Table of Contents | Previous | Next |
INTERNALIZING/EXTERNALIZING BEHAVIOR PROBLEMS
NATIONAL INSTITUTE OF CHILD HEALTH AND HUMAN DEVELOPMENT
STUDY OF EARLY CHILD CARE AND YOUTH DEVELOPMENT
Measure: Behavior and Emotional Problems Scales (includes Internalizing, Externalizing, and Total Behavior Problems, as well as more behavior specific subscales)
Background
The National Institute of Child Health and Human Development Study of Early Child Care and Youth Development (NICHD SECC-YD) used the Achenbach Child Behavior Checklist (CBCL) to measure behavior problems (Total, Externalizing, and Internalizing) longitudinally. The CBCL is a much-used measure of negative child behavior and is widely considered reliable and valid in the field. Two versions of the CBCL were used. Children were assessed with the CBCL for ages 2 to 3 (CBCL/2-3) at 24 and 36 months, and then assessed at 54 months and during their kindergarten year with the CBCL for ages 4 to 18 (CBCL/4-18).
The NICHD SECC-YD was initiated by and is funded by the NICHD and is directed by a steering committee and advisory board. The research team comprises researchers from a variety of child development and policy disciplines and represents over 24 institutions in the United States and London.
Population Assessed
The NICHD SECC-YD is not a nationally representative sample. Participants were recruited shortly after the birth of a child, in 10 sites across the United States. Within selected 24-hour periods, all women who had given birth in specific hospitals were screened for willingness to participate and eligibility. The conditional sampling plan for inclusion was based on the creation of a sample where 60 percent of mothers planned to work or go to school full-time in the child’s first year, 20 percent planned to go part-time in the child’s first year of life, and 20 percent planned to stay at home with their child. Families were also selected to reflect demographic diversity of the sites (e.g., economic, educational, and ethnic), and both single- and two-parent families were included. Families in which mothers were less than 18 years of age, planned on leaving the study site within 3 years, or were not conversant in English were excluded from the study. Families with children who were born with obvious physical or mental disabilities or who remained at the hospital for more than 7 days were also excluded from the study.
Of the original 8,986 women who were first approached in the hospitals at the time of their respective child’s birth, 5,416 met the criteria and agreed to be contacted 2 weeks later, as well as met eligibility requirements (see above). From this number, 3,015 women were “conditionally, randomly” sampled to meet proposed sample specifications (see above), and 1,526 met eligibility requirements and agreed to participate. When contacted, 1,364 mothers participated in the first round of data collection (NICHD Early Child Care Research Network, 1996, 1997, 2003a). The original sample was diverse, including 24 percent minority children. Similarly, maternal characteristics showed some variability, with 11 percent of mothers not completing high school, and 14 percent of the sample being composed of single mothers. On average, families showed incomes of 3.24 times the poverty rate (i.e., 1.0 being poverty). Because the NICHD SECC-YD is an ongoing, longitudinal study, some attrition was expected. From the original sample of 1,364 when the child was 1 month old, the most recent NICHD SECC publication reports that 1,058 (NICHD Early Child Care Research Network, 2003a) families were still involved in the study once the children had reached kindergarten. The most current available sample data (children in kindergarten) significantly differed from the sample at the study’s inception when children were 1 month old. Mothers remaining in the sample when the children reached kindergarten were found to be more educated, more likely to have a husband, and less likely to be Black, non-Hispanic than those in the original sample. The remaining sample also showed a higher income-to-needs ratio.2
Periodicity
Maternal reported and caregiver/teacher reported CBCL/2-3 (that is, CBCL for children ages 2 to 3) questionnaires were given in their entirety when the children were at the ages of 24 (1993) and 36 (1994) months. The CBCL/4-18 was given at three time points and used multiple reporters. At 54 months (1996) and during the child’s first-grade year, both maternal and paternal, as well as teacher reported CBCL questionnaires, were given. Only mother-reported CBCL scales were collected during the children’s kindergarten year.
Subscales/Components
Both age versions of the CBCL have multiple scales. The most relevant to this construct are the CBCL summary scales, Internalizing, Externalizing, and Total Problem Behaviors, but various other behaviorally specific subscales can be derived when the CBCL is given in its entirety (i.e., emotionally reactive, anxious/depressed, somatic complaints, withdrawn, and aggressive behavior; sleep problems; ;affective problems; anxiety problems; pervasive developmental problems; attention deficit/hyperactivity problems; and oppositional defiant problems). The CBCL ratings, for which each of these subscales can be derived, were assessed at 24, 36, and 54 moths and at kindergarten and first grade.
Because of copyright restrictions, item-level information regarding which CBCL items were used for the Internalizing, Externalizing, and Total Behavior Problems scales in the NICHD SECC-YD cannot be reproduced. The NICHD SECC-YD probably used the specific items in the configuration suggested by the authors, which are available in the CBCL manuals (Achenbach & Rescorla, 2000b). The CBCL is a widely used measure of these constructs and is regarded as a reliable and valid measure of these constructs. A detailed review of reliability and validity information for the CBCL 1½-5 (an updated version of the CBCL/2-3) is available in the work by Bridges and colleagues (2003).
