Projects

• Role: Principal Investigator (PI)

• Partner & Sponsoring Organization: Iowa Measurement Research Foundation (IMRF)

• Focus Population: Secondary School Students

• Timeline: 2026–2028

Many school-based suicide prevention programs train students to notice when a peer may be struggling, respond with care, and connect that peer to a trusted adult. Yet most evaluations still rely on what students say they feel ready to do. Confidence, preparedness, and self-efficacy are important, but they do not fully show how a student might respond when faced with a realistic peer distress situation.

This project develops an AI-simulated peer interaction assessment to better understand students’ demonstrated suicide gatekeeping skills. Instead of only asking students whether they feel prepared, the assessment presents brief, structured peer scenarios involving emotional distress, social withdrawal, hopelessness, or suicide risk cues. Students respond as they would in a peer-support situation, and their responses are evaluated using a structured scoring rubric.

The assessment focuses on four core skills: recognizing warning signs, responding supportively, understanding the boundaries of a student helper role, and connecting a peer to appropriate adult or professional support. The goal is not to train students to act as counselors or conduct clinical suicide risk assessments. Rather, the project examines whether students can identify concerning cues, communicate support, and take developmentally appropriate helping actions within a school-based safety and support system.

A major purpose of this project is to examine whether AI-simulated interactions can be used as a reliable and meaningful measurement tool. The study includes the development of standardized scenarios, scoring criteria, expert review, rater training, and preliminary analyses of score consistency. This allows the project to ask not only whether students report feeling prepared, but also whether their responses show observable gatekeeping skills.

This study is part of a broader research agenda on performance-based assessment in youth mental health, school safety, and peer support. By using AI simulation as a structured assessment method, the project aims to strengthen how researchers and educators evaluate applied mental health competencies beyond self-report. The long-term goal is to build more authentic, ethical, and evidence-informed ways to assess whether students can translate suicide prevention training into supportive action.