Inventory (ECB I) is a widely used parent rating scale measuring disruptive behavior disorders in child ren ages 2 to 16 (Eyberg & Pincus, 1999). The validity and reliability of this measure is well-established (Boggs, Eyberg, & Reynolds, 1990; Funderburk, Eyberg, Rich, & Behar, 2003). The E CBI has also been correlated with diagnostic interview ratings specifically in a preschool sample (Doctoroff & Arnold, 2004). The ECBI was originally designed as a unidimensional measur e of disruptive behavior (Eyberg & Robinson, 1983), and some studies confirm that a one-factor model provides the best fit for the structure of the ECBI (Eyberg, 1992; Colvin, Eyberg, & Adam s, 1999; Gross et al., 2007). However, other researchers have found evidence for three sepa rate factors within the ECBI representing oppositional defiant behavior, conduct problems, and attention difficulties (Burns & Patterson, 1991; Burns & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). If a three factor structure is valid, then component scores may be useful in screening for the specific disorders. However, current data are mixed, and more research is warranted before th ese three factors are used as separate diagnostic indicators. No tably, the two recent studies that examined the factor structure of the ECBI specifically in preschool samples fo und conflicting results. In their predominantly Caucasian sample of 2 to 6-year-old childre n, Weis, Lovejoy, & Lundahl (2005) found support for the three-factor structure, whereas a more ethnically diverse sample of 2 to 4-year-olds resulted in a one factor model (Gross et al., 2007). Both of these studies i nvolved community,

PAGE 23

23 rather than clinical, samples. Future research examining the underlying factor structure of the ECBI therefore should focus on clinic-referred preschool populations. Developmental Modifications to Current DSM Standards The field is increas ing its focus on the e xpression of psychological disorders in preschoolers. Along with recognition that certain behaviors listed in the DSM are, in fact, found in normally developing children, comes the reali zation that a more appr opriate approach to diagnosis in children may involve developmentally modified criteria. So me researchers argue that criteria should be less stringent for younger children because it would be unrealistic for young children to demonstrate stable patterns of problem behaviors (Luby et al., 2002); others argue for more stringent criteria, given that fleeting problem behavior s can characterize normal development (Wakschlag, Leventhal, & Thomas, in press). Limited research has systematically explored developmental modifications to DSM cr iteria, yet this research is increasingly recognized as an important future direction to improve the developmental sensitivity of current diagnostic tools (Wakschlag, Leventhal, & Th omas, in press; Egger & Angold, 2004; Eyberg, Schuhmann, & Rey, 1998). One mechanism that may increase developm ental sensitivity invol ves the use of agegraded criteria. According to the DSM-IV, a six-year-old and a fourteen-year-old must demonstrate the same number of symptoms to warrant a diagnosis of ADHD. This standard seems somewhat counterintuitive given that the six-year-old will natura lly be more active and less attentive than the fourteen-year-old. The young child is therefore more likely to receive the diagnosis, even if the adolescent deviates more se verely from age-appropriate behaviors (Silk et al., 2000). The concept of age-graded criteria is particularly relevant in diagnosing preschool ADHD; the high rates of ADHD-lik e behaviors in typically deve loping children has led to the recommendation of longer symptom duration for very young child ren (Barkley, 2006). Thus, the

PAGE 24

24 field is beginning to consider the possibility of di fferent criteria for vari ous developmental levels. Although the DSM-IV advises its us ers to consider what is developmentally appropriate for a child of a certain age, this is a vague reco mmendation, and it is unclear how much credence clinicians give to this warning. A more formal age-graded system based on developmental norms would concretize the importance of the childs developmental leve l in assigning diagnoses. This concept is included in the Di agnostic and Statistical Manual for Primary Care (DSM-PC; Wolraich, Felice, & Drotar, 1996), developed to facilitate the identification of psychological disorders in the primary care setting. The fram ework of the DSM-PC reflects the idea that psychopathology occurs along a spectrum: the clinician can classify problematic child behaviors as developmental variations, problems, and disord ers, with each category denoting an increasing level of problem severity. The category of devel opmental variation allows the clinician to recognize that a child behavior viewed as problematic to the caregiver may reflect a normative variation of child development. Problems include those behaviors that may be distressing or causing some impairment in the childs overall functioning but do not warrant a formal DSM-IV diagnoses. The final category of disorders reflects those ca tegorized in the DSM-IV. The DSM-PC also includes concrete examples of de velopmental variations, problems, and disorders across different age groups. Thus, the DSM-PC recognizes that psychopathology occurs along a continuum, and that the level at which a behavior is reflective of psychopathology is influenced by child age (Wolraich, Felice, & Drotar, 1996). Research Questions This study was designed to exam ine the valid ity of DSM-IV symptoms for ODD, CD, and ADHD in preschoolers and explor e the validity of current dia gnostic assessment measures for this age group. The study also examined relations between DSM symptoms, symptom severity,

PAGE 25

25 and child age, to serve as a preliminary examin ation of age-graded criteria for the disruptive behavior disorders and ADHD. The first aim of this study was to examine th e validity of the DSM constructs of ODD, CD, and ADHD in a clinical presc hool sample. Specifically, we focused on the differentiation between the different subtypes of ADHD, as well as separate factors of ODD and CD. The majority of previous research supports two separate factors relating to the inattentive and hyperactive domains within ADHD, even within a preschool sample. Add itionally, a recent study found support for collapsing ODD and CD symptoms into one single construct of disruptive behavior. We therefore hypothesi zed that a three-factor model, with constructs representing ADHD-Inattentive, ADHD-Hyperactive, and ODD/CD w ould provide the best fit in a sample of clinic-referred preschoolers. The second specific aim further examined curre nt assessment techniques with a preschool sample. We compared the validity of three measures commonly used to assign DSM-IV symptoms and diagnoses. Specifically, symptom counts from the Diagnostic Interview Schedule for Children (DISC-IV; Schaffer et al., 2000), th e Young-Child version of the DISC (YC-DISC; Lucas, Fisher, & Luby, 2000), and the Early Ch ildhood Inventory (ECI-4; Gadow & Sprafkin, 2000) were correlated with other measures of ch ild disruptive behavior, including the Attention Problems and Aggressive Behavior subscales from the Child Beha vior Checklist (CBCL) and the ECBI. We hypothesized that the three assessment methods would demonstrate similar correlations with other measures of child behavior, suggesting similar validity across diagnostic interview type. To examine further a commonl y used dimensional measure of disruptive behavior in preschoolers, we c onducted a confirmatory factor anal ysis of the ECBI to determine whether a three factor model whic h maps onto the three externaliz ing behavior diagnoses can be

PAGE 26

26 confirmed in a diverse clinical sample of preschoolers. In light of the range of literature that supports a three-factor model of the ECBI, we hypothesized that the three factors would provide good fit in a diverse sample of clinic-referred preschoolers. The final set of specific aims was related to the idea that children of different ages may vary in the number of sympto ms required to warrant a clinic al diagnosis. Specifically, we examined whether number of symptoms would be a significant predictor of severity on the relevant subscale of the CBCL and whether age would moderate that relationship. We hypothesized that number of ADHD symptoms would predict attention problem scores on the CBCL and that age would moderate the relationshi p such that the positive relationship between ADHD symptom count and attention problems would be stronger for older children. We also hypothesized that number of ODD symptoms would predict scores on the externalizing subscale of the CBCL and that age would moderate the re lationship such that the relationship between ODD symptom count and externalizing behavior w ould be stronger for older children. Finally, we hypothesized that number of CD symptoms would predict scor es on the aggressive behavior subscale of the CBCL and that age would moderate the relationship such that the relationship between CD symptom count and aggressive beha vior would be stronger for older children.

PAGE 27

27 CHAPTER 2 METHODS Participants Participants were 207 children between the ag es of 3 and 6, inclusive, who received a diagnosis of ADHD, ODD, CD, or a com binati on of ADHD and ODD/CD. To increase sample size and to allow adequate power for the type of statistical analyses propo sed, participants were drawn from several studies, each of which is described below. All children were involved in research on the effectiveness of Parent-Child Interaction Therapy (PCI T). Exclusion criteria included immediate crisis requiring hospitaliz ation or out-of-home placement. Project GIFT A subset of participants was drawn from a st udy of PCIT and the maintenance of treatment gains (Project GIFT). One hundred boys and girls be tween the ages of 3 and 6 were enrolled in the study after being referred to a university psychology outpatient clinic for treatment of behavior problems; data were available from 99 of the original 100 participants. Families were referred for treatment by pediatricians, teachers, other mental health providers, or were selfreferred. Inclusion criteria for Pr oject GIFT stated that the child must meet DSM-IV diagnostic criteria for the diagnosis of ODD and must live with at least one parent able to participate in treatment. Children were excluded from the study if they scored below a standard score of 70 or if their parent scored below a standard score of 75 on cognitive screening measures, if they had a major sensory impairment such as blindness, or if they had been diagnosed with a pervasive developmental disorder. Children who were taking psychotropic medication were required to be stabilized on this medication for at le ast one month prior to entering the study.

PAGE 28

28 Project GANA Another sub set of participants was drawn from a study of the efficacy of PCIT with Mexican American families (Project GANA). Participants were 58 Mexican American families of 3to 7-year-old children w ith clinically significant behavior problems seen for treatment at a community mental health clinic; the three 7-ye ar-old children were dropped from the current study, leaving data from 55 participants. Families were referred from schools/teachers, hospitals or other clinics, social workers, or were self-referred. Families were eligible if they met the following criteria: (1) Parent identified the child as a Mexican American child between the ages of 3 and 7, (2) Child received a score above the clinical cutpoint on the Intensity Scale of the Eyberg Child Behavior Inventory (ECBI; Eyberg & Pincus, 1999), and (3) Neither parent nor child were participating simultaneously in any ot her psychosocial treatment targeting the childs behavior problems. Project SHAPE The final group of participants was d rawn from an ongoing study on the efficacy of PCIT in treating children with ADHD and comorbid ADHD + ODD (Project SHAPE). Data from 53 boys and girls between the ages of 4 and 6 were available for the current study. Families were referred for treatment by pediatricians, teachers, other mental health providers, or were selfreferred. Inclusion criteria for Pr oject SHAPE stated that the child must meet DSM-IV diagnostic criteria for the diagnosis of ADHD, either hyperactive/impulsive or combined type. Due to the need to establish the presence of ADHD sympto ms across two or more settings (APA; 1994), children were required to attend preschool or daycare for some part of each week. They were also required to live with a fema le primary caregiver able to participate in treatment. Along with a diagnosis of ADHD, some of the children a dditionally met criteria for ODD. Children were excluded from the study if they scored below a st andard score of 70 or if their parent scored

