You are about to be buried in pseudostatistics and factoids purporting to demonstrate just how bad Michigan roads are. We have already demonstrated that the TRIP vehicle maintenance factoids fabricated by a devious road contractors’ business league are a complete fraud.

Now State of Michigan government entities are unleashing a wave of propaganda intended to drive your vote on Proposal 2015-01. Two State of Michigan government entities have prepared campaign flyers on behalf of Proposal 15-1 using Michigander’s tax dollars: MDoT and the Transportation Asset Research Council (TAMC).

Both MDoT and TAMC have spent years preparing slick pamphlets decrying the sad condition of Michigan’s roads, citing PAvement Surface Evaluation and Rating (PASER) data collected by TAMC with MDoT and local road agency help. These pamphlets claim that 30 – 40 % of Michigan roads are in poor condition. But are they really?

Michigan road conditions are currently measured by four different methodologies, not just PASER alone. Two of the rating methodologies are based upon subjective (personal visual observation from a moving vehicle) ratings, the other two are based upon objective (instrumental) ratings. The four methodologies are:

Sufficiency Index (SI)

The sufficiency index is an annual, subjective, ‘windshield’ rating of the 9,655 mile State Trunkline road system performed from a moving automobile by a chauffeured MDoT pavement engineer. This survey rates road pavement for ride and distress using a five point index, with ‘1’ being best and ‘5’ being worst. This is MDoT’s original pavement rating system from 1961.

It is not clear what MDoT uses its Sufficiency Index ratings for today, but since the data set goes back to 1961 it would be a useful tool for normalizing current State Trunkline road condition data measured by other methodologies.

PAvement Surface Evaluation and Rating (PASER)

PASER is the newer, subjective ‘windshield’ rating of Michigan’s road system contracted out by the Transportation Asset Management Council in conjunction with MDoT. It is performed from a moving vehicle by a representative from MDoT, a representative from the Regional or Metropolitan Planning Organization (RPO/MPO), and a representative of the local jurisdiction being surveyed. The PASER survey rates road pavement for ride and distress using a ten point index, with ’10’ being best and ‘1’ being worst. TAMC has been collecting PASER data since 2004.

Over 100 teams of these PASER raters assess 50% minimum of Michigan’s 39,700 miles of federal aid-eligible roads and something less than 9% of Michigan’s 40,000 miles of paved, non federal-aid eligible road miles each year. So all of Michigan’s State Trunkline roads are PASER rated over each two year period, but the non federal-aid eligible roads are PASER rated on a selective basis which is not in any way statistically valid.

Proponents of Proposal 2015-01 are using PASER data exclusively to make their case that Michigan’s roads are in deplorable condition. There is no indication that TAMC or MDoT use PASER ratings for any purpose other than political talking points.

Note that the Sufficiency Index rating is conducted on a more limited number of roads than PASER ratings: State Trunkline roads are a 9,655 mile subset of Michigan’s approximately 39,700 miles of federal aid-eligible roads. Michigan has approximately 120,000 miles of roads in total, 40,000 of which are unpaved. So there are about 40,000 miles of paved, and 40,000 miles of unpaved, Michigan roads which are neither in the State Trunkline system nor federal aid-eligible. These non federal aid-eligible roads, mostly local and subdivision streets, are not rated by any State standardized (or statistically valid) process.

International Roughness Index (IRI)

IRI is an objective measurement of cumulative pavement height changes in the wheel paths of an instrumented vehicle equipped with laser or mechanical, vertical measurement (profiling) sensors. The instrumented vehicle is driven down the road and the height of the pavement is measured continuously, capturing the fine height changes in the wheel paths. IRI data is collected on half the 9,655 mile State Trunkline road system each year by an MDoT contractor.

The IRI pavement measurements are expressed in inches (vertical height) divided by miles (horizontal length). A value of zero would be perfectly smooth pavement and IRI numerical values increase as pavement roughness increases. MDoT interprets IRI values below 95 in/mi (inches per mile) to represent pavement in good condition, values between 95 – 170 in/mi to represent pavement in fair condition, and values above 170 in/mi to be pavement in poor condition. This is a pass through of the Federal Highway Administration recommended IRI interpretation.

Michigan’s IRI data is reported to the Federal Highway Administration, which uses this data to assess the condition of the American highway system. FHWA collects and tabulates IRI data from all the state transportation/roads agencies. IRI data can also be used to determine how susceptible roads are to heavy truck damage, since pavement roughness increases the number of high force impacts delivered to the pavement by truck tires. It should be noted that some states, like Wisconsin, collect their IRI data in metric format – meters per kilometer – so there is some processing necessary on the part of FHWA to insure comparability of ISI data.

Distress Index (DI) and Remaining Service Life (RSL)

Distress Index is the other objective, instrumental methodology and is quite important because it is a key component of the road work decision making index actually used by MDoT: Remaining Service Life. DI is a continuously recorded measure of road damage such as spalls, cracks, and potholes in the path of a vehicle equipped with downward aimed video cameras. The vehicle is driven down the road and the condition of the road is imaged continuously, capturing the surface damage on the lane. DI data is collected on the same half of the State Trunkline road system as IRI, each year, by an MDoT contractor. The same contracted vehicle which records IRI data usually records DI data simultaneously, so the road sections evaluated by the two instrumental methodologies each year are almost identical.

