Objective: To determine the difference in self-esteem scores among high school students who participate in artificial intelligence counseling and those who do not when controlling for pretest scores.
Methods:
- Quasi-experimental, nonequivalent control-group, pretest-posttest design
- 74 high school students from a private Christian school in Maryland
- Treatment group received 20 minutes of AI counseling weekly for 4 weeks
- Rosenberg Self-Esteem Scale (RSES) used to measure self-esteem
- One-way ANCOVA used for data analysis
Key Findings:
- No statistically significant difference between treatment and control groups
- Treatment group showed slight increase in mean scores (28.46 to 28.84)
- Control group showed slight decrease in mean scores (29.97 to 29.21)
- ANCOVA results: F(1,71) = 1.015, p = 0.317, partial eta squared = 0.014
Implications:
- AI counseling may serve as an affordable alternative to traditional counseling
- Results suggest potential positive impact on student self-esteem, though not statistically significant
- Schools may consider implementing AI as a supplementary counseling tool
- Study provides framework for future research on AI counseling in education
Limitations:
- Small sample size (74 participants)
- Short intervention duration (4 weeks)
- Single school setting
- Limited diversity in sample
- Self-reported questionnaire bias
- Assumption of normality violation
Future Directions:
- Conduct studies with larger sample sizes
- Increase diversity of student participants
- Extend study duration
- Use non-self-reported instruments
- Examine longitudinal effects
- Investigate different implementation designs
Title and Authors: "Improving the Self-Esteem Scores of High School Students Through Artificial Intelligence Counseling: A Quantitative, Quasi-Experimental Study" by Alexis Breanne Bolen
Published on: 2024
Published by: Liberty University (Dissertation)