PAGE 29

29 below a standard score of 75 on cognitive screening measures, if they had a major sensory or mental impairment such as blindness, or if they had been diagnosed with a pervasive developmental disorder. One of the specific aims of Project SHAPE involved the investigation of the effectiveness of psychosocial, rather than pharmacological, treatment of ADHD. Therefore, children using psychotropic medications were excluded from this particular study. Measures Demographic and Background que stionnaire: Parent questionnaires were used to obtain descriptive information about the child and family including sex, age, race/ethnicity, occupation, education level, and medical history. Eyberg Child Behavior Inventory: (ECBI; Eyberg & Pincus, 1999). The ECBI is a 36item parent rating scale of disrup tive behavior. The ECBI Intensity Scale measures the frequency with which disruptive behavior occurs on a likert scale from 1 to 7, and the Problem Scale asks whether the parent finds the beha vior problematic for themselves on a yes-no scale. The Intensity and Problem Scales of the ECBI yield internal consistency coefficients of .95 and .93 (Colvin, Eyberg, & Adams, 1999), interrater (mother-f ather) reliability coefficients of .69 and .61 (Eisenstadt, McElreath, Eyberg, & McNeil, 1994), and test-retest reliability coefficients of .80 and .85 across 12 weeks and .75 and.75 across 10 months, respectively (Funderburk, Eyberg, Rich, and Behar, 2003). Studies documenting co nvergent, discriminant, and discriminative validity of the scales have been summarized in the test manual (Eyberg & Pincus, 1999). The alpha coefficient for the E CBI in this study was .89. Child Behavior Checklist: (CBCL; Achenbach, 1991; Achenbach, 1992; Achenbach & Rescorla, 2000; Achenbach & Rescorla, 2001). The CBCL is a commonly used measure of child behavior. The CBCL has different forms accord ing to child age and recently underwent a revision. Therefore, analysis invo lved four different versions of the CBCL. In Project GIFT,

PAGE 30

30 parents of children aged 4 through 6 completed the CBCL/4-18, which consists of 118 behaviorproblem items rated on a 3-point scale from (0) not true to (2) very true or often true. The items have been factor analyzed into two broadband scal es of internalizing and externalizing behavior problems, as well as narrow band scales assessing specific problem areas Mean one-week testretest reliability has been reported at .93 for the Externalizing scale, .91 for the Aggressive behaviors subscale, and .90 for the Atten tion Problems subscale (Achenbach, 1991). The CBCL/2-3, administered to parents of 3-year-old ch ildren in Project GIFT, is similar in format to the CBCL/4-18 but contains 99 items rated by the pa rent for frequency in the past 2 months on the 3-point scale. Test-retest reliability of the CBCL/2-3 has been reported at .90 for the externalizing scale and .91 for the Aggressive Behavior subscale over a three-week period (Koot et al., 1997). Notably, the CBCL/2-3 did not include an Attention Problems subscale. Both age forms of the CBCL underwent a rece nt revision; parents in Projects GANA and SHAPE therefore completed the appropriate age form of the most recent version. The CBCL/618 (Achenbach & Rescorla, 2001) is a 120-item pa rent rating scale designed to measure the frequency of childrens behavior or emo tional problems on a 3-point scale from (0) not true to (2) very true or often true One-week test-retest reliabilities have been reported as .92 for the Externalizing Scale, .90 for the Aggressive behavior subscale, and .92 for the Attention Problems subscale (Achenbach & Rescorla, 2001). The CBCL/1.5-5 (Achenbach & Rescorla, 2000) is comprised of 99 items describing childrens behavi oral and emotional problems, which are rated by parents on a 3-point scale from (0) not true to (2) very true or often true. One week test-retest reliabilities have been reported as .87 for the Externalizing scale, .87 for the Aggressive behavior subscale, and .78 for the Attention Problems subscale (Achenbach & Rescorla, 2000).

PAGE 31

31 Diagnostic Interview Schedule for Children: (DISC-IV; Schaffer et al., 2000).The DISC-IV is a structured diagnostic interview of child psychopathology based on Diagnostic and Statistical Manual of Mental Disorders criteria (DSM-IV; American Psychiatric Association, 1994). Test-retest reliabilities have been repo rted at .54 for ODD, .54 for CD, and .79 for ADHD on this instrument (Schaffer et al., 2000). Th e DISC-IV-P was administered to the childs primary caregiver at pre-treatment by a trained gr aduate research assist ant in Project GIFT. Young-Child DISC. (Lucas, Fisher, & Luby, 2000) Project SHAPE involved the use of a recent adaptation of the DISC interview for young children; the Young-Child DISC (YC-DISC). The YC-DISC is highly similar to the DISC-IV in its format and administration. However, the language has been modified to increase its se nsitivity, and the time fr ame for each symptom has been uniformly set at three months. The YC-D ISC was administered to the childs primary caregiver at pre-treatment by a trained graduate research assistant in Project SHAPE. Early Childhood Inventory: (ECI-4; Gadow & Sprafkin, 2000). The ECI-4 is a parent behavior rating scale that screens for DSM-IV emotional and behavioral disorders in children between 3 and 6 years old. Parents are asked to rate the frequency of specific behaviors on a 4point scale from (0) never, (1) sometimes, (2) of ten, and (3) very often. Symptoms generally map onto DSM-IV criteria; however, four symptoms of CD are not included: breaking and entering, missing curfew, running away, and truancy. Add itionally, the ECI-4 assesses for stealing behaviors only, with no distinction between w ith and without confr ontation. Thus, the CD module on the ECI-4 includes 10 symptoms, comp ared to the 15 listed in the DSM-IV. The wording of certain symptoms is altered to incr ease developmental sensitivity; for example, the ECI-4 asks the parent to rate their preschoolers preoccupation with sexua l activity rather than the DSM symptom of forcing sexual activity. A recent study reported internal consistency

PAGE 32

32 coefficients of .91 for ADHD-Inattentive, .90 for ADHD-Hyperactive, .93 for ODD, and .87 for CD. Study results also documented adequate convergent, discriminant, and discriminative validity of the ECI-4, particularly for th e ADHD and ODD scales (Sprafkin, Volpe, Gadow, Nolan, & Kelly, 2002). According to the scoring guidelines in the ECI-4 manual, behaviors rated as often or very often were included in the symp tom count for this study to help ensure that the behavior was truly reflective of a frequent and pervasive sy mptom (Gadow & Sprakfin 2000). The ECI-4 was completed by the childs primary caregiver at the pre-treatment assessment in Project GANA. Procedures Following the inf ormed consent process, familie s in all studies completed a pre-treatment assessment that included completion of the de mographic questionnaire, ECBI, and CBCL. The presence of ODD, CD, and ADHD symptoms was determined by the female primary caregivers responses on the DISC-IV interview in Projec t GIFT and the YC-DISC interview in Project SHAPE. In Project GANA, the female primary car egiver completed the ECI-4 to assess child symptoms of ODD, CD, and ADHD. Table 2-1 provides an overview of the assessment measures across the three different studies. To obtain access to data from all three datase ts, the investigator first sought IRB approval through the University of Florida Health Science Center. After re ceiving IRB approval, all data were first de-identified and then transmitted through data-encrypted files. Confirmatory Factor Analysis and Model Evaluation This study exam ined the fit of the factor structure implied by the DSM-IV in a preschool population. We therefore employed confirmatory f actor analysis (CFA) to test the current diagnostic system. A review of the literature provides support for the different disorders represented by the DSM-IV, partic ularly in school-aged and a dolescent samples. However,

PAGE 33

33 differentiation between the subtypes of ADHD and between the two disruptive behavior disorders, ODD and CD, has been more difficult to establish. We ran a series of three models: (a) The four-factor model inherent in the DS M-IV: ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD; (b) A combined ADHD model that colla pses symptoms across the inattentive and hyperactive domains to create three factors of ADHD, ODD, and CD, and (c) A disruptive behavior model that combines symptoms of ODD /CD into one factor, le aving three factors of: ADHD-Inattentive, ADHD-Hyperactive, and ODD/CD. The observed indicators involved in this an alysis are dichotomous and consequently violate assumptions of multivariate normalit y. Maximum likelihood estimation, commonly used in factor analysis, is therefore inappropriate in this study because it uses dichotomous indicators. Analyses were run using the Mplus program (version 4.2; Muthn & Muthn, 2006). The program addresses dichotomous variables through analysis of the tetrac horic correlation matrix, using Weighted Least Squares with Mean and Va riance Adjustment (WLSMV) to adjust for nonnormal data. To evaluate the three models, we examined the absolute fit (chi-square statistic), relative fit indices (Comparative Fit Index, CFI; Tucker-Lewis Index, TLI), the Root Mean Square Error of Approximation (RMSEA), which estimates discre pancy between model-implied and population covariance matrices, and the Weighted Root Mean Squared Residual (WRMR). Traditional guidelines for establishing good fit have been cr iticized as somewhat arbitrary (Marsh & Hau, 1996).Yu and Muthn (2002) evaluated fit indices w ith different types of data (i.e., continuous, nonnormal continuous, and dichotomous) in small, moderate, and large sample sizes to determine the cutoff values that would mi nimize both Type I and Type II errors. For dichotomous indicators and N > 250, their results indicated that least error would occur using the

PAGE 34

34 following guidelines: CFI > .95, TLI > .95, RMSEA < .06, and WRMR < 1.0 (Yu & Muthn, 2002). The fit indices of each of the three models were held to these guidelines. Additionally, adjusted 2 difference tests allowed direct comparison of the models. Specif ically, the DSM-IV model was compared to the combined ADHD mode l. The better-fitting of these two was then compared to the disruptive behavior disorder model. A significant difference in the 2 suggested that the more differentiated model fit the data bett er than the less different iated (Sterba, Egger, & Angold, 2007).

PAGE 35

35 Table 2-1. Overview of relevant measures across datasets Project GIFT ( N = 99) Project SHAPE ( N = 53) Project GANA ( N = 55) Measure of DSM-IV Symptoms DISC-IV-P X YC-DISC X ECI-4 X CBCL CBCL/2-3 and CBCL/4-18 X CBCL/1.5-5 and CBCL/6-18 X X ECBI X X X Demographic Questionnaire X X X Note. All measures are from the pre-treatment assessment.

PAGE 36

36 CHAPTER 3 RESULTS Participant Characteristics The 207 fa milies who participated in th e study included 147 boys (71%) and 60 girls (29%) with a mean age of 4.48 years ( SD = 1.01). Fifty six percent of children were Caucasian (by parent report); 25% Hispanic, 7% African-American, 0.5% Asian, 0.5% Native American, 6% biracial, and 4% identified as Other. Mean yearly inco me for the entire sample was $36, 278.83 ( SD = $29,779.31); however, income varied signifi cantly across the three studies. Families enrolled in Project SHAPE reported significantly higher income than families in both GIFT ( t (142) = 3.68, p < .001) and GANA ( t (105) = 5.60, p < .001). Additionally, families in Project GIFT reported higher income than families in Project GANA ( t (145) = 3.18, p < .001). Full information on participants is presented in Table 3.1. The clinical presentation of these three groups was also compared. Specifically, a between factor MANOVA was conducted with three depende nt variables: CBCL externalizing score, CBCL internalizing score, and the ECBI intens ity score. The GIFT sample was significantly larger than the other two groups Multicollinearity among the three dependent variables (CBCL Externalizing score, CBCL Internalizing score, a nd the ECBI intensity score) was examined and was within acceptable limits. The Box-M test for the homogeneity of variance-covariance matrices produced a significant result, F (12, 78260.34) = 5.33, p < .01. However, Levenes te st for the Externalizing score was the only significant result, F (2, 190) = 6.32, p < .01. Due to this violation of the assumption of homogeneity of variance, the relativel y conservative Pillais trace was used in the following MANOVA.

