How we report
Our Methodology
Formative Spaces investigates what is changing in education, why it is happening and what the evidence shows. We connect public records, data and attributed reporting so readers can examine the basis of each finding.
Editorial responsibility
Eric Gutierrez is the founder and principal author and directs the publication’s editorial work. Formative Spaces Research is credited as an institutional author. Every article receives human editorial review before publication and continued human review afterward. These bylines identify responsibility for the publication; the process below explains how its work is produced.
We do not accept money from donors, interest groups or lobbyists, and we do not seek to steer education policy in a particular direction. Editorial decisions are guided by the significance of a story, the quality of its evidence and its value to readers. Read about our founder and purpose.
We begin with a question and a verifiable record
We follow decisions that affect students, families and school systems: policy changes, funding, school operations and approaches to learning. Reader questions and public discussions can point to a subject worth investigating. Claims about what happened must rest on an official record or substantive, attributed reporting.
We distinguish an enacted law from a proposal, optional guidance from a binding rule, and a published policy from evidence of classroom practice. A deadline, effective date or local decision keeps its stated scope.
Our contribution is the finding the sources support
A story should add a useful comparison, document analysis, calculation, implementation consequence or researched answer beyond the announcement that prompted it. We identify the evidence for that contribution and keep its limits visible. When the record supports only a brief announcement, we retain it as a research lead rather than publish a standalone article.
District research combines education data with geographic, economic and demographic context. The sources can include federal and state education records, district budgets, housing data and local reporting. Each report states its coverage and observation periods. We avoid single-number school ratings and unsupported causal claims.
Human review, AI assistance and adversarial checks
We use AI-assisted research and writing to assemble sourced briefs, draft reporting and examine comparative evidence. A separate, adversarial automated review challenges claims, citations, source support and historical context. Automated checks also test source availability, citation references, required article structure and factual risks. Guardrails are designed to minimize falsehoods, bias, unsupported conclusions and misleading framing.
Human editorial review accompanies those checks before publication and continues afterward. The workflow can request revisions, hold a story or identify it as developing when further verification is needed. A developing label identifies ongoing verification; it does not establish that an unresolved claim is true.
We make the evidence and material qualifications visible so readers can examine our findings. Review is an ongoing responsibility: new evidence and substantiated reader challenges can lead to further reporting or a correction.
Sources readers can examine
Source links and endnotes connect consequential claims to named records. We prefer original documents and datasets for legal provisions, official decisions and numerical findings. When we rely on another publication’s reporting, we attribute that work. A source’s observation date, our research date and the article’s publication date may differ.
Policy records are reference entries with scope, status and source dates. Notices summarize announcements. Neither format automatically establishes national adoption, an independent investigation or a finding about a school’s quality. Read about our data sources and coverage.
Updates and corrections
We update reporting when verified evidence changes a consequential fact or explains what happened next. Substantive updates retain the earlier publication date and identify the new evidence. Rechecking an unchanged document does not make an old event new.
Time-sensitive features carry observation dates. A trend snapshot that has passed its observation window is presented as a dated record and removed from the current homepage feature. Missing evidence remains a gap, rather than an invented outcome.
Send corrections to the publication with the page URL, statement at issue and supporting record. We assess the evidence, correct verified errors and identify material changes on the affected story.
Use and attribution
Consult the originating agencies for source datasets and their terms. Formative Spaces’ prose, analysis and graphics are its published work. Contact us about permissions or research collaboration.