Teachers raiting symptoms ADHD

[ Pobierz całość w formacie PDF ]
2008, Vol. 23, No. 3, 418 – 435
Copyright 2008 by the American Psychological Association
1045-3830/08/$12.00 DOI: 10.1037/a0012668
Teacher Ratings of ADHD Symptoms in Ethnic Minority Students:
Bias or Behavioral Difference?
Shelley J. Hosterman and
George J. DuPaul
Lehigh University
Asha K. Jitendra
University of Minnesota
Disproportionate placement of African American and Hispanic students into disability
and special education categories may result from true behavioral and cognitive differ-
ences, bias in assessment and referral, or some combination of the two. Studies of
commonly used ADHD rating scales suggest teacher bias may contribute to placement
discrepancies. This investigation compared teacher ratings of ADHD symptoms on the
Conner’s Teacher Rating Scale—Revised Long Version (CTRS-R:L; Conners, 1997)
and the ADHD-IV: School Version (DuPaul, Power, Anastopoulous, & Reid, 1998),
with objective classroom observations from the Behavioral Observation of Students in
Schools code (BOSS; Shapiro, 2003). Participants were first through fourth grade
students (
N
172; 120 male) classified as Caucasian (
n
112) or ethnic minority (17
African American, 38 Hispanic, 5 African American and Hispanic). Contrary to
hypothesis, results showed teacher ratings of ethnic minority students were more
consistent with direct observation data than were ratings of Caucasian students.
Findings suggest teacher ratings of ethnic minority students may more accurately
reflect true behavioral levels.
Keywords:
ADHD, ethnicity, teacher ratings, direct observations, bias
African American and Hispanic students are
disproportionately diagnosed and placed into
categories of special education in the United
States (Coutinho & Oswald, 2000; Oswald,
Coutinho, Best, & Singh, 1999). Government
education data show overrepresentation of Af-
rican American students in categories of emo-
tional disturbance (ED) and mental retardation
(MR) for each year since 1968 (Donovan &
Cross, 2002; Hosp & Reschly, 2004). Among a
group of one million American students,
160,000 more African Americans students than
Caucasian students will be placed in special
education (Hosp & Reschly, 2004).
Higher rates of identification among ethic
minority students may stem from true cognitive
or behavioral differences, bias in the referral
and assessment process, or some combination
of these two. For the purposes of the current
investigation, “bias” is defined as variation in
teacher ratings of behavior based on student
ethnicity (Chang & Stanley, 2003). If teachers’
perceptions of “typical” and “atypical” behavior
vary across ethnic groups, resulting in different
appraisals or actions based on the same behav-
ior sample, then “bias” is present (Chang &
Stanley, 2003).
Teacher referrals and ratings are central de-
terminants in special education evaluations. Ev-
idence shows referral may represent one source
of placement differences as the majority of stu-
dents referred to special education are ulti-
mately classified and placed (Artiles & Trent,
1994; Ysseldyke, Vanderwood, & Shriner,
1997). Hosp and Reschly (2003) found that 132
African American and 106 Hispanic students
are referred for every 100 Caucasian students.
Studies also show that teacher tolerance is a
primary indicator for identification of behavior
problems and teachers are less tolerant of be-
haviors that are inconsistent with their cultural
expectations (Gerber & Semmel, 1984; Lam-
bert, Puig, Lyubansky, Rowan, & Winfrey,
Shelley J. Hosterman and George J. DuPaul, Department
of Education and Human Services, Lehigh University; and
Asha K. Jitendra, Department of Educational Psychology,
University of Minnesota.
The preparation of this article was supported by NIMH
Grant R01-MH62941.
Correspondence concerning this article should be addressed
to Shelley Hosterman, Lehigh University, Department of Ed-
ucation and Human Services, Iacocca Hall, 111 Research
Drive, Bethlehem, PA 18015. E-mail: sjh6@lehigh.edu
418
School Psychology Quarterly
TEACHER RATINGS OF ADHD IN ETHNIC MINORITY STUDENTS
419
2001). Because 90% of teachers in the U.S. are
Caucasian, it is important to explore the possi-
ble influence of cultural bias on identification of
behavior problems (U.S. Department of Educa-
tion, Office of Educational Research & Im-
provement, 1998).
