Metopen Kuantitatif - Review Artikel 16

fqs November 28, 2025 #PFIS258005 #kuliah

Nama: Firman Qashdus Sabil
NIM: 250321830676

Gambar/tabel/instrumen yang direferensi dapat dilihat langsung di dalam artikel dan lampiran artikel.

Identitas Artikel

Judul: Impacts of an AI-based chatbot on college students’ after-class review, academic performance, self-efficacy, learning attitude, and motivation.
Penulis: Yen-Fen Lee, Gwo-Jen Hwang, Pei-Ying Chen
Jurnal: Education Tech Research Dev (2022) 70:1843–1865
Tautan: https://doi.org/10.1007/s11423-022-10142-8

Objective

The article aims to examine whether integrating an AI-based chatbot into after-class review activities improves:

Academic performance
Self-efficacy
Learning attitude
Learning motivation

(Research questions listed on page 5–6). The chatbot is used as a tool for personalized, immediate feedback during review of public health infectious diseases.

Methods

Desain

Quasi-experimental design (page 9)
Two groups:
- Experimental: chatbot-assisted review (n=18)
- Control: traditional teacher-assisted review (n=20)

Duration

2-week intervention, two sessions/week (page 10)

Instruments

Learning motivation questionnaire (Wang & Chen, 2010)
Learning attitude questionnaire (Hwang et al., 2013)
Self-efficacy scale (Pintrich et al., 1991)
Semi-structured interviews

(page 9–10)

Analysis

ANCOVA for post-test comparisons while controlling for pre-test scores (page 12–14).
Qualitative thematic coding for interview data (page 15–16).

Results

Significant Improvements for Experimental Group The chatbot group outperformed the control group in all measured areas:

Variable	F-value	p-value	Effect size (η²)	Notes
Learning achievement	7.70	< .01	0.18	Large effect (Table 1, page 12)
Self-efficacy	9.25	< .01	0.21	Large effect (Table 2, page 13)
Learning attitude	8.60	< .01	0.19	Large effect (Table 3, page 13)
Learning motivation	5.39	< .05	0.13	Medium effect (Table 4, page 14)

Qualitative Themes

Students reported:

Increased engagement and willingness to ask questions
Ubiquitous access to answers
Personalized feedback enhancing intrinsic motivation

Some limitations noted by students:

Chatbot lacks human nuance, sometimes answers “I don’t know”
Incomplete or overly short replies on complex cases

Kesimpulan

The study demonstrates that integrating an AI-based chatbot into after-class review significantly enhances performance, self-belief, attitudes, and motivation. The authors attribute this to:

Instant feedback
Reduced shyness / hesitation in asking questions
Rich knowledge base about public health
Personalized learning pathway

However, limitations include novelty effect, short duration, small sample size, and chatbot’s inability to answer complex queries.

Advantages of the Article

Proper ANCOVA checks: Levene’s test + homogeneity of regression.
Clear theoretical grounding: constructivism, feedback theory, AI in education literature.
Quantitative + qualitative triangulation strengthens credibility.
Uses a real government-developed chatbot (Disease Stewardship App).

Gaps and Limitations

Extremely small sample (n=38)
Short-term intervention
No learning analytics, the study does not analyze: Chat logs, Misconceptions, Cognitive processes.
The article itself doesn’t investigate: why some students benefit more, what profile (prior knowledge, attitudes) influences success.

Novel Article Ideas

How Do Student Characteristics Moderate the Effectiveness of AI-Generated Feedback in Physics Learning?
AI-Supported After-Class Review for Electricity and Magnetism: Effects on Problem-Solving and Conceptual Understanding