Even a "Value-added" Leveling of the Academic
Playing Field
Does Not Strengthen the Single Snapshot Quality of a Set of Assessments
Rather innocently, the public tends to compare high performing
schools with low performing schools without considering differences in student
characteristics. The point of using statistical modeling is to establish an achievement
benchmark that acknowledges the challenges that can affect children's readiness to learn.
However, any performance assessment -- or academic test -- measures only how well a child
performs on a particular test, on a particular day. Circumstances having nothing to
do with the testing can have profound effects on how children perform. A classmate's
illness or car accident can affect the scores of that grade's test-takers, just as the
news of an impending divorce can affect the scores of an individual test-taker.
Similarly, one year's class of students might be unusually challenged while the following
year's happens to be especially talented, which will also create swings in the scores.
Only over time, with a good deal more data, will we be able to see how an individual
school "usually" performs in relation to its counterparts.
The lists accessible by the buttons above are, therefore, indications of how groups of
students (and by extension, schools) perform as compared to other students like themselves
in the rest of the state, at a very particular moment in time. The value-added exercise
attempts to account for some of the uneven playing field on which children and schools are
measured, but can not possibly account for all unevenness. Only over the course of time
will the statistically generated model be able to demonstrate the effectiveness of a given
school.
Still, we recommend visiting the consistently high-performing schools to determine why
they tend to get superior results.
RI Elementary Schools
View in HTML || View or Download in PDF format (57KB) 
RI Middle Schools
View in HTML || View or Download in PDF format (47KB) 
RI High Schools
View in HTML || View or Download in PDF format (50KB) 
(Please note: These charts are available in both
HTML and PDF format. For on-screen viewing, we recommend using the HTML
version. For printing, however, the PDF files are the best choice in order to keep
the look and feel of the documents we created intact.)
Statistically Generated Performance
Models - A brief layperson's explanation
Technical Brief on the RI
Statistically Generated Model
Some Resource Materials on "Value-Added":
National Center for Research on
Evaluation, Standards, and Student Testing
University of Tennessee
Institute of Assessment and Evaluation