ISEDJ

Abstract: Data analytics education increasingly incorporates generative AI tools, yet AI output can be inconsistent or misleading. This teaching case develops critical thinking by challenging students to evaluate AI-generated interpretations of data visualizations through Robert Ennis' three dimensions: logical, criterial, and pragmatic. Using a wage-by-tenure visualization that subtly violates linear regression assumptions, students prompt ChatGPT with structured queries, assess responses, and refine their prompts while engaging in peer reflection. The case targets the Data Understanding phase of CRISP-DM, where students must determine whether relationships are linear or require feature engineering. Structured rubrics guide assessment of prompt engineering and reasoning. Designed for undergraduate analytics, MIS, and statistics courses (adaptable to graduate programs), this case prepares students for AI-augmented workplaces by strengthening both technical skills and professional judgment in evaluating AI outputs. Supporting materials include complete datasets, sample deliverables, and validated implementation guides.

Download this article: ISEDJ - V24 N6 Page 31.pdf

Recommended Citation: Larson, B.E., Bohler, J.A., Bolekar, N., (2026). Developing Critical Thinking in Data Analytics Education: A Teaching Case Evaluating ChatGPT Responses to a Visualization. Information Systems Education Journal 24(6) pp 31-43. https://doi.org/10.62273/RLDU1198