JSM 2013: Simultaneous Inferences

08/04/2013 at 10:39 am


In a longitudinal cluster randomized trial, we investigate the impact of a new inquiry-based approach to teaching science, the Science Writing Heuristic, on students’ mathematics, reading, and science test scores. The study includes 48 elementary schools (clusters) from Iowa participating, with a total of roughly 2,300 students per grade (third, fourth and fifth) per year enrolled in these schools. We are primarily interested in the effect of the Science Writing Heuristic on student-level outcomes over time. A multivariate longitudinal regression model (Shah et al., 1997; Wu et al., 2009) is selected to simultaneously address correlation in the responses, non-ignorable missing data, and time dependence in the covariates.

Slides: pdf

JSM 2013: A multilevel mediation model

08/04/2013 at 10:30 am


In a longitudinal cluster randomized study, we investigate the role of the quality of instruction in mediating the effect of a new inquiry-based learning approach for teaching science, the Science Writing Heuristic (SWH), to Iowa students in third, fourth, and fifth grades. There are 48 schools (clusters) participating in the study, with a total of roughly 2300 students per grade per year enrolled in the schools. The students’ standardized test scores in math, reading, and science were collected each year. We are primarily interested in student-level outcomes and in quality of implementation at the teacher level as a mediator of those outcomes. The method for teaching science was assigned randomly to the school building; given this configuration, we will use a 3-2-1 mediation design (Pituch et al., 2010). We will also investigate whether other socioeconomic, sociodemographic, or learning variables moderate the indirect or direct effects of the SWH approach …

AERA 2013 Presentation: Two-level Quantile Regression

04/29/2013 at 10:27 am

Are Teachers to Blame?: Two-Level Quantile Regression of Teacher Quality on Student Performance

The impact of a new science curriculum on students’ critical thinking skills was investigated using a randomized controlled study. A structural equation model is used to model students’ abilities and evaluate the impact of the quality of the teacher on student performance as measured by their improvement on the Cornell Critical Thinking Test. A two-level quantile regression is used to dissect the effect of the teacher and the Science Writing Heuristic curriculum on students at various quantiles. The quality of the teacher is seen to have the greatest impact on students at lower quantiles. However, the overall impact of the quality of the teacher on the performance of the students is marginal. The impact of the curriculum is seen to have an impact at many quantiles, particularly for students at lower quantiles.

Disclaimer: After this submission …

NARST 2013 Presentation: Longitudinal Analysis of Student Achievement

04/11/2013 at 8:25 pm

Our research group at ISU had two proposals accepted to the 2013 NARST meetings in Puerto Rico. Our colleagues over at the University of Iowa also had two proposals accepted. We have had many positive results and are excited to share these results with the science teaching community.

It is the opinion of our research group that an inquiry based approach to teaching science is superior to many traditional approaches. It is essential that we move away from teaching students to memorize what they need to know from science books and move towards teaching students how to develop an understanding of the knowledge-building process by learning how to develop explanations and predictions about our world. As members of society are expected to process information that is updated constantly and rapidly, it is critical to understand how ideas are developed and processed. An inquiry based approach to teaching science facilitates this …

SREE Fall 2012 Presentation: Multilevel Model Evaluating Teacher Impact

09/07/2012 at 10:17 pm

Balancing Self-Directed Learning with Expert Mentoring: The Science Writing Heuristic Approach

Abstract: A multi-level model was used to evaluate the impact of an inquiry based learning approach to teaching science on student’s critical thinking abilities. The level of implementation of the learning approach was incorporated into the model through video ratings of the individual teachers using the Reformed Teaching Observation Protocol instrument. The results show that the inquiry based approach does have a significant positive impact on student’s critical thinking skills. There level of implementation is not significant.

Slides: SREE Fall 2012 Presentation