The manual allows experimentation—but users rarely depart

GAO interviewed practitioners in 26 state and local jurisdictions about the Manual on Uniform Traffic Control Devices. Almost all said they never or rarely deviate from it, and many cited liability concerns as a reason for strict adherence.

The sample was not designed to represent every road agency. Its value is qualitative: it shows how professional risk perceptions can make a minimum federal standard function like a practical ceiling, even when experimentation is legally available.

Sources: GAO-27-108431, November 16, 2026

An experiment ledger without enough public detail

The Federal Highway Administration maintains a database of approved traffic-control experiments, but GAO found that the public entries contain few details. More information is available on request, yet the website did not make that route clear.

That design weakens reuse. A city considering a safer sign, marking or signal needs to know what another jurisdiction tried, how it was evaluated and what FHWA concluded. A title and status alone cannot support that comparison.

One recommendation, accepted by DOT

GAO recommended that FHWA improve public access to information about traffic-control experiments. The Transportation Department agreed. The meaningful completion test is a searchable record with project purpose, design, evaluation and disposition—not simply a notice that fuller material can be requested.

The Scoop's analysis distinguishes the evidence: GAO established a transparency weakness and recurring liability concerns in its interview sample. It did not conclude that the manual causes crashes or that any specific deviation is safer.

Sources and further reading

GAO traffic-control report ↗

Traffic-control logo and public-domain record ↗