Megan Heyman & Eric Reyes (Rose-Hulman Institute of Technology)
In traditional final exams, students work alone to answer a specific set of questions, potentially by analyzing a "toy" dataset. Even if real data is used during the exam, this format does not assess how STEM students will utilize statistics after the course. Moreover, a revised emphasis was added to the Guidelines for Assessment and Instruction in Statistics Education (GAISE) in 2016, recommending that statistics be taught as an investigative process of problem-solving and decision-making. Our goal is to incorporate the revised GAISE recommendation to the final course assessment, while simultaneously creating an authentic experience for students who will eventually produce their own statistical studies and analyses.
Similar to a scientific practicum, students collaborate in small groups to design a statistical study which addresses a question of interest, collect relevant data, perform the statistical analysis, and interpret the results during the time when the final exam is traditionally held. The assessment technique is novel from traditional final projects, since students complete the study during the allotted exam time, while the instructor is present. We will describe how we have implemented this exam format over the last 5 terms, challenges we have faced, and positive outcomes we've observed.