Evidence for the Validity and Reliability of Performance Clusters Within the Academic Support Index

• In Event: Advances in Estimating Reliability: Innovative Approaches and Applications

Abstract

Reliable predictors of student performance are critical to efficiently directing educational resources. This study provides evidence for the validity and reliability of two specific performance clusters within the Academic Support Index (ASI): one cluster where students are highly likely to meet or exceed standards and one cluster where students tend to fail to meet standards. Analysis of four years of Smarter Balanced Assessments for two school districts across seven grade levels yielded a large average effect size between the clusters (mean d=1.22). The ASI cluster of lower performing students identified up to 89% of those students who failed to meet standards. Early identification of potentially underperforming students should be addressed with additional monitoring and intervention resulting in improved achievement.

Author

• David Stevens, University of California - Berkeley