# What a road condition survey can and can't tell you

*A condition survey is a measurement, and every measurement has a spread. Knowing how large that spread is, and where it comes from, stops a council mistaking a change of rater for a change in the road.*

## Not all condition numbers are equally firm

Condition data comes in two broad kinds. Sensor measures, such as roughness (IRI) and rut depth from a survey vehicle, are repeatable when the equipment is calibrated. Visual distress ratings, whether done on foot, from a windscreen or from images at a desk, involve judgement about the type, severity and extent of cracking, ravelling or patching. The US Federal Highway Administration's 2022 Pavement Management Roadmap notes that sensor data is typically considered less variable than manual distress surveys, and that practitioners find distress results difficult to compare.

How different can ratings be? A 2020 Transportation Research Board webinar on certifying condition data showed repeat PCI ratings of the same US test sections by trained staff over seven years. On sections in good condition, the coefficient of variation was around 5%. On sections in poor condition it reached 35–45%. The worse the road, the less two ratings of it agree.

The same webinar showed an airfield runway in good condition rated 89 by manual survey, 90 from HD camera images and 94 from 3D laser imaging. The methods agreed reasonably on good pavement and diverged as distress increased.

## Three common traps

**A new vendor looks like a new road.** When the survey contractor, the equipment or the rating method changes, condition indices often step up or down across the whole network. The FHWA Roadmap lists “stabilizing performance indices after changes in data collection protocols or vendors” as a standing problem. Treat that step as a method change, not a trend.

**Results arrive too late to use.** US agencies reported a lag of three to six months between survey and usable results, by which time budgets may be set. Agree delivery dates and data formats in the survey contract.

**Nobody checks the data before it is loaded.** Quality checks are easy to skip under deadline. Errors loaded once then flow into every report.

## How agencies check survey quality

The webinar described checks at three stages:

1. **Before the survey:** certify the equipment and the raters against reference measurements.
2. **During the survey:** re-rate a sample, or include control sections whose condition is already known to the agency but not to the contractor.
3. **After the survey:** run acceptance checks for completeness, plausible ranges and sudden jumps before the data is used.

One example of a pass mark: the Metropolitan Transportation Commission in the San Francisco Bay Area certifies automated rating systems on about 24 control sites. To pass, more than half the sections must be within 5 PCI points of the reference, and no more than 12% may differ by more than 15 points.

A small council does not need a formal certification programme to borrow the idea. Choose a handful of sections that represent your network, rate them carefully once, and compare each new survey against them.

## Questions to ask your survey contractor

- What equipment and rating method will you use, and has either changed since our last survey?
- How are raters trained and checked for consistency?
- Will you rate our control sections without knowing their reference values?
- When will we receive the data, and in what format?
- Will raw images be supplied, and for how long can we access them?

## How PaveKeep handles this

PaveKeep records which survey each condition figure came from, and every number can be traced back to that survey, the setting used and the arithmetic applied. It also shows how the network has changed between surveys, rather than a single snapshot, so a sudden network-wide step is visible and can be questioned instead of being read straight as deterioration.

PaveKeep is built for small and regional councils and is not yet in use at any council.

## Sources

- Transportation Research Board (2020), webinar *How to Certify and Verify Pavement Surface Condition Data*, 10 June 2020: presentations by FHWA Resource Center, Quality Engineering Solutions and Applied Pavement Technology.
- Federal Highway Administration (2022), *Pavement Management Roadmap*, FHWA-HIF-22-054, Themes 1, 2 and 4.

---

Source: https://pavekeep.com.au/knowledge/what-a-survey-can-tell-you · PaveKeep · bound content