MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIES RESEARCH REVIEWS 11: 226237 (2005)

ASSESSMENT OF YOUNG CHILDREN USING THE ACHENBACH SYSTEM OF EMPIRICALLY BASED ASSESSMENT (ASEBA)
Leslie A. Rescorla*
Department of Psychology, Bryn Mawr College, Bryn Mawr, Pennsylvania

After providing a brief review of three other approaches to assessment of preschool children (DSM-IV diagnoses, Zero to Three diagnoses, and temperament scales), this paper focuses on the Achenbach System of Empirically Based Assessment (ASEBA). The empirically based assessment paradigm provides user-friendly, cost-effective, reliable, and valid procedures for assessing childrens behavioral/emotional problems from the perspectives of multiple informants. The ASEBA preschool forms, the Child Behavior Checklist for ages 1.55 (CBCL/1.55) and the CaregiverTeacher Report Form (CTRF), are usable by many different kinds of professionals in diverse settings. The CBCL/1.55 also includes the Language Development Survey (LDS), which provides a quick screen for delays in vocabulary and word combinations. The problem items of the CBCL/1.55 and the CTRF are scored on both empirically based syndromes and DSM-oriented scales, which are normed on the same general population sample. Variations in childrens functioning across contexts and interaction partners make it essential to obtain and integrate data from multiple sources. Therefore, ASEBA software provides side-by-side comparisons of item and scale scores from up to eight assessment forms per child. Clinical and research applications of ASEBA preschool forms are summarized in the paper, and strengths 2005 Wiley-Liss, Inc. and limitations are discussed.
MRDD Research Reviews 2005;11:226 237.

Central features of the ASEBA include (a) assessment of behavioral and emotional problems with user-friendly forms, (b) proles of scores on statistically derived, empirically based syndromes, (c) proles of scores on DSM-oriented scales, (d) norms by age and gender based on national probability samples, (e) systematic comparison of ratings from multiple informants, and (f) scores on Internalizing, Externalizing, and Total Problems scales. Research conducted on ASEBA forms by scholars around the world has yielded more than 5,500 published reports. Although most ASEBA research has focused on school-age children and adolescents, there are more than 200 published studies of the ASEBA preschool forms [Be rube and Achenbach, 2005]. APPROACHES TO ASSESSMENT OF YOUNG CHILDREN Assessment of young childrens emotional and behavioral problems utilizes many different procedures. These include obtaining information from parents and caregivers via structured or semi-structured interviews, written questionnaires, and rating forms; observing the child in naturalistic or quasi-naturalistic interaction with caregivers or peers; and conducting child interviews using verbal and play-based procedures. Most assessments use a combination of such methods. However the information is obtained, the professional must summarize the data using some descriptive system. For structured assessment tools such as behavior checklists, item ratings are easily summarized via scores on scales. Professionals summarize data from naturalistic observations and from clinical interviews by writing narrative reports, by rating the child on a list of dimensions/characteristics, by coding discrete behaviors, or by evaluating the presence/absence of symptoms for different diagnoses. DSM Diagnoses One of the most common ways to assess young children is to make psychiatric diagnoses using the American Psychiatric

Key Words: CBCL/1.55; empirically based assessment; preschool assessment

his paper describes use of the Child Behavior Checklist for Ages 1.55 (CBCL/1.55) and the Caregiver Teacher Report Form (CTRF) [Achenbach and Rescorla, 2000] to assess behavioral and emotional problems in young children. The CBCL/1.55 is a revision of the CBCL/ 23 [Achenbach et al., 1987; Achenbach, 1992]; and the CTRF for ages 1.55 is a revision of the CTRF/25 [Achenbach, 1997]. The CBCL/1.55 and the CTRF are components of the Achenbach System of Empirically Based Assessment (ASEBA); a comprehensive approach to the assessment of individuals from 18 months to 90 years of age [Achenbach and Rescorla, 2000, 2001, 2002; Achenbach et al., 2004]. When the rst ASEBA forms for preschool children were published in the 1980s; they were based on more than two decades of research on parallel forms for school-age children [Achenbach, 1966, 1991; Achenbach and Edelbrock, 1981]. Norms for the 2000 versions of the CBCL/1.55 and the CTRF were based on data collected as part of the 1999 National Survey of Children, Youth, and Adults [Achenbach and Rescorla, 2000].

*Correspondence to: Leslie A. Rescorla, Ph.D., Professor of Psychology and Director of Child Study Institute, Department of Psychology, Bryn Mawr College, 101 North Merion Avenue, Bryn Mawr, PA 19010-2988. E-mail: [email protected] Received 11 July 2005; Accepted 12 July 2005 Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/mrdd.20071

2005 Wiley-Liss, Inc.

Associations [1994] Diagnostic and Statistical Manual (DSM-IV). Concepts of childhood disorders have often been shaped by adult diagnostic categories, many of which originated in the 19th century (e.g. schizophrenia, bipolar disorder). For many disorders, such as depression, children are diagnosed using the adult categories and criteria, with little modication. One of the rst diagnostic categories specic to very young children was early infantile autism [Kanner, 1943]. The DSM-IV includes ve Pervasive Developmental Disorder (PDD) or autism spectrum diagnoses (Autistic Disorder, Aspergers Disorder, Childhood Integrative Disorder, Retts Disorder, and Pervasive Developmental Disorder NOS). Other DSM-IV diagnoses that are commonly used for young children include Attention Decit/Hyperactivity Disorder (ADHD), Oppositional Deant Disorder, Separation Anxiety Disorder, Specic Phobia, Generalized Anxiety Disorder, Selective Mutism, Major Depressive Disorder, Reactive Attachment Disorder, Feeding/Eating Disorders, and Elimination Disorders. Recently, Bipolar Disorder has become an accepted diagnosis for young children. In their review of psychiatric diagnoses for preschool children, Angold and Egger [2004] suggest that research on psychiatric disorders in preschoolers is about 30 years behind such research on older children and adolescents. They indicate that only a few epidemiological studies of DSM diagnoses in preschoolers have been published [Earls, 1982; Shaw et al., 1997; Lavigne et al., 1998; Keenan and Wackschlag, 2000]. Most of the studies had small samples, low response rates, and incomplete reports of ndings. Diagnostic procedures varied widely across the studies, as did prevalence rates. Across the three community studies cited by Angold and Egger [2004], prevalence rates ranged from 14 to 26% for any Axis I diagnosis. Comorbidity between internalizing and externalizing disorders was high when it was reported. Structured diagnostic interviews have not been validated for preschoolers. Nonetheless, Angold and Egger [2004] conclude that instruments and nosologies designed and tested for use with older children appear to be applicable to younger children, and that prevalence rates for the presence of any diagnosis appear to be rather similar to those found for older children and adolescents. Epidemiological studies of DSM-IV disorders in large general population samples of preschool children using well-specied, uniform diagnostic procedures are
MRDD RESEARCH REVIEWS

