AI evidence extraction

The Perception by University Students of the Use of ChatGPT in Education

Thi Thuy An Ngo · International Journal of Emerging Technologies in Learning (iJET) · 2023

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

7/10

Relevance

1/4

Quality (LMQS)

Evidence

257

Citations

9.11

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.3991/ijet.v18i17.39019

Methodology & findings

Study design

Mixed-methods study combining quantitative survey and qualitative semi-structured interviews.

Sample

N = 200, 3 groups

Primary method

SPSS software version 25 employed for statistical analysis. Methods included: (1) Descriptive statistics (means, standard deviations) calculated for Likert scale responses; (2) One-sample t-tests conducted to compare mean scores against test value of 3.00 (midpoint of 5-point scale) to determine if perceptions were above average; (3) Cronbach's alpha test for internal consistency/reliability assessment (α > 0.9); (4) Qualitative thematic analysis for interview data (transcription, coding, categorization).

Main result

The study found that "students had generally positive attitudes toward using ChatGPT for learning," with an overall mean score of 3.61 on a 5-point Likert scale. The research revealed that "the benefits of ChatGPT, according to students, included saving time, providing information in various areas, providing personalized tutoring and feedback, and illuminating ideas in writing." Additionally, "the most concerning issues for students while using ChatGPT were inability to assess the quality and reliability of sources, inability to cite sources accurately, and inability to replace words and use idioms accurately."

Reports effect sizes and confidence intervals.

Research paradigm

Pragmatist/Mixed Methods

Author conclusions

"The purpose of this study was to determine what university students thought about using ChatGPT for learning. Data analysis revealed that students had generally positive attitudes toward using ChatGPT for learning. Aside from the benefits of ChatGPT, such as saving time, providing information in various fields, providing personalized tutoring and feedback, and illuminating ideas in writing, students encountered several barriers, such as inability to examine the quality and reliability of sources, inability to cite sources correctly and accurately, inability to replace words and use idioms wisely." The authors recommend "verifying ChatGPT responses with reliable sources; using ChatGPT as a reference source or consultant tool; providing guidelines for ChatGPT use; and most importantly, promoting academic integrity among students to ensure ethical uses of ChatGPT in an academic context."

Risk of bias

Selection bias: Convenience sampling from Vietnamese university students with prior ChatGPT experience only; Self-report bias: Questionnaires and interviews vulnerable to social desirability and response bias; Lack of control condition: No comparison group or alternative conditions; Limited generalizability: Single country, single education level (university only); Sample composition bias: 59% male, concentrated in Information Technology (36.5%); Selection bias: Random sampling of Vietnamese university students only; limited to single institution context; Response bias: Self-reported measures via questionnaire and interviews susceptible to social desirability bias; Sampling bias: 15% of original sample (30/200) selected for interviews may not be representative; Temporal bias: Data collected in February 2023, shortly after ChatGPT launch (November 2022); limited user experience; Geographic/cultural bias: Study limited to Vietnamese university students only; Attrition: Study does not address dropout rates or missing data handling; Selection bias: Random sampling used but limited to Vietnamese university students with prior ChatGPT experience; Response bias: Self-report measures susceptible to social desirability bias; Recall bias: Participants asked to report on past experiences with ChatGPT; Temporal bias: Data collected in February 2023, shortly after ChatGPT launch (November 2022), may reflect novelty effects; Geographic/cultural bias: Limited to one country (Vietnam) and FPT University context

Limitations

The authors acknowledged several limitations: "Firstly, the focus on university students may limit the generalizability of the results to other educational levels and age groups." Furthermore, "the reliance on self-report measures, such as questionnaires and interviews, introduces the possibility of response biases and subjective interpretations." Finally, "the study primarily focused on exploring students' perceptions and recommendations for using ChatGPT in an educational context, but it did not extensively examine the feasibility or practical implementation of the proposed solutions."

Open questions raised

The authors identify: (1) limited research on ChatGPT application in higher education; (2) gap in understanding students' actual experiences with ChatGPT in educational contexts; (3) need to examine feasibility and practical implementation of proposed solutions; (4) need for research across diverse educational levels and age groups beyond university students.
Scarcity of studies examining students' actual experiences with ChatGPT in educational contexts
Need for understanding implications of integrating ChatGPT into learning processes
Limited research on ChatGPT applications in higher education despite growing interest
Lack of feasibility and practical implementation studies of proposed solutions
Need for research on ChatGPT use across other educational levels beyond university students

Extracted from: pdfAgreement 64%

Explore related topics

Responsible AI Use Academic Writing Mixed Methods