Leveraging Students' Background Characteristics to Predict Performance on Conversation-Based Assessments of Mathematics

• In Event: The Role of Child, Families, and Communities in Early Childhood Education

Abstract

Students enter educational settings with wide variations in relevant prior knowledge and experiences, social-emotional skills, and personal qualities, and these individual differences have implications for students’ performance on learning and assessment tasks (Duckworth & Yaeger, 2015). The current study examined whether social-emotional and affective variables predicted students’ performance on an interactive conversation-based assessment (CBA) for mathematical argumentation (Authors, 2015a) after controlling for demographics and prior achievement. CBA performance was significantly and strongly predicted by students’ prior achievement in the domain, after controlling for demographic and other background variables. Students’ grade level, cognitive flexibility, and affective dimensions of school failure tolerance were also related to CBA performance. Implications for the design of personalized “caring assessments” are discussed.

Authors

• Jesse R. Sparks, Educational Testing Service

• Jonathan Steinberg, EurekaFacts

• Blair Lehman, Educational Testing Service

• Diego Zapata-Rivera, Educational Testing Service