sorely needed. Although DSM-IV disorders such as ADHD and PDD have been widely researched with young children, systematic epidemiological research is lacking on the prevalence, patterning, and discriminative power of the symptoms that the DSM-IV uses to dene most disorders in this age group. Zero to Three Classication System Dissatisfaction with the DSM-IV approach led to the development of the Zero to Three Diagnostic Classication of Early Mental Disorders (DC: 0 3). [Zero to Three: National Center for Infants, Toddlers, and Families, 1994]. The categories of the DC: 0 3 were based on expert consensus and distillations from 224 case reports [Lieberman et al., 2004]. Five axes are assessed in this multiaxial system. First, a primary diagnosis is made using the general categories of Traumatic Stress Disorder, Disorder of Affect (including anxiety, bereavement/grief, depression, mixed, reactive attachment, and gender identity), Adjustment Disorder, Regulatory Disorder (hypersensitive, underreactive, motorically disorganized/ impulsive, or other), Sleep Behavior Disorder, Eating Behavior Disorder, or Disorder of Relating and Communicating (PDD-type problems). Second, a Relationship Disorder Classication can be made using one of six categories (overinvolved, underinvolved, anxious/tense, angry/hostile, mixed, or abusive). Axis III deals with Medical and Developmental Disorders/Conditions, including DSM-IV diagnoses. Axis IV assesses Psychosocial Stressors, and Axis V measures Functional, Emotional Developmental Level (5-level scale). In their review chapter, Angold and Egger [2004] noted that the DC: 0 3 has been used in only a few research studies. Some of these studies [Dunitz et al., 1996; Thomas and Guskin, 2001] have demonstrated partial but not complete overlap between the DSM-IV and the DC: 0 3 systems. That is, in clinical samples of preschoolers, some children received comparable diagnoses in both systems (PTSD, Anxiety Disorder, Adjustment Disorder), but others received different diagnoses in the two systems (e.g., Regulatory Disorder in DC: 0 3, but Oppositional Deant Disorder or Adjustment Disorder in DSM-IV). This raises the issue of whether Regulatory Disorder is a useful separate category, or whether symptoms associated with Regulatory Disorder should be incorporated into DSM criteria for various diagnoses. Reliability and validity data on the DC: 0 3 are limited. Lieberman et al.
OF

[2004] described an unpublished report [Barnard et al., 2002] that involved eight clinical cases. A primary clinician assessed each case using a commonly agreedupon protocol. Raw data from each case (interviews, observations, and test scores) were then reviewed by the team and a consensus was reached for both DC: 0 3 and DSM-IV diagnoses. The clinicians apparently agreed that the DC: 0 3 primary diagnosis was a better t to the child than the DSM-IV diagnosis. Five of the eight cases received a DC: 0 3 diagnosis of Regulatory Disorder; while four of the eight received an Axis II diagnosis of parent child relationship problem. The raters concluded that the DC: 0 3 diagnosis was always validated by the reported history, observations of the child during play or snack time, symptom screening tests, and parent child interaction observations, a fact that Lieberman et al., [2004] presented as an important nding. The DC: 0 3s focus on the parent child relationship taps important aspects of young childrens functioning that are not captured in the DSM-IV. In addition, the DC: 0 3 emphasizes regulatory disorders more than the DSM-IV does. However, the DC: 0 3 system has not been studied using the procedures for assessing reliability and validity that are standard in the eld. These standard procedures include using a large sample that includes cases with diverse problems as well as noncases measuring agreement between independent raters, and validating the diagnoses with data that were not used in the diagnostic process. Temperament Scales Another widely used approach to the assessment of young children involves the rating of temperamental characteristics. Commonly used approaches for assessing temperament include parent ratings, home observation, and lab-based assessment [Stifter and Wiggins, 2004]. Because they are the simplest and most efcient, parent report scales are the most widely used means for measuring temperament. The Revised Infant Temperament Questionnaire [Carey and McDevitt, 1978], the Buss and Plomin [1984] EAS Temperament Survey, and the Rothbart [1981] Infant Behavior Questionnaire are among the parent rating scales commonly used for assessing temperament. Such scales vary in length, scoring (continuum or dichotomous scoring), and dimensions tapped. Studies of parent temperament ratings have often demonstrated associations between characteristics of the rater and 227

ASSESSMENT

YOUNG CHILDREN USING ASEBA

RESCORLA

child temperament scores [Stifter and Wiggins, 2004]. For example, mothers who are depressed, anxious, overly stressed, or of low socioeconomic status (SES) typically report more difcult temperament in their young children. Similarly, mothers prenatal working model of the infant and their ratings of aversiveness or tape-recorded crying predict their subsequent ratings of temperament. These ndings indicate that it is important to obtain temperament ratings from additional informants to determine the degree to which temperamental characteristics reported by the mother are unique to her perspective or are corroborated by others. In studies examining this question; such as the report by Seifer et al., [1994], agreement between maternal ratings and trained observers has been low. Researchers have used a variety of observational methods for assessing temperament, such as those by Garcia Coll et al., [1992] and by Goldsmith and Rothbart [1991]. However, such procedures require extensive training, and they provide a restricted range of observational contexts. Additionally, observational measures must be collected over numerous sampling occasions to be reliable [Seifer et al., 1994; Sifter and Wiggins, 2004], precluding their use in clinical settings. GOALS/OBJECTIVES OF THE ASEBA PRESCHOOL FORMS The ASEBA preschool instruments were designed to be user-friendly, standardized tools that can be utilized by different kinds of professionals assessing children in diverse settings. An important feature of ASEBA instruments is that they do not require professional time or training for either administration or scoring. Young children are evaluated in different centers, including pediatric primary care centers, early intervention services, child development clinics, tertiary medical centers, community mental health centers, private practices, child care settings, child welfare/protective service agencies, foster care services, and research settings. Reasons for assessment may include planning therapeutic and educational interventions, mounting prevention efforts, deciding on placements, determining custody, and studying etiologies, correlates, interventions, and outcomes of problems. Assessment must be able to take account of possible variations related to ethnicity, family constellation, SES, and languages spoken in the home. The assessment methods need to be us228

able by many kinds of professionals, such as nurses, educators, child development specialists, pediatricians, child and family service workers, social workers, psychiatrists, and psychologists. Although direct observation of the child and caregivers is always important, young childrens behavior is often highly variable. For example, time of day, physical state, environmental context, interaction partner, and presence of observers may all affect a childs behavior. Even an in-depth clinical assessment over multiple settings and days can sample only a small fraction of the childs possible behaviors. Furthermore, young children cannot be expected to report on their own functioning. Thus, it is essential to obtain reports of the childs behavioral and emotional characteristics by adults who know the child well, such as parents, caregivers, and teachers. Because young children may manifest many different kinds of problems, assessment instruments need to be comprehensive enough to detect possible problems of many types. Based on these considerations, the CBCL/1.55 and the CTRF were designed to provide normed scores on a wide array of behavioral and emotional problem scales in young children. Because language is so important for young childrens adaptive development, an additional goal in developing the CBCL/ 1.55 was to obtain a quick screen of language development for children from 1.5 to 3 years of age. Thus, the CBCL/ 1.55 includes the Language Development Survey (LDS) [Rescorla, 1989]. Inclusion of the LDS with the CBCL/ 1.55 makes it easy to simultaneously determine whether children who appear to have language delays are also reported to have more behavioral/emotional problems than normative samples of peers. Conversely, when assessing behavioral/emotional problems, practitioners can simultaneously screen for language delay. HISTORICAL BACKGROUND OF THE ASEBA PRESCHOOL INSTRUMENTS The empirically based paradigm was rst applied to young children when the CBCL for ages 23 was constructed in 1981 for assessment of low birth weight children [Achenbach et al., 1987]. This was followed by the CTRF for ages 25 [Achenbach, 1997]. In 2000, revisions of these instruments were published [Achenbach and Rescorla, 2000]. Two items from the CBCL/23 [Achenbach, 1992] were replaced by new items

