PnB-T06 - Comparing Alternative AI Syllabus Policies for Teaching Introductory Statistical Computing 

Information

Our research compares two syllabus policies for using generative artificial intelligence (AI) resources during SAS coding assessments in an introductory statistics course. The first policy, "must cite", allows use of AI on most coding assessments but requires students to clearly cite their use of such resources for each assessment. The second policy, "AI prohibited", prohibits the use of AI on all coding assessments. Our objective was to determine which syllabus policy was most effective for teaching SAS-naïve (and mostly coding-naïve) introductory statistics students to code using SAS statistical software. Four different instructors each enrolled two sections of the same introductory course into the study. Each instructor assigned one section to the "must cite" policy and the other to the "AI prohibited" policy. Four sections (two of each type) completed the study during fall 2025, and the remaining four sections completed the study during spring 2026. Accuracy of student code for two standardized SAS coding assessments was compared between AI policy groups. Students also completed an anonymous post-semester survey asking about their confidence for using SAS to conduct analyses in future classes or research, enjoyment in learning to code in SAS, pride in the SAS knowledge attained, and excitement to tell friends and family about their new coding skills, among other things. Survey responses were also compared between AI policy groups. Elon University is a private, mid-sized university in North Carolina. Statistics in Application is our advanced introductory statistics class that is required for many STEM and social science majors but also serves as a first-year foundations math course for students in any major. The vast majority of students in this class are first-years or sophomores, although each section typically also includes a small number of juniors and seniors. This class is generally the first exposure to coding for the majority of students. We offer 8-10 sections of Statistics in Application every semester, and almost every section typically fills at or near the capacity of 30 students. The course is frequently taught in a computer lab.