PAGE 37

37 The overall MANOVA reveal ed a significant main effect for study source, F (6, 378) = 10.94, p < .01, 2 = .148. To further understand the m odel, univariate ANOVAs were conducted for each dependent variable. The main effect of study source was signifi cant for Externalizing score, F (2, 190) = 7.32, p < .01, 2 = .071, and ECBI intensity score, F (2,190) = 13.65, p < .001, 2 = .126. However, the three groups did not differ in Intern alizing scores, F (2, 190) = 1.32, p = .271. Post hoc tests with Bonferroni correc tion were conducted to decompose further the significant main effect of st udy source. Children enrolled in GIFT demonstrated higher Externalizing scores than GANA participants ( p < .01), but GIFT and S HAPE participants did not differ in their Externalizing scores ( p = .06). Additionally, GANA participants demonstrated higher ECBI scores than both GIFT ( p < .01) and SHAPE ( p < .001) participants. GIFT participants also exhibited higher ECBI scor es as compared to SHAPE participants ( p < .05). Table 3-1 includes the means and standard devi ations for the three studies on each of the three dependent variables. Overall, the results suggest some differences in the clinical presentation of participants across different st udies. Children enrolled in GIFT demonstrated higher rates of externalizing behaviors than children in GANA. However, GANA participants demonstrated higher total ECBI scores than both the GANA and SHAPE participants. Thus, it seems that GANA children cannot uniformly be char acterized as less severe in their clinical presentation and rates of overall disruptive behavior. Symptom and Diagnostic Patterns Participants were drawn from three different studies, and DSM symptoms were assessed with a different assessment measure in each stud y. Symptom frequencies a nd rates of diagnosis are therefore presented by assessment and study s ource (See Tables 3-2 and 3-3). Notably, six of the fifteen DSM symptoms for CD occurred be tween zero and two percent of families in the

PAGE 38

38 entire sample. These items were: (a)forcing se xual activity, (b)fire-se tting, (c)breaking and entering, (d)missing curfew, (e)r unning away, and (f)truancy. Th ese six items were removed from further analysis. The ECI-4 does not include breaking and en tering, missing curfew, running away, or truancy, and given the low fr equency of these items when assessed in the DISC-IV and DISC-YC, their exclusion seem s warranted. Additionally, the ECI-4 assesses stealing behaviors only, while the DISC-IV and DISC-YC adhe re to DSM-IV criteria and differentiate between stealing wi th and without confrontation. As can be seen in Table 3-2, stealing with confrontation occurred at low fr equencies in both the GIFT (5%) and SHAPE (2%) samples. Further analysis therefore used a stealing composite symptom which included stealing with and without confrontation. Confirmatory Factor Analyses of DSM Symptoms CFA was us ed to determine which of three models provided the best fit for the 35 DSM symptoms of ADHD, ODD, and CD. The first model, referred to as the DSM-IV model, included four separate factors representing ADHD-Ina ttentive, ADHD-Hyperactive, ODD, and CD. The second model, the combined ADHD model, colla psed the inattentive and hyperactive ADHD symptoms and subsequently resulted in three factors of ADHD, ODD, and CD. The disruptive behavior model also included three factors bu t combined ODD and CD symptoms, resulting in the constructs of ADHD-Inattentive, ADHD-Hyperactive, and ODD/CD. Fit indices for each of three models were compared and are presented in Table 3-4. All three models resulted in a significant chi squa re; however, the ratio of the chi-square to the degrees of freedom did not exceed the cutoff value of 2 for any of the three models. Therefore, all three models provided an adequate fit fo r the data. However, as described above, the conservative cutoff scores recommended for binary indicators and a small sample size are: CFI > .95, TLI > .95, WRMR < 1.0, RMSEA < .06 (Yu & Muthn, 2002). Using these guidelines, the

PAGE 39

39 DSM four-factor model provided better fit for th e data than either the combined ADHD threefactor model or the disruptive behavior three-factor model. Chi square difference testing confirmed these results (Muthn & Muthn, 2005). Specifically, when the DSM model was compared to the combined ADHD model, the DS M model provided signif icantly better fit, 2 (2, N = 207) = 14.02, p < .001). Similarly, when the DSM model was compared to the disruptive behavior model, the DSM model pr ovided significantly better fit, 2 (2, N = 207) = 11.09, p < .001). Thus, results fail to find support for collapsing the ADHD symptoms across inattentive and hyperactive domains or for combining ODD and CD symptoms in this clinical sample of preschoolers. Table 3-5 shows that the loadings of the individual items onto their factors were generally high; the path diag ram is shown in Figure 1. The two ADHD factors were highly correlated, as were the ODD and CD factors. Correlations between the ADHD and ODD, as well as ADHD and CD, demonstrated a moderate rela tionship. Correlations among the six factors are included in Table 3-6. The CFA results provide support for the four-f actor structure curren tly implied with the DSM-IV. The modification indices provide further evaluation of model fit, as they provide the expected reduction in chi square if a paramete r were freely estimated. Notably, no modification indices were above the minimum value of ten (Muthn & Muthn, 1998-2006). Validity of the Three Assessment Methods To provide som e measure of validity of the three assessment measures included in this study, ADHD, ODD, and CD symptom counts from the DISC-IV, DISC-YC, and ECI-4 were correlated with the ECBI, CBCL A ttention Problems and CBCL Aggr essive Behavior subscales, as well as the broadband Externalizing scale from the CBCL (See Table 3-7). Examination of the correlation matrix revealed that symptom count s across different assessment methods correlated

PAGE 40

40 most highly with their conceptual counterpar t scale of the CBCL. Additionally, correlations across the three different assessment measures tended to be of similar magnitude. Confirmatory Factor Analysis of th e Eyberg Child Behavior Inventory CFA was us ed to confirm the three-factor so lution of the ECBI (B urns & Patterson, 1991; Burns & Patterson, 2000; Weis, Lovejoy, & Lundah l, 2005). Although previous studies have all reported factors reflecting oppositional behavior, conduct problems, and attention difficulties, there are some differences in previous results regarding which items load onto the factors. The current study will confirm the structure reported by Weis, Lovejoy, & Lundahl (2005) as theirs is the only study that found support for a three-factor model in a sample of preschoolers, and they sought to confirm the three-factor model previ ously supported in an older sample (Burns & Patterson, 2000). The factor structure described in their study was hi ghly similar to those presented in other research suppor ting a three-factor model and required several ECBI items to be dropped from the analysis, leaving 22 items in the model. The observed indicators involved in this anal ysis are ordinal data. Due to the clinical nature of the sample, ECBI items tended to be positively skewed. The MPlus program addresses ordinal variables through analysis of the polychoric correlation matrix; the current analysis also employed Weighted Least Squares with Mean and Variance Adjustment (WLSMV) to adjust for non-normal data. Results suggested that the th ree-factor solution was a relati vely poor fit for the data. The model resulted in a highly significant chi square and failed to meet the cut off scores proposed by Yu and Muthn (2002) for any of the fit indices (see Table 3-8 for fit indices and Table 3-9 for the factor structure). Examination of the modi fication indices suggested model improvement by allowing several items to load onto all three fact ors. Additionally, results suggested that several items from the oppositional behavior factor s hould be allowed to load onto the attention

PAGE 41

41 difficulties factor. Overall, findings failed to prov ide strong evidence for a three-factor model, and the modification indices were extensive a nd difficult to justify conceptually. Thus, an exploratory factor analysis seem ed the appropriate next step. All 36 ECBI items were therefore subjected to an exploratory factor analysis (EFA). Because other researchers found support for a thre e-factor model, we examined a one, two, and three-factor EFA of the ECBI items. Results from all three of these analyses are included in the Appendix. MPlus provides RMSR values for EFA an alyses; significant decreases with increasing number of factors suggest better fit of the more differentiated models. As can be seen in the Appendix, the decrease in the RMSR from the one to two-factor model was small, and the decrease from the RMSR from the two to three -factor was minimal. Perhaps more importantly, the two and three-factor models did not concep tually match the factors of oppositional behavior, attention difficulties, and conduct problems found in previous studies (Burns & Patterson, 1991; Burns & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). Indeed, it was difficult to interpret the twoand three-factor models. In the two factor model, the first factor incl uded most of the items and seemed to reflect a more general disruptive behavior factor that included a range of oppositional and attentionseeking behaviors. The second factor resemble d the conduct problems factor found in previous studies, reflecting overt aggressi on and provocation. This second factor was comprised of five items: (a) destroys toys and othe r objects, (b) teases or provokes other children, (c) verbally fights with sisters and brothers, (d) physically fights with sist ers and brothers, and (e) steals. Thus, the two-factor EFA seemed to reflect a general disruptive behavior construct with a smaller factor of severe conduct problems and a ggressive behaviors. The three-factor EFA was exceedingly difficult to understand conceptually. Th e first factor again included the majority of

PAGE 42

42 the ECBI items, representing oppos itional and attentionseeking behaviors. The only significant loadings onto the second factor demonstrat ed a negative relationship with the underlying construct. The third factor included five items seem ing to reflect a lack of respect for rules, such as having poor table manners and refusing to do chores when asked. None of the items demonstrated particularly strong loadings onto the third factor. The one-factor model was then subjected to CF A to examine the fit of the model. Items with factor loadings below .4 on the EFA were dropped from the analysis. This included many items that failed to demonstrate significant loadings onto any factors in previous studies, such as whining and having poor table manners. The CFA also excluded several items that have previously been included on the conduct problems fact or in previous studies, as these items seemed to be reflective of a more severe beha vior than most other ECBI items. Thus, 25 items were subjected to a one-factor CFA. All meas ures of model fit im proved and the results suggested good fit according to the TLI. The CFI va lue was marginal according to the guidelines proposed by Yu and Muthn (2002), although comp arable to the value reported in previous studies (Weis, Lovejoy, & Lunda hl, 2005), and the WRMR was also marginal. However, the RMSEA continued to suggest poor fit (see Table 3-8). The factor structure of the one-factor CFA is presented in Table 3-10. Exploration of Age-Graded Criteria As a prelim inary examination of the possibili ty of age-graded criteria, a series of hierarchical regressions were conducted in which scores on the corresponding scale of the CBCL were regressed upon total number of symptoms, child age, and an interact ion between those two variables. To test the hypothesized moderator effects of age, we evaluated the significance of the change in R2 after the addition of the interaction term.

PAGE 43

43 The first model examined the predictive power of ADHD symptoms and child age in inattention scores. Specifically, scores from th e attention scale of the CBCL were regressed upon total number of ADHD symptoms, child age, and an interaction term between ADHD symptoms and child age to test the hypothesis that age would modera te the relationship between ADHD symptom count and severity of inattention. Tota l number of ADHD symptoms was entered in the first block and acted as a significant predictor, F( 1, 169) = 81.28, p < .001, R2 = .325. Child age was entered into the model and did not a dd to the predictive power of the model, R2 = .001, nor was the interaction term between numb er of symptoms and age significant, R2 = .001. Thus, number of ADHD symptoms was a significant predictor of attention scores on the CBCL. However, the current analysis failed to support the hypothesis that age would moderate that relationship between symptom count and severity of inattention. The second model examined the predictive power of ODD symptoms and child age in externalizing behavior scores. In this analysis, scores from th e externalizing scale of the CBCL were regressed upon total number of ODD symptoms, child age, and an inte raction term between ODD symptoms and child age to test the hypothesi s that age would moderate the relationship between ODD symptom count and severity of externalizing behavior. Total number of ODD symptoms was entered in the first block and acted as a significant predictor, F( 1, 198) = 63.04, p < .001, R2 = .241. Child age was entered into the model and did not a dd to the predictive power of the model, R2 = .011, nor was the interaction term between number of symptoms and age significant, R2 = .001. Thus, number of ODD symptoms was a significan t predictor of externalizing scores on the CBCL. However, the analysis failed to support the hypothesis that age would moderate the relationship between sy mptom count and severi ty of externalizing behavior.