for the CBCL/1.55. No items were changed from the CTRF/25 [Achenbach, 1997]. The CBCL 1.55 and the CTRF [Achenbach and Rescorla, 2000] improved on the previous versions in the following ways: (a) the age range of both the CBCL and the CTRF were extended so that both forms now span ages 1.55, (b) the LDS [Rescorla, 1989], a parent-report screening tool for language delay, was added to the CBCL/1.55, (c) new empirically based scales were developed for the CBCL/1.55 and CTRF, (d) new DSM-oriented scales were constructed for both instruments, and (e) new computerized scoring programs were developed to provide cross-informant comparisons for up to eight forms per child. Data for norming the ASEBA preschool forms were collected as part of the 1999 National Survey of Children, Youth, and Adults, which involved over 10,000 participants ages 1.590. Temple Universitys Institute for Survey Research conducted the National Survey using its national sampling frame of 100 Primary Sampling Units (PSUs), chosen to be collective representative of the 48 contiguous states. Interviewers within each PSU were assigned listing areas of 150 households. They visited each household in their listing areas to determine age and gender of residents; 99% of households visited provided such information. After eligible preschool children were identied (age between 18 and 71 months, with at least one parent who spoke English), candidate children were selected by stratied random sampling. Of the 781 eligible preschool children recruited, the CBCL was completed for 738 (94.4% completion rate). Children who had received mental health or special education services in the past year were excluded from the normative sample (38 children, 5.1%), yielding a nal sample of 700 children who were healthy in epidemiological terms. This CBCL/1.55 normative sample consisted of 362 boys and 338 girls. The sample was representative of the ethnic diversity of the United States, with 56% NonLatino white, 21% AfricanAmerican, 135 Latino, and 10% mixed or other (South Asian, Asian, and Native American). Non-English languages were spoken in about 25% of the homes. The other language was typically Spanish, but Tagalog, Chinese, Vietnamese, Tamil, American Sign Language, Bengali, Japanese, and Farsi were also spoken. The SES of the sample was 33% upper, 49%, middle, and 17% lower, based on an up

MRDD RESEARCH REVIEWS

ASSESSMENT

OF

YOUNG CHILDREN USING ASEBA

RESCORLA

dated version of Hollingsheads [1975] 9-step scale for coding parental occupation. Regional distribution was 17% Northeast, 22% Midwest, 40% South, and 21% West. The CTRF normative sample contained 203 children whose parents completed the CBCL/1.55 who attended daycare or preschool, whose parents gave consent to contact a caregiver or teacher, and whose caregiver or teacher completed and returned the form. To augment this sample, 989 children from the 1997 CTRF/25 normative sample [Achenbach, 1997] were added. The 1997 CTRF normative sample consisted of 753 children recruited in the National Institute of Child Health and Development Study of Early Child Care [NICHD Early Child Care Research Network, 1994] recruited in 9 states, plus 36 children drawn from 14 daycare and preschool programs in 12 states. The nal 2000 CTRF normative sample thus consisted of 1,192 children, 588 boys and 604 girls. Ethnicity was 48% NonLatino white, 36% African American, 8% Latino, and 9% mixed or other. SES distribution was 47% upper, 43% middle, and 10% lower. Regional distribution was 29% Northeast, 17% Midwest, 32% South, and 22% West. DESCRIPTION OF ADMINISTRATION PROCEDURES The CBCL/1.55 is completed by parents, parent surrogates, and others who see children in home-like contexts. The CTRF is completed by daycare providers and preschool teachers. The CBCL/1.55 and the CTRF have 82 similar problem items, plus 17 items that are specic to home versus daycare and preschool contexts, and an open-ended item for adding other problems that are not listed on the forms. Both forms request respondents to rate each item as follows: 0, not true; 1, somewhat or sometimes true; or 2, very true or often true of the child now or within the past 2 months [Achenbach and Rescorla, 2000]. Both forms also request descriptions of behavior, illnesses, disabilities, what concerns the respondent most about the child, and the best things about the child. Thus, the forms not only provide quantitative scores for each problem item, but they also yield descriptions of the childs functioning in the respondents own words. The LDS portion of the CBCL/1.55 requests respondents to provide information about possible risk factors for language delays, to report ve of the childs best word combinations if
MRDD RESEARCH REVIEWS

the child is combining words, and to circle on a 310-word vocabulary list the words used by the child. Completion of the ASEBA forms requires only 5th grade reading skill. If there are questions about a respondents ability to complete a form independently, the following procedure is recommended: An interviewer hands the respondent the form while retaining a second copy. The interviewer then says, Ill read you the questions on this form and Ill write down your answers. Respondents with adequate reading skills will often start answering without waiting for the questions to be read. However, for respondents who cannot read well, this procedure avoids embarrassment while maintaining the standardization of the assessment process. ADMINISTRATION TIME The ASEBA preschool forms can be completed independently by most respondents in about 10 min. Completion of the LDS also takes about 10 min, and for a total CBCL/1.55 completion time of 20 min. TRAINING REQUIRED No professional training is required for administration of the ASEBA preschool forms. Instructions on the forms make them self-explanatory. In many settings, the CBCL/1.55 is sent to parents by mail along with various other forms, such as consent forms and developmental history forms. The CTRF is also routinely sent to teachers and daycare providers, after parents have provided consent. Parents, teachers, and daycare providers can then mail back the forms, which can be scored by a clerical worker or by a professional, using the ASEBA computer scoring program. Computer-scoring takes about 5 min per form. It is also possible to hand-score the forms, using a set of scoring templates and proles. However, hand-scoring takes more time than computer scoring (10 15 min per form) and does not yield the same cross-informant displays that computer scoring generates. ASEBA also offers a web-based utility called WebLink, whereby forms can be sent, completed, returned, and scored via the Internet. Although administering and scoring the ASEBA forms require no professional training, interpretation of the information obtained does require professional training related to childrens behavioral and emotional problems. Interpretation of the LDS also requires some familiarity with the process of lanOF

guage acquisition in young children. Graduate training at the masters level or the equivalent is the appropriate user qualication for ASEBA forms. DESCRIPTION OF SCORING SYSTEM LDS Scores The LDS portion of the CBCL/ 1.55 is scored by adding up the number of words circled and by calculating the number of words in each phrase/sentence provided. Simple scoring rules are provided for determining the number of words in multiword utterances (e.g., thank you is counted as a single word utterance). Number of vocabulary words is scored in relation to national norms for children at ages 18 23 months, 24 29 months, and 30 35 months. The average length of word combinations is only scored for the ages 24 29 and 30 35 months, because before 24 months children often do not combine words [Achenbach and Rescorla, 2000]. Figure 1 displays the scores obtained from the LDS completed for 30month-old Kenny Randall by his mother. Based on the ratings from Kennys mother, both his vocabulary score (55th percentile) and his average length of phrases score (50th percentile) were above the cutpoints for concern (15th and 20th percentiles, respectively). Although Ms. Randall wrote ve sentences for Kenny, she added that the sentence Doesnt open, Ben was one for the only times she had ever heard Kenny speak in a meaningful, communicative way with another child. The paucity of communicative sentences raised questions about Kennys pragmatic language skills. CBCL/1.55 and CTRF Empirically Based Scores Both the CBCL/1.55 and the CTRF yield scores on empirically based syndromes. These syndromes were derived by conducting exploratory and conrmatory factor analyses of item scores for large samples of children. Factor analysis of problems works best if the forms utilized have non-zero scores on a substantial number of items. Thus, the CBCL/1.55 and CTRF factor analytic samples included all children from the 1999 National Survey sample scoring above the median for Total Problems, plus many other children from clinical and nonclinical settings whose Total Problems scores were at or above the National Survey median (for factor analyses of the CBCL/1.55 total N 229

ASSESSMENT

YOUNG CHILDREN USING ASEBA

RESCORLA

Fig. 1.

