Data Visualization of Science Test Score Distribution Across Districts

• Activity Number: 558
• Type: Contributed
• Date/Time: Wednesday, August 3, 2016 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Education
• Author(s): Leila Jamoosian* and Kathryn Hayes and Eric Suess
• Companies: and California State University at East Bay and California State University at East Bay
• Keywords: Visualization ; Analytics Report ; R ; Tableau ; Data Mining ; Educational Data

Abstract

This study presents statistical visualization techniques for IMSS(Integrated Middle School Science) data focused on trends in student science performance. Because information visualization plays an important role in both the research and market areas, we used several different techniques to summarize the data for reporting purposes, which report on here. Our methods consisted of two phases. First, we cleaned the data via Excel and R. Since the collected data was gathered from teacher and student records covering 2005-2013, analysis required unifying teacher and school codes and merging districts, along with other complex large data set manipulations. In the second phase, we used data visualization methods such as charts, maps and time series plots to illustrate trends in student scores over six years, as well as the distribution by gender, ethnicity and a poverty index (Free and Reduced Lunch). The latter was completed using R and Tableau software. Overall we present methods for cleaning, analyzing and visualizing primary and key factors in student science score trends.