The DI data is used by MDoT to calculate the Remaining Service Life (RSL) of roads, the measure MDoT actually uses to prioritize and scale roadwork in Michigan. DI measurements are denominated as numbers from 0 to 100 which can thought of as the extent of damage to the road, sort of a percentage but not quite. MDoT considers a road to have exceeded its useful life at a DI of 50, where RSL is zero by definition. MDoT calculates remaining service life from DI data and commercial vehicle censuses (daily truck traffic data) using a logistic regression model. This creates an RSL number denominated in years of remaining life. A road has a positive remaining service life when its DI is less than 50.

RSL is the most important measure of road condition from MDoT’s perspective. It is the measure MDoT uses to prioritize road repair and replacement work. It is based upon DI data and truck traffic data. This reliance on RSL for decision making tells you that MDoT places little credence in PASER data, Sufficiency Index, and the International Roughness Index as measures of overall road condition. It also tells you that MDoT recognizes heavy truck traffic as the actual cause of road deterioration, to the exclusion of all other types of vehicles – such as passenger cars and light trucks.

So how does the politically useful PASER data being shilled by Proposal 2015-01 proponents such as TAMC compare with MDoT’s all important internal standard: Remaining Service Life? Not well, but this is not readily apparent because TAMC and MDoT have gone to some length to prevent any valid comparison of the four road condition assessment methodologies. Fortunately, the Office of the Auditor General in Michigan audited the performance of state pavement condition measurement in 2012. While the auditor(s) were not engineers and got smoked in a couple of areas, they did a credible job of extracting data from both TAMC and MDoT. And their unflattering findings regarding PASER data collection were spot on, at least as far as they went.

Let’s compare PASER data to DI / RSL data for the State Trunkline road system. We cannot compare these two measures beyond the 9,655 State Trunkline system because no DI data is collected on Michigan’s non federal aid-eligible roads. We also have to use MDoT’s interpretation of ‘good’, ‘fair’, and ‘poor’ because the ratings scales for each methodology are so different. But if there is a consistent bias in PASER pavement condition determinations relative to DI / RSL determinations on the State Trunkline, it is then reasonable to assume the same bias exists in PASER data collected from the non federal aid-eligible roads.

Using the 2010 data in the Michigan Auditor General’s performance audit, the last comparable data set available, let’s look at the percent of State Trunkline roads found to be in poor condition by each methodology:

The TAMC PASER methodology reported 55% more poor and very poor condition roads in the State Trunkline system than the MDoT DI / RSL methodology did.

Next, let’s look at the percent of State Trunkline roads found to be in fair condition by each methodology:

The TAMC PASER methodology reported a 67% higher incidence of fair condition roads in the State Trunkline system than the MDoT DI / RSL methodology did.

Finally, let’s look at the percentage of State Trunkline roads found to be in good and excellent condition by each methodology:

The TAMC PASER methodology reported only half the good & excellent condition roads in the State Trunkline system that the MDoT DI / RSL methodology did.

Cumulatively, the politically useful PASER pavement rating methodology finds Michigan’s State Trunkline road system to be in far worse condition than the DI / RSL methodology that MDoT actually uses to prioritize road work. A logistic regression – MDoT’s preferred analytical tool – shows that there is a statistically significant variance between the methodologies as they are applied to the State Trunkline system. Either the PASER or the DI / RSL methodology is not properly evaluating Michigan pavement conditions.

Most civil engineers consider RSL to be the ‘gold standard’ of pavement condition evaluation, so PASER ratings, as performed for TAMC, are likely wrong.  Should you be inclined to think that the PASER evaluations better represent the condition of Michigan’s roads, ask yourself why MDoT and all the other State DoT’s do not use PASER evaluations to prioritize road work.  You would also have to ask yourself why the other instrumented pavement rating methodology, FHWA’s IRI, shows Michigan’s State Trunkline roads to be in even better condition than the DI / RSL methodology.  Bear in mind that the PASER raters all represent government entities which stand to directly benefit from the avalanche of revenue that Proposal 2015-01 will generate.

The two objective, instrumental methodologies, IRI and DI / RSL, agree with each other fairly well. The differences between them probably reflect differences in heavy truck traffic census counts, which modify the DI data to create the RSL numbers. It could also be that Michigan roads are fairly smooth between potholes, or that IRI does not capture some prevalent defect, such as pavement breaks in the direction of travel, outside of instrumented vehicles’ wheel paths.

So here is the question you should now ask yourself: is the politically useful PASER condition rating of Michigan roads, not in the State Trunkline system, valid? PASER data suggest that these are the Michigan roads in the worst condition. But how bad are they really? You only have PASER data to go on, and PASER data doesn’t accurately characterize the State Trunkline roads where we can actually evaluate the validity of this methodology.

The 2012 performance audit of state highway measurement systems conducted by the Michigan Office of the Auditor General found three reportable conditions in the quality control and quality assurance of the PASER methodology as performed for TAMC circa 2012.  Had the MOAG auditors made a direct comparison of the rating methodologies on State Trunkline roads, they would have considered these QA / QC issues to be material conditions, a far more serious category of audit finding.  But they weren’t engineers, so they missed this.  They also found a timeliness issue with some DI / RSL data utilization in the same time frame, but this is a far less severe issue than the multiple QA / QC issues found in PASER rating as performed for TAMC.

It only takes a quick look at the charts above to realize that there is something quite wrong with the TAMC PASER road ratings being touted by Proposal 2015-01 supporters. Michigan’s mainstream media are regurgitating these politically useful PASER data as authoritative without any further analysis, so many voters in Michigan are being deceived. Deceit seems to be the modus operandi of Proposal 2015-01 proponents. It is long past time to clear the air by conducting DI / RSL evaluations of non federal aid-eligible roads in Michigan. Then we can discuss a time-limited plan to remediate Michigan’s roads.