SELECTING USEFUL WORK ZONE PERFORMANCE MEASURES

In this section, several possible performance measures are described that can be useful for monitoring and characterizing safety and mobility impacts at highway work zones. The value of these measures is described, along with key considerations in obtaining and monitoring them.

EXPOSURE PERFORMANCE MEASURES

Table 1 presents a list of possible performance measures pertaining to work zone exposure. Exposure measures describing vehicle throughput and miles of travel through work zones can be valuable for establishing safety and mobility rate-based measures as well as for tracking motorist diversion decisions. One of the challenges in using vehicle-miles-travelled as an exposure measure is in defining what length or limits should be used in the computations. For major roadway rehabilitation and reconstruction projects, temporary geometric changes and traffic control features throughout the work zone may suggest that project limits be used for estimating vehicle miles traveled (VMT) of exposure. For other projects, though, vehicle exposure during the times when temporary lane closures are occurring (such as during hours of a hot-mix asphalt overlay job) may be the exposure numbers of interest. In this case, the length of actual lane closures each day or night, rather than the total length of the project, is likely to be a more useful exposure measure.

Exposure measures for tracking contractor activity and efficiency, as well as the amount of roadway space closed to traffic because of work zones, can also be useful for agencies. Examples of these types of measures include the percent of days (or nights) when work activity is occurring, the average number of work activity hours per day (or night), and the percent of work hours when one or more lanes are temporarily closed.

Actions taken, or technologies used, to improve safety or operations by an agency can also be measured as illustrated below (this list is not exhaustive):

• The number or percent of projects using portable concrete barrier to protect work spaces and/or pavement drop-offs,
• The number or percent of projects where the agency employed a reduced speed limit (possibly stratified by the amount of speed limit reduction), and
• The number or percent of projects where work zone intelligent transportation system technology was used.

Table 1. Potential Exposure Measures

Measure	Why It is Important	Considerations in Using It
Vehicles passing through the work zone, possibly stratified by work activity or lane closure hours.	Required to compute other performance measures on a per vehicle basis.	Work zones with significant diversion require continuous counts in order to be accurate.
% change in vehicles passing through the work zone	Can represent traffic diverting from the work zone due to public information or other strategies	Requires continuous counts in order to be accurate.
Vehicle-miles-of travel through the work zone, possibly stratified by work activity/inactivity periods and /or lane closure times.	Standard measure of exposure for safety assessments	For some projects, total project length may be the appropriate length to use; for other projects, the length of each lane closure may be more appropriate for computing vehicle-miles-traveled of exposure.
% of days or nights when work activity occurs.	Describes intensity of effort being made to complete the job	Some projects are based on total calendar days, while other projects restrict when work can occur and should consider only the allowable work days/nights.
Average hours of work per day or night.	Useful for assessing contractor utilization of time windows allowed for work	Overlap of subcontractor work efforts on complex jobs can make it difficult to define an overall start and end time each day or night.
% work activity hours with (1,2,3, etc.) lanes closed	Allows queue and delay measures to be stratified by amount of roadway capacity reduction involved	Lanes are sometimes closed in sequence over the course of a work shift (especially at night). It can be difficult to accurately capture these data in detail.
Average lane closure length	Useful for evaluating compliance to maximum lane closure length specifications (if included in contract). Also, useful for extrapolating to similar projects that are not being monitored as closely.	Lane closure lengths can change significantly during a shift as additional lanes are closed (especially at night).
Lane-mile-hours of closures	May be useful in explaining high levels of delay or crash rate increases at certain projects	This measure is less intuitive than other measures.
Number or % of projects employing certain types of strategies or technologies to reduce safety or mobility impacts	It is a program-level measure of agency efforts to improve safety or mobility	Decisions to implement these are typically site specific. It may be necessary to stratify the measures by type of facility, type of work, or other site characteristics in order to be meaningful to the agency.

SAFETY PERFORMANCE MEASURES

Table 2 lists several work zone safety-related performance measures that agencies can use. Many states currently track overall work zone fatalities and injuries statewide on an annual basis. This approach provides a program-level indication of the safety impacts in work zones statewide, and is tracked over time to observe trends in crash frequency over time.

The percent of work zone crashes by manner of collision or other crash characteristics is a common type of performance measure used by agencies to identify potential problem areas for countermeasure emphasis. These types of measures are dependent upon differences in traffic volumes and other factors from project to project (5) . Consequently, examination of these types of measures at a program level (for all work zone crashes in the state, for example) can yield different conclusions from year to year as the characteristics of the projects changes each year.