PAGE 44

44 The third model examined the predictive power of CD symptoms and child age in rates of aggressive behavior. In this analysis, scores from the aggr ession scale of the CBCL were regressed upon total number of CD symptoms, child age, and an interaction term between CD symptoms and child age to test the hypothesis th at age would moderate the relationship between CD symptom count and severity of externalizin g behavior. Total number of CD symptoms was entered in the first block and act ed as a significant predictor, F( 1, 199) = 51.51, p < .001, R2 = .206. Child age was entered into th e model and did not add to the predictive power of the model, R2 = .002, nor was the interaction term between number of symptoms and age significant, R2 = .001. Thus, number of CD symptoms was a signi ficant predictor of aggression scores on the CBCL. However, the analysis failed to support the hypothesis that age would moderate that relationship between symptom count and severity of aggressive behavior. In sum, this study failed to find support for the hypothesis that ag e would significantly moderate the relationship between symptom count and the severity of problem behaviors. However, it should be noted that the sample re presented a limited age ra nge, and thus it may have been difficult to find differences by age.

PAGE 45

45 Table 3.1. Participant characteristics. Total Sample GIFT SHAPE GANA Age (years) M = 4.48 SD = 1.01 M = 4.38 SD = 1.09 M = 4.85 SD = 0.75 M = 4.33 SD = 1.31 Gender 71% male 69% male 74% male 80% male Ethnicity Caucasian 56% 76% 76% 7% African American 7% 8% 14% 0% Hispanic 25% 3% 1% 91% Asian 0.5% 1% 0% 0% Native American 0.5% 0% 4% 0% Biracial/Other 10% 12% 5% 2% Family yearly income M = 36,278.83 SD = 29,779.31 M = 34,308.52 SD = 25,302.84 M = 54,251.54 SD = 39,681.04 M = 22,582.25 SD = 13,299.15 CBCL Externalizing T Score M = 70.57 SD = 8.95 M = 72.82 SD = 6.91 M = 69.00 SD = 9.05 M = 67.49 SD = 11.07 CBCL Internalizing T score M = 61.91 SD = 9.91 M = 62.54 SD = 9.58 M = 59.83 SD = 9.95 M = 62.67 SD = 9.63 ECBI Intensity Raw score M = 169.56 SD= 28.41 M = 168.34 SD= 25.06 M = 154.08 SD= 29.68 M= 182.58 SD= 26.95

PAGE 46

46 Table 3-2. Individual symptom percentages across the DISC-IV, YC-DISC, and ECI-4 Symptom DISC-IV (GIFT; n = 99) YC-DISC (SHAPE; n = 53) ECI-4 (GANA; n = 55) ADHD-Inattentive Careless mistakes 43 67 24 Short attention span 55 87 35 Doesnt listen 93 93 46 Fails to finish tasks 70 91 51 Disorganized 53 76 40 Avoids concentrating 52 82 27 Loses items 44 65 36 Distractible 70 89 64 Forgetful 41 72 33 ADHD-Hyperactive Fidgets 74 91 66 Leaves Seat 81 85 66 Runs/Climbs 90 91 76 Too Loud 63 67 55 Driven by motor 77 85 66 Excessive talking 71 78 56 Blurts out answers 48 59 31 Difficulty waiting turn 78 85 58 Interrupts 90 94 49 ODD Loses temper 95 74 36 Argues 96 80 51 Defiant 95 90 75 Annoys others 76 63 56 Blames others 51 56 49 Easily annoyed 75 59 47 Angry 84 70 40 Spiteful 61 39 43 CD Bullying 42 22 11 Fighting 27 24 16 Use of weapon 26 15 4 Cruel to people 24 9 9 Cruel to animals 27 15 5 Stole with confrontation 5 2 ** Forced sex* 0 0 2 Firesetting* 1 0 2 Destruction of property 50 50 10 Breaking and entering* 0 0 ** Lying 23 33 15 Stole without confrontation 29 31 ** Missed curfew* 0 0 ** Runaway* 0 0 ** Truancy* 0 0 ** Stealing composite*** 29 31 2 *Dropped from further analyses due to low frequency: **Not asked on ECI-4; ***Combination of stealing with and without confrontation

PAGE 47

47 Table 3-3. Percentage of diagnos is across assessment method DISC-IV (GIFT; N = 99) YC-DISC (SHAPE; N = 53) ECI (GANA; N = 55) ADHDInattentive 1 5 4 ADHDHyperactive 24 11 32 ADHDCombined 46 76 22 ODD 91 79 56 CD 42 29 9

PAGE 48

48 Table 3-4. Absolute, relative, and component model fit for three DSM-IV CFA models Model 2 DF p CFI TLI RMSEA WRMR DSM-IV (4 factor) 170.66 82 < .001 .95 .96 .06 1.0 Combined ADHD (3 factor) 177.05 88 < .001 .93 .95 .07 1.1 Disruptive behavior (3 factor) 174.90 89 < .001 .93 .95 .07 1.1

PAGE 49

49 Table 3-5. Four-factor solution of DSM-IV symptoms ADHD-I ADHD-H ODD CD Symptom ADHD-I Careless mistakes .74 Short attention span .90 Doesnt listen .89 Fails to finish tasks .86 Disorganized .73 Avoids concentrating .80 Loses items .74 Distractible .90 Forgetful .76 ADHD-H Fidgets .81 Leaves Seat .88 Runs/Climbs .88 Too Loud .82 Driven by motor .90 Excessive talking .68 Blurts out answers .67 Difficulty waiting turn .69 Interrupts .72 ODD Loses temper .67 Argues .84 Defiant .89 Annoys others .72 Blames others .71 Easily annoyed .68 Angry .74 Spiteful .65 C D Bullying .79 Fighting .63 Use of weapon .77 Cruel to people .68 Cruel to animals .46 Destruction of property .71 Lying .73 Stealing composite .71

PAGE 50

50 Table 3-6. Correlations among DSM-IV factors ADHDInattentive ADHDHyperactive ODD CD ADHD-I 1.00 ADHD-H .81 1.00 ODD .49 .54 1.00 CD .35 .39 .74 1.00

PAGE 51

51 Table 3-7. Pearson correlations between symptom counts and CBCL and ECBI scores ADHD symptoms ODD symptoms CD symptoms DISCIV YCDISC ECI4 DISCIV YCDISC ECI4 DISCIV YCDISC ECI4 Attention Scale .49** .43** .52** .02 .09 .32* .08 .02 .26* Aggression Scale .29** .20 .33* .37** .41** .49**.45** .46** .34* Externalizing Scale .31** .24 .43** .37** .41** .51**.41** .43** .35** ECBI intensity score .42** .47** .30* .27** .61** .19 .42** .51** .33* *p< .05, **p<.01.

PAGE 52

52 Table 3-8. Absolute, relative, and co mponent model fit for ECBI models Model 2 DF P CFI TLI RMSEA WRMR Three-factor CFA 805.24 53 < .0001 .75 .86 .27 2.28 One-factor CFA 356.84 66 <.0001 .92 .97 .15 1.2

PAGE 53

53 Table 3-9. The CFA of the thr ee-factor model of the ECBI ECBI item Factor One Factor Two Factor Three 12. Gets angry when doesnt get own way .94 10. Acts defiant when told to do something .93 8. Does not obey house rules on own .90 14. Sasses adults .82 9. Refuses to obey until threatened with punishment .76 13. Has temper tantrums .50 11. Argues with parents about rules .41 5. Refuses to do chores when asked .47 7. Refuses to go to bed on time .45 17. Yells or screams .36 30. Is easily distracted .96 34. Has difficulty concentrating on one thing .90 32. Fails to finish task or projects .88 31. Has short attention span .56 24. Verbally fights with friends own age .87 22. Lies .85 26. Physically fights with friends own age .84 25. Verbally fights with si sters and brothers .55 27. Physically fights with sisters and brothers .53 19. Destroys toys and other objects .21 23. Teases or provokes other children .15 21. Steals -.47

PAGE 54

54 Table 3-10. One-factor CFA of the ECBI ECBI item 10. Acts defiant when told to do something .89 8. Does not obey house rules on own .89 13. Has temper tantrums .89 32. Fails to finish task or projects .89 6. Slow in getting ready for bed .83 30. Is easily distracted .83 34. Has difficulty concentrating on one thing .83 36. Wets the bed .83 2. Dawdles or lingers at mealtime .80 22. Lies .79 4. Refuses to eat food presented .77 18. Hits parents .76 26. Physically fights with friends own age .75 9. Refuses to obey until thre atened with punishment .74 28. Constantly seeks attention .73 24. Verbally fights with friends own age .72 20. Is careless with toys and other objects .61 31. Has short attention span .53 7. Refuses to go to bed on time .51 5. Refuses to do chores when asked .45 14. Sasses adults .44 16. Cries easily .41 12. Gets angry when doesnt get own way .36 35. Is overactive or restless .06 29. Interrupts .01

PAGE 55

55 Figure 3-1. Path diagram of fou r-factor solution of DSM-IV sy mptoms of ADHD-Inattentive, ADHD-Hyperactive, ODD, and CD. ADHD Inattentive Careless mistakes Short attention Difficulty listening Fails to finish Avoids concentrating Loses items Distracted Disorganized Forgetful Too Loud Runs/ Climbs On the go Leaves Seat Talkative Difficulty waiting Interrupts Fidgets Blurts out ADHD Hyperactive .82 .68 .69 .67 .72 .90 .88 ODD Annoys others Defiant Argues Blames others Easily annoyed Spiteful Loses temper .67 .84 .89 .72 .71 CD .81 .54 Angry .68 .74 .65 .74 .46 .71 .73 .71 Stealing Fighting Lying Destroys property Cruel to animals Cruel to people Use of weapon Bullying .88 .81 .68 .77 .63 .79 .35 .76 .90 .74 .80 .73 .86 .89 .90 .74

PAGE 56

56 CHAPTER 4 DISCUSSION The present study provided im portant inform ation regarding the validity of DSM-IV symptoms of ODD, CD, and ADHD in a relatively diverse, clinical preschool sample. Findings also provide support for the validity of current assessment measures for preschoolers. Results can be used to inform future diagnostic methodol ogy and assessment techniques in this population. Ultimately, findings from the current study can support effective assessment and treatment of psychopathology in preschoolers. Validity of the DSM-IV for Preschoolers Results suggest that the four constructs of ODD, CD, ADHD -Inattentive and ADHDHyperactive implied by the DSM-IV are valid in younger children. Even when using the relatively conservative standards outlined by Yu and Muthn (2002), indices suggested good fit. The factor loadings of each symptom onto its corresponding underlying construct were generally high. None fell beneath the commonly used leve l of .4 (Floyd & Widaman, 1995) and only one symptom had a factor loading below .6. Results generally support the growing body of research supporting the structure of the DSM-IV in older children (Lahey et al., 2004; Lahey et al., 2008) and adds to the limited body of research ex amining the validity of DSM-IV nosology in preschoolers. The four-factor DSM-IV model provided the be st fit in this sample of clinic-referred preschoolers when compared to a three-factor model that combined the in attentive and hyperactive domains into a single ADHD factor, with separate constructs of ODD and CD. The DSM-IV model also proved superior when compared to a three-factor model of ADHDInattentive, ADHD-Hyperactive, and a disruptiv e behavior factor that collapsed ODD and CD symptoms. The two alternate models resulted in adequate fit according to certain fit indices;

