The following four changes to the SafeStat methodology are proposed to
improve the performance of the current SafeStat algorithm (Version 8.6).
1. New Traffic Violations Indicator (TVI)
Proposed Improvement
Replace the Moving Violations Indicator (MVI) with the Traffic Violations
Indicator (TVI) in the Driver Safety Evaluation Area (SEA) Value
calculation.
The TVI differs from the MVI in four major respects:
- Applies a different measure of exposure: The TVI will
be normalized by the average number of power units a carrier operates at the
end of the three time periods used for time weighting in the TVI calculation
(0-6 months, 7-18 months, and 19-24 months) rather than the current number
of drivers a carrier employs.
- Expands use of traffic violation data: The TVI will
expand the use of traffic violations in the Driver SEA by considering 392.2
(general/local laws) violations in addition to the twelve specified moving
violations currently used in calculating the MVI. The 392.2 general
violations will be given less weight than specified violations.
- Emphasizes driver on-road performance in the calculating Driver
SEA Value and SafeStat Score: The TVI will be assigned the same
weight as the DRI and DII. The highest of the TVI, DRI, and DII will be
assigned to the Driver SEA Value. This change will allow the TVI to impact a
carrier’s Driver SEA Value more than the MVI. This approach makes the
treatment of the TVI in the Driver SEA consistent with the use of the other
two indicators and with the treatment of indicators in the other SEAs of assigning
the highest indicator to the SEA value.
- Compares performance of similarly sized carriers: Each
carrier receiving a Traffic Violations Measure (TVM) will be assigned to one
of four peer groups by average number of power units (i.e., Group 1: <=5,
Group 2: >5 and <=50, Group 3: >50 and <=500, Group 4: >500).
Each carrier’s TVM will be ranked against the TVMs of all of the other
carriers within the same peer group and assigned a percentile from 0
(representing the lowest TVM) to 100 (representing the highest TVM). As with
the other indicators, the percentile value determined by the rank of the
measure (the TVM) will be the indicator value (TVI).
The methodology on how the TVI will be calculated can be found by click here.
Reason for Change
Issues such as SafeStat’s ability to identify high crash risk carriers,
fairness in assessing a carrier’s safety status, data quality, algorithm
consistency, and carrier coverage were examined in developing the TVI. Results
of these analyses show that incorporating the TVI accomplishes the
following:
- Identifies more carriers with higher crash risk: The
SafeStat effectiveness study has been used to measure SafeStat’s ability to
identify high crash risk carriers by observing the subsequent crash rates of
SafeStat-identified (Category A & B) carriers.Click here to see the SafeStat Effectiveness
study. By applying the methodology of the SafeStat effectiveness
study to the algorithm with the proposed TVI changes, the number of Category
A & B carriers increased by 30% and their subsequent group crash rate
rose 12% over the results of the current algorithm. These increases indicate
that the incorporation of the TVI will improve SafeStat’s ability to find
unsafe carriers.
- Improves geographical coverage: The inclusion of 392.2
(general/local laws) violations in the TVI increases the amount of traffic
violations considered by SafeStat, particularly in states with relatively
low reporting of the specified moving violations used in the MVI, thereby
reducing the chances of geographical bias occurring.
- Simplifies and improves consistency of the SafeStat algorithm:
The TVI uses the same measure of exposure, average power units,
that is already used in calculating the Accident Involvement Indicator
(AII). Also, assigning the highest driver indicator (i.e., DII, DRI, or TVI)
to the Driver SEA Value makes the Driver SEA Value calculation consistent
with the Vehicle and Safety Management SEA Value calculations.
- Uses Better Data: SafeStat requires an accurate measure
of exposure to properly calculate the TVI. The data used to calculate the
number of power units are more complete than the data used to calculate the
number of drivers. Also, the use of average power units more
accurately reflects changes in exposure over the period of time that
carriers are being evaluated by the TVI.
- Covers more carriers: Twice as many carriers have
sufficient data to receive a TVI than have sufficient data to receive an
MVI.
2. Shortened Data Exposure Time Period
Proposed Improvement
Reduce the amount of time considered from 30 months of data to 24 months
for:
- State-reported crash data when calculating the Accident Involvement
Indicator (AII),
- Inspection data when calculating the Vehicle Inspection Indicator (VII)
and Driver Inspection Indicator (DII),
- Traffic violation data when calculating the Traffic Violations Indicator
(TVI), and
- Recently closed enforcement case data when calculating the Enforcement
History Indicator (EHI).
Reason for Change
In 1996, a 30-month time frame of data was set for the various SafeStat
indicators to obtain enough safety information to assess the safety status of
a large portion of the carrier population. Since then, the amount of safety
data available to SafeStat has increased and the timeliness of these data has
substantially improved, thereby providing more safety data to assess the
safety status of individual motor carriers. Adjusting the 30-month time frame
to 24 months:
- Increases emphasis on more recent events: This change
will reduce the instances in which a carrier has made recent substantive
changes to it safety and compliance performance, yet the results from the
older events prevent SafeStat from recognizing these changes quickly.
- Has little impact on the effectiveness of identifying carriers
with high crash risk: The SafeStat effectiveness study has been
used to measure SafeStat’s ability to identify high crash risk carriers by
observing the subsequent crash rates of SafeStat-Scored (Categories A, B,
and C) carriers, carriers with high SEA Values, and carriers with Indicators
of 75+. By applying the methodology of the SafeStat effectiveness study to
the algorithm with the proposed 24-month change, the effectiveness of
identifying high-crash risk carriers does not change substantially at the
Category, SEA, or Indicator levels.
- Reduces the number of carriers assessed by
SafeStat:SafeStat requires that a carrier have a sufficient amount
of data before its safety status can be evaluated. Reducing the length of
time from 30 to 24 months will reduce the number of carriers with sufficient
data to be evaluated by SafeStat by 10%. According to the SafeStat
effectiveness study results, the group of carriers that no longer had
sufficient data due to the reduction in time frame had a crash rate that was
half the crash rate of the group of carriers that had sufficient data. Thus,
while this change will cause many carriers to fall under the “SafeStat
radar,” the effectiveness study shows that these carriers, as a whole, have
a low crash risk.
- Aligns the SafeStat time frame with other FMCSA systems and
reports that use two years of safety event data (e.g., SAFER, Motor Carrier
Profile).
3. Simplifying the Calculation of the Accident SEA Value
Proposed Improvement
Simplify the calculation of the Accident SEA value to be equal to the
Accident Involvement Indicator (AII).
Reason for Change
By assigning the Accident SEA value to the AII and no longer using the RAI
will:
- Establish one standard of measuring crash rate, thereby avoiding
potentially conflicting results of multiple standards.
- Provide a stable value to better monitor safety
performance over time.
4. Applying More Vehicle Out-of-Service (OOS) Violation Data for the
Vehicle Inspection Indicator (VII)
Proposed Improvement
Include the results of Level 3 (Driver only) inspections that resulted in
vehicle OOS violations when calculating the VII.
Reason for Change
While conducting Level 3 inspections of the driver, enforcement officer may
discover vehicle OOS violations. Although these vehicle OOS violations are
currently not being used in the VII, which considers only the result of
vehicle inspections (Levels 1, 2, and 5), they are indicative of unsafe
vehicle conditions. Including the results of Level 3 inspections with vehicle
OOS violations will improve SafeStat’s ability to identify carriers with
vehicle safety problems.