1 Ghent University (BELGIUM)
2 University of Gothenburg (SWEDEN)
About this paper:
Appears in: INTED2021 Proceedings
Publication year: 2021
Pages: 1249-1256
ISBN: 978-84-09-27666-0
ISSN: 2340-1079
doi: 10.21125/inted.2021.0291
Conference name: 15th International Technology, Education and Development Conference
Dates: 8-9 March, 2021
Location: Online Conference
Instructional quality is considered an important predictors of learning outcomes. Different approaches were used to evaluate the instructional quality in the classroom, for instance, peer observations, classroom artefacts, student surveys and teacher self-reported data, but the information from different sources could reflect their perceptions of instructional quality. Wagner et al. (2016) observed that the association between teachers’ self-reported instructional quality and student perceived instructional quality was low to moderate due to different idiosyncrasies and the strength of the correlation between teacher and student ratings might not be stable over time.

Several studies have verified the three-dimensional framework of instructional quality, namely, classroom management, supportive climate and cognitive activation. However, the three- dimensional framework is mainly regarded as the generic subject, and it does not cover to a specific school subject, for example, mathematics education, with the subject-specific characteristics, such as the usage of appropriate mathematical language, problem-solving, or the focus on mathematical concepts.

Therefore, this study is to validate the three-dimensional framework of mathematics instructional quality using information from both a teacher and student perception and test the measurement properties of the construct instructional quality in mathematics is comparable across different educational systems. The research questions need to investigate: (1) What is the measurement property of mathematics instructional quality from teacher and student perspectives? (2) To what extent can measurement invariance across seven countries be confirmed? (3) How consistent are teachers’ and students’ rating of instructional quality?

The linking data from the Teaching and Learning International Survey 2013 (TALIS 2013) and the Programme for International Student Assessment 2012 (PISA 2012) were applied. We selected the mathematics teacher data from TALIS 2013 who participated PISA 2012 mathematics assessment and taught the majority of PISA 2012 students in the class, linking with the PISA 2012 data by anchor variable PISASCHOOL ID, including seven countries: Australia, Finland, Latvia, Portugal, Romania, Singapore and Spain. Structural equation modelling (SEM) technique, particular in confirmatory factor analysis (CFA) used to measure the factor structure of instructional quality and to test the measurement invariance when comparing instructional across seven countries.

The results present that five underlying dimensions were found in student perceptions about mathematics instructional quality when looking at the PISA data (i.e., classroom disciplinary climate, teacher support, cognitive activation, classroom management, and student-orientated instruction), while three dimensions were identified (i.e. classroom disciplinary climate, teacher support and cognitive activation) from teacher perspectives in TALIS data. The five-dimensional factor structure was stable in all country data, and the measurement properties were metric invariant. The three-factor model of instructional quality with teachers’ data varies across countries. The findings imply that teachers and students hold different perspectives on mathematics instructional quality within and between countries, and the concepts should not be used interchangeably to investigate school effectiveness in mathematics education across countries.
Mathematics education, instructional quality, measurement invariance, structural equation modelling, TALIS-PISA Link.