PAGE 57

57 however, inferential chi-square testing allowed direct comparis on of the three models and found support for the four-factor model. This is an impo rtant finding that adds to the limited research on the validity of separate ODD and CD factors in younger children. Two previous studies found similar results in large, nonclin ical samples of school-aged childr en and adolescents (Lahey et al., 2008; Burns et al., 1997). However, a recent study of DSM symptoms in 307 preschoolers found support for collapsing ODD and CD symptoms into a single disruptive behavior factor (Sterba, Egger, & Angold, 2007). Thus, the present study is the first to find support for the differentiation of CD from ODD sp ecifically in a preschool populati on. The clinical nature of the sample may explain the discrepancy from th e results by Sterba, Egger, and Angold (2007). Given the extreme nature of many of the CD criteria, it may be di fficult to establish a clear CD factor in a community sample, particularly in younger children. Althoug h more research is needed, the present study adds to a small but gr owing body of literature that supports the validity of CD in preschoolers (Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998; Keenan & Wakschlag, 2000; Wilens et al., 2002). Although results of this study s upport a separate CD factor, it is notable that several CD symptoms were dropped from analyses due to lo w frequency: forcing sexu al activity, fire-setting, breaking and entering, missing curf ew, running away, and truancy. Th is is hardly surprising in light of preschoolers developmental abilities an d limited exposure to these types of situations. However, the current DSM-IV provides no developm ental modifications to current CD criteria. Amidst claims that the diagnosis of CD is i rrelevant to preschoolers, it seems important to increase the developmental sensitivity of CD criteria for this age group. The Task Force on Research Diagnostic Criteria for Infancy and Preschool (RDC-PF; 2003) has dropped breaking and entering, breaking curfew, r unning away, and truancy; thus, it seems likely that the DSM-V

PAGE 58

58 will exclude these symptoms in the assessment of preschoolers. Further research is needed on potential modifications to CD criteria to in crease developmental sensitivity while also maintaining the clinical essence of the symptoms (Egger & Angold, 2006). Results also confirm the validity of two sepa rate factors of inatte ntion and hyperactivity within the ADHD domain, specifically in a preschool sample. A substantial body of research supports this distinction in older children a nd adolescents (Glutting, Youngstrom, & Watkins, 2005; Willoughby, Curran, Costello, & Angold, 2000). Some research has suggested that a single ADHD factor is most appropriate for younger child ren (Bauermeister, 1992). Yet results from our study replicate recent findings by Sterba, Egge r, and Angold (2007) and add to the validity of the DSM-IV model of ADHD in young children. Nevertheless, furt her study is needed. Given the necessity of ADHD symptoms in the school setting for diagnosis, it will be particularly important to validate the two domains of inattent ion and hyperactivity with teacher ratings. Data from the Preschoolers with Attention Deficit/H yperactivity Disorder Treatment Study (PATS) found conflicting results between parent and t eacher ratings of ADHD symptoms. Specifically, the two-factor model of inattention and hyperact ivity provided adequate fit when applied to parent responses on a behavior rating scale of ADHD symptoms, but neither a one-factor nor two-factor model provided adequate fit for teacher ratings (Hardy et al., 2007). The authors note that the use of a behavior rating scale may not have provided the most accurate representation of DSM symptoms; thus, the failure of the study to find support for the two-factor model may reflect a methodological artifact. However, it also may be that the structure of ADHD symptoms is different when applied to classr oom behavior (Hardy et al., 2007). This study focused on the validity of specifi c diagnostic categories in preschoolers. The validation of psychopathology for preschoolers is an important step. However, more research is

PAGE 59

59 needed to demonstrate the clinical utility of these specific disorders in young children. As noted by Huffine (2002), treatment recommendations for CD are virtually identic al to those for ODD, leading some to question the utility of two separa te disorders. This argument seems particularly relevant to preschoolers because the more inte nsive, multimodal treatment programs sometimes used to treat CD, such as Multisystemic Th erapy (MST; Henggeler & Borduin, 1990) would be inappropriate for preschoolers. A similar argument may be relevant when discussing the clinical diagnosis of ADHD. Treatme nt recommendations for ADHD in school-aged children and adolescents frequently involve a medication trial and more individualized behavioral interventions, such as organiza tional skills training. Results from the Preschoolers with Attention Deficit/Hyperactivity Disorder study (PATS) suggest that stimulant medications can be effective in this age group (Greenhill et al., 2006; Vitiello et al., 2007). However, the efficacy of stimulant medication tends to be less (Gr eenhill et al., 2006) and the side effects more severe when compared to school aged children and adolescent s (Wigal et al., 2006). Notably, no medication is currently approved by the Food and Drug Administ ration (FDA) for the treatment of ADHD in preschool aged children. A recent study examined ma ternal perceptions of me ntal health services for their preschool-aged children. The majority of mothers surveyed expressed interest in receiving assistance in managing their childs be havior problems; however, most also indicated that they would not be intere sted in pursuing medication (Har wood, OBrien, Carter, & Eyberg, under review). Due to the controversy surrounding medication use in this age group, and parental concerns regarding the safety of medicati on for young children, first line treatment recommendations for preschoolers with ADHD ofte n involve behavioral interventions such as parent training and contingenc y management, the same treatments used for ODD/CD in young children (dosReis, et al., 2003; Rushton, Fant, & Clark, 2004; Verduin, Abikoff, & Kurtz, 2008).

PAGE 60

60 The efficacy of these types of be havioral interventions, and specifi cally parent training programs, is well-established for the tr eatment of ODD/CD in young ch ildren (Eyberg, Nelson, & Boggs, 2008). The results of our study do not support the lesser differentiation in the presentation of psychopathology in preschoolers. However, treatment of externalizing disorders, specifically for this population, may involve many of the same co mponents. Due to their developmental level and the significance of the parent-child relations hip at this young age, a range of presenting problems in preschoolers may be improved with comprehensive parent training programs that apply behavioral strategies to a broad spectru m of child problem behaviors (Chase & Eyberg, 2008). Validity of Assessment Measures Results also provided validity for the DI SC-IV, YC-DISC, and ECI-4 in a clinical preschool sample. Regardless of the diagnostic approach, sympto ms across diagnostic categories correlated most highly with their conceptual counterpart on the Child Behavior Checklist (CBCL). These correlations tended to be of similar magnitude for each assessment method. Additionally, symptom counts across the three studies generally demonstrated similar correlations with the ECBI intensity score. Some variability was seen in the correlations between the ECBI and ODD symptom counts. Specifically, ODD symptoms as assessed by the DISC-YC demonstrated the highest correla tion with the ECBI. ODD symptoms from the DISC-IV resulted in a small but significant correl ation with the ECBI, while ODD sy mptoms from the ECI-4 were not significantly correlated with the ECBI intensity score. However, this result is likely related to the restricted range of ODD symptoms in Project GIFT and ECBI scores in Project GANA. As noted previously, the inclusion criteria for Project GIFT n ecessitated higher rates of ODD symptoms and the inclusion criteria for Projec t GANA resulted in higher ECBI scores. Thus, the weaker relationship between DISC-IV ODD symptoms and the ECBI and the lack of relationship

PAGE 61

61 between ECI-4 ODD symptoms most likely reflects a methodological weakness of our study rather than a comment on the validity of either the DISC-IV or ECI-4. Overall, this study provided evidence of c onvergent validity for the DISC-IV, YC-DISC, and ECI-4 in a preschool sample. This is among the first studies to assess the validity of the DISC-IV specifically in this age group and th e validity of the newly developed YC-DISC. Results suggest the two interviews are highly simila r in their relationships with other measures of child behavior. Additionally, symptom counts from the ECI-4, although lower than both the DISC-IV and YC-DISC, demonstrated significant co rrelations with the a ppropriate subscale of the CBCL. The ECI-4 is a relatively new measure, and this study provides important information regarding its validity as a measur e of DSM-IV disorders. Future research on preschool diagnosis should address additional indicators of validity. A particularly impor tant direction is the study of predictive validity of assessment measures for this age group. Although ADHD and the disruptive behavior disorders diagnosed in early childhood have been shown to persist into later childhood and adolescence, symptoms remit in approximately half of children diagnosed in the preschool years (Campbell, 2002). Instruments sp ecifically designed for this age group may relate to improved predictive validity. For exam ple, the YC-DISC has altered the time duration of each symptom and asks about the presence of a behavior over a three-month time span. Longitudinal research will dete rmine whether modifications such as these allow better identification of those children who will continue to exhibit symptoms into later childhood and adolescence. Additionally, modifying the la nguage of certain symptoms to increase developmental sensitivity may improve face valid ity with this population, as well as overall consumer satisfaction.

PAGE 62

62 Factor Analysis of the Eyberg Child Behavior Inventory Our study failed to support the three-factor structure of the ECBI re ported in the literature (Burns & Patterson, 1991; Burn s & Patterson, 2000; Weis, L ovejoy, & Lundahl, 2005). The model with factors of oppositional behavior, in attention, and conduct problems yielded a highly significant chi square and did not indicate good fit acco rding to any of the fit indices. Most ECBI items demonstrated significant lo adings onto one factor of child behavior problems, consistent with previous studies reporting a one-factor structure of the ECBI (Eyberg, 1992; Colvin, Adams, & Eyberg, 1999; Gross et al, 2007). Several different reasons may e xplain the failure to replicate the three-factor model. Notably, this is the first study to factor analyze the ECBI with a clinical sample of preschoolers. The two studies that have examin ed the factor structure of the ECBI specifically with preschoolers both used a community sample, and those two studies led to conflicting results (Weis, Lovejoy, & Lundahl, 2005; Gross et al., 2007). Certainly, the use of a clinical sample limits the generalizability of the current results. However, it seems particularly important to replic ate the three-factor stru cture in clinic-referred children. The importance of using a clinical sample is highlighted by suggestions in the literature that component scores from the three ECBI factor s could be used for diagnostic purposes (Burns & Patterson, 1991; Burns & Patterson, 2000; We is, Lovejoy, & Lundahl, 2005). Weis, Lovejoy, & Lundahl (2005) found support for the three-factor model in a community sample of preschoolaged children. However, the component scores from the individual factors were largely unable to differentiate among the three differe nt disorders in a clinical sa mple of children aged 4 to 6 years. Specifically, the Oppos itional Defiant component did not distinguish children with oppositional behavior from children with attention problems, and the Inattentive component did not differentiate inattentive children from oppos itional children. Indeed, the authors found that the full ECBI intensity scale was more useful in differentiating these disorders than the

PAGE 63

63 individual component scores. They suggested that the failure of the component scores to differentiate among the different disorders may refl ect the lack of behavi oral differentiation in childrens symptom expression (W eis, Lovejoy, & Lundahl, 2005). Ho wever, our findings from the DSM-IV confirmatory factor analysis ar gue against a lesser differentiation of symptom presentation in younger children. Results from the current study suggest that the ECBI component scores fail to repr esent the distinct disorders in the same way that they are conceptualized in the DSM-IV. Notably, the ECBI was developed prior to the development of the DSM-III and was developed as a unidimensional measure of ch ild disruptive behavior. Although several ECBI items appear to map onto DSM-IV criteria, some may not be clear indi cators of DSM symptoms, particularly in a clinical pres chool sample. For example, the ECBI item has short attention span appears very similar to the DSM-IV ADH D inattentive symptom, often has difficulty sustaining attention in tasks or play activities. However, ex pectations for a preschoolers attention span are generally lowe r as compared to school-aged ch ildren and adolescents (Silk et al., 2000). A short attention span in a 4 year-old may be viewed as problematic by the childs parent or caregiver; however, it also may be less indicative of pervasive attention difficulties in a younger child. As a child matures and expectations for the ability to attend increase, short attention span may be more likely to reflect a significant problem with attention, and thus a stronger indicator of the constr uct of ADHD-Inattentive Type. The wording of DSM-IV items may be particularly important when assessing fo r these specific symptoms in preschoolers. Indeed, when discussing developmental modifications of current DSM-IV criteria for preschoolers, researchers have highlighted th e importance of developmentally appropriate modifications that maintain th e clinical meaning of a par ticular symptom (Egger & Angold,

