Online Program

Return to main conference page

All Times EDT

Friday, June 5
Education
Statistics for the Engaged Citizen: Revising Educational Practices to Increase Relevance in Everyday Life
Fri, Jun 5, 11:15 AM - 12:50 PM
TBD
 

Real Data Analysis in the Classroom Through the Use of Case Studies (308106)

*Leah Jager, Johns Hopkins Bloomberg School of Public Health 

Keywords: case studies, real data, education

When teaching statistics, we often present students with clean data analysis examples designed to demonstrate one concept at a time. This allows students to focus on a single method while they are learning, but can lead students to think that data analysis is clean — there’s a clear analysis pathway with only a single right way to analyze a data set. Real data analysis, however, can be messy and ambiguous. Analysts make decisions at each stage in the process and often there isn’t only one right choice, although there may be some wrong ones! It’s important to expose students to this real process as well as help them understand the reasons for different analytic choices.

Case studies provide one way to bring real data analysis into the classroom. Case studies are in-depth data analyses derived from interesting real-world questions and have non-trivial solutions that leave room for different analysis paths. Through case studies, instructors model the data analytic process, allowing students to see “why” analysts make certain choices and helping them improve their data analytic skills. Case studies also help students to make connections between the theory they learn in class and the application of that theory to a real-world problem.

A main drawback to teaching with case studies is that constructing meaningful case studies takes considerable time and effort. To help with this, we have been developing a library of complete data analyses that start with a motivating question of interest and then work through the steps of accessing, importing, and wrangling the data, followed by data visualization, modeling, and communicating results. At each step, we try to communicate now only “what” choice the analyst made but also “why” the choice was made.