Attention-deficit/hyperactivity disorder (ADHD)
is among the most common childhood disorders. In
fact, an estimated 3% to 5% of children exhibit
symptoms of ADHD (APA, 2000). Almost 50%
of children with ADHD are eventually placed in
special education programs for behavioral dis-
orders or learning disabilities (Reid, Maag,
Vasa, & Wright, 1994). Students with ADHD
are also assigned to the growing category of
“other health impaired” (National Center for
Education Statistics, 2001).
Evaluations of ADHD symptoms typically
employ a behavioral assessment approach uti-
lizing multiple methods of data collection
across settings and informants (e.g., Anastopou-
los & Shelton, 2000; Barkley, 2006). Major
components include parent and teacher inter-
views, rating scales completed by parents and
teachers, and observations of the child in mul-
tiple settings and task situations (DuPaul &
Stoner, 2003). Thus, responsible diagnosis and
placement decisions rely on a comprehensive
and complex collection of data rather than a
single source (Shapiro & Kratochwill, 2000).
Teacher ratings are a valued aspect of ADHD
assessment because they summarize extensive,
accumulated observations of child behavior
from individuals who are familiar with devel-
opmental expectations (Busse & Beaver, 2000).
These ratings contribute to diagnostic decision-
making by clarifying whether ADHD symp-
toms are inconsistent with developmental level
and associated with impairment across two or
more settings (APA, 2000). Teacher ratings of
behavior are among the most commonly used
methods in school-based assessment of ADHD
(Barkley, 2006; DuPaul & Stoner, 2003). If bias
occurs in teacher ratings of ADHD symptoms,
this may be one source of incongruity in special
education placements across ethnic groups.
Three cross-cultural studies using videotaped
vignettes of scripted ADHD symptoms and dis-
ruptive behaviors have demonstrated the influ-
ence of rater culture on perceptions of ADHD
symptoms. Mann et al. (1992) showed that Chi-
nese and Indonesian clinicians rated hyperac-
tive-disruptive behaviors more severely than
did their Japanese and American colleagues. In
a second study, American and Japanese teachers
rated student behavior more moderately than
did teachers from China, Indonesia, and Thai-
land (Meuller et al., 1995). Similarly, teachers
and student teachers from mainland China
rated videotaped samples of ADHD symp-
toms higher than did teachers from Hong
Kong and the United Kingdom (Alban-
Metcalfe, Cheng-Lai, & Ma, 2002). These
studies emphasize variation in standards for
child behavior across cultures and highlight
the influence of rater’s culture on perception
of ADHD symptoms. A teacher’s personal
history and culture may impact their view of
student behavior.
Other investigations have highlighted the in-
fluence of student ethnicity on teacher percep-
tions of behavior concerns. Stevens (1980)
found teacher ratings of both positive and neg-
ative student characteristics were more influ-
enced by student ethnicity and SES than by
observable behaviors. Results of Prieto and
Zucker (1981) showed that teachers rated a stu-
dent with symptoms of emotional disturbance
as appropriate for services more often when the
student was Mexican American than Caucasian.
Similarly, African American students described
as “difficult to teach” were rated more appro-
priate for special education referral than their
Caucasian peers, regardless of teacher ethnicity
(Bahr, Fuchs, Stecher, & Fuchs, 1991). Middle
school teachers rated students with movement
styles common in African American culture
(e.g., a nonstandard “stroll” style of walking
with a swagger and knees bent) lower in
achievement, higher in aggression, and more in
need of special education than students with
standard movement styles (Neal, McCray,
Webb-Johnson, & Bridgest, 2003). In contrast,
two recent investigations showed that neither
teacher recommendations for placement (Frey,
2002) nor rating scale reports (Cullinan &
Kauffman, 2005) for students with emotional
and behavioral disorders demonstrated bias to-
ward overrepresentation of ethnic minority stu-
dents after SES was controlled. Nonetheless,
evidence suggests student ethnicity may influ-
ence teacher ratings of ADHD symptoms.
