Investigating Factors Contributing Constructed-Response Performance When Paired Selected Responses Are Correct (Poster 14)

• In Event: Friday Poster Session 7:45 am
  In Poster Session: Considering Teachers, Students, and Contexts in Science Education

Abstract

The present study investigated the factors contributing to science performance on constructed-response (CR) questions when paired selected-response (SR) performance was good. Using 13 paired SR-CR items from PISA 2018 simulated science assessments, this study analyzes contributions to CR performance from country, item (e.g., scientific knowledge, scientific competency, item proficiency level), student characteristics (e.g., age, gender, immigration status, socioeconomic status (SES)), reading proficiency, and language features. Results revealed significant relationships between CR performance and country, scientific knowledge, item proficiency level, gender, reading, and specific language use. However, reading performance at the school level, SES at individual and school levels, immigration status, age, and scientific competencies were unrelated to CR performance. Implications for addressing science assessment inequality are discussed.

Author

• Haiying Li, University of Pennsylvania