Kenny Randalls LDS scoring prole.

1,728; for the CTRF, N 1,113). Analyses conducted for each gender separately on each instrument identied the following six syndromes for both genders scored on both instruments: Aggressive Behavior, Anxious/Depressed, Attention Problems, Emotionally Reactive, Somatic Complaints, and Withdrawn. In addition, analyses of the CBCL/1.55 identied a syndrome designated as Sleep Problems, which consists of sleep-related items that are not assessed by the CTRF. A childs score on a syndrome is obtained by summing the ratings for the items that comprise the syndrome. Scores for each item of each syndrome and the total score for each syndrome are displayed on a prole, as shown for 30month-old Kenny in Figure 2. The prole displays Kennys scores in relation to scores for the 700 children from the CBCL/1.55 normative sample. For each syndrome scale, the broken lines printed across the prole indicate a nor230

mal range (93rd percentile), a borderline clinical range (93rd to 97th percentile), and a clinical range (97th percentile). Figure 2 indicates that Kennys mothers ratings yielded scores in the clinical range (above the top broken line) on the Withdrawn syndrome and in the borderline clinical range (between the broken lines) on the Emotionally Reactive syndrome. CBCL/1.55 and CTRF Aggregate Scales The ASEBA preschool forms also yield three aggregate (broad-band) scores. When the preschool syndromes were submitted to a second order-factor analysis, two global groupings emerged, which are labeled Externalizing and Internalizing. Similar scales have been found in multivariate analyses of childrens behavioral/emotional problems for many different instruments over several decades of research [Achenbach and Rescorla, 2000]. For the preschool

forms, the Emotionally Reactive, Anxious/Depressed, Somatic Complaints, and Withdrawn syndromes loaded on the Internalizing scale, whereas the Aggressive Behavior and Attention Problems syndromes loaded on the Externalizing scales Finally, all the items on the CBCL/ 1.55 and the CTRF are summed to yield a Total Problems score for each form. For Internalizing, Externalizing, and Total Problems, the clinical range is dened as T scores 64 (about the 90th percentile), the borderline range consists of T scores from 60 to 63 (84th to 90th percentiles), and the normal range consists of scores below the 84th percentile (T 60). The reason for selecting lower clinical and borderline range cutpoints than those on the syndromes is that the Internalizing, Externalizing, and Total Problems scales encompass more numerous and more diverse problems than any single syndrome scale.

MRDD RESEARCH REVIEWS

ASSESSMENT

OF

YOUNG CHILDREN USING ASEBA

RESCORLA

Fig. 2.

Kenny Randalls CBCL syndrome prole.

DSM-Oriented Scales Because practitioners and researchers are often expected to evaluate childrens problems in terms of diagnostic categories, there is a need to link empirically based assessment with such categories. The empirically based paradigm takes a bottom-up approach whereby syndromes are derived statistically to reect patterns of problems that are found to co-occur in large samples of individuals rated by various kinds of informants. The DSM, by contrast, takes a topdown approach whereby experts formulate diagnostic categories and then select symptoms and other criteria for dening each category. Although numerous studies have shown statistically signicant associations between DSM diagnoses and empirically based syndromes, the obtained associations vary greatly according to the basis for making DSM diagnoses, the sources of data, the particular diagnostic categories, and the analytic methods (e.g., Edelbrock and
MRDD RESEARCH REVIEWS

Costello, 1988; Rey and Morris-Yates; 1992; Gould et al., 1993; Arend et al., 1996; Kasius et al., 1997]. To help practitioners and researchers view the problem items of the CBCL/1.55 and CTRF in relation to DSM diagnostic categories as well as empirically based syndromes, we constructed DSM-oriented scales [Achenbach et al., 2000]. Sixteen experienced child psychiatrists and psychologists from 10 cultures were invited to rate each of the CBCL/1.55 and CTRF problem items as not consistent, somewhat consistent, or very consistent with each of nine DSM diagnostic categories. Items that were rated as very consistent with a DSM category by at least 10 of the 16 raters (63%) were deemed to be sufciently consistent with DSM categories to be included in the DSM-oriented scales. Because of major overlaps in DSM diagnostic criteria, as well as in the obtained ratings of the problem items, the
OF

nine DSM categories were collapsed into the following ve scales: Affective Problems (including Major Depressive Disorder and Dysthymic Disorder), Anxiety Problems (including Generalized Anxiety Disorder, Separation Anxiety Disorder, and Specic Phobia), Attention Deficit/Hyperactivity Problems (including HyperactiveImpulsive and Inattentive types), Pervasive Developmental Problems (including Aspergers Disorder and CBCL/1.55 versus CTRF). There are some differences between the items of the CBCL/1.55 and CTRF versions of the DSM-oriented scales. For example, the CTRF version of the Attention Decit/Hyperactivity Problems scale has more items than the CBCL/1.55 version, because more attention problem items are appropriate for rating by daycare providers and teachers than by parents and parent surrogates. The proles of DSM-oriented scales are analogous to the proles of empirically based syndromes, as illus231

ASSESSMENT

YOUNG CHILDREN USING ASEBA

RESCORLA

Fig. 3.

Kenny Randalls CBCL DSM-oriented prole.

trated by the prole of DSM-oriented scales shown for Kenny in Figure 3. Like the scales for the empirically based syndromes, the score for each DSM-oriented scale is computed by summing the 0 12 ratings of the items that comprise the scale. The proles of DSM-oriented scales display childrens scores in relation to the same national normative samples as for the empirically based scales, with percentiles, T scores, and normal, borderline, and clinical ranges displayed in the same way as for the empirically based scales. As can be seen in Figure 3, Kennys DSM-oriented prole for the CBCL, completed by his mother, yielded scores in the clinical range for both the Affective Problems and the Pervasive Developmental Problems scales. It should be noted that a high score on a DSM-oriented scale is not directly equivalent to a DSM diagnosis for the following reasons: (a) the items of the DSM-oriented scales 232

do not correspond precisely to DSM criteria, (b) the DSM-oriented scales are quantitative, whereas the DSM is based on yes-or-no judgments, and (c) DSM criteria are the same regardless of the informant who provides the data, whereas the DSM-oriented scales are normed based on the type of informant. However, the DSM-oriented scales are useful in suggesting diagnoses that may be relevant to consider for a particular child. INTER-RATER RELIABILITY: THE IMPORTANCE OF CROSSINFORMANT COMPARISONS Children often behave differently in different settings and with different interaction partners. Furthermore, two people seeing the same behavior may report it differently. Meta-analyses of correlations between scores obtained from informants reports of problems for 1.5 19-year-olds found an average correla

tion of 0.60 between pairs of informants who play similar roles with respect to children, such as pairs of parents and pairs of teachers [Achenbach et al., 1987]. Between informants who play different roles with respect to children, such as parents versus teachers and teachers versus mental health workers, the correlations averaged 0.28. In the 1999 National Survey, a mean correlation of 0.61 was found when z-transformed correlations between scores obtained from 72 pairs of mothers and fathers on all CBCL syndromes, DSM-oriented scales, and aggregate scores were averaged. When z-transformed correlations between scores provided by 102 pairs of caregivers/teachers across all CTRF scales were averaged, the mean correlation was 0.65. When z-transformed correlations between scores obtained on all CBCL/ CTRF scales from 226 parent caregiver/ teacher pairs were averaged, the mean

