Reading-grounded guidance for analyzing education policy cases.
The question
How can students use course readings to reason through a policy case, weigh competing interests, and make the assumptions behind a recommendation visible?
My role and collaborators
I worked with Fernando M. Reimers to curate course sources and build agents for stakeholder analysis, policy memo development, and multidimensional reform evaluation. I also tested an integrated assistant with A801 policy cases and activities.
A decision that shaped the work
I built specialist guidance around the tasks students were doing in the course and connected those functions through a single assistant. For a memo, the response can ask which source supports a claim. For stakeholder analysis, it can ask who might bear the costs of a proposed reform.
How the activity shapes the scaffold
A policy case can invite different kinds of reasoning. The response should depend on whether the student is identifying stakeholders, connecting evidence to a recommendation, or comparing a reform across dimensions.
Bot walkthrough
I wrote this fictional dialogue to illustrate the design decisions.
Map stakeholders
Learner
“Our case proposes moving to a centralized school assignment process. Who should I consider?”
Sample agent reply
“Start with families, school leaders, and the district office. For each group, write down what could change for them. Which families might have trouble using the new process?”
Design annotation
The prompt asks students to consider differences within a stakeholder group.
Develop a memo
Learner
“My recommendation is to pilot the policy in a few schools.”
Sample agent reply
“Which evidence in the case supports a pilot? Name one assumption you are making and explain how the pilot would help you test it.”
Design annotation
The reply asks students to connect an activity-specific recommendation to evidence and an assumption.
Evaluate a reform
Learner
“The new process could make assignment easier to administer.”
Sample agent reply
“Compare administrative simplicity with access for families. What additional information would help you decide how much weight to give each concern?”
Design annotation
The reply brings a trade-off into view while leaving the policy judgment with the learner.
Artifacts
Draft media spaces for approved examples of course integration.
Screenshot pending
Course activity. Show where the agent appears in an approved case activity or guide.
Screenshot pending
Agent guidance. Show how the assistant guides evidence use or the consideration of trade-offs.
Video demo pending
A801 course activity demo. A captioned recording could demonstrate the handoff from the course prompt to task-specific guidance.
Evidence, status, and limitations
Completed work
Specialist agents, a tested integrated assistant, and availability within A801.
Proposed analysis
An exploratory analysis would look at help-seeking, the guidance offered, and the reasoning visible in later turns. Analysis status is pending confirmation.
Findings and limitations
Findings are pending confirmation. Testing agent responses with course cases examines the guidance they offer. Student reasoning requires evidence from student work or interactions.