Table 2 also lists changes in work zone crash rates (or crash costs) as a potential safety performance measure. The use of crash rates reduces the effects of differences in exposure and other external factors between projects and allows for a more accurate assessment of work zone safety by the agency. The crash rate per million-vehicle-miles-traveled (MVMT) is a commonly-used measure, and can be stratified by crash severity (injury crashes per MVMT) or other crash characteristic. Other exposure measures can be used as well (e.g., crashes per hour of work activity). Program-level evaluation of agency safety policies and procedures will be best accomplished through crash rate measures.

Table 2. Potential Safety Measures

Measure	Why It is Important	Considerations in Using It
Number of crashes per year or per project: Fatal Injury PDO	Actual crash counts represent the direct effects of work zones on safety.	Crash counts are dependent upon exposure (volumes, work zone length, and work hours at the project level; number of projects at the program level). Changes in funding from year to year could influence exposure and thus program level work zone crash counts.
% Crashes of certain types such as manner of collision (e.g., rear-end crashes), contributing factors (e.g., DWI crashes)	Agency priorities for countermeasure implementation are often linked to the major contributors of work zone crashes.	Percentages of various crash types can also be affected by exposure.
% Change in crash rate or absolute change from the expected no-work zone crash rate; possibly stratified by roadway, work zone type, severity	Changes in crash rates account for key differences in exposure that may also affect crash frequencies, thereby improving the comparability of the measure between sites.	This measure is computed for each individual project of interest; it requires information on exposure before and during each work zone
% of projects that exceed an acceptable crash rate in the work zone	A program-level measure to evaluate compliance with agency-established policies or targets	A reasonable number of projects must be evaluated for this to be meaningful to agencies. It can create substantial workload for the agency, especially if different roadway and project types are to be evaluated separately.
% Change in work zone crash costs from the expected no-work zone crash costs	Crash costs combine frequency and severity together in one measure. Results from multiple projects could be combined for program-level analysis by the agency.	This measure is also computed for each individual project of interest; it requires information on exposure before and during each work zone.
Number of highway worker injuries or worker injury rate per hours worked	Workers are a particularly vulnerable population in work zones.	Obtaining worker injury data from contractors can be difficult. The use of rates will require data on hours of worker exposure.
Work zone inspection scores	Work zone inspections or reviews are under the control of the agency. This can simplify data collection and analysis.	Compliance to agency inspection standards is believed to be correlated to improved work zone safety, but research has not yet validated this.
Number of work zone inspections performed	The number of inspections is a program-level measure that agencies can use in conjunction with other safety measures to assess current policies and procedures.	Same as above.
Frequency or change in frequency of service patrol or fire department dispatches to a work zone	Crashes (especially severe crashes) will involve emergency responders. Agency access to this type of data can be faster than access to crash data.	Interagency agreements may be needed in order to obtain the data. For service patrols, non-crash event dispatches (stalls, motorist assistance with fuel, debris in road) may need to be excluded from the dataset to have a meaningful work zone measure.

The use of crash costs allows an agency to incorporate the effects of both crash frequencies and crash severities into a single measure. It is possible, for example, for the number of property-damage-only (PDO) crashes to increase in a work zone (relative to a no-work zone condition) and the number of injury crashes to decrease slightly because of slower speeds through the work zone. The combined effect of those changes can be captured in the change in crash costs. Given that most states have a minimum damage cost reporting threshold for PDO crashes, increases in very minor crashes will not be captured in this measure, however.

Highway worker injuries or injury rates are another potentially useful work zone safety performance measure. Most agencies already track injuries of their own employees, although not always distinguishing between injuries that are traffic related and those that are work related. Basic information on non-agency highway worker injuries can be obtained through the Bureau of Labor Statistics (BLS) database (6) .

Many agencies perform regular inspections or reviews of work zones under their jurisdiction. Standardizing the review forms and assessment criteria has allowed some state DOTs to establish safety performance measures using these data. Both project-level rating scores and program-level percentages of projects meeting agency-established score thresholds can be established (3) . In the absence of a formal rating system, other agencies will typically track the number of project inspections or reviews done as a type of work zone safety performance measure.

