Purpose: American elementary mathematics classrooms have been characterized as having a low level of mathematical discourse and a dearth of opportunities for students to explain mathematics. International comparisons suggest there is a great deal of room for improvement in these dimensions and that student competence would likely be improved could we do so. The research team aims to develop effective methods to help teachers monitor and improve classroom discussion in mathematics.

Project Activities: The project has three related parts:

1. Adapting and validating an automated technique (the Language ENvironment Analysis, or LENA) for giving teachers daily feedback on the extent on distribution of classroom talk,
2. using video-based online professional development to provide teachers with methods for leading mathematical discussions and keeping a focus on important mathematical ideas; and
3. developing a test bed for understanding how students attend to classroom mathematical talk and comparing the relative effectiveness of teacher and student explanations.

Products: The team will produce peer-reviewed publications describing the findings from the proposed studies.

Structured Abstract

Setting: This study will take place in elementary schools in Southeast Michigan.

Sample: Experienced and beginning elementary teachers from Southeast Michigan will take part in the project, along with 5th-6th students in the test bed experiments.

Research Design and Methods: The team will complete a 3-year iterative process of 1) validating the LENA against conventional measures of classroom talk; 2) developing, deploying, and refining an online professional development system based around the LENA; 3) conducting test bed experiments on student use of peer explanations, with results incorporated into the professional development system. In Year 3, the fully developed system will be compared to a business as usual control group.

Control Condition: In Year 3, the researchers will compare the development of our targeted teachers with a business as usual condition. In Year 1, the research team will have a control group that undergoes a very limited form of the intervention that will provide a more realistic comparison for Years 2 & 3.

Key Measures: Outcome measures will include changes in distribution and quality of classroom discussion as measured by LENA and a validated measure of classroom mathematics instruction (the MQI), as well as both proximal and distal student outcome data. For the test bed experiments, changes in student understanding and in the quality of explanations will be measured.

Data Analytic Strategy: There are three main analytic procedures that will be used. The team will validate the LENA parameters with hand-coded video measures of discussion quality. The team will rely on latent growth specification with a multi-level framework in order to measure the effects of PD with LENA and other records, looking at developmental trajectories. Finally, the researchers will use mixed model MANOVAs to evaluate student change in the test bed experiments.

Products and Publications

Journal article, monograph, or newsletter

Samudra, P.G., Min, I., Cortina, K.S., and Miller, K.F. (2016). No Second Chance to Make a First Impression: The 'Thin-Slice' Effect on Instructor Ratings and Learning Outcomes in Higher Education. Journal of Educational Measurement, 53(3), 313–331.

Wang, Z., Miller, K.F., and Cortina, K.S. (2013). Using the LENA in Teacher Training: Promoting Student Involvement Through Automated Feedback. Unterrichtswissenschaft, 41(4): 290–302.

Wang, Z., Pan, X., Miller, K.F., and Cortina, K.S. (2014). Automatic Classification of Activities in Classroom Discourse. Computers and Education, 78: 115–123.

** This project was submitted to and funded under Teacher Quality: Mathematics and Science Education in FY 2010.