PAGE 64

64 2004). Differentiating components of the ECBI may be more relevant for older children, when certain behaviors are inherently more problematic and more indicative of specific disorders. Exploration of Age-Graded Criteria Our study failed to find support for the use of ag e-graded criteria in assigning diagnoses am ong 3to 6-year-olds. Symp tom counts of ADHD, ODD, and CD served as significant predictors of behavior severity, as measured by the corresponding scale on the CBCL. However, age failed to moderate that re lationship. The failure to find su pport for our hypothesis is likely related to the truncated age range of the sample. The study of agerelated criteria would be more appropriate across larger age ra nges. In their comprehensive re view of the presentation and epidemiology of psychopathology in preschooler s, Egger and Angold (2006) systematically evaluated the DSM qualification th at behaviors included in ODD cr iteria must occur often in order to represent a true symptom. ODD sympto ms naturally occur at a higher frequency in preschool populations as compared to school-a ged children. Thus, applying the cutoff criteria previously used for school-aged children (above the 90th percentile) for pres choolers would have overestimated rates of ODD. This example highl ights the importance of using established developmental norms in determining the pres ence of a symptom (Egger & Angold, 2006), and supports further study of the use of developmentally modified criteria for different age groups. Study Limitations In an effort to increase sample size and the dem ographic diversity of the study sample, this study included participants from three differe nt studies. Although all studies examined the efficacy of Parent-Child Interaction Therapy (PCIT) in the tr eatment of disruptive behavior disorders and/or ADHD in preschoole rs, each study varied in its in clusion criteria. Consequently, participants differed in diagnostic presentati on and demographic characteristics, introducing variability that may have affected the results. Notably, to provide th e sample size necessary,

PAGE 65

65 symptoms entered into the factor analysis were assessed with different instruments. However, all items mapped onto DSM-IV criteria. Indeed, the confirmation of the DSM-IV structure despite the methodological difference may argue for the stre ngth of the DSM-IV model. Also of note, we found few differences in the convergent valid ity of the different measures, suggesting that they may provide similarly accu rate diagnostic information. Drawing participants from differe nt studies also introduced vari ability in the versions used for study measures. Specifically, children from Proj ect GIFT were assessed with an older version of the CBCL (Achenbach; 1991; Achenbach, 1992) than children in Projects SHAPE and GANA (Achenbach; 2000; Achenbach; 2001). Although these ve rsions were highly similar, particularly in the subscales of interest for this study, ther e were some differences across measures. Perhaps most importantly, the 1992 version of the CBCL fo r 2 and 3-year-old children did not include an Attention Problems subscale. Thus, 3 year-old chil dren from Project GIFT were excluded from the analyses involving this measure. However, this represented a re latively small group of children relative to the entire sample, and it s eems unlikely that their inclusion would have significantly changed the results. A relatively small sample size is another lim itation of the current study. Our sample of 207 was enough to meet the recommendation of 5 cases per item when conducting factor analysis (Floyd & Widaman, 1995). However, larger sample sizes are considered optimal for factor analysis. The sample size also prevented crossvalidation of either the DSM-IV or ECBI model by age, race, or gender. Overall, previous research with older children and adolescents supports the factors implied by the DSM-IV across age groups and gender (Lahey et al., 2008). However, it will be important to confirm f actor invariance in younger children.

PAGE 66

66 The sample size may have affected the clarity of our results, particul arly for the factor analysis of the ECBI. Our findings resulted in an unclear factor structure, which was inconsistent with some earlier research in this area (Burns & Patterson, 1991; Burns & Patterson, 2000; Weis, Lovejoy, & Lundahl, 2005). However, it is notable that our result s are more consistent with a recent factor analytic study of the ECBI in a sample of 682 preschoolers (Gross et al., 2007). Overall, results in this area of the literature are conflicted, highlighting the need for future studies that examine the factor structure of the ECBI in larger samples. Despite these limitations, findings provide va luable information on current diagnostic standards and measures for the preschool population. The clinical nature of the sample is a particular strength of the current study. Much of the literature examini ng issues of diagnostic validity for this age group often involves comm unity samples of non-referred children. Although results from this study may not be generalized to non-clinical children, it seems particularly important to focus on the children that exhibit th e specific disorders. This may be particularly important in examining diagnosis of CD in pr eschoolers, as our resu lts found support for two separate factors of ODD and CD in contrast to findings by Sterba, Egger, & Angold (2007). Their sample of 307 preschoolers recruited from a primary care clinic may have been less representative of children with externalizing diso rders. The severity of the CD criteria may make it difficult to establish a clear CD construct in non-clinical samples, particularly in younger children, for whom certain behaviors are deve lopmentally irrelevant to diagnosis. The discrepancy of these results hi ghlights the importance of future research that examines CD symptoms in both clinical and nonclinical samples. Specifically, furthe r study should involve a broader range of clinical severi ty, perhaps including children from psychiatric inpatient units or day treatment facilities.

PAGE 67

67 Future Directions This study adds to the lim ited literature on the validity of current diagnostic standards and assessment methods for preschoolers. Overall, results suggest that DSM-IV categories and common assessment measures are appropriate for this age group. Future research should continue to examine the validity and reliability of preschool diagnosis. An important area of research includes the predictive validity of preschool diagnosis. A lthough a growing body of research supports DSM disorders in presc hoolers, it is unknown which preschoolers will continue to demonstrate sympto ms into childhood and adolescence. Longitudinal research is needed on long-term outcomes of children diagnose d with disruptive behavior disorders in early childhood. Further research in th is area will increase our unders tanding of the factors that distinguish developmental variations from true psychopathology and promote the continuing evolution of the DSM.

PAGE 68

68 APPENDIX STATISTICAL ANALYSES NOT INCLUDED IN MAIN TEXT Table A-1. One-factor EFA of the Eyberg Child Behavior Inventory(E CBI) (RMSR = .10) ECBI item 12. Gets angry when doesnt get own way .89 8. Does not obey house rules on own .88 10. Acts defiant when told to do something .87 30. Is easily distracted .87 28. Constantly seeks attention .84 6. Slow in getting ready for bed .83 34. Has difficulty concentrating on one thing .82 32. Fails to finish task or projects .82 35. Is overactive or restless .80 14. Sasses adults .79 2. Dawdles or lingers at mealtime .79 20. Is careless with toys and other objects .79 16. Cries easily .78 4. Refuses to eat food presented .75 22. Lies .72 9. Refuses to obey until thre atened with punishment .71 24. Verbally fights with friends own age .69 26. Physically fights with friends own age .67 18. Hits parents .63 36. Wets the bed .61 29. Interrupts .56 31. Has short attention span .46 13. Has temper tantrums .46 7. Refuses to go to bed on time .45 21. Steals -.44 5. Refuses to do chores when asked .42 15. Whines .40 17. Yells or screams .36 11. Argues with parents about rules .35 1. Dawdles in getting dressed .26 3. Has poor table manners .22 19. Destroys toys and other objects .12 23. Teases or provokes other children -.06 33. Has difficulty entertaining self alone .05 25. Verbally fights with sisters and brothers .01 27. Physically fights with sisters and brothers .01

PAGE 69

69 Table A-2. Two-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR = .08) ECBI item Factor One Factor Two 12. Gets angry when doesnt get own way .93 .12 8. Does not obey house rules on own .89 .05 10. Acts defiant when told to do something .89 .08 30. Is easily distracted .88 .03 6. Slow in getting ready for bed .85 .08 2. Dawdles or lingers at mealtime .84 -.16 32. Fails to finish task or projects .83 .02 14. Sasses adults .81 .10 16. Cries easily .81 -.09 34. Has difficulty concentrating on one thing .80 .06 4. Refuses to eat food presented .78 .11 35. Is overactive or restless .77 .08 20. Is careless with toys and other objects .76 .12 9. Refuses to obey until thre atened with punishment .70 .03 22. Lies .68 .13 24. Verbally fights with friends own age .67 .07 18. Hits parents .65 .08 36. Wets the bed .64 -.09 26. Physically fights with friends own age .63 .14 29. Interrupts .47 .32 28. Constantly seeks attention .46 .32 7. Refuses to go to bed on time .40 .19 31. Has short attention span .39 .28 13. Has temper tantrums .39 .25 5. Refuses to do chores when asked .35 .25 15. Whines .34 .17 11. Argues with parents about rules .30 .16 17. Yells or screams .28 .25 1. Dawdles in getting dressed .24 .08 3. Has poor table manners .11 .38 33. Has difficulty entertaining self alone .01 .15 19. Destroys toys and other objects -.07 .68 23. Teases or provokes other children -.20 .65 25. Verbally fights with si sters and brothers -.18 .66 27. Physically fights with sisters and brothers -.17 .61 21. Steals -.61 .57

PAGE 70

70 Table A-3. Three-factor EFA of the Eyberg Child Behavior Inventory(ECBI) (RMSR = .06) ECBI item Factor One Factor Two Factor Three 12. Gets angry when doesnt get own way .91 .05 .01 26. Physically fights with friends own age .84 -.31 -.23 28. Constantly seeks attention .83 .04 .05 24. Verbally fights with friends own age .81 -.20 -.15 30. Is easily distracted .78 .03 .17 14. Sasses adults .76 .06 .09 34. Has difficulty concentrating on one thing .74 -.07 .15 20. Is careless with toys and other objects .74 .14 .11 10. Acts defiant when told to do something .72 .12 .26 32. Fails to finish task or projects .72 .03 .20 22. Lies .72 -.19 .02 8. Does not obey house rules on own .71 .08 .28 35. Is overactive or restless .68 -.06 .21 18. Hits parents .67 .01 -.03 36. Wets the bed .66 .02 -.04 6. Slow in getting ready for bed .66 .13 .29 2. Dawdles or lingers at mealtime .61 .23 .33 16. Cries easily .60 .15 .31 4. Refuses to eat food presented .60 .15 .26 9. Refuses to obey until threatened with punishment .45 .09 .42 29. Interrupts .34 -.22 .33 31. Has short attention span .28 -.20 .29 23. Teases or provokes other children .00 -.70 -.05 19. Destroys toys and other objects .02 -.65 .12 25. Verbally fights with sist ers and brothers -.06 -.64 .06 27. Physically fights with sisters and brothers -.01 -.62 .00 21. Steals -.59 -.46 .17 3. Has poor table manners -.19 -.16 .63 5. Refuses to do chores when asked .09 -.07 .52 7. Refuses to go to bed on time .14 -.02 .50 1. Dawdles in getting dressed -.05 .10 .49 15. Whines .09 -.02 .46 11. Argues with parents about rules .08 -.01 .43 17. Yells or screams .14 -.14 .34 13. Has temper tantrums .26 -.16 .31 33. Has difficulty entertaining self alone -.11 -.05 .24

PAGE 71

71 LIST OF REFERENCES Abikoff, H. & Klein, R. G. (1992). Attenti on-deficit hyperactivity and conduct disorder: Com orbidity and implications for treatment. Journal of Consulting and Clinical Psychology, 60 881-892. Achenbach, T. M. (1991). Manual for the Child Behavior Ch ecklist/4-18 and the 1991 profile. Burlington, VT: University of Verm ont Department of Psychiatry. Achenbach, T. M. (1992). Manual for the Child Behavior Checklist/2-3 and 1992 profile. Burlington, VT: University of Vermont Department of Psychiatry. Achenbach, T.M. (1995). Diagnosis, assessment, and comorbidity in ps ychosocial treatment research. Journal of Abnormal Child Psychology, 23 45-65. Achenbach, T. M., & Rescorla, L. A. (2000). Manual for ASEBA Preschool Forms & Profiles Burlington, VT: University of Vermont, Research Center for Children, Youth, & Families. Achenbach, T. M., & Rescorla, L. A. (2001). Manual for ASEBA School-Age Forms & Profiles Burlington, VT: University of Vermont, Research Center for Children, Youth, & Families. American Psychiatric Association. (1952). Diagnostic and Statistical Manual of Mental Disorders (1st ed.). Washington, D.C.: Author. American Psychiatric Association. (1968). Diagnostic and Statistical Manual of Mental Disorders (2nd ed.). Washington, D.C.: Author. American Psychiatric Association. (1980). Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). Washington, D.C.: Author. American Psychiatric Association. (1987). Diagnostic and Statistical Manual of Mental Disorders (3rd ed., Rev.). Washington, D.C.: Author. American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, D.C.: Author. Baker, B., & Heller, T. (1996). Pr eschool children with externaliz ing behaviors: Experience of fathers and mothers. Journal of Abnormal Child Psychology, 24 513-532. Barkley, R.A. (2006). Attention-deficit hyperactivity disorder: A handbook for diagnosis and treatment (3rd ed.). New York: Guilford Press. Bauermeister, J. (1992). Factor analyses of teach er ratings of attention-deficit/hyperactivity and oppositional defiant symptoms in children aged four through thirteen years. Journal of Clinical Child Psychology, 21 27-34.

