Home / Editorial Standards

Editorial Standards

How RefDat sources, ranks, and corrects the data we publish, and how readers can hold us to it.

RefDat is an independent Australian open-data publication. We pull from authoritative public sources, build ranking frameworks where useful ones did not exist, and explain what the data actually means. This page documents how we do that work, who is accountable for it, and how readers can hold us to it.

Why we do not publish bylines

RefDat does not publish individual names. The decision is deliberate.

Our NDIS Trust Index work has named providers operating questionably across more than 26,000 organisations. Journalists and researchers who do similar work in this space have faced sustained harassment, doxxing, and physical threats. We protect contributors by keeping them anonymous. The trade-off is that readers cannot verify our credentials by individual name.

We substitute the byline signal with something we think is stronger: every framework is published with full methodology, every claim ties back to a named primary public dataset, and every ranking is reproducible by anyone who follows the documented method. The work is auditable even when the people behind it are not named. If a number on this network is wrong, anyone can demonstrate that with the source data and the published methodology, and we will correct it.

Google quality raters explicitly recognise that some content domains require anonymity, particularly investigative reporting on subjects where attribution creates safety risks. RefDat sits in that category by design.

If you need to reach the editorial team, the address is admin@refdat.com.

How a RefDat piece moves from idea to published

  1. Source identification. Every piece starts with a primary public dataset. ABS Census, BOM weather stations, FSANZ food standards, ACARA school data, the NDIS Commission register, USDA FoodData Central, Coeliac Australia certifications, BIPM and NIST measurement factors, Open-Meteo, GHCN. If we cannot point at a primary source for a claim, we either find one or we do not publish the claim.
  2. Methodology drafting. Where we build a ranking framework (the NDIS Trust Index, the Weather Comfort Score, the Gluten traffic-light verdict, the RefDat Score for product reviews), the scoring rubric is written down before the data is scored. The methodology page goes up alongside the rankings, not after.
  3. Internal review. Every framework change is reviewed by a second contributor against a defensibility test: does this signal trace directly to the individual record, or is it borrowing from a neighbour, a proxy, or an address-based assumption? Indirect signals are display-only context, never score inputs. This rule is a direct outcome of our May 2026 NDIS audit. The audit log is public at ndis.refdat.com/methodology/.
  4. Fact-check. Every numeric claim on every page is generated from the underlying dataset at build time, not typed by hand. Stat strips, ranking distributions, suburb counts and provider totals are produced by the data pipeline and updated on every refresh.
  5. Publication. Pages ship via a single deploy script that snapshots the source state to git first, then deploys, then pings IndexNow, then re-submits the sitemap to Google Search Console. Every deploy stamps the per-site record so we know what shipped, when, and how many files moved.

Corrections policy

We publish reference data and rankings about real people, real organisations, and real places. If we get something wrong, we want to know.

Email admin@refdat.com with the claim, the page URL, and the source data that contradicts what we published. We aim to acknowledge within five business days and to correct or remove the disputed content within ten business days of confirming the error. Significant corrections are logged publicly. For NDIS Trust Index changes specifically, the audit trail lives at ndis.refdat.com/methodology/.

Corrections do not get silently overwritten. If a page changed materially in response to a correction, the change is noted on the methodology log so the record is preserved.

Sources policy

We use primary public datasets first, and we name them on every methodology page. The sources currently in production:

Secondary sources are cited where used. Wikipedia is a navigation aid, never a citation. Press releases are sources for the press release, not for the claim it makes.

Conflicts of interest

RefDat earns revenue from two sources, and only two:

  1. Google AdSense display advertising across most subdomains. AdSense placement is automated; we do not control which advertisers appear on which pages. Auto Ads is enabled on the apex; explicit ad slots are used on some subdomains.
  2. Affiliate links on topproducts.refdat.com, currently through the eBay Partner Network (Australia). When a reader clicks a buy link and purchases from eBay, RefDat receives a small commission. The commission does not change the price the reader pays, and it does not influence which products receive higher RefDat Scores. The RefDat Score methodology is independent of affiliate availability. We have published positive scores for products with no buy link, and negative scores for products with buy links.

What we do not do: no sponsored content. No paid product placements. No brand partnerships. No payment for inclusion in best-of lists. No exchange of editorial coverage for access. No advertorial. No "as seen on" badges sold to advertisers.

Revenue scale is modest. RefDat is currently a self-funded independent publication. If that changes (a research grant, a substantial donation, an institutional partnership), it will be disclosed here.

Independence and what we do not do

RefDat is not a government agency. We are independent of the bodies that publish the data we use. We do not accept funding from any of the data publishers we cite.

We do not provide medical advice. The Gluten Guide is reference data, not clinical guidance. The Weather Comfort Score is liveability research, not a health warning. For medical, financial, or legal decisions, consult a qualified professional.

Reader feedback

admin@refdat.com is the address for everything: corrections, methodology questions, source disputes, missing data, broken pages, anything. Replies are slower than a newsroom but faster than a government department.

Published by RefDat Editorial. Last reviewed 2026-07-19. Corrections: admin@refdat.com