Timely access to crash data is a problem that many agencies face when attempting to measure work zone safety performance. At least one agency has attempted to circumvent this problem by examining service patrol and fire department dispatch frequency to locations within their work zones (3) . Although information regarding crash location (and possibly crash severity) will be available through these sources, other crash characteristics will not. In addition, non-crash events (stalls, debris in road, etc.) may also be captured by service patrol dispatch logs, which can reduce the quality of this measure if these events cannot be separated from the dataset.

MOBILITY (TRAFFIC OPERATIONS) PERFORMANCE MEASURES

Queuing Performance Measures

Both queues and traffic delays reflect the effect of work zones on traveler mobility, and are correlated with each other. Both the safety considerations associated with the formation of queues and the fact that they can be directly measured more easily than delays has led several agencies to establish policies as to the maximum acceptable length and duration of work zone queues. Table 3 presents several work zone queue-related performance measures for agencies to consider.

As can be seen, the measures shown in Table 3 are more detailed than simply whether or not a queue formed at a given work zone. Queue frequency, length, and duration are all important attributes to be captured in work zone queue performance measures. Generally speaking, it is preferable to have multi-level performance measures rather than simple present/not present or yes/no performance measures whenever possible (9) .

Table 3. Potential Queuing Measures

Measure	Why It is Important	Considerations in Using It
Number or % of days or work activity periods when queuing occurred	Some work zones have only occasional queues, whereas others will experience queues on almost daily basis	It is more difficult to detect and record infrequent queuing unless traffic conditions are being continuously monitored and recorded.
Average queue duration	Useful for road user cost computations	Very short-duration queues cannot always be detected. Also, crashes or other external influences can skew work zone queue duration values.
Average queue length	Useful for road user cost computations	Defining the beginning and ending points of a queue can be a challenge. Also, crashes or other external influences that occur can skew work zone queue lengths.
Maximum queue length	Maximum queue lengths can help agencies assess whether advance warning signage is placed far enough upstream of the lane closure to adequately warn approaching motorists.	Queue lengths can change rapidly over time, and may be at a maximum for only a very short period of time. Also, queue lengths may differ by lane, depending on the geometrics of the roadway and driver behavior.
% Time when work zone queue length exceeds xx mi	Combines queue frequency, length, and duration into a single performance measure	Queues due strictly to external influences (weather, crashes) should be examined separately from queues that are due strictly to work zone operations and temporary lane closures. This measure will typically require continuous monitoring of traffic conditions.
Amount (or % of ADT) that encounters a queue	The number of vehicles that encounter a queue is useful for evaluating appropriate beginning and ending times of temporary lane closure periods.	Queues can cause diversion from the work zone, the amount of which is difficult to predict.

ADT = average daily traffic (vehicles per day)
xx = threshold level as defined by the agency

Agencies will likely be interested in examining queue performance measures that are due to certain work activities (temporary lane closures, close proximity of work equipment to travel lanes) separately from those measures that include external queuing influences in the work zone. For example, adverse weather conditions can slow vehicle speeds and increase gaps between vehicles, thereby increasing the likelihood of queues developing where one typically does not occur. The occurrence of a traffic crash in the limits of a work zone may also create a queue where one normally would not occur, or significantly increase the length or duration of a queue that normally exists. While it can be argued that those types of queues are also related in some fashion to the design of the work zone, queues that occur strictly because of work operations reflect directly upon agency work zone policies and procedures in terms of how many lanes are allowed to be closed or hours when lanes can be closed.

Queuing performance measures in Table 3 are defined relative to a no-queue condition during the periods of work activity. The assumption is that any queues that develop are the direct result of the work activity and temporary lane closures required. If recurrent queues were occurring at the site before the project began, these queue performance measures would need to be defined in terms of changes from their pre-work zone levels. Likewise, if long-term lane closures were used at a project that resulted in queues only during peak periods, it may be more appropriate to evaluate queue performance measures strictly over the peak period at that site.

Delay Performance Measures

Table 4 describes several possible work zone delay performance measures for use. Delays are needed to estimate road user costs caused by work activities. Road user costs drive decisions regarding bidding approaches and contracting strategies employed, incentive and/or disincentive provisions used in the contract, techniques used to accelerate construction, and the traffic impact mitigation strategies that may be implemented. Separate delay measures are often used to describe the impacts of the work zone to an individual vehicle (delay per vehicle), and to the collective motoring public overall (total vehicle-hours of delay).

Table 4. Potential Delay Measures

Measure	Why It is Important	Considerations in Using It
Vehicle-hours of delay per: Work period Work period when queues are present Peak period Project	Multiplied by the value of user travel time, these measures define road user costs that are attributable to