Procedures for Administration
Each version of the CBCL includes approximately 100 items, in which the caregiver/teacher, mother, and/or father (depending upon assessment point, see periodicity above) are required to answer questions regarding characteristic behavior of the child over the past 2 months. The location in which the CBCL was administered differed by point in time and reporter. At 24 and 36 months, mothers responded in the lab, while caregivers responded from the child care environment. At 54 months, mothers responded in the home, fathers in the lab, and caregivers in child care. In first grade the venue for mothers and fathers reversed, with mothers responding in the lab and fathers from home. Caregivers responded from the after-school care environment. For the kindergarten collection, mother report of the CBCL was classified as being obtained at an “Other” location (NICHD Early Child Care Research Network, 2003c). Although the NICHD SECC-YD does not state the amount of time needed, the CBCL manual reports that the CBCL takes between 10 and 15 minutes to administer (Achenbach & Rescorla, 2000a).
Psychometrics/Data Quality
Because of copyright restrictions and restricted access data, study-level psychometrics are not publicly available for the CBCL (NICHD Early Child Care Research Network, 2003c, 2003d) . All psychometric information on the CBCL is based on information available from the CBCL technical manual.
Reliability
The CBCL, based on manual information provided by the NICHD SECC-YD authors shows strong test-retest reliability (.71 -.93), and inter-parent agreement (.63 at age 2, .60 at age 3) for the 2–3 year version of the measure (NICHD Early Child Care Research Network, 2003b). It is unclear whether the cited test-retest correlations are across all subscales or the Total, Externalizing, and Internalizing scales, or whether inter-parent ratings differed across scales. Psychometrics for children at older ages are not summarized in the NICHD SECC-YD Phase II Instrument document, but are reported to be strong by the NICHD SECC documentation (NICHD Early Child Care Research Network, 2003c, 2003d). Consulting the CBCL/4-18 manual confirms this, showing inter-parent agreement ranging from .57 to.71 for Internalizing, .70 to .86 for Externalizing, and .69 to.82 for Total Problem Behaviors, across ages. (Achenbach, 1991).
Validity
The CBCL is widely used and is a reportedly highly valid measure of internalizing, externalizing, and problem behaviors (as well as the more specific problems assessed in the detailed subscales). The CBCL manual (based on more recent version of CBCL) reports strong evidence of convergent, discriminant, and predictive validity of the measure and has been useful in discriminating clinical levels of internalizing and externalizing problems from nonclinical (Achenbach & Rescorla, 2000b). Validity information based on the NICHD SECC sample is unavailable.
Languages Available
Available NICHD SECC-YD documentation does not note whether the CBCL was administered in any language other than English. However, the CBCL has been used in over 700 cross-cultural studies, spanning many languages (Achenbach System of Empirically Based Assessment, 2003).
Items Included
The data are restricted.
References and Source Documents
Achenbach System of Empirically Based Assessment. (2003). Cross-Cultural Applications of ASEBA. ASEBA. Retrieved July, 31, 2003, from http://www.aseba.org/ABOUTUS/cross_cultural.html
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry.
Achenbach, T. M., & Rescorla, L. A. (2000a). Manual for the ASEBA preschool forms & profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth, & Families.
Achenbach, T. M., & Rescorla, L. A. (2000b). Manual for the ASEBA preschool forms & profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth, & Families.
Bridges, L. J., Berry, D. J., Calkins, J., Zaslow, M. J., Margie, N. G., Cochran, S. W., & Ling, T. J. (2003). Early Childhood Measures Profiles. In M. J. Zaslow (Ed.), Early Childhood Measures Profiles. Washington, DC: Child Trends.
Gresham, F. M., & Elliot, S. N. (1990). The Social Skills Rating System. Circle Pines, MN: American Guidance Systems.
NICHD Early Child Care Research Network. (1996). Characteristics of Infant Child Care: Factors Contributing to Positive Care Giving. Early Childhood Research Quarterly, 11, 269–306.
NICHD Early Child Care Research Network. (1997). Familial factors associated with characteristics of non-maternal care for infants. Journal of Marriage and the Family, 59, 389–408.
NICHD Early Child Care Research Network. (2003a). Does amount of time in child care predict socio-emotional adjustment during the transition to kindergarten? Child Development, 74, 976–1005.
NICHD Early Child Care Research Network. (2003b). NICHD Study of Early Child Care and Youth Development. Research Triangle Institute. Retrieved August, 8, 2003, from http://secc.rti.org/summary.cfm
NICHD Early Child Care Research Network. (2003c). NICHD Study of Early Child Care: Phase I Instrument Document. Research Triangle Institute. Retrieved August, 7, 2003, http://secc.rti.org/instdoc.doc
NICHD Early Child Care Research Network. (2003d). NICHD Study of Early Childcare and Youth Development: Phase II Instrument Document. Research Triangle Institute. Retrieved 7/31, 2003, http://secc.rti.org/Phase2InstrumentDoc.pdf
http://secc.rti.org
2 Current comparisons are based on the most recent available sample characteristics to the original (statements about specific study analyses that may have occurred between these two points should be directed to the sample characteristics at the actual time of analyses). (back)
| Table of Contents | Previous | Next |