Tu-20: The value of Log files of students’ interaction with software applications: Performance Evaluation in Correlation Guessing


By Shunqi Zhang, Dennis Pearl, Matthew Beckman, Neil Hatfield, & Yiyun Gong (Penn State University)


Abstract

The Book Of Apps for Statistics Teaching (BOAST: https://sites.psu.edu/shinyapps/), is a collection of interactive and open access educational R Shiny applications designed to augment instruction for a variety of undergraduate statistics courses. Over the past several years, we’ve collected log files of students' interactions with these apps. Participants will see the expanded opportunity provided by the analysis of log files from interaction with educational software to evaluate student learning and improve software to better meet educational goals. As an example, we analyzed a popular BOAST application on guessing the correlation form a given scatter plot, some of which have noticeable outliers. The poster presents data for 124 individuals with at least 30 responses per individual. We evaluated both the overall improvement on response correctness by students as trials proceed as well as the effect of outliers on their correctness. We hope this will inspire participant usage of log files for studying other sorts of web-based pedagogy. 


Recording

Tu-20 - The value of Log files of students’ interaction with software applications- Performance Evaluation in Correlation Guessing.pdf

register