MRDD RESEARCH REVIEWS

ASSESSMENT

OF

YOUNG CHILDREN USING ASEBA

RESCORLA

Fig. 4.

Kenny Randalls cross-informant bar graph.

correlation was 0.40. For Total Problems scores, the mean cross-informant correlations were 0.65 for pairs of parents on the CBCL, 0.72 for pairs of teachers on the CTRF, and 0.50 for parents with teachers on the CBCL and CTRF, respectively [Achenbach and Rescorla, 2000]. Even though reports by each kind of informant may be reliable and valid in their own right, the modest agreement among informants indicates that no single informant can substitute for all others. Because childrens behavior varies from one context and interaction partner to another and because informants differ in what they notice, remember, and report, comprehensive assessment requires data from multiple informants. Therefore, the ASEBA preschool forms and proles provide several ways to systematically compare data from multiple informants. One way information from multiple informants is compared is by the crossinformant bar graph. Figure 4 illustrates a cross-informant bar graph of Kenny
MRDD RESEARCH REVIEWS

Randall for CBCLs completed by his mother and father and CTRFs completed by his two teachers. As can be seen in Figure 4, ratings by all four adults yielded scores in the clinical range for the Withdrawn syndrome and in the borderline range for the Emotionally Reactive syndrome. On the other hand, only teachers ratings placed Kenny in the clinical range for the Attention Problems syndrome and the borderline range for the Somatic Complaints syndrome. Kennys cross-informant bar graph for the DSM-oriented scales (not shown here) indicated that ratings by all four adults placed him in the clinical range on the Pervasive Developmental Problems scale, whereas only his teachers yielded elevated scores on the Anxiety Problems and Attention Decit Hyperactivity Problems scales. Another way that the ASEBA software facilitates cross-informant comparisons is by printing correlations between ratings by each pair of informants and comparing these correlations with correOF

lations obtained for large reference samples of similar pairs of informants. This enables users to evaluate how well particular informants agree with one another in rating a particular child. The software also prints side-by-side comparisons of the scores obtained from each informant on each item and each scale, enabling users to see at a glance the items and scales on which there are consistencies versus inconsistencies among the informants. PSYCHOMETRIC PROPERTIES OF THE ASEBA PRESCHOOL FORMS Reliability To assess testretest reliability of the CBCL/1.55, maternal ratings for 68 nonreferred children obtained about one week apart were compared using both correlations and t-tests. Similar analyses were done on CTRF ratings for 59 children. Across the various ASEBA syndromes and scales, most testretest cor233

ASSESSMENT

YOUNG CHILDREN USING ASEBA

RESCORLA

relations were in the 0.80s and 0.90s, with a mean r of 0.85 and 0.81 across all scores for the CBCL and CTRF, respectively. The Total Problems r was 0.90 for the CBCL and 0.88 for the CTRF. Testretest reliability for the LDS was 0.99 in a sample of 30 toddlers assessed over a one-week interval and 0.97 for 33 toddlers assessed over a onemonth interval [Achenbach and Rescorla, 2000]. When t-tests were used to compare CBCL and CTRF problem scores across the rst and second administrations, there was a small decline in scores (about 1% of the variance on both forms). A decline in scores when an assessment procedure is repeated after a short interval is called testretest attenuation. The testretest attenuation effects found for ASEBA forms are smaller than those typically obtained on structured diagnostic interviews [Edelbrock et al., 1985]. Validity Validity of the ASEBA preschool forms has been examined in several ways. In one set of analyses, ASEBA scores were compared for referred versus nonreferred children (N 563 for each) matched on age, gender, SES, and ethnicity. In regression analyses, the effect size (ES) for referral status ranged from 3% (DSM-oriented Anxiety Problems) to 25% (Total Problems) of the variance for CBCL scales and from 2% (Somatic Complaints) to 24% (Total Problems) of the variance for CTRF scales. Referred children obtained signicantly higher scores on all problem scales of the CBCL/1.55 and CTRF, with 16 of the 29 effects meeting Cohens [1988] standards for medium size (13% of the variance). Referred children also obtained higher scores on all but two items of the CBCL/1.55 and CTRF. Age effects on CBCL/1.55 scores were generally very small, and this is the reason that the same norms for the ASEBA preschool forms span the entire age 1.55. Gender effects were few and very small on the CBCL, and so genders were pooled for the norms. Because 8 of the 14 gender effects on the CTRF were signicant, albeit small (23% of the variance), separate CTRF norms for boys and girls were developed. There were no signicant SES effects on the CTRF, but small SES effects (13% of the variance) were found for 8 of the 15 CBCL scales. For the LDS, signicant gender effects for vocabulary score necessitated separate norms by gender. However, the 234

gender difference for average length of combinations was not signicant, so genders were pooled for this measure. Age differences were signicant for both LDS measures, which is why norms are provided for the three age groups for vocabulary score and for the two older age groups for average length of combinations. The relations between referral status and scores in the deviant range (above the borderline cutpoint) on each CBCL scale were analyzed using odds ratios (ORs) for the same demographicallymatched samples described above [Achenbach and Rescorla, 2000]. The OR indicates the odds of the child being in the referred versus the nonreferred group given a score in the deviant range on an ASEBA scale. The highest OR was for the DSM-oriented Pervasive Developmental Problems scale (OR 11); children with a Pervasive Developmental Problems T score 65 were 11 times more likely to be in the referred group. When the predictor was having at least one CBCL syndrome in the deviant range, the OR was 9; 77% of referred children but only 26% of nonreferred children had at least one CBCL syndrome with T score 65. Similarly, 57% of the referred group and 18% of the nonreferred group had a Total Problems score in the deviant range (T 60) on the CBCL (OR 6); the corresponding gures for the CTRF were 63% versus 18% and an OR 8. Criterion-related validity has also been demonstrated for the LDS [Achenbach and Rescorla, 2000]. Correlations between LDS vocabulary score and other language measures across 11 samples ranged from 0.56 to 0.87, with 11 of the 15 correlations 70. The language measures used in these 11 samples included number of different objects named on Bayley [Bayley, 1969], number of different pictures named on the Stanford Binet IV [Thorndike et al., 1986], number of pictures named on the Reynell Expressive Language Scale [Reynell and Gruber, 1985], total scores on the Reynell Expressive and Receptive Language Scales, language score on the Mullen Scales [Mullen, 1993], mean length of utterance from a speech sample, parent report on a Spanish version of the MacArthur CDI [Fenson et al., 1993], and Communication scores on the Vineland Adaptive Behavior Scale [Sparrow et al., 1984]. Sensitivity of the LDS for identifying children found to be language-delayed on formal testing was 87% in one study [Rescorla, 1989], 91% in another study [Klee et al., 1998], and

