PnB-M07 - Exploring Students' Perceptions of Generative AI Use in Mathematical Statistics

Information

This research explores the affordances and limitations of generative AI undergraduate students in a mathematical statistics course identified when interacting with generative AI tools to solve and craft statistical problems. The research complements current discussions around generative AI in postsecondary education, delving into the experiences of students in a senior-level course. Students' opinions were captured mainly through the comments and reflections they wrote when working on two assignments related to moment generating functions and sampling distributions. Two follow-up interviews were conducted and treated as complementary sources for the investigation. Through this research study, we center students' ideas/voices within the context of an advanced statistics class where most students have prior experience with statistics and data science topics. This research was conducted within an undergraduate mathematical statistics course at a Research 1 institution. The course is the second semester of a two-semester sequence of courses that focuses on statistical inference theory (e.g., sampling distributions, point and interval estimation, and hypothesis testing). Upper-level undergraduate students majoring in statistics, data science, mathematics, or other related fields often take the course, with approximately 30-40 students in a section. This research was conducted within a section of this course that did not assess students using exams; instead, students completed individual assignments (i.e., Questions of the Day and Learning Checks that are similar to entry tickets and open-book week-long quizzes), as well as collaborative group assignments (homework assignments and two projects). Students were asked to cite any instances of their generative AI use, and throughout the semester, the course explicitly integrated discussions, assignment activities, and assignment questions that encouraged students to interact with and reflect on their use of generative AI. Guided by the Technology Acceptance Model (Davis et al., 1989), we conducted a thematic analysis of individual student responses to course assignments, a course project completed in groups, and individual student interviews. In the course assignments and projects, students were prompted to use generative AI throughout the problem and reflect on the output given by generative AI. In the interviews, students were asked questions about their use of generative AI before and during the mathematical statistics course. Themes resulting from this thematic analysis provide insight into students' perceived affordances and limitations of using generative AI.