PAGE 72

72 Bemporad, J.R., & Schwab, M.E. (1986). The DSM-III and clinical child psychiatry. In T. Millon & G.L. Klerman (Eds.), Contemporary directions in psychopathology: Toward the DSM-IV (pp. 135-150). New York: Guilford Press. Boggs, S.R., Eyberg, S., & Reynolds, L.A. (1990). Concurrent validity of the Eyberg Child Behavior Inventory. Journal of Clinical Child Psychology, 19 75-78. Burns, G.L., Boe, B., Walsh, J.A., Sommer s-Flangan, R., & Teegarden, L.A. (2001). A confirmatory factor analysis on the DSM -IV ADHD and ODD symptoms: What is the best model for the organization of these symptoms? Journal of Abnormal Child Psychology, 29, 339-349. Burns, G.L., & Patterson, D.R. (1991). Factor stru cture of the Eyberg Child Behavior Inventory: One-dimensional or multidimensional measure of disruptive behavior? Journal of Clinical Child Psychology, 20 439-444. Burns, G.L., & Patterson, D.R. (2000). Factor stru cture of the Eyberg Child Behavior Inventory: A parent rating scale of oppositional defian t behavior towards adults, inattentive behavior, and conduct problem behavior. Journal of Clinical Child Psychology, 29 569577. Burns, G., Walsh, J., Patterson, D., Holte, C., Sommers-Flanagan, R., & Parker, C. (1997). Internal validity of the disr uptive behavior disorder symptoms: Implications form parent ratings for a dimensional a pproach to symptom validity. Journal of Abnormal Child Psychology, 25, 307-319. Campbell, S. (2002). Behavior problems in preschool children: Clinical and developmental issues (2nd ed.). New York: Guilford Press. Chase, R.M. & Eyberg, S.M. (2008). Clinical presentation and treatmen t outcome for children with comorbid externalizing and internalizi ng symptoms. Journal of Anxiety Disorders, 22, 273-282. Christiansen, H., Chen, W., Oades, R.D., Ashe rson, P., Taylor, E.A., Lasky-Su, J., Zhou, K., Banaschewski, T., Buschgens, C., Franke, B., Gabriels, I., Manor, I., Marco, R., Muller, U.C., Mulligan, A., Psychogiou, L., Rommelse, N. N.J., Uebel, H., Buitelaar, J., Ebstein, R.P., Eisenberg, J., Gill, M., Miranda, A., Mulas, F., Roeyers, H., Rothenberger, A., Sergeant, J.A., Sonuga-Barke, H.C., Steinha usen, M., Thompson, M., & Faraone, S.V. (2008). Co-transmission of conduct problems with attention-deficit/hyperactivity disorder: Familial evidence for a distinct disorder. Journal of Neural Transmission, 115 163-175.

PAGE 73

73 Cole, D., Truglio, R., & Peeke, L. (1997). Relati on between symptoms of anxiety and depression in children: A multitrait-multimethod-multigroup assessment. Journal of Consulting and Clinical Psychology, 65 110-119. Colvin, A., Eyberg, S.M., & Adams, C. (1999). St andardization of the Eyberg Child Behavior Inventory with chronically ill children. Gaines ville: University of Florida, Child Study Laboratory. Available online at http://www.pcit.org Craighead, W.E. (1991). Cognitive factors and cl assification issues in adolescent depression. Journal of Youth & Adolescence, 20 311-326. Doctoroff, G.L., & Arnold, D.H. (2004). Parent-ra ted externalizing behavior in preschoolers: The predictive utility of structured interviews, teacher reports, and classroom observations. Journal of Clinical Child and Adolescent Psychology, 33 813-818. dosReis, S., Zito, S.M., Safer, D.J., Soeke n, K.L., Mitchell, J.W., & Ellwood, L.C. (2003). Parental perceptions and satisfaction with stimulant medication for attention-deficit hyperactivity disorder. Journal of Developmental & Behavioral Pediatrics, 24 155-162. DuPaul, G.J., McGoey, K.E., Eckert, T., & VanB rakle, J. (2001). Preschool children with attention-deficit/hyperactivity disorder: Impairments in be havioral, social, and school functioning. Journal of the American Academy of Child and Adolescent Psychiatry, 40 508-515. Egger, H.L. & Angold, A. (2006). Common emoti onal and behavioral disorders in preschool children: Presentation, nosology, and epidemiology. Journal of Child Psychology and Psychiatry, 47 313-337. Egger, H.L. & Angold, A. (2004). The presc hool age psychiatric assessment (PAPA): A structured parent interview fo r diagnosing psychiatric disorder s in preschool children. In R. Delcarmen-Wiggens and A. Carter (Eds.) A Handbook of Infant and Toddler Mental Health Assessment. New York, NY: Oxford University Press. Eisenstadt, T.H., McElreath, L.S., Eyberg, S.M. & McNeil, C.B. (1994). Interparent agreement on the Eyberg Child Behavior Inventory. Child & Family Behavior Therapy, 16, 21-28. Eley, T.C., Bolton, D., OConnor, T.G., Perrin, S., Smith, P., & Plomin, R. (2003). A twin study of anxiety-related behaviors in pre-school children. Journal of Child Psychology and Psychiatry, 44 945-960. Eyberg, S.M. (1992). Parent and teacher behavi or inventories for the assessment of conduct problem behaviors in children. In L. Va ndeCreek, S. Knapp, & T.L. Jackson (Eds)., Innovations in Clinical Practice: A Source Book (Vol. 12, pp.377-382). Sarasota, FL: Professional Resource Exchange.

PAGE 74

74 Eyberg, S.M., Nelson, M.M., & Boggs, S.R. (2008). Evidence-based psychosocial treatments for children and adolescents w ith disruptive behavior. Journal of Clinical Child & Adolescent Psychology, 37 215-237. Eyberg, S. M., & Pincus, D. (1999). Eyberg Child Behavior Inventory and Sutter-Eyberg Student Behavior Invent ory: Professional manual. Odessa, FL: Psychological Assessment Resources. Eyberg, S.M., & Robinson, E.A. (1983). Conduct problem behavior: Standardization of a behavior rating scale with adolescents. Journal of Clinical Child Psychology, 11 130137. Eyberg, S. M., Schuhmann, E. M., & Rey, J. (1998). Child and adoles cent psychotherapy research: Developmental issues. Journal of Abnormal Child Psychology 26, 71-82. Floyd, F.J., & Widaman, K.F. (1995). Factor anal ysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7 286-299. Funderburk, B.W., Eyberg, S.M., Rich, B.A., & Behar, L. (2003). Further psychometric evaluation of the Eyberg and Behar rating scal es for parents and teach ers of preschoolers. Early Education and Development, 14 67-81. Gadow, K.D. & Sprafkin, J. (2000). Early Childhood Inventory 4 Screening Manual. Stony Brook, NY: Checkmate Plus. Glutting, J., Youngstrom, E., & Watkins, M. ( 2005). ADHD and college students: Exploratory and confirmatory factor structures with student and parent data. Psychological Assessment, 17, 44-55. Greene, R. W., Biederman, J., Zerwas, S., Monuteaux, M., Goring, J. C., & Faraone, S. V. (2002). Psychiatric comorbidity, family dysfunction, and social impairment in referred youth with oppositional defiant disorder. American Journal of Psychiatry, 159 12141224. Greenhill, L., Kollins, S., Abikoff, H., McCr acken, J., Riddle, M., Swanson, J., McGough, J., Wigal, S., Wigal, T., Vitiello, B., Skrobala, A., Posner, K., Ghuman, J., Cunningham, C., Davies, M., Chuang, S., & Cooper, T. (2006). Efficacy and safety of immediate-release methylphenidate treatment fo r preschoolers with ADHD. Journal of the American Academy of Child and Adolescent Psychiatry, 45, 1284-1293. Gross, D., Fogg, L., Young, M., Ridge, A., Co well, J., Sivan, A., & Richardson, R. (2007). Reliability and validity of the Eyberg Child Behavior Inventory w ith African-American and Latino parents of young children. Research in Nursing & Health, 30, 213-223. Hardy, K.K., Kollins, S.K., Murray, D. W., Riddle, M.A., Greenhill, L.,

PAGE 75

75 Cunningham, C., Abikoff, H.A., McCracken, J.T ., Benedetto, V., Davies, M., McGough, J.J., Posner, K., Skrobala, A.M., Swanson, J.M., Wigal, T., Wigal, S.B., Ghuman, J.K., & Chuang, S.Z. (2007). Factor structure of parentand teacher-rated AttentionDeficit/Hyperactivity Disorder symptoms in the preschoolers with AttentionDeficit/Hyperactivity Disorder treatment study (PATS). Journal of Child and Adolescent Psychopharmacology, 17, 621-634. Harwood, M.D., OBrien, K.A., Carter, C.G., & Eyberg, S.M. (under review). Mental health services for preschool children in primary care: A survey of maternal attitudes and beliefs. Journal of Pediatric Psychology. Henggeler, S.W. & Borduin, C.M. (1990). Fa mily therapy and beyond: A multisystemic approach to treating the behavior problems of children and adolescents. Pacific Grove: CA: Brookes and Cole. Huffine, C.W. (2002). Conduct disorder should be eliminated from the diagnostic and statistical manual of mental disorders. In L. Flaherty (ed). Adolescent Psychiatry: The Annals of the American Society for Adolescent Psychiatry, Vol. 26 Hillsdale: NJ: The Analytic Press. Jensen, P.S., Watanabe, H.K., Richters, J.E., Roper, M., Hibbs, E.D., Salzberg, A.D., & Liu, S. (1996). Scales, diagnoses, and child psychopa thology: II. Comparing the CBCL and the DISC against external validators. Journal of Abnormal Child Psychology, 24 151-168. Keenan, K. & Wakschlag, L. S. (2002). Can a valid diagnosis of disruptive behavior disorder be made in preschool children? American Journal of Psychiatry, 159, 351-358. Keenan, K. & Wakschlag, L.S. (2000). More than th e terrible twos: The na ture and severity of behavior problems in clinic -referred preschool children. Journal of Abnormal Child Psychology, 28, 33-46. Koot, H. M, Van Den Oord, E. J. C. G., Verhulst F. C., & Boomsma, D. I. (1997). Behavioral and emotional problems in young preschoolers: Cross-cultural testing of the validity of the child behavior checklist/2-3. Journal of Abnormal Child Psychology, 25, 183-196. Krueger, R. F., Caspi, A., Moffitt, T. E., & Silv a, P. A. (1998). The structure and stability of common mental disorders (DSM-III-R): A longitudinal-epidemiological study. Journal of Abnormal Psychology, 107, 216-227. Lahey, B.B., Rathouz, P.J., Van Hulle, C., Urbano, R. C., Krueger, R.F., Applegate, B., Garriock, H.A., Chapman, D.A., & Waldman, I.D. (2008) Testing structural models of DSM-IV symptoms of common forms of ch ild and adolescent psychopathology. Journal of Abnormal Child Psychology, 36 187-206. Lahey, B., Applegate, B., Waldma n, I., Loft, J., Hankin, B., & Rick, J. (2004). The structure of child and adolescent psychopathol ogy: Generating new hypotheses. Journal of Abnormal Psychology, 113 358-385.

