Validating the inferences made from the 2012 mathematics PISA

• Main Author: Shivraj, Pooja
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01-01-2014
• Subjects: Academic Achievement; Data Analysis; Educational Environment; Educational tests & measurements; Grade 4; International Assessment; Outcomes of Education; Teaching Methods
• ISBN: 9781321092134; 132109213X

The Programme of International Student Assessment (PISA) has been administered to 15-year olds every three years since 2000. Since then, the U.S. has performed below the OECD mathematics average, and significantly below countries such as Finland, China, Japan, Korea, and Singapore. The objective of this study was to use an argument-based approach to validity to examine whether fairness could be an explanatory factor in the mediocre rankings of the U.S. on the PISA. Fairness, defined in this study as opportunity to learn and lack of bias, were investigated as the assumptions underlying the inferences made from the PISA scores, and an alignment study and a differential item functioning analysis were the pieces of validity evidence used to support these assumptions. Although evidence supports the underlying assumption of the lack of bias on the PISA items using the reference groups and criterion specified, the evidence for the assumption of students' opportunity to learn is weak. While students in the U.S. have the opportunity to learn the content on the PISA, the depth to which they learn it is insufficient. These findings are also supported by studies on opportunity to learn conducted by the OECD on the 2012 mathematics PISA data.