A closer look at US schools

What characteristics are associated with scientific literacy? A multivariate multilevel analysis using Pisa 2015

Bibliographic Data

ID21392724
AuthorsHyesun You (0000-0002-5835-5053, Department of Physical Sciences, Science Education Arkansas Tech University Russellville Arkansas USA, corresponding author), Sunyoung Park (0000-0002-6292-5514, Graduate School of Psychology, Master of Science in Clinical Psychology California Lutheran University Thousand Oaks California USA), Cesar Delgado (0000-0002-4961-7736, Department of STEM Education, College of Education North Carolina State University Raleigh North Carolina USA)
Year2021
Volume105
Issue2
Pages406-437
Publication date2021-03-01
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueScience Education (JOURNAL)
Journal identifiersISSN: 0036-8326 • E-ISSN: 1098-237X
PublisherWiley (PUBLISHER • GB)
DOI10.1002/sce.21609
OpenAlexW3112900607
LanguageEN
Citations received9
References cited103

The purpose of this study is to examine the characteristics of US schools associated with two measures of scientific literacy (content knowledge and “procedural and epistemic” knowledge) using the 2015 Programme for International Student Assessment (PISA) data. Because outcomes are nested within students, and students within schools, a multivariate three‐level modeling method was employed. About 21% of the total variance in science performance lies between schools, indicating that school characteristics are important in predicting scientific literacy. The results revealed clearly significant and positive relationships of the student‐level variables of grade, enjoyment, motivation, and economic/social/cultural status (ESCS) with both measures of science literacy, after controlling for school factors. A significant gender difference is seen in science content knowledge in favor of males. At the school level, the results from the full model suggest that school ESCS, climate, and school type are significant predictors of all students' performance. Surprisingly, school size, teaching experience, professional development (PD) participation, and science‐specific resources are not found to contribute significantly to achievement. Considerable variability is also evident among schools on student performance in both science knowledge domains as a result of school variables. For both low‐ and mid–high‐achieving students, the most significant school factor is school mean ESCS. PD participation and school climate are significantly associated with student performance only for the average and high‐performing groups. This paper can inform policymakers, researchers, and educators on how US schools can be supported to improve science learning

Academic achievement · Literacy · Mathematics education · Multilevel model · Multivariate analysis · Multivariate analysis of variance · Pedagogy · Science education · Scientific Literacy · Variance (accounting) · Education, Achievement, and Giftedness · Mathematics · Parental Involvement in Education · Psychology · School Choice and Performance

Tools

Open DOI