Student-Level Reliability, Classification Accuracy, and Convergent Validity Studies
Student-Level Reliability, Classification Accuracy, and Convergent Validity Studies are used to describe the quality of students' test scores.
HumRRO has conducted reliability analyses for the following states:
- Kentucky
- Florida
- Utah
- Idaho
- Pennsylvania
Because students are typically assigned to performance levels based on test scores (typically 4 or 5 categories that signify below average to exceptional achievement), the likelihood of the assignment being correct or incorrect is also an important statistic. HumRRO has calculated this data for the following states:
- Kentucky
- Idaho
In addition to these test reliability issues, HumRRO has also conducted convergent validity studies that examine relationships between scores from state customized tests and scores from nationally available norm-referenced tests. HumRRO has conducted convergent validity studies with K-12 NRTs and college admissions tests (ACT and SAT) for the following tests:
- Kentucky
- Massachusetts
- Pennsylvania
In addition, HumRRO has also conducted convergent validity studies with Pennsylvania's assessment and college placement exams. Students' high school state assessment scores have also been correlated with students' GPA in their first college-level courses in a predictive validity study.
