Project overview
Next Level Teaching Blueprint is a multi-phase design-based research initiative examining how generative AI can function as a structured discourse partner rather than a shortcut to an answer. The project focuses on undergraduate and graduate research methods courses, where students must learn to articulate reasoning, evaluate evidence, question assumptions, and make defensible methodological decisions.
I co-lead the project as Student PI and Graduate Researcher. Our goal is to develop and test scalable instructional principles that strengthen higher-order thinking, critical thinking, and AI literacy while preserving learner agency.
Constraint-first, dialogic learning
The learning activities use a constraint-first sequence designed to keep student thinking visible:
1. Commit to initial reasoning
Students articulate an initial interpretation, methodological choice, or solution before consulting an LLM.
2. Interrogate through dialogue
Students use AI to surface alternatives, question assumptions, test claims, and identify limitations rather than simply request a finished answer.
3. Revise and reflect
Students revise their response, explain what changed, and document how they accepted, rejected, or reframed the AI’s contribution.
This structure treats the LLM as a discourse partner while requiring learners to remain accountable for the reasoning and final judgment.
Multi-phase research design
The project connects two stages of implementation:
- Spring 2025 pre-pilot: Quantitative analyses of statistical learning and AI literacy outcomes, including data validation, demographic reporting, pre/post comparisons, and publication-ready tables and visualizations.
- Spring 2026 CEHHS Phase 1 pilot: An integrated mixed-methods study across research methods courses, combining learning outcomes with process data and classroom evidence.
| Evidence stream | Analytic purpose |
|---|---|
| Pre/post measures | Examine changes in disciplinary learning, AI literacy, and critical thinking |
| Student learning diaries | Trace reflection, metacognition, and changes in learner judgment |
| AI chat logs | Study discourse sequences, prompting behavior, and interaction patterns |
| Assignment artifacts | Compare initial reasoning, revisions, and final disciplinary products |
| Classroom observations | Document implementation, participation, and instructional context |
| Faculty feedback | Identify feasibility, adaptations, and requirements for scaling |
Computational analysis of student–AI discourse
The computational analysis plan examines not only what students produce, but how they position themselves in relation to AI.
Interaction sequences
Modeling how students move among proposing, questioning, validating, revising, and reflecting across an AI-supported task.
Dialogue acts
Classifying requests, explanations, challenges, evaluations, and metacognitive statements within student–AI exchanges.
AI positioning
Examining whether learners treat AI primarily as an answer generator, a validator, or a co-thinker.
My contributions
My work includes:
- Co-leading the multi-phase design-based research initiative with the PI and interdisciplinary research team.
- Contributing to the design and classroom implementation of constraint-first, dialogic learning activities.
- Developing the integrated mixed-methods workflow connecting outcome measures, learning diaries, chat logs, artifacts, observations, and faculty feedback.
- Conducting quantitative analyses for the Spring 2025 pre-pilot study, including validation, descriptive reporting, pre/post comparisons, tables, and visualizations.
- Designing computational analyses of student–AI discourse, interaction sequences, dialogue acts, learner authority, and AI positioning.
- Working with participating faculty to translate findings into scalable pedagogical principles and implementation resources.
Scaling and dissemination
The team is using empirical findings to refine faculty resources and scalable implementation principles for research methods instruction. Scholarly dissemination is in development, including a Phase 1 manuscript targeted for CHI 2027.