From <scp>AI</scp> to <scp>TA</scp> : How to use <scp>ChatGPT</scp> to quickly create statistics and analytics assessments
Abstract This paper presents templates to use with ChatGPT to quickly create statistics and analytics assessments. The templates are a form of prompt engineering. They were derived through extensive trial and error. The templates are presented following the organizational structure found in a typical introductory statistics course: data classification, descriptive statistics, probability, and inferential statistics. Every template can be customized, and multiple examples are shown. These templates help to keep content fresh, reduce plagiarism, and align assessments with instructor learning outcomes. They also reduce writing time, which allows instructors to focus on higher‐level aspects of assessment development. It is important to note that ChatGPT's output can contain errors. ChatGPT's weaknesses and errors are discussed throughout the paper.
Authors
- Lauren Kenyo
- Raymond D. Frost (ORCID: https://orcid.org/0000-0001-6163-0178)
- Janna Chimeli
- Ellen R. Gordon (ORCID: https://orcid.org/0009-0001-6822-9501)
Institutions
- Ohio University (US)
Publication Details
- Journal
- Teaching Statistics
- Published
- 2025-09-01
- DOI
- https://doi.org/10.1111/test.70013
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00