Education Research Question Generation
Goal
Produce focused, answerable, evidence-aware research questions for education papers.
Inputs
- Topic or selected direction
- Study type: qualitative, quantitative, mixed methods, review, intervention, design-based research
- Target population and context
- Available data or planned data
- Expected paper level: course paper, master's thesis, PhD dissertation, journal article
Question Frameworks
Choose the framework that fits:
- Quantitative: PICOT, PICO, PECO, variable relationship model
- Qualitative: phenomenon + participants + context + meaning/process
- Mixed methods: quantitative relationship + qualitative explanation
- Intervention: population + intervention + comparison + outcome + time
- Review: population/context + concept/intervention + evidence scope
Workflow
- Classify the topic by method type and likely evidence base.
- Generate 8-12 candidate research questions:
- 3 descriptive/exploratory
- 3 explanatory/relationship
- 2 intervention/effectiveness if applicable
- 2 qualitative mechanism/process questions if applicable
- Search scholarly databases for each candidate or keyword cluster.
- Score questions on:
- clarity
- answerability
- novelty
- data feasibility
- theoretical contribution
- method fit
- Rewrite top questions into publishable form.
- Generate aligned hypotheses or subquestions.
Tool Calls
Elicit
Use Elicit manually or through API if available. Best for evidence-grounded research reports and extraction tables.
- Web: https://elicit.com/
- Query pattern: "What is known about [topic] in [population/context]?"
- Extraction fields:
research question, population, intervention, variables, outcomes, method, limitations.
Semantic Scholar Search
curl "https://api.semanticscholar.org/graph/v1/paper/search?query=student%20engagement%20AI%20feedback%20education&limit=20&fields=title,year,abstract,citationCount,authors,venue,fieldsOfStudy"
Semantic Scholar Recommendations
Use after selecting a seed paper:
curl "https://api.semanticscholar.org/recommendations/v1/papers/forpaper/PAPER_ID?fields=title,year,abstract,citationCount,authors"
OpenAlex Query
curl "https://api.openalex.org/works?search=student%20engagement%20AI%20feedback%20education&filter=from_publication_date:2020-01-01&sort=cited_by_count:desc&per-page=20&mailto=YOUR_EMAIL"
Output Format
Return:
| Candidate RQ |
Type |
Variables/Phenomenon |
Population/Context |
Evidence Base |
Feasibility |
Risk |
Improved Version |
Then provide:
- Recommended main research question
- 2-4 subquestions
- Optional hypotheses
- Required data
- Suggested analysis method
- Why this RQ is stronger than alternatives
Quality Rules
- Avoid questions that are too broad, e.g. "How does AI affect education?"
- Use measurable verbs for quantitative questions: predict, mediate, moderate, influence, compare.
- Use meaning/process verbs for qualitative questions: experience, perceive, negotiate, construct, adapt.
- Ensure each question implies a feasible method.
1---2name: education-research-question-generation3description: Use when generating, refining, comparing, or validating education research questions and hypotheses. Supports qualitative, quantitative, mixed-methods, intervention, design-based research, survey, classroom observation, and systematic review questions using Elicit-style evidence search, Semantic Scholar/OpenAlex, and LLM reasoning.4---56# Education Research Question Generation78## Goal910Produce focused, answerable, evidence-aware research questions for education papers.1112## Inputs1314- Topic or selected direction15- Study type: qualitative, quantitative, mixed methods, review, intervention, design-based research16- Target population and context17- Available data or planned data18- Expected paper level: course paper, master's thesis, PhD dissertation, journal article1920## Question Frameworks2122Choose the framework that fits:2324- Quantitative: PICOT, PICO, PECO, variable relationship model25- Qualitative: phenomenon + participants + context + meaning/process26- Mixed methods: quantitative relationship + qualitative explanation27- Intervention: population + intervention + comparison + outcome + time28- Review: population/context + concept/intervention + evidence scope2930## Workflow31321. Classify the topic by method type and likely evidence base.332. Generate 8-12 candidate research questions:34 - 3 descriptive/exploratory35 - 3 explanatory/relationship36 - 2 intervention/effectiveness if applicable37 - 2 qualitative mechanism/process questions if applicable383. Search scholarly databases for each candidate or keyword cluster.394. Score questions on:40 - clarity41 - answerability42 - novelty43 - data feasibility44 - theoretical contribution45 - method fit465. Rewrite top questions into publishable form.476. Generate aligned hypotheses or subquestions.4849## Tool Calls5051### Elicit5253Use Elicit manually or through API if available. Best for evidence-grounded research reports and extraction tables.5455- Web: https://elicit.com/56- Query pattern: "What is known about [topic] in [population/context]?"57- Extraction fields: `research question`, `population`, `intervention`, `variables`, `outcomes`, `method`, `limitations`.5859### Semantic Scholar Search6061```bash62curl "https://api.semanticscholar.org/graph/v1/paper/search?query=student%20engagement%20AI%20feedback%20education&limit=20&fields=title,year,abstract,citationCount,authors,venue,fieldsOfStudy"63```6465### Semantic Scholar Recommendations6667Use after selecting a seed paper:6869```bash70curl "https://api.semanticscholar.org/recommendations/v1/papers/forpaper/PAPER_ID?fields=title,year,abstract,citationCount,authors"71```7273### OpenAlex Query7475```bash76curl "https://api.openalex.org/works?search=student%20engagement%20AI%20feedback%20education&filter=from_publication_date:2020-01-01&sort=cited_by_count:desc&per-page=20&mailto=YOUR_EMAIL"77```7879## Output Format8081Return:8283| Candidate RQ | Type | Variables/Phenomenon | Population/Context | Evidence Base | Feasibility | Risk | Improved Version |84|---|---|---|---|---|---|---|---|8586Then provide:8788- Recommended main research question89- 2-4 subquestions90- Optional hypotheses91- Required data92- Suggested analysis method93- Why this RQ is stronger than alternatives9495## Quality Rules9697- Avoid questions that are too broad, e.g. "How does AI affect education?"98- Use measurable verbs for quantitative questions: predict, mediate, moderate, influence, compare.99- Use meaning/process verbs for qualitative questions: experience, perceive, negotiate, construct, adapt.100- Ensure each question implies a feasible method.
ilog3