An interpretive guide.
goal here is to apply some common-sense descriptions of accuracy to the
classifications that are used to indicate student achievement in STAR .
Version 2.0 August, 2003: CAT/6 Survey,
Version 1.0 August, 1999: Stanford 9, Harcourt
CRESST Technical Report 509A
Week Article October 6, 1999
Stanford Report Questions Accuracy of Tests
New York Times Column, September 13, 2000
How Tests Can Drop The Ball
Standards Tests (CST)
Version 3.0 August, 2004: California Standards Tests: Math, ELA 2002, 2003.
direct downland: llclassify computational function
|Also. If you have a liking for reliability coefficients, try out the shoe-shopping example|
Materials: Percentile Rank Scores
Many informal statements and references in the NPR Accuracy Guides are given explicit, mathematical definition in two CRESST Technical Reports released in concert with the Version 1.0 Accuracy Guide.
So for those who might be interested in technical details and underpinnings for these calculations, the two related CRESST Technical Reports are
1. Accuracy of Individual Scores Expressed in Percentile Ranks: Classical Test Theory Calculations
CRESST Technical Report 509, September, 1999
2. Accuracy of Year-1, Year-2 Comparisons Using Individual Percentile Rank Scores: Classical Test Theory Calculations
CRESST Technical Report 510, September, 1999