Source-linked AI summary
Exploring Teachers' Perceptions of ChatGPT Through Prompt Engineering
Dimitrios Gousopoulos
TL;DR
This pilot study asks whether prompt-engineering training changes secondary science teachers’ perceptions of ChatGPT as an educational assistant. After training, teachers’ attitudes improved, and the authors describe potential educational applications while urging cautious interpretation.
Problem
The study asks whether prompt-engineering training improves secondary science teachers’ perception of ChatGPT as an educational assistant.
Source: Discussion · p. 5Method
The pilot study trained science teachers in prompt engineering and compared questionnaire responses before and after the intervention.
Source: Results · p. 3, Discussion · p. 5Results
Z = -3.296, p < .001; teachers’ median attitude scores were higher after training than before it.
Source: Results · p. 3Takeaways & Limitations
The authors describe prompting as a practical classroom tool for personalized scenarios, physics inquiry activities, identifying learning difficulties, and tailored assessment questions.
Source: Discussion · p. 5The pilot had few participants, a short intervention, and a questionnaire not validated for the Greek context.
Source: Discussion · p. 5
Abstract
from arXiv · showhide
Artificial Intelligence and especially Large Language Models (LLM), such as ChatGPT has revolutionized the way educators work. The results we get from LLMs depend on how we ask them to help us. The process and the technique behind an effective input is called prompt engineering. The aim of this study is to investigate whether science educators in secondary education improve their attitude toward ChatGPT as a learning assistant after appropriate training in prompt engineering. The results of the pilot study presented in this paper show an improvement in the previously mentioned teachers perceptions.
Results
After prompt-engineering training, teachers’ attitude scores toward ChatGPT as an educational assistant were higher, with statistical tests indicating a significant change. A bootstrap analysis estimated a positive pre/post mean-score difference.
Results
- Z = -3.296, p < .001; teachers’ median attitude scores were higher after training than before it.The comparison used mean questionnaire-item responses and the Wilcoxon signed-rank test.Source: Results · p. 3, Figure 1 · p. 4
- The bootstrap estimated the pre/post mean-score difference from 10,000 resamples with replacement, using a percentile-method 95% confidence interval.The authors used bootstrapping because the study had few participants.Source: Results · p. 4
- The confidence interval for the mean difference between pre- and post-intervention scores ranged from 0.682 to 0.909; because it excluded zero, the change following intervention was statistically significant.Source: Results · p. 4
- The authors conclude that the findings support their hypothesis that prompt-engineering training improves secondary teachers’ perception of ChatGPT as an educational assistant.Source: Results · p. 4
Discussion
This pilot study reports improved teacher attitudes toward ChatGPT after prompt-engineering training, including greater confidence and appreciation of it as an educational assistant. The authors describe potential classroom applications while noting constraints that call for cautious interpretation.
Discussion
- Following the intervention, teachers showed greater confidence in what they could achieve with an LLM and greater appreciation of LLMs as educational assistants.Source: Discussion · p. 5
- Teachers could use prompting to personalize learning scenarios, develop physics inquiry activities, identify learning difficulties, and tailor assessment questions.The authors connect these uses with differentiated instruction and inclusive learning.Source: Discussion · p. 5
- Interpret the pilot cautiously because it had few participants, a short intervention, and a questionnaire not validated for the Greek context.The questionnaire’s validity and reliability were not tested in Greece because of time constraints.Source: Discussion · p. 5
- The authors suggest future qualitative research to explore factors influencing changes in teachers’ attitudes toward ChatGPT in education.Source: Discussion · p. 5