94% in a third study [Rescorla and Alley, 2000], based on the scores of either the Reynell Expressive Language Scale [Rescorla, 1989; Rescorla and Alley, 2000] or the Mullen Scales [Klee et al., 1998]. Specicity, or the percent of typically developing children who passed the LDS, was 86, 87, and 67% across the same three studies. In the Rescorla and Alley [2001] study, an OR of 34 was found: Children who failed the LDS by having fewer than 50 words or no word combinations at 24 months were 34 times more likely to obtain a score 10th percentile on the Reynell Expressive Language Scale than toddlers who passed the LDS. UTILIZING ASEBA FORMS IN SERVICE SETTINGS The CBCL/1.55 and CTRF are appropriate for use in diverse service settings. The standardized rating forms can be completed by people such as parents, parent-surrogates, daycare providers, and preschool teachers when children are evaluated in various contexts. In many settings, practitioners can have parents routinely complete the CBCL/1.55 at home prior to their scheduled appointment. Once parental consent is obtained, CTRFs can also be obtained from teachers and daycare providers. By looking at the childs computer-scored or hand-scored proles and the LDS scoring form, the practitioner can quickly see whether the child is in the normal range with respect to problem scores and language development. If the child is not in the normal range, the practitioner can use the prole and the descriptive comments written on the CBCL/1.55 as a basis for interviewing the parent about the childs functioning. If syndrome scores are very elevated or if there is evidence for deviance in multiple areas, the practitioner may elect to conduct a more extensive evaluation or to refer the family to a specialist. If referral to a specialist is indicated, the completed CBCL/ 1.55, LDS, and prole forms can be sent, with parents consent, to the specialist to provide intake information. The ASEBA preschool forms are a useful component of the intake and evaluation process in special education, child development, and mental health settings. Whether referrals arise from concerns about language development, cognitive functioning, stressful experiences such as abuse or neglect, or behavior problems, it is always helpful to have a standardized picture of the childs behavioral/emotional problems and language development, as seen by people who live with

MRDD RESEARCH REVIEWS

ASSESSMENT

OF

YOUNG CHILDREN USING ASEBA

RESCORLA

the child. Such a picture is also essential for evaluations related to adoption and foster placement. USING ASEBA FORMS IN RESEARCH ASEBA preschool forms have been widely used in research. One such use is to evaluate effects of interventions. The CBCL/1.55, LDS, and CTRF can be administered before, during, and after interventions to evaluate changes and outcomes, as exemplied in studies of interventions for low birth weight children from low SES families [Brooks-Gunn et al., 1994] and for children at-risk because of psychosocial adversities [Kitzman et al., 1997]. Many studies have demonstrated the sensitivity of ASEBA preschool scales to the effects of interventions. For example, Blair [2002] reported ndings from the Infant Health and Development Program for low birth weight children, which involved both home visiting and child care components. Twothird of the mothers had a high-school education or less, and 63% of the sample was AfricanAmerican or Hispanic. Results indicated that infants rated by their mothers as high in negative emotionality showed signicant intervention effects by age 3, with half as many intervention children scoring in the deviant range (T 64) on CBCL/23 Internalizing or on Externalizing problems as control group children. When an elevated score on both Internalizing and Externalizing was used as the outcome measure, control group children who had showed negative emotionality as infants were four times more likely to be deviant at age 3 than intervention children. ASEBA forms can be combined with other assessment instruments in large research studies investigating multiple aspects of development. For example, the NICHD Study of Early Child Care used the ASEBA preschool forms in its longitudinal investigation of more than 1,000 children in 10 sites ([NICHD Early Child Care Research Network, 1998]. In a study reported by Belsky and Fearon [2002] from this project, age 3 scores on the CBCL/23 showed a signicant linear dose-response relation with level of cumulative risk, a composite based on measures of SES, race, maternal education and psychopathology, parenting stress, social support, marital status, and marital quality. In another report from this larger study, La Paro et al. [2002] examined developmental predictors of eligibility for special services as determined by a medical professional. Results indicated that high CBCL/23 scores
MRDD RESEARCH REVIEWS

and problematic health history from 16 to 36 months were the only signicant predictors of special needs identication out of 9 candidate predictors tested, with the full model correctly predicting status of 73% of the children. Scores on ASEBA forms have also been used in studies designed to analyze the mechanisms by which a general risk factor such as low SES produces negative outcomes for children. For example, Linver et al. [2002] found that the effect of low SES on CBCL behavioral/emotional problems was mediated by two major factors, low cognitive stimulation and poor parenting practices. Poor parenting practices, but not low cognitive stimulation, were in turn mediated by povertyrelated maternal emotional distress. Longitudinal studies from preschool to school-age can also be done with ASEBA forms. As an example of longitudinal analyses of this sort, the preschool CBCL was used in a study of children who had participated in an experimental short-term intervention designed to facilitate the development of low birth weight infants [Achenbach et al. 1993]. Children were assessed at ages 2 and 3 by having the parents complete the preschool CBCL and at older ages by having them complete the CBCL/6 18. The longitudinal correlations between scores for the Aggressive Behavior syndrome at age 2 and later ages were substantial, ranging from 0.65 between ages 2 and 4 to 0.50 between ages 2 and 9 [Achenbach and Rescorla, 2000]. The ASEBA preschool forms have also been used for research in other cultures. For example, a Dutch translation of the CBCL/23 was used with 420 children in a community sample, 426 clinic referred children, and 1,306 twin pairs [Koot et al., 1997]. Factor analysis of item scores yielded the same six factors across the three samples: Oppositional, Withdrawn/Depressed, Aggressive, Anxious, Overactive, and Sleep Problems. These are quite similar though not identical to the American syndromes, despite the fact that different factor analytic methods were used. A second-order factor analysis yielded Externalizing (Aggressive, Oppositional, and Overactive syndromes) and Internalizing (Anxious, Withdrawn/Depressed) aggregate scales, similar to those found in the USA. Testretest reliabilities were high (0.87 for Total Problems), and motherfather inter-rater reliability was 0.66, both comparable with data from the USA. When demographically-matched referred and nonreferred children were compared, all problem scales were signicantly higher in the referred children (ES of 27%
OF

