Methods & transparency
What does a finding measure, whose work is it, and what can a reader reasonably conclude? These are the standards for public Acropolis releases. The methods and limitations for a specific result belong with that result.
Start with what was measured
A report should identify the source, population, geography, and years behind its findings. The years covered by a dataset are not the date the report was published. For a rate, readers also need to know what was counted and who was included in the denominator.
If definitions, county boundaries, or collection methods changed, the methods note should explain how those differences were handled and which comparisons still make sense.
Show the limits of a finding
A difference on a chart or map is not, by itself, evidence that one exposure caused an outcome. We distinguish observed values from model-based estimates and describe relevant limitations, such as missing data, survey sampling, or small counts.
Where appropriate, include intervals or other measures of uncertainty. If a value is withheld, label it as withheld rather than displaying a zero. Report-specific suppression rules and access conditions belong with the data they govern.
Distinguish the source of the work
- Independent institute research: Acropolis develops the question and describes the contributors and methods in the release.
- Commissioned or collaborative work: identify the institute’s role and distinguish it from that of the commissioning organization, partners, or other authors. A partner engagement does not imply permission to publish its results.
- Prior scholarship: work completed at another institution keeps its original authorship and affiliation. A library link does not turn it into an institute publication.
See the independence statement and the founder’s academic profile for context.
Publishing a result is not publishing the underlying records
The research library is for approved public materials only. Survey-derived findings require review for disclosure risk. Raw participant responses, direct identifiers, verification records, and unreviewed free text do not belong in this public repository.
Study-specific participation and data-handling information belongs on the study page before recruitment opens. See the survey directory and the website privacy notice.
Make the version identifiable
Use a release’s suggested citation and version, or the publisher’s citation when linking to a journal article. For a substantive correction to an institute release, keep its permanent landing page, identify what changed, and distinguish the revised file from the version it replaces.
Found a possible error? Request a correction with the release title or URL, its date or version, the page or figure, and a short description of the issue. Please do not include personal health information or confidential records.