Jim Popham (emeritus UCLA), who was responsible for a huge amount of
state mandated testing in the 70s (albeit criterion-referenced) and who
was summoned to trial in Florida to account for the disproportionate
numbers of kids of color failing the Florida test (which he helped
build) says:
[in addressing the relevance of testing as evidence of
learning]<color><param>FFFF,0000,0000</param> "In order to produce such
score differences, test developers include many items that are not
likely to be influenced by classroom instruction. In truth, such items
actually assess children's socioeconomic status and their inherited
academic aptitudes. They measure what children bring to school, not
what they learn there. Moreover, test makers can only sample course
content. As a results, the particular items might not coincide with the
content addressed by a specific school. Taken together, these problems
typically lead to the mismeasurement of a school's effectiveness."
</color>This is, imho, equivalent in significance to Lauren Resnick's
comments in her life time achievement address at AERA several years
ago, when she admitted "transfer is a red herring."
I love these moments.
Think anyone heard him?
Linda
ps
Don't get me wrong; I love Jim. His text on stat (co-authored with
Sirotnik) is the ONLY understandable text I've ever encountered on the
subject.