Center for Education Statistics, 1996,
1997, 1998).
This coverage commonly doesn’t
highlight the fact that student performance within the United States varies
greatly from state to state and from
district to district. Although U.S. 8th
grade performance was the same as the
international mean in the 1999 TIMSS
benchmark study, the performance of
the 13 U.S. states and 14 U.S. school
districts that participated as “countries”
in the study virtually spanned the
performance range (Mullis et al.,
2001). Seven states and eight
districts performed significantly
higher than the international
mean, three districts performed
significantly lower than the international mean, and the rest
scored at about the international
mean. Thus, U.S. states and
districts differ not only in the access
to mathematics content that they offer
students, but also in their student
performance outcomes.
One of the most important findings
from our analysis of the 1995 TIMSS
(Schmidt et al., 2001) was that achievement differences from country to
country were significantly related to
what was taught. This conclusion was
possible because of the rich portrait of
math and science instruction available
from the 1995 TIMSS curriculum
analysis. For each country, we looked at
the intended content (what officials
intended for teachers to teach) and the
enacted content (what teachers actually
taught in their classrooms). In most
countries, we determined the intended
content by looking at the national
curriculum (or, in the handful of countries without a national curriculum, by
looking at other formal statements of
intended content at the regional or local
The real issue
behind differences in
student performance
is unequal access
to a high-quality,
challenging
curriculum.
level). In all the countries, we determined the enacted content by surveying
teachers about what they had covered.
Analysis of these rich curriculum
data, along with our more curriculum-sensitive measures of student achievement, revealed that the mathematics
content teachers covered in their classrooms was significantly related to their
students’ performance even when
researchers adjusted this relationship for
student background factors (ethnicity,
parent education level, socioeconomic
status, and so on). This relationship was
evident at every level—classroom,
district, and state.
Schooling does make a difference in
student achievement. Specifically, the
curriculum itself—what is taught—
makes a huge difference.
Need for Curriculum-Sensitive
Assessment
No Child Left Behind (NCLB) explicitly
affirmed the fundamental democratic
goal of schooling in the United States.
In the wake of this legislation, issues of
suitable standards for all students and
equitable access to adequate learning
opportunities have acquired a new
urgency in education reform. States and
individual districts are being compelled
to make explicit what it means for all
students to have equitable opportunities
to learn essential and challenging
content (Achieve, 2002).
NCLB brought attention to some
important issues, such as the need to
establish high learning standards for
all students. But the fundamental
flaw in NCLB is the disconnect
between the assessments that are
used to determine education
outcomes and the content standards
that guide and inform classroom
instruction and learning.
Using the 20 specific mathematics
topic scores we created from the 1995
TIMSS data, for example, we found that
when students were given access to
specific curricular content, there was a
significant benefit for student performance in those content areas (Schmidt et
al., 2001). The total mathematics score,
however, was insensitive to important
differences in curricular emphasis
(Schmidt, Jakwerth, & McKnight,
1998).
Is this an important distinction, or is
it just another dart to throw at the
accountability endeavor? Clearly, we
believe that this distinction is fundamentally important. Unless assessments
are sensitive to important differences in
instructional content coverage, student
achievement gaps can be misattributed
to individual background factors that