Two studies, each examining the correspon-
dence between teacher ratings and direct obser-
vations of varied problem behaviors in cross-
cultural settings, suggested congruence between
420
HOSTERMAN, D
U
PAUL, AND JITENDRA
teacher and student cultures as a key influence.
Puig et al. (1999) compared teacher ratings with
direct observation data in a sample of African
American students and Jamaican students of
African descent. Although Jamaican students
displayed significantly higher levels of observ-
able problem behaviors, teacher ratings of Af-
rican American students were nearly double
those of their Jamaican peers. Puig et al. sug-
gested that Caucasian teachers working in the
U.S. may have lower thresholds of tolerance for
problem behaviors in African American stu-
dents and provide exaggerated reports of these
symptoms. In a second study, direct observa-
tions showed that levels of problem and off-task
behavior in American students were nearly
twice those of Thai students, but reports of
problem behaviors from Thai teachers nearly
doubled those of American teachers (Weisz,
Chaiyasit, Weiss, Eastman, & Jackson, 1995).
Clearly, the match between student and teacher
cultures must be considered when interpreting
rating scale data.
Studies of common ADHD rating scales re-
veal patterns of overidentifying students from
certain ethic minority backgrounds. Initial re-
search in this area emerged after early versions
of scales were developed and interest recently
resurged in response to concerns about cultur-
ally fair assessment and overrepresentation.
Langsdorf, Anderson, Waechter, Madrigal, and
Juarez (1979) examined distributions of scores
on an early version of the Abbreviated Conner’s
Teacher Rating Scale (CTRS; Conners, 1973)
across a large (
n
twice those of Caucasian peers. These three
studies provide evidence of potential bias in
teacher ratings of ADHD but cannot distinguish
rater effects from true behavioral variation
across ethnic groups.
The commonly used ADHD-IV Rating
Scale-School Version (DuPaul, Power, Anasto-
poulos, & Reid, 1998) has also been examined
across ethnicity with similar results. In a large
sample of males ages 5 to 18, Reid et al. (1998)
found teachers rated African American students
higher on all symptoms relative to their Cauca-
sian peers. Results showed that screening all
students using norms based on a primarily Cau-
casian standardization sample would lead to
twice as many positive screenings in African
American students. Reid et al. (2000) found an
interaction effect for gender and ethnicity; Af-
rican American males received the most severe
and Caucasian females the least severe ratings.
One additional study (Arnold et al., 2003) found
that African American students scored signifi-
cantly higher on the Swanson, Nolan, and Pel-
ham Rating Scale-IV (SNAP-IV; Swanson,
1992) compared to their Hispanic and Cauca-
sian peers, even after controlling for SES.
The studies discussed above establish that
mean scores on common teacher rating scales of
ADHD vary across ethnic groups. However, the
possibility remains that such ratings may be
accurate reflections of behavioral variation.
Several investigators have suggested comparing
teacher ratings of ADHD symptoms with more
objective direct observations to explore the im-
pact of student ethnicity relative to true behav-
ior difference (Epstein et al., 1998; Langsdorf
et al., 1979; Reid et al., 2001, 1998). Only three
studies in the literature have used this method to
investigate the possibility of ethnic bias in
teacher ratings of ADHD symptoms.
Ramirez and Shapiro (2005) examined the
effects of both teacher and student ethnicity on
ratings of ADHD symptoms. Caucasian and
Hispanic teachers completed ADHD-IV rating
scales after viewing videotaped vignettes of stu-
dents exhibiting similar levels of overactive,
inattentive, or disruptive behaviors. Results
showed effects of teacher ethnicity, specifically
that Hispanic teachers may hold students of
their own culture to stricter standards of behav-
ior. Student ethnicity had no effect on the rat-
ings of Caucasian teachers in this analog situa-
tion. Although this study provided excellent
1719) sample. Results
showed that in schools with non-Caucasian ma-
jorities, African American students were signif-
icantly more, and Mexican American students
significantly less likely to be rated hyperactive
compared to Caucasian peers. Both SES and
ethnicity had significant effects on CTRS rat-
ings, with poor African American students far-
ing the worst (Langsdorf et al., 1979). Epstein,
March, Conners, and Jackson (1998) found that
African American children were rated higher on
CTRS-R (Conners, 1997) externalizing problem
factors, including conduct disorder and hyper-
activity problems, compared to Caucasian
peers. A third study (Reid, Casat, Norton, Anas-
topoulos, & Temple, 2001) found African
American students would screen positive for
ADHD on the IOWA Conner’s (Pelham,
Milich, Murphy, & Murphy, 1989) at levels
TEACHER RATINGS OF ADHD IN ETHNIC MINORITY STUDENTS
421
controls of ethnicity and behavior, it may not
accurately represent the use of ADHD rating
scales in school settings, where teachers typi-
cally base behavioral ratings on accumulated
interaction with students rather than discrete
samples.