PAGE 76

76 Lahey, B. B., Pelham, W.E., Stein, M A., Loney, J ., Trapani, C., Nugent, K., Kipp, H., Schmidt, E., Lee, S., Cale, M., Gold, E., Hartung, C.M., Willcutt, E., & Baumann, B. (1998). Validity of DSM-IV attention-deficit/ hyperactivity disorder for younger children. Journal of the American Academy of Child and Adolescent Psychiatry, 37 695-702. Lavigne, J. V., Arend, R., Rosenbaum, D., Binns, H. J., Christoffel, K K., & Gibbons, R. D. (1998). Psychiatric disorders w ith onset in the preschool ye ars, part 1: stability of diagnoses. Journal of the American Academy of Child and Adolescent Psychiatry, 37 1246-1254. Lilienfeld, S. O., Waldman, I. D., & Israel, A. C. (1994). A critical examination of the use of the term and concept of comorbid ity in psychopath ology research. Clinical Psychology Science and Practice, 1 71-83. Loeber, R., Burke, J.D., Lahey, B.B., Winter s, A., & Zera, M. (2000). Oppositional defiant disorder and conduct disorder: A review of the past 10 years, part I. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1468-1484. Luby, J.L., Heffelfinger, A.K., Mrakotsky, C., Hessler, M.J., Brown, K.M., & Hildebrand, T. (2002). Preschool major depressive disorder: Preliminary validation for developmentally modified DSM-IV criteria. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 928-937. Lucas, C.P., Fisher, P., Luby, J. (2000) The Young Child DISC New York: Columbia DISC Development Group. Marsh, H.W. & Hau, K.H. (1996) Assessing goodness of fit: Is parsimony always desirable? The Journal of Experime ntal Education, 64, 364-390. McClellan, J., & Speltz, M. (2003). Psychi atric diagnosis in preschool children. Journal of the American Academy of Child and Adolescent Psychiatry, 42 127-128. Muthn L. & Muthn B. (2005). Chi-square di fference testing using the S-B scaled chi-square. Note on Mplus website, www.statmodel.com. Newcorn, J. H., Halperin, J. M., Jensen, P. S ., Abikoff, H. B., Arnol d, E., Cantwell, D. P., Conners, C. K., Elliott, G. R., Epstein, J. N., Greenhill, L. L., Hechtman, L., Hinshaw, S. P., Hoza, B., Kraemer, H. C., Pelham, W. E., Severe, J. B., Swanson, J. M., Wells, K. G., Wigal, T., & Vitiello, B. (2001). Symptom prof iles in children with ADHD: Effects of comorbidity and gender. Journal of the American Ac ademy of Child & Adolescent Psychiatry, 40, 137-146. Nolan, E. E., Gadow, K. D., & Sprafkin, J. ( 2001). Teacher reports of DSM-IV ADHD, ODD, and CD symptoms in schoolchildren. Journa l of the American Academy of Child and Adolescent Psychiatry, 40, 241-249.

PAGE 77

77 Nottelmann, E. D. & Jensen, P. S. (1995). Comorbid ity of disorders in ch ildren and adolescents: Developmental perspectives. In T. H. Ollendick & R. J. Prinz (Eds.), Advances in Clinical Child Psychology: Vol. 17 (pp. 109-155). New York: Plenum Press. Pillow, D.R., Pelham, W.E., Hoza, B., Molina, B.S.G., and Stultz, C.H. (1998). Confirmatory factor analyses examining attention defic it hyperactivity disorder symptoms and other childhood disruptive behaviors. Journal of Abnormal Child Psychology, 26, 293-309. Ross, C.N., Blanc, H.M., McNeil, C.B., Eyberg, S.M., & Hembree-Kigin, T.L. (1998). Parenting stress in mothers of young children with opposi tional defiant disorder and other severe behavior problems. Child Study Journal, 28, 93-110. Rushton, J.L., Fant, K.El, & Clark, S.J. (2004). Us e of practice guidelines in the primary care of children with attention-defi cit/hyperactivity disorder. Pediatrics, 114, E23-E28. Schuhmann, E., Foote, R., Eyberg, S.M., Boggs, S., & Algina, J. (1998). Parent-child interaction therapy: Interim report of a randomized trial with short-term maintenance. Journal of Clinical Child Psychology, 27, 34-45. Shaffer, D., Fisher, P., Lucas, C. P., Dulca n, M. K., & Schwab-Stone, M. E. (2000). NIMH diagnostic interview schedule for children version IV (NIMH DISC -IV): Description, differences from previous versions, and reliability of some common diagnoses. Journal of the American Academy of Ch ild & Adolescent Psychiatry, 39, 28-38. Silk, J. S., Nath, S. R., Siegel, L. R., & Kenda ll, P.C. (2000). Conceptualizing mental disorders in children: Where have we been and where are we going? Development and Psychopathology, 12 713-735. Speltz, M., McClellan, J., DeKylen, M., & Jones, K. (1999). Preschool boys with Oppositional Defiant Disorder: Clinical pres entation and diagnostic change. Journal of the American Academy of Child and Adolescent Psychiatry, 38 838-845. Spence, S., Rapee, R., McDonald, C., & Ingram, M. (2001). The structure of anxiety symptoms among preschoolers. Behavior Research and Therapy, 39, 1293-1316. Sprafkin, J., Volpe, R.J., Gadow, K.D., Nolan, E.E., & Kelly, K. (2002). A DSM-IV-referenced screening instrument for preschool ch ildren: The early childhood inventory-4. Journal of the American Academy of Child and Adolescent Psychiatry, 41 604-612. Sterba, S., Egger, H.L., & Angold, A. (2007). Di agnostic specificity and nonspecificity in the dimensions of preschool psychopathology. Journal of Child Psychology and Psychiatry, 48, 1005-1013. Task Force on Research Diagnosti c Criteria: Infancy and Preschool. (2003). Research diagnostic criteria for infants and pr eschool children: The process and empirical support. Journal of the American Academy of Child and Adolescent Psychiatry, 43 1504-1512.

PAGE 78

78 Thomas, J. & Guskin, K. (2001). Disruptive be havior in young children: What does it mean? Journal of the American Academy of Child and Adolescent Psychiatry, 40, 44-51. Verduin, T.L., Abikoff, H., & Kurtz, S.M.S. (2008). Evidence-based treatment of attention deficit/hyperactivity disorder in a preschool-age child: A case study. Journal of Clinical Child & Adolescent Psychology, 37 477-485. Vitiello, B., Abikoff, H.B., Chuang, S.Z., Kollins, S.H., McCracken, J.T., Riddle, M.A., Swanson, J.M., Wigal, T., McGough, J.J., Ghum an, J.K., Wigal, S.B., Skrobala, A.M., Davies, M., Posner, K., Cunningham, C., & Greenhill, L.L. (2007). Effectiveness of methylphenidate in the 10-month continua tion phase of the preschoolers with ADHD treatment study (PATS). Journal of Child and Adolescent Psychopharmacology, 17, 593603. Wakefield, J.C., Pottick, K. J., & Kirk, S.A. ( 2002). Should the DSM-IV diagnostic criteria for conduct disorder consider social context? American Journal of Psychiatry, 159 380-386. Wakschalg, L. S., Leventhal, M. D. & Thomas, M. D. Disruptive behavior disorders & ADHD in preschool children: Characterizing hetero typic continuities for a developmentallyinformed nosology for DSM V. In D. Reig er, M. First & W. Narrow (Eds). Age and gender considerations in psychiatric di agnosis: A research agenda for DSM-V. Washington, D.C.: American Psychiatric Publishing Inc. (in press). Weis, R., Lovejoy, M.C., & Lundahl, B.W. (2005). Factor structure and discriminative validity of the Eyberg Child Behavior Inventory with young children. Journal of Psychopathology and Behavioral Assessment, 27 269-278. Wigal, T. Greenhill, L., Chuang, S., McCough, J., V itiello, B., Skrobala, A., Swanson, J., Wigal, S., Abikoff, H., Kollins, S., McCracken, J., Riddle, M., Posner, K., Ghuman, J., Davies, M., Thorp, B., & Stehli, A. (2006). Safety and tolerability of methylphenidate in preschool children with ADHD. Journal of the American Academy of Child and Adolescent Psychiatry, 45 1294-1303. Wilens, T., Biederman, J., Spencer, T., & Monut eaux, M. (2003). Psychiatric diagnosis in preschool children: Reply. Journal of the American Academy of Child and Adolescent Psychiatry, 42 128-129. Wilens, T. E., Biederman, J., Brown, S., Tanguay, S., Monuteaux, M.C., Blake, C., & Spencer, T. J. (2002). Psychiatric comorbidity and f unctioning in clinically referred preschool children and school-age youths with ADHD. Journal of the American Academy of Child and Adolescent Psychiatry, 41 262-268. Willoughby, M., Curran, P., Costello, E., & Angold, A. (2000). Implications of early versus late onset of attention-deficit/hype ractivity disorder symptoms. Journal of the American Academy of Child and Adolescent Psychiatry, 39 1512-1519.

PAGE 79

79 Wolraich, M.L., Felice, M.E., & Drotar, D. (E ds). (1996). The classification of child and adolescent mental diagnosis in primary care: Diagnosis and Statistical Manual for Primary Care (DSM-PC) Child and Adolescent Version. Elk Gove, IL: American Academy of Pediatrics. World Health Organization (1993). The ICD-10 classification of mental and behavioural disorders: Diagnostic criteria for research, 10th edition. Geneva: World Health Organization. Yu, C., & Muthn, B. (2002). Evaluation of model fit indices for latent variable models with categorical and continuous outcomes. U npublished Dissertation available online at: www.statmodel.com.

PAGE 80

80 BIOGRAPHICAL SKETCH Rhea Marisa Chase was born in Manchest er, Connecticut on October 6, 1979. She graduated magna cum laude with a B.A. in psychology in 2001 from Bates College in Lewiston, Maine. She then worked in clinical research at Massachusetts Gene ral Hospital in Boston, Massachusetts for two years. He r graduate work was done in th e Department of Clinical and Health Psychology at the University of Florida, and she completed her predoctoral internship at Duke University Medical Center. Upon completion of her Ph.D. program, Rhea will begin a post doctoral position through the Center for Anxiety and Related Disord ers at Boston University.