Chatbot Test Anxiety Intervention Study

ISEF Category: Behavioral and Social Sciences

Subcategory: Other  ·  Difficulty: Intermediate  ·  Setup: Home Setup  ·  Time: 1 to 2 Months

The Hook

Test anxiety can spike fast, even when you know the material. A short chatbot message may calm that stress in the moment, like a pocket-sized coach. That makes it a strong research idea, because you can test whether a tiny intervention changes real scores. You also get a clean control group, which helps you tell signal from noise.

What Is It?

This project asks a simple question: can a short chatbot conversation lower test anxiety more than a neutral chatbot conversation? You are not testing whether chatbots are “good” or “bad.” You are testing one very specific behavior change tool, a micro-intervention, which means a tiny, focused support prompt.

Think of it like two versions of the same pep talk. One version gives coping steps, like reframing worried thoughts, slowing breathing, or planning the next action. The other version sounds friendly but does not teach those skills. If the first version lowers STAI scores more, you have evidence that the coping content matters.

The STAI, or State-Trait Anxiety Inventory, is a survey that measures how anxious someone feels right now. That makes it useful here because you want to see short-term change, not just general personality.

Why This Is a Good Topic

This is a strong science fair topic because it has a clear independent variable, a measurable outcome, and a real-world use. Test anxiety affects grades, confidence, and stress, so your project connects directly to student life. You can also learn core research skills, like random assignment, survey design, control groups, and basic statistical analysis, without needing a lab bench.

Research Questions

• How does a chatbot-delivered CBT micro-intervention change pre-to-post STAI scores compared with a sham-chatbot control?
• What is the effect of adding a coping-skills prompt versus a neutral check-in on immediate test anxiety?
• Does the intervention work better for students with higher baseline STAI scores?
• To what extent does message length change the size of the anxiety reduction?
• Which chatbot tone, reassuring, formal, or peer-like, produces the lowest post-intervention STAI scores?
• Does including a plan-for-the-next-step prompt improve anxiety reduction more than relaxation-only guidance?

Basic Materials

• Phone, tablet, or laptop with chatbot access.
• Google Forms or another survey tool for pre and post STAI responses.
• Random assignment sheet or spreadsheet.
• Google Sheets for scoring and data entry.
• Consent and assent forms approved by your school or fair rules.
• Quiet space for participants to read and answer without interruption.
• Stopwatch or timer for keeping each session consistent.

Advanced Materials

• A secure chatbot platform or locally hosted interface for version control.
• REDCap, Qualtrics, or another compliant survey system for participant tracking.
• Exportable CSV logs from the chatbot sessions.
• R or Python for data cleaning and statistical tests.
• JASP or jamovi for noncoding analysis and effect sizes.
• A document editor with version history for intervention scripts.
• Institutional review materials, if your school or mentor requires them.

Software & Tools

• Google Forms: Collects pre and post STAI responses in a simple survey format.
• Google Sheets: Stores randomization, scores, and basic summaries.
• R: Runs t-tests, ANCOVA, effect sizes, and plots for your results.
• JASP: Lets you analyze group differences without writing code.
• Python: Helps you automate scoring, clean exports, and check for missing data.

Experiment Steps

1. Define the exact anxiety skill your chatbot will teach, such as reframing, breathing, or planning the next step.
2. Build two matched versions of the chatbot script, one with the intervention and one with neutral filler.
3. Set your random assignment rules so each participant has the same chance of ending up in either group.
4. Choose your measurement plan, including when you will collect pre and post STAI scores and how you will score them.
5. Plan your analysis before collecting data, including the main comparison and any subgroup checks.
6. Write your ethics and privacy rules so participants know what data you will keep, store, and delete.

Common Pitfalls

• Letting the intervention chatbot sound much warmer than the control chatbot, which makes tone a hidden variable.
• Changing the script between participants, which breaks consistency and weakens the comparison.
• Mixing up state and trait anxiety scores, which can blur short-term change with long-term baseline anxiety.
• Letting students discuss their assigned version before the post survey, which can contaminate the results.
• Ignoring missing or rushed survey answers, which can distort the average STAI change.

What Makes This Competitive

A class-level version of this project tests whether one chatbot lowers anxiety more than another. A stronger version controls message length, tone, and timing, then tests whether the CBT element itself drives the effect. You can also add a harder analysis, such as checking whether the effect changes with baseline anxiety or prior test experience. Careful group matching and clean preplanned statistics will make the project much stronger.

Project Variations

• Test a chatbot micro-intervention for math test anxiety instead of general test anxiety.
• Compare text-only chatbot support with a chatbot that includes short breathing prompts.
• Measure whether the intervention works better for students before quizzes, finals, or practice exams.

Learn More

• PubMed: Search for review articles on test anxiety, cognitive behavioral therapy, and digital mental health interventions.
• NIH PubMed Central: Read free full-text studies on anxiety interventions and school-based mental health.
• APA Dictionary of Psychology: Look up terms like state anxiety, trait anxiety, and cognitive restructuring.
• OpenStax Psychology 2e: Review the basics of anxiety and stress in an accessible textbook.
• JASP documentation: Find free guides for t-tests, ANCOVA, and effect size analysis.
• NIH RePORTER: Explore funded research projects on adolescent anxiety and technology-based interventions.