6.1.3
Target Setting Methods
This subsection outlines methods for setting transportation performance targets, focusing on safety, infrastructure condition, reliability, and traffic congestion. It introduces quantitative target-setting approaches, such as policy-based methods, historical trends, probabilistic approaches, statistical models, and summarizes their ease of application, technical robustness, ease of communication, and support for agency policies. The guide also offers suggestions for selecting an appropriate target-setting approach.
Introduction
NCHRP Research Project 23-07, Guide to Effective Methods for Setting Transportation Performance Targets, presents several approaches for setting performance targets to support a TPM framework. It focuses on target setting for the national measures that are required under federal TPM requirements, including:
- Safety measures: Number of fatalities, rate of fatalities, number of serious injuries, rate of serious injuries, number of nonmotorized fatalities and nonmotorized serious injuries.
- Infrastructure condition measures: Percentage of the Interstate system pavements in good and poor condition, percentage of the non-Interstate NHS pavements in good and poor condition, and percentage of the NHS bridges in good and poor condition.
- Reliability (travel time and freight) measures: Percentage of person miles traveled on the Interstate and Non-Interstate NHS that is reliable and truck travel time reliability index.
- Congestion measures: Annual hours of peak hour excessive delay per capita and percentage of non-single-occupancy vehicle travel.
- Target setting for nonrequired measures, such as accessibility, greenhouse gas emissions, active transportation, transit ridership, and customer satisfaction are included in the final section of the guide.
TPM Webinar #12 - Target Setting
Target Setting Approaches
The guide recognizes that transportation agencies take different approaches to target setting. Some may prioritize setting realistic targets based on fiscal constraints. Other agencies may set aspirational (or fiscally unconstrained) or conservative targets.
Regardless of the approach to target setting, the guide recognizes that both quantitative and qualitative approaches may be considered. In some instances, qualitative approaches that are heavily influenced by politics or agency leadership may be appropriate. An annual target to reduce fatalities to address a long-term Vision Zero goal is an example of a qualitative target. Other approaches may use statistics or probabilities to define a quantitative target. The use of travel demand forecasts to set a mobility target illustrates the use of a quantitative approach to target setting. A combination of qualitative and quantitative approaches may also be used to set effective targets.
The following five quantitative methods for setting targets are presented in detail within the guide (Grant, M., et.al. 2023A):
- Policy-based approaches, such as establishing a maximum rate of change (e.g., annual increase of at least 2 percent).
- Historical trends (e.g., set a value based on a 5-year trend).
- Probabilistic and risk-based approaches that consider performance variability (e.g., performance based on the likelihood of increased storm frequency and intensity).
- Statistical models (e.g., rates of deterioration based on regression models).
- Other tools and models (e.g., output from a bridge or pavement management system).
When selecting the appropriate method for target setting, the guide offers the following tips (Grant, M., et al. 2023A):
- Understand the complexity of the methods - some methods require sophisticated data that may not be readily available and may result in marginal improvements in the target's effectiveness.
- Consider combining methods - by using several approaches, an agency has the benefit of considering the results from multiple methods in setting the final target.
The Guide (NCHRP Report 1035) presents several target-setting methods for each of the Federal performance measure categories along with a summary of their ease of use, robustness, ease of communication, and support for agency policies. A high-level synopsis of the information presented in the Guide is presented in the following table.