for Total Problems). As in the USA, age, gender, and SES effects were signicant but quite small. A deviant score on the Total Problems scale (85th percentile) yielded an OR of 8.7 for clinic versus nonreferred status, with 73% of the children being correctly classied. Correlations of CBCL/ 23 scores with comparable scores from the Dutch CBCL/4 16 two years later were 0.41, 0.54, and 0.60 for Internalizing, Externalizing, and Total Problems scores, respectively, which are quite similar to American ndings. Finally, when scores based on the Dutch scales were compared with those obtained for the same children using the American scales, correlations were 0.90 for Internalizing and 0.97 for Externalizing, and they ranged from 0.80 to 1.00 for individual syndromes. The psychometric properties of an Arabic translation of the CBCL/23 have been examined in a sample of 694 three-years-old children in the United Arab Emirates (Yunis F, Eapen V, and Zoubeidi, T (personal communication) The CBCL and several other measures were administered by an interviewer in the home. Items with a reported prevalence of less than 5% were excluded from further analyses, leaving 71 problem items on six empirically based syndromes. Internal consistency indexed by Cronbachs alpha was very high for the Total Problems score (0.93), and quite high for Internalizing (0.76) and Externalizing (0.88); alpha values for syndromes ranged from 0.55 for Withdrawal to 0.84 for Aggressive Behavior. Testretest reliability over a one week interval was 0.82 for Total Problems, and ranged from 0.60 to 0.75 for the six syndromes. Children with a family history of psychiatric illness or psychosocial stress had signicantly higher CBCL scores. When children scoring two standard deviations above the mean on Total Problems (high scorers) and a contrasting low scoring group were interviewed by a clinician unaware of their CBCL scores, 94% of high scorers and 0% of low scorers were judged to have a clinically significant problem. Thus, the CBCL/23 demonstrated excellent external validity, despite the many cultural differences between the USA and the United Arab Emirates. CAUTIONS AND LIMITATIONS ASEBA preschool forms are designed to improve clinical assessment and services by providing standardized rating forms usable by multiple informants who see the child in different contexts. ASEBA forms are intended to be integrated with rather than to substitute for 235

ASSESSMENT

YOUNG CHILDREN USING ASEBA

RESCORLA

cognitive assessment, physical examinations, psychiatric diagnosis, observations, and developmental histories. Each of these is essential for comprehensive assessment and understanding of young children. Detailed knowledge of family functioning and relationships is also needed to understand both the liabilities and strengths that characterize families. Evaluation of cognitive, physical, and family functioning requires considerable time and effort by skilled clinicians. At low cost, empirically based assessment adds much useful data without requiring clinicians time. However, clinicians must ultimately integrate empirically based assessment data with other data. A major challenge is to help clinicians weigh and integrate various kinds of data to make the best decisions for each child. STRENGTHS AND BENEFITS OF THE ASEBA APPROACH The ASEBA preschool forms are standardized assessment instruments that are user-friendly, cost-effective, and usable by a wide range of professionals in different settings. The CBCL/1.55 and CTRF yield scores on normed empirically based syndrome scales, DSM-oriented scales, and Internalizing, Externalizing and Total Problems aggregate scales. The norms are based on large and diverse general population samples. Findings based on ASEBA forms for preschool children have been reported in more than 200 published research studies. The CBCL/1.55 also contains the LDS, which can be used for reliable and valid screening of language delay in children from 18 months to 3 years of age. Because there is no single gold standard for assessing behavioral and emotional problems, the empirically based approach employs parallel assessment forms to obtain data from multiple informants who see children in different contexts and from different perspectives. ASEBA software systematically compares item and scale scores obtained from up to eight informants per child. The software also computes correlations between ratings by different informants and compares these correlations with those obtained for large reference samples of informants. Because ASEBA DSM-oriented scales are scored from the same rating forms and are normed on the same national samples as the empirically based syndromes, users can directly compare reports of childrens problems grouped according to DSM categories and grouped according to empirically based syndromes. Categorical cutpoints on all 236

the scales enable users to categorize children as being in the normal, borderline, or clinical range. Clinical applications of empirically based instruments include routine use in health care settings to identify children with deviant problem scores and to advise parents, daycare providers, and teachers who are concerned about a childs problems. Other clinical applications include use of the empirically based instruments to evaluate children seen in special education, child development, and mental health settings; to evaluate changes during interventions, as well as outcomes following interventions; and to evaluate children involved in adoption and foster placement. Research applications for ASEBA preschool forms include identifying correlates of different kinds of problems, tracking the developmental course of problems, testing the efcacy of pharmacotherapy and other interventions, identifying etiological factors, and evaluating prevention programs. Because the CBCL/1.55 and CTRF have counterpart forms for assessing school-age children and adolescents, it is easy to test predictive relations between scores obtained on the forms for ages 1.55 and the forms for ages up to 18. f REFERENCES
Achenbach TM. 1966. The classication of childrens psychiatric symptoms: a factor-analytic study. Psychol Monogr 80:137. Achenbach TM. 1991. Integrative Guide for the 1991 CBCL/4 18, YSR, and TRF Proles. Burlington, VT: Department of Psychiatry, University of Vermont. Achenbach TM. 1992. Manual for the Child Behavior Checklist/23 and 1992 Prole. Burlington, VT: Department of Psychiatry, University of Vermont. Achenbach TM. 1997. Guide for the Caregiver Teacher Report Form for Ages 25. Burlington, VT: Department of Psychiatry, University of Vermont. Achenbach TM, Dumenci L, Rescorla LA. 2000. Ratings of Relations Between DSM-IV Diagnostic Categories and Items of the CBCL/ 1125 and CTRF. Burlington, VT: Department of Psychiatry, University of Vermont. (Available at: www.ASEBA.org). Achenbach TM, Edelbrock C. 1981. Behavioral problems and competencies reported by parents of normal and disturbed children aged four to sixteen. Monogr Soc Res Child Dev 46:1 82. Achenbach TM, Edelbrock C, Howell CT. 1987. Empirically based assessment of the behavioral/emotional problems of 23-year-old children. J Abnorm Child Psychol 15:629 650. Achenbach TM, Howell CT, Aoki M, et al. 1993. Nine-year outcome of the Vermont intervention program for low birthweight infants. Pediatrics 91:4555. Achenbach TM, McConaughy SH, Howell CT. 1987. Child/adolescent behavioral and emotional problems: implications of cross-infor

mant correlations for situational specicity. Psychol Bull 101:213232. Achenbach TM, Newhouse PA, Rescorla LA. 2004. Manual for the ASEBA Older Adult Forms & Proles. Burlington, VT: Research Center for Children, Youth, and Families, University of Vermont. Achenbach TM, Rescorla LA. 2000. Manual for ASEBA Preschool Forms & Proles. Burlington, VT: Research Center for Children, Youth, and Families, University of Vermont. Achenbach TM, Rescorla LA. 2001. Manual for the ASEBA School-age Forms & Proles. Burlington, VT: Research Center for Children, Youth, and Families, University of Vermont. Achenbach TM, Rescorla LA. 2002. Manual for the ASEBA Adult Forms & Proles. Burlington, VT: Research Center for Children, Youth, and Families, University of Vermont. American Psychiatric Association. 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th Ed. Washington, D.C.: American Psychiatric Association. Angold A, Egger HL. 2004. Psychiatric diagnosis in preschool children. In: DelCarmen-Wiggins R, Carter A, editors. Handbook of Infant, Toddler, and Preschool Mental Health Assessment. Oxford, UK: Oxford University Press. Arend R, Lavigne JV, Rosenbaum D, et al. 1996. Relation between taxonomic and quantitative diagnostic systems in preschool children: emphasis on disruptive disorders. J Clin Child Psychol 25:388 397. Barnard K, Wieder S, Carter A, et al. 2002. The DC: 0 3 Assessment Protocol Project: Final Report to the Commonwealth Fund. Washington, D.C.: Zero to Three: National Center for Infants, Toddlers, and Families. Bayley N. 1969. Bayley Scales of Infant Development. New York: Psychological Corporation. Belsky J, Fearon RM. 2002. Infant-mother attachment security, contextual risk, and early development: a moderational analysis. Dev Psychopathol 14:293310. Be rube RL, Achenbach TM. 2005. Bibliography of Published Studies Using ASEBA Instruments: 2005 edition. Burlington, VT: Research Center for Children, Youth, and Families, University of Vermont. Blair C. 2002. Early intervention for low birthweight, preterm infants: the role of negative emotionality on the specication of effects. Dev Psychopathol 14:311332. Brooks-Gunn J, McCarton CM, Casey PH, et al. 1994. Early intervention in low-birth-weight premature infants. Results through age 5 years from the Infant Health and Development Program. JAMA 272:12571262. Buss AH, Plomin R. 1984. Temperament: Early Developing Personality Traits. Hillsdale, NJ: Lawrence Erlbaum. Carey WB, McDevitt SC. 1978. Revision of the infant temperament questionnaire. Pediatrics 61:735739. Cohen J. 1988. Statistical Power Analysis for the Behavioral Sciences, 2nd Ed. New York: Academic Press. Dunitz M, Scheer PJ, Kvas E, et al. 1996. Psychiatric diagnoses in infancy: a comparison. Infant Ment Health J 17:1223. Earls F. 1982. Application of DSM-III in an epidemiological study of preschool children. Am J Psychiatry 139:242243. Edelbrock C, Costello AJ. 1988. Convergence between statistically derived behavior problem

