Table 4.1 presents national NHSDA estimates
of means and standard errors for the eight syndrome scales for 1994, 1995, and
1996 by gender and age group. Achenbach (1991) gives means and standard errors
for two samples of youth, aged 11 to 17 or 18. One sample, the referred or clinical
sample, consisted of youths who were referred to psychiatric clinics for help
with a problem. The other sample, the nonreferred or nonclinical sample, was
a matched sample of youth who were not referred to a clinic. The national NHSDA
estimates are based on a nonclinical sample and are generally similar to those
of the nonclinical samples reported by Achenbach (1991). However, adolescents
in the national sample were generally less withdrawn, and less likely to report
attention problems or anxiety/depression. Compared to adolescents in the Achenbach
nonclinical samples, males in the national sample tended to report less aggressive
behavior and females tended to report more somatic complaints, thought problems,
and delinquent behavior. These differences may be due to the composition of
specific groups. All other national estimates are generally consistent with
those of the nonclinical samples reported by Achenbach.
Interitem consistency. To examine
construct validity, the extent to which the measures represent a theoretical
concept, coefficient alpha for each scale was calculated for each gender
group. Results are shown in Table 4.2. Reliability coefficients for the NHSDA
data range from .53 to .84, comparable to those reported by Achenbach. The
coefficients are generally higher for females than for males. As expected,
coefficients are generally higher for scales with greater numbers of items.
The following section presents psychometric
properties of the widely-used YSR in a nonreferred nationally representative
sample. Generally, the measures of emotional and behavioral problems in the
national samples approximated those of the nonclinical reference sample reported
by Achenbach (1991); however, the nonclinical sample males differed from the
comparable national samples more than did the nonclinical sample. Internal
consistency measures for the subscales were close to those reported by Achenbach
(1991). The one exception was for Social Problems, which showed substantially
less internal consistency in the combined national samples than it did in
the norming sample. Compared to the norming samples, the internal consistency
coefficients for the various syndromes were somewhat lower in the national
samples, with the exception of Withdrawn, which showed slightly higher internal
consistency in the national samples. Less independence among scales is indicated
in the national samples than was reported for the norming sample. On average,
scales were correlated in the .52 range, although the average intercorrelation
decreased slightly for the older groups. Withdrawn and Anxious/Depressed were
the most closely related for all age and gender combinations; the relationships
between Somatic Complaints, Social Problems, and Delinquent Behavior were
generally lowest.
SAMHSA, an agency in the Department of Health and Human Services, is the Federal
Government's lead agency for improving the quality and availability of
substance abuse prevention, addiction treatment, and mental health
services in the United States.
