SAG Research · Editorial standards
Research Standards
& Disclosures
The method, the evidence and the limits belong in the published record.
1. Identify the work accurately.
Each publication must state whether it is a protocol, original empirical research, evidence synthesis, research note, practitioner perspective or commentary. It must name its authors and identify the date, version, sources, funding and review status.
A planned study must remain labeled as planned. Drafts must remain labeled as drafts. Publisher-hosted articles are part of Joseph De Gregorio’s published portfolio; they are not automatically SAG Research Papers or independently reviewed empirical studies.
2. Make the research inspectable.
Empirical work requires a defined question, documented source selection, inclusion and exclusion rules, a coding or measurement plan, and an analysis plan. Publish a dated protocol before full-scale coding or confirmatory analysis. Record amendments and explain whether a decision was made before or after examining the results.
Release a source inventory, codebook, relevant analytic code and appropriately shareable data with the findings. Explain missing information, retrieval failures, exclusions, access restrictions and deviations from the protocol. Do not substitute citations to secondary summaries for checking the underlying evidence where the primary source is available.
3. Separate observation, inference and advice.
Distinguish what a source states from what an author infers. Descriptive associations do not establish causation. A court’s discussion of evidence does not establish that the evidence caused a particular sentence. An individual outcome does not establish a typical result.
Report contrary evidence, mixed findings and uncertainty. Research must not be designed to validate a predetermined sales claim. Public reports may conclude that an anticipated relationship is absent, weak or not measurable with the available data.
4. Require real external review.
A completed SAG Research Paper must receive substantive review by a qualified methods reviewer and a qualified federal sentencing legal reviewer before release under that series. Reviewers must be outside the author team, disclose relevant conflicts and have no contingent compensation tied to favorable findings or client acquisition.
The research lead must keep the review comments, author responses and unresolved limitations. The publication must describe what was reviewed, when and by whom where the reviewer consents to identification. Otherwise, explain reviewer qualifications and why the identity is not public. Do not use “peer reviewed” as shorthand for an editorial read, AI review or informal endorsement.
At launch, external study reviewers and a second coder have not been appointed. The first protocol remains a draft. The independent-review requirement is a release condition, not a completed credential.
5. Disclose the commercial relationship.
SAG Research is an initiative within Sentencing Advocacy Group, a commercial federal sentencing advisory business. Joseph De Gregorio has an economic interest in that business. This relationship must be disclosed in SAG research publications relevant to its services.
Each study must identify its financial and in-kind support, relevant consulting interests, author and reviewer relationships, and any SAG involvement in included matters. No funder or client may make favorable findings a condition of publication. Unfavorable or null findings must not be suppressed because they conflict with commercial positioning.
SAG Research is not presently represented as a separate nonprofit, accredited academic institution or independently governed institute. Existing professional relationships and publication credits do not establish research partnerships.
6. Protect people and source integrity.
The first proposed study is limited to lawfully accessible public written opinions. It does not call for private client files, sealed records, Presentence Investigation Reports or confidential communications.
Use only information necessary for the research question. Do not republish personal identifiers, victim details or sensitive narratives merely because they appear in a public source. Any later project involving interviews, private information or identifiable human participants requires a separate ethics and privacy assessment, an appropriate consent process and a determination of applicable review requirements before collection begins.
7. Document any use of AI.
AI may assist with discovery, organization, drafting and exploratory classification. Record the tools, versions or access dates where available, their purposes and material limitations. Generated citations, quotations and extracted facts require verification against the source.
AI output is not an independent legal review, methods review or second human coder. Do not upload confidential client material into research workflows without an appropriate, separately authorized process. Disclose material AI assistance in each publication.
8. Correct the record visibly.
Send a correction request to joseph@sentencingadvocacygroup.com, identifying the publication number, affected passage, proposed correction and supporting source. Do not send confidential case material.
Substantive corrections must receive a dated notice explaining what changed and whether the interpretation or conclusions changed. Preserve the original publication date and increment the version. Maintain an accessible version history for substantive revisions. If a central finding cannot be supported, publish a withdrawal or retraction notice rather than silently replacing it. Minor typographical fixes may be recorded without a new substantive version.
Public reference points
Standards that inform the design.
RAND’s public standards emphasize rigor and transparency. Brookings publishes research-independence and conflict-disclosure policies. These are methodological reference points; neither organization has reviewed, endorsed or partnered with SAG Research.