|
Student performance
adjusted for “value-added” Download
an illustrated guide to
understanding the value-added charts
Select a school level to view and/or print the charts.
What you are looking at:
You are looking at selected
results of the statistical modeling that compares the
performance of a school’s student body with the
performance of similar students statewide. (See below
for explanation, rationale and methodology.) The charts
are sorted into levels -- high, middle and elementary. The state-level charts use four subtests, two for
math and two for English, as indicators of student
academic achievement as compared to similar students
statewide. Those whose actual scores are consistently
above the model are in the top band; those consistently
below are in the bottom; all other schools are spread
along a continuum in between. RI high schools and middle
schools can each be represented on single pages. The
elementary chart occupies four pages.
Cautionary note: These charts use only one year
of performance data, so they are sensitive to the
differing abilities of individual classes of students
and versions of the tests. There is also a 5% chance
that a school’s students perform better or worse than
expected due solely to chance. The charts are most
useful when viewed in conjunction with prior years’
value-added charts, which are available through the
website.
What you are looking for:
Having adjusted for differing student characteristics
such as poverty, you are looking for those schools that
appear to have techniques for successfully helping their
unique student body, techniques from which the rest of
us can learn. Remember that even the students in the
high-performing schools – according to the model – are
not reaching 100% proficiency, so all schools are still
in the process of re-designing themselves. But two
schools with similar populations that score very
differently according to the model prompt the question:
why?
Leveling the Playing
Field
If you take the raw performance scores by district and
sort them high to low, you would find that you have also
sorted almost perfectly by the median family income of
each district. Without the strong intervention of
schools, students tend to achieve according to their
socioeconomic backgrounds. This pattern is by no means
peculiar to RI, or even just the United States.
Schools with high concentrations of low-income or
special needs children have always complained about
being unfairly compared with schools whose less
challenged children perform at high levels on
standardized tests. The public tends to compare
high-performing schools with low-performing schools
without considering differences in student
characteristics. In fact, poverty is the strongest
single predictor of student achievement, except for a
student’s prior achievement. (Without a Universal
Student Identifier system in place that would enable
RIDE to know students’ grade point averages, RI is not
able to factor prior achievement into its research.)
Statistical Modeling
In recent years educational researchers have begun
building statistically generated models that can
calculate what results schools are likely to achieve
when taking into consideration the characteristics of
their student body. For over 40 years, researchers have
known that the achievement results of different sets of
students, such as those from different schools, vary in
association with several specific key factors,
including:
-
Poverty (by far the
strongest predictor of student achievement, with the
exception of prior achievement)
-
Non-English speaking
background
-
Educational background of
the parents
-
Having special learning
needs, and
-
Having a minority racial
group identity
While individuals with one
or more of these characteristics can and do perform well
on state assessments, the majority tend to perform less
well than children who do not have these
characteristics. The many reasons for these historic
patterns of achievement include such things as school
expectations, the availability of flexible grouping and
different types of instruction, inadequate funding and
support to the schools these children attend, and the
quality of social services offered to students.
Statistical models make it possible to establish an
achievement benchmark that acknowledges the challenges
that can affect children’s readiness to learn. The
models help us look at the same assessment data through
a lens that filters out some of the students’
challenges. This lens provides a different, but newly
uniform and, in some ways, more realistic benchmark
against which to measure actual performance. For
example, some RI schools are categorized as
“low-performing,” but their students are out-performing
similar students statewide according to the value-added
modeling. Clearly such schools add considerable value to
their students education, even though they haven’t yet
mastered helping their students to 100% proficiency.
Through the value-added exercise, such schools signal
that they have lessons for schools with similar students
who are not doing as well.
Please note: In general, Information Works!
does not report data cells smaller than 10 students
because of the possibility of identifying or guessing at
the identity of individual students. Schools that
sometimes do not appear in these charts often have
especially small students bodies – most frequently RI
School for the Deaf and Block Island School – which
leaves them vulnerable to overly-small tested grades.
The only exception to this rule is when three years of
assessment results are used to create disaggregations of
student performance, where we use 5 students as the
criterion.
The value-added lens is most powerful over
time.
To fully understand the value-added lists and/or your
school’s position on them, we recommend you examine the
lists from prior years. With only one year’s assessment
data, these lists are vulnerable to movement resulting
from the abilities of a given class of students and the
match between the curriculum and that year’s version of
the test. Looking across the years gives a stronger
sense of whether or not the school is making true
statistical gains, on similar students statewide, or is
stuck or losing ground.
Prior lists are also available: Infoworks!
2001
І
2000
І
1999
Please note: ‘No-score’ results are calculated in
the over-all proficiency of the school’s children,
starting with the 2001 Information Works! The
charts prior to 2001 are important, but not directly
comparable since the old model only considered those
children who received scores. By including these
eligible but untested children, RIDE is emphasizing its
insistence that schools account for All Children.
The relationship between the value-added charts and
the Performance Progress charts
While the Performance Groupings list contains
information that triggers consequences for schools and
the value-added chart does not, these charts provide
different looks at the same school which aids, and in
some cases tempers, our understanding of either one of
the charts. For example, Asa Messer School in Providence
is considered “low-performing,” though “improving,”
according to the Performance Progress criteria. On the
value-added charts, that school has performed
consistently above statistical prediction for the third
year in a row. Nearly 100% of its children are eligible
for subsidized lunch – a poverty indicator – so this
school is obviously having more success with
socioeconomically challenged children than its
counterparts statewide. While it needs and deserves the
state support for low-performing schools, it also has
lessons to teach about improving the outcomes for
low-income children.
At the opposite end of the spectrum, we find a school in
suburban, RI who is at the bottom of the value-added
lists, meaning its students are consistently
under-performing compared with their relatively affluent
counterparts statewide. This same school is both
‘high-performing’ and ‘improving’ according to the
Performance Progress lists. This difference is likely
due to the fact that there are so few students at the
extreme end of the model. This school is to be commended
for its improvement, certainly, but while it stays out
of RIDE’s intervention focus, its community has strong
reasons to investigate the low achievement of these
students as compared to their statewide counterparts.
Back to Table of Contents
|
