Year: 2024
Type: article
Abstract: ABSTRACT We investigate the accuracy of ChatGPT's answers to multiple‐choice and short essay questions from a widely used OER economics textbook. The questions are classified by their difficulty and cognitive level. In our tests, the bot scored a high D in multiple‐choice questions and a low A in essay questions. Most errors were found in questions of higher cognitive levels, and with calculations and graphs. While a student who has attended lectures would not gain much, if at all, from using the bot for multiple‐choice questions, the gain for a higher grade in essays is considerably larger.
Source: Bulletin of Economic Research
DOI: 10.1111/boer.12481
Authors: Rachel Faerber‐Ovaska, Tomi Ovaska, Joseph Palardy, Yogesh Uppal
Institution: Youngstown State University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
Estimated APC paid: $2,760
OpenAlex ID: W4404870021