Data Corrections Log

July 13, 2026 — Per-state braces and veneer figures were derived, not observed

Severity: high. Affected our published dataset, four pages, and one downloadable report.

What was wrong

Our per-state figures for braces and veneers were not observed prices. They were computed from our implant series by a fixed multiplier — braces were exactly 0.80 × the state implant average, veneers exactly 0.25 × it, in all 51 markets. The same applied to the low/high bands (0.70× and 1.40× the average).

The practical effect: every procedure on this site shared the same state ranking. Whatever we published as "the cheapest states for veneers" was, mathematically, just the cheapest states for implants wearing a different label. Our "composite" Cost Index had the same flaw — because two of its three inputs were multiples of the third, it reduced algebraically to a function of the implant price alone.

This was not disclosed anywhere. Our methodology page described these as observed averages and named the ADA, FAIR Health and CMS as sources. That was inaccurate.

How we found it

An internal consistency check flagged that our published dataset (national braces average $3,606) contradicted our own article pages (~$5,000). Investigating which was right, we found the ratio between the braces and implant columns was constant to four decimal places across all 51 states — the signature of a derived column, not a measured one.

Why it matters — the data refutes the shortcut

We then compared five independently sourced state-level price series (implants, braces, veneers, crowns and dentures) across all 51 markets. Dental prices do not share a common geography. They fall into at least three regimes that barely correlate:

• Restorative work (implants, crowns, dentures) is tightly coupled: 0.93–0.96. A state cheap for one really is cheap for the others.
• Orthodontics is independent of everything else (0.04). It is a specialist market with its own economics.
• Cosmetic work is nearly independent too (~0.18). Veneer prices correlate with a state's cost of living at just r = 0.07, against r = 0.836 for implants.

So deriving braces from implants was not merely imprecise — it was structurally incapable of being right.

What was actually affected, and what was not

Our implant series turned out to be genuine. Validated against an independent source (ASQ360°/CareCredit), it agrees at a Spearman rank correlation of 0.945. Because restorative procedures cluster so tightly, the rankings we published for crowns and dentures — although derived — happen to land within 0.91 of the real thing.

What changed

• Removed the derived braces_avg_usd and veneer_avg_usd columns from the published dataset. The open dataset now contains only our own observed implant series.
• Renamed the composite index to the Restorative Cost Index, which is what it actually measures. It is no longer used to rank orthodontic or cosmetic procedures.
• Rebuilt the braces, veneers, crown and denture state figures on observed prices from the ASQ360° / CareCredit Average Procedural Cost Study, cited and linked wherever displayed.
• Rewrote the cheapest states for veneers page entirely — its ranking and its central explanation were both wrong.
• Corrected the methodology page, which had listed data sources we do not in fact use.
• Issued a corrected version of the dataset on Zenodo. The DOI is versioned; the original record remains citable, and the correction is documented in its changelog.

If you cited our data

If you cited the national braces average ($3,606) or any per-state braces or veneer figure from our dataset published before July 13, 2026, those figures were wrong and should not be used. The correct sourcing is set out on each page. We are sorry — and we would rather tell you than let it stand.

Questions or a correction to report: research@realdentalcosts.com.