Frequently Asked Questions
About SafeInsights | Data & Privacy | Research Process | Technical | Getting Help
About SafeInsights {#about}
What is SafeInsights?
SafeInsights is education research infrastructure that enables you to analyze learning data from multiple platforms while protecting student privacy. Your code runs securely within data enclaves—you receive aggregated results without ever seeing individual student records.
What kinds of research can I do?
Version 1 supports post-hoc analysis—studying existing learning data to understand patterns, relationships, and predictions. You can analyze engagement, performance relationships, subgroup differences, predictive models, and longitudinal trends.
You cannot yet run interventions, A/B tests, or cross-platform studies.
Data and Privacy {#privacy}
Will I ever see individual student records?
No. You receive only aggregated outputs that meet privacy thresholds (typically minimum cell sizes of 10-20 observations). Individual student records never leave the secure enclave and are never visible to researchers.
Learn about privacy protections →
What data is available through SafeInsights?
Currently: homework systems, digital textbooks, learning management systems, and courseware platforms. Data types include event logs, assessment responses, resource usage, notes and highlights, and (where available) limited demographics.
Specific availability varies by platform—some have rich longitudinal data, others have deep item-level detail.
Can I see what variables are available before I request data?
Yes. Each platform provides:
- Simulated data files showing structure and variable names
- Data dictionaries documenting all available fields
- Example queries showing typical analyses
Explore OpenStax data example →
Research Process {#process}
What are the main steps to conduct research?
- Check feasibility (data catalog + research guide)
- Write and submit proposal
- Proposal review (2-4 weeks)
- Develop analysis container with simulated data (2-6 weeks)
- Container execution on real data (hours)
- Output review and release (1-2 weeks)
Typical total timeline: 2-4 months
How do I know if my research question is feasible?
Start by:
- Reviewing the research patterns guide
- Checking the data catalog for platforms with relevant data
- Asking yourself: "Can this be answered with observational data that already exists?"
Still unsure? Schedule a feasibility consultation →
How much does it cost to use SafeInsights?
Studies on SafeInsights incur computational costs data processing resources—these are charged by the data organization at cost. Additional costs charged by data organizations correspond to administrative effort of reviewing study proposal, code, and outputs.
Technical {#technical}
What programming languages can I use?
Python, R, and SQL. You write your analysis using standard libraries and packages, then package it as a container for execution in the enclave.
I've never used containers—is this too technical for me?
Many researchers learn containerization in a few hours using our templates and guides. If you can write Python or R scripts, you can learn to package them as containers.
How do I test my code before submitting?
Use the simulated data provided by each platform. This has the same structure and variable names as real data, allowing you to develop and debug locally before submitting your container.
Getting Help {#help}
How do I get support?
- Email: support@safeinsights.org (response within 2 business days)
- Office hours: Weekly sessions for live Q&A (see schedule)
- Feasibility consults: 30-minute meetings to discuss your research idea
What if I have a question not answered here?
Contact us → with your question. Common questions get added to this FAQ to help future researchers.
Last updated: December 2025
Have a question not answered here? Contact support →