Working in the United Kingdom, Sonuga-
Barke, Minocha, Taylor, and Sandberg (1993)
compared teacher reports from interviews and
questionnaires with direct observations and me-
chanical data on student behavior in analog and
regular classroom settings. Teacher reported hy-
peractivity levels of Asian (parents from India,
Pakistan, or Bangladesh) students were overes-
timates as compared to objective measures of
student behavior. In fact, direct observations
showed the behavior of Asian students rated
“hyperactive” was comparable to that of En-
glish students in the control group. This study
suggests that comparisons of subjective teacher
ratings to more objective systematic observa-
tions may reveal teacher bias toward certain
ethnic groups (Sonuga-Barke et al., 1993).
The third study, Epstein et al. (2005), com-
pared teacher ratings on three common scales to
observation data of African American and Cau-
casian students. Results showed that overall,
teacher ratings of ADHD symptoms were
higher for the African American group (Epstein
et al., 2005). However, direct observation data
revealed a similar pattern, with African Amer-
ican students exhibiting higher levels of observ-
able ADHD symptoms than their Caucasian
peers (Epstein et al., 2005). Correlations of
teacher ratings and observation data were sim-
ilar for both groups. Thus, unlike Sonuga-Barke
et al. (1993), these results did not evidence bias
in teacher ratings of ADHD symptoms for stu-
dents from ethnic minority backgrounds.
Clearly, the literature examining the question
of bias in teacher ratings of ADHD symptoms
by comparing teacher rating scale data with
direct observations of classroom behavior is
both sparse and inconclusive. The purpose of
the current study was to expand this literature
by comparing teacher ratings on two commonly
used ADHD ratings scales to direct observation
data from a well-established observational code.
The study aimed to extend the work of Sonuga-
Barke et al. (1993) to a sample of American
students. Also, it extended the U.S. based work
of Epstein et al. (2005) by including a sample
with an ethnic composition more representative
of the U.S. population.
The current study aimed to answer the ques-
tion: Are teacher ratings of ADHD symptoms in
African American and Hispanic students less
consistent with objective observations than are
teacher ratings of Caucasian students? The hy-
pothesis was that teacher ratings of ADHD
symptoms would correlate more strongly with
behavioral observations for Caucasian students
than for a group comprised of African American
and Hispanic students.
Method
Participants
Sample of Larger Investigation
Participants were selected from a larger study
examining academic interventions for elemen-
tary school students with ADHD (for details,
see DuPaul, Jitendra et al., 2006). All partici-
pants met several initial inclusion criteria. Re-
cruited from schools in eastern Pennsylvania,
all students were enrolled in a first through
fourth grade, general or special education class-
room in the public school system. Participants’
families also indicated that they did not plan to
move from the area within 2 years following the
start of the study.