MRDD RESEARCH REVIEWS

ASSESSMENT

OF

YOUNG CHILDREN USING ASEBA

RESCORLA

syndromes and child psychiatric diagnoses. J Abnorm Child Psychol 16:219 231. Edelbrock C, Costello AJ, Dulcan MK, et al. 1985. Age differences in the reliability of the psychiatric interview of the child. Child Dev 56:265275. Fenson L, Dale PS, Reznick JS, et al. 1993. Guide and technical manual for the MacArthur Communicative Inventories. San Diego, CA: Singular Press. Garcia Coll CT, Halpern LF, Vohr BR, et al. 1992. Stability and correlates of change of early temperament in pre-term and full-term infants. Infant Behav Dev 15:137153. Goldsmith HH, Rothbart MK. 1991. Contemporary instruments for assessing early temperament by questionnaires and in the laboratory. In: Strelau J, Angleitner A, editors. Explorations in Temperament. New York: Plenum. Gould MS, Bird H, Jaramillo BS. 1993. Correspondence between statistically derived behavior problem syndromes and child psychiatric diagnoses in a community sample. J Abnorm Child Psychol 21:287313. Hollingshead A. 1975. Four-Factor Index of Social Status. New Haven, CT: Department of Sociology, Yale University. Kanner L. 1943. Autistic disturbances of affective contact. Nerv Child 2:217250. Kasius MC, Ferdinand RF, van den Berg H, et al. 1997. Associations between different diagnostic approaches for child and adolescent psychopathology. J Child Psychol Psychiatry 38: 625 632. Keenan K, Wakschlag LS. 2000. More than the terrible twos: the nature and severity of behavior problems in clinic-referred preschool children. J Abnorm Child Psychol 28:33 46. Kitzman H, Olds D, Henderson C, et al. 1997. Effect of prenatal and infancy home visitation by nurses on pregnancy outcomes, childhood injuries, and repeated childbearing. JAMA 278:644 652. Klee T, Carson DK, Gavin WJ, et al. 1998. Concurrent and predictive validity of an early

language screening program. J Speech Lang Hear Res 41:627 641. Koot HM, van den Oord EJ, Verhulst FC, Boomsma DI. 1997. Behavioral and emotional problems in young preschoolers: crosscultural testing of the validity of the Child Behavior Checklist/23. J Abnorm Child Psychol 25:183196. La Paro KM, Olsen K, Pianta RC. 2002. Special education eligibility: developmental precursors over the rst three years of life. Except Child 69:55 66. Lavigne JV, Arend R, Rosenbaum D, et al. 1988. Psychiatric disorders with onset in the preschool years. I. Stability of diagnoses. J Am Acad Child Adolesc Psychiatry 37:1246 1254. Lieberman AF, Barnard KE, Wieder S. 2004. Diagnosing infants, toddlers, and preschoolers: the zero to three diagnostic classication of early mental disorders. In: DelCarmen-Wiggins R, Carter A, editors Handbook of Infant, Toddler, and Preschool Mental Health Assessment. Oxford, U.K.: Oxford University Press. Linver MR, Brooks-Gunn J, Kohen DE. 2002. Family processes as pathways from income to young childrens development. Dev Psychol 38:719 734. Mullen EM. 1993. Mullen Scales of Early Learning. Circle Pines, MN: American Guidance Service. NICHD Early Child Care Research Network. 1998. Early child care and self-control, compliance, and problem behavior at 24 and 36 months. Child Dev 69:11451170. NICHD Early Child Care Research Network. 1994. Child care and development: the NICHD study of early child care. In: Friedman SL, Haywood HC, editors. Developmental Follow-Up: Concepts, Domains, and Methods. New York: Academic Press. p 377 395. Rescorla L. 1989. The Language Development Survey: a screening tool for delayed language in toddlers. J Speech Hear Disord 54:587 599.

Rescorla L, Alley A. 2001. Validation of the Language Development Survey: a parent report tool for identifying language delay in toddlers. J Speech Lang Hear Res 44:434 445. Rey JM, Morris-Yates A. 1992. Diagnostic accuracy in adolescents of several depression rating scales extracted from a general purpose behavior checklist. J Affect Disord 26:716. Reynell J, Gruber C. 1985. Reynell Developmental Language Scales. Los Angeles, CA: Western Publishing Company. Rothbart MK. 1981. Measurement of temperament in infancy. Child Dev 52:569 578. Seifer R, Sameroff AJ, Barrett LC, et al. 1994. Infant temperament measured by multiple observations and mother report. Child Dev 65:1478 1490. Shaw DS, Keenan K, Vondra JI, et al. 1997. Antecedents of preschool childrens internalizing problems: a longitudinal study of low-income families. J Am Acad Child Adolesc Psychiatry 36:1760 1767. Stifter CA, Wiggins CN. 2004. Assessment of disturbances in emotion regulation and temperament. In: DelCarmen-Wiggins R, Carter A, editors. Handbook of Infant, Toddler, and Preschool Mental Health Assessment. Oxford, U.K.: Oxford University Press. Sparrow S, Cicchetti DV, Balla D. 1984. Vineland Adaptive Behavior Scales-Revised. Circle Pines, MN: American Guidance Service. Thomas JM, Guskin KA. 2001. Disruptive behavior in young children: what does it mean? J Am Acad Child Adolesc Psychiatry 40:44 51. Thorndike RL, Hagen EP, Sattler JM. 1986. Stanford-Binet Intelligence Scale, 4th ed. Chicago, IL: Riverside Press. Zero to Three: National Center for Infants, Toddlers, and Families. 1994. Diagnostic Classication 0 3. Diagnostic Classication of Mental Health and Developmental Disorders of Infancy and Early Childhood. Washington, D.C.: Zero to Three.

MRDD RESEARCH REVIEWS

ASSESSMENT

OF

YOUNG CHILDREN USING ASEBA

RESCORLA

237