Table 6.2 - Features associated with different target-setting approaches (based on Grant, M., et.al. 2023A)
| Federal Performance Measure Category | Target Setting Approach | Description | Ease of Application | Technical Robustness | Ease of Communication | Allows for Policy Preference |
|---|---|---|---|---|---|---|
| Safety | Targeted Reduction | Defined decrease from baseline regardless of past trends | H | L | H | H |
| Safety | Time-Series Trend | Based solely on historical performance data | H | M | H | L |
| Safety | Trend Plus Other Factors | Adjustments made to results from other approaches | H | L | M | H |
| Safety | Multivariable Statistical Model | Statistical analysis considering multiple variables | L | H | M | L |
| Infrastructure Condition | Target Based on Change in Condition | Consensus decision | H | None | L | H |
| Infrastructure Condition | Time-Series Trend | Based solely on historical performance data | H | L | M | M |
| Infrastructure Condition | Time-Series Trend Plus Future Funding | Historical trends extrapolated into the future | H | L | M | M |
| Infrastructure Condition | Asset Management System | Condition forecasts based on expected funds | L | H | M | M |
| Infrastructure Condition | Scenario Analysis | Management system analysis of multiple scenarios | L | H | M | M |
| Reliability | Building off the Baseline with Assumptions | Qualitative approach to adjusting baseline values | H | L | H | M |
| Reliability | Time-Series Trend Analysis | Based solely on historical performance data | H | M | M | M |
| Reliability | Trend Plus Other Factors | Adjustments made to results from other approaches | H | M | M | M |
| Reliability | Performance Risk Analysis | Statistical analysis of variations due to risks | M | H | M | M |
| Reliability | Segment Risk Analysis | Analysis of individual segments to determine those that shift between reliable and unreliable | L | H | M | M |
| Reliability | Multivariable Statistical Model | Statistical analysis considering multiple variables | L | H | L | L |
| Traffic Congestion | Building off the Baseline with Assumptions | Qualitative approach to adjusting baseline values | H | L | H | H |
| Traffic Congestion | Time-Series Trend Analysis | Based solely on historical performance data | M | M | M | M |
| Traffic Congestion | Trend Plus Other Factors | Adjustments made to results from other approaches | M | M | H | H |
| Traffic Congestion | Travel Forecasting Model | Model used to estimate excessive delay for the base year and forecasted year | H | M | M | L |
| Traffic Congestion | Policy Based | Model based on regional policy goals | H | L | H | H |
Features Summary (L=Low, M=Medium, H=High)
New Jersey DOT
NCHRP Web-Only Document 358, which is a supplemental report to the Guide to Effective Methods for Setting Transportation Performance Targets, provides examples documenting how target-setting methods are being used by various transportation agencies. One of the examples illustrates how the New Jersey DOT piloted the use of a scenario analysis approach for setting its pavement infrastructure condition targets. An important consideration in the use of this approach was the availability of a pavement management system (PMS) capable of forecasting future pavement conditions. However, the New Jersey DOT had not established prediction models for the Federal cracking metric, so methods were developed to correlate forecasted conditions from existing models to the Federally-required cracking metric. Correlations were developed using three years of condition data collected in accordance with both the agency’s legacy Condition Status rating and the Federal measures in 0.1-mi segment lengths. Using both sets of data, correlations were developed based on the likelihood that a pavement section rated “Good” using the agency’s CS rating would also be classified as a “Good” pavement using the Federal definitions. The correlations found that 88.43 percent of the segments were rated “Good” based on both approaches (Grant, M., et.al. 2023B). A similar approach was used to correlate “Poor” conditions, but the analysis showed more variability in correlating segments at this condition level. To address the variability, New Jersey DOT decided to use 3-year averages to establish the final correlations, which are presented below (Grant, M., et.al. 2023B).
Table 6.A - Correlation between NJDOT CS and Federally-mandated condition ratings (Grant, M., et.al. 2023B)
| Federal Good | Federal Fair | Federal Poor | |
|---|---|---|---|
| NJDOT Good | 87.5% | 13.78% | 0.00% |
| NJDOT | 23.74% | 76.25% | 0.01% |
| NJDOT | 6.36% | 86.02% | 7.62% |
The correlations were applied to the predicted conditions generated by the PMS to determine the expected conditions using the Federally-mandated performance measures. Several scenarios were generated, allowing the NJDOT to use the results to set realistic Federal targets.
Addressing Disruptions to Performance in Target Setting
It is inevitable that agencies will face events that disrupt what might be expected to be typical performance. Sometimes these events alter performance for a short period of time before performance returns to typical patterns. In other situations, the event may alter performance for a long period of time, as the changes associated with working from home following the COVID-19 pandemic have had on traffic patterns, congestion, and safety. The guide provides examples illustrating times when an agency might choose to include or exclude the disruption or use the disruption to alter previous performance patterns.