Students in the ADHD group met four addi-
tional inclusion criteria. Children were consid-
ered for the ADHD group if they either had a
previous ADHD diagnosis from an outside ser-
vice provider or were reported by a teacher to
exhibit significant problems with inattention,
impulsivity and/or hyperactivity. All potential
participants were then screened based on the
following criteria. All students in the ADHD
group scored at or above the 90th percentile on
one or both of the Inattention or Hyperactivity-
Impulsivity subscales of the ADHD-IV (DuPaul
et al., 1998) across both parent and teacher
ratings. Higher ratings and percentiles on the
ADHD-IV represent more severe ADHD symp-
toms. Each student also met
DSM–IV
(APA,
2000) criteria for one of the three ADHD sub-
types based on a parent interview with the
NIMH Diagnostic Interview Schedule for Chil-
dren-IV (DISC-IV; Shaffer, Fisher, & Lucas,
1998), administered by trained graduate stu-
dents. This interview was used to ensure that
422
HOSTERMAN, D
U
PAUL, AND JITENDRA
students met a sufficient number of individual
symptoms to warrant diagnosis. Finally, each
child was also experiencing achievement prob-
lems in either math or reading, according to
teacher report. Inclusion in the ADHD group
was based on reports from multiple respondents
and settings. These criteria addressed symptom
severity by using stringent cut-points on the
ADHD-IV and ensuring that participants met
DSM–IV
criteria for ADHD.
Students in the non-ADHD control group
were recruited from the same schools, but from
different classrooms than children in the ADHD
group. Their teachers indicated that these stu-
dents were experiencing no behavioral or aca-
demic problems. Each student in the control
group scored below the 90th percentile on all
subscales of the ADHD-IV (DuPaul et al.,
1998) and did not meet criteria for ADHD,
Oppositional Defiant Disorder (ODD), or Con-
duct Disorder (CD) based on the DISC-IV par-
ent version (Shaffer, Fisher, & Lucas, 1998).
These students were also matched to those in
the ADHD group by grade and gender.
For the purpose of the current investigation,
all Caucasian, African American, and Hispanic
participants from both the ADHD and non-
ADHD control groups of the larger study were
considered for inclusion. That is, the current
study considered students classified into both
the ADHD (
n
Sample of Current Investigation
Using this sample, two groups were created:
a Caucasian group and an ethnic minority group
comprised of all African American and His-
panic students in the sample. Inclusion was fur-
ther limited to ensure independence of raters.
That is, if a classroom teacher had two or more
students enrolled in the larger study, one student
was randomly selected for the current investi-
gation. Inclusion was also limited by teacher
completion of behavior rating scales. For exam-
ple, data obtained from a direct observation in
math were included only if the student’s math
teacher completed the behavior ratings scales.
Because separate analyses were conducted for
reading and math, the same student could be
represented in analyses for both academic areas.
Table 1 displays demographic data for the
final sample of the current study. These data are
shown separately for the Caucasian (
n
112)
60) groups. The ethnic
minority group consisted of 17 (28.3%) African
American, 38 (63.3%) Hispanic, and 5 (8.3%)
students of both African American and Hispanic
descent. The Caucasian group was 75% (
n
85) male and the ethnic minority Group 90%
male (
n
54). At enrollment, 21% of partici-
pants were taking psychotropic medication (pri-
marily psychostimulants). Based on results of
the DISC-IV (Shaffer et al., 1998), 17.7% of
participants met criteria for ADHD: Inattentive
subtype, 6.9% for ADHD Hyperactive-Impul-
sive subtype, 46% for the combined subtype,
and the remainder were control students.
Among the 171 students included in the reading
analyses, 54 (31.6%) displayed academic difficul-
ties in reading only, 35 experienced difficulty in
both reading and math (20.4%), and 82 (48%)
175) and non-ADHD control
66) groups of the larger investigation for
inclusion, working from a combined original
sample of 241 students. The sample of the
larger study represented the ethnic composition
of the geographic area and included students
classified as Caucasian (
n
146), African
American (
n
22), Hispanic (
n
64), and
other (
n
6).
Table 1
Demographic Characteristics of the Caucasian and Ethnic Minority Groups
Group
Measure
Caucasian (
n
112)
Ethnic minority (
n
60)
t
(170) or
2
(1)
p
Age (in years)
8.48
a
(1.25)
b
8.49 (1.13)
.052
0.959
Male (%)
75
90
5.543
.019
2.596 .010
Group (% ADHD group) 69.6 76.7 .958 .328
Note.
SES
social economic status determined by highest Hollingshead parent occupation; ADHD
attention deficit
hyperactivity disorder.
59.71 (22.62)
43.62 (26.90)
a
Mean.
b
Standard deviation.
p
.05.
and ethnic minority (
n
(
n
SES
[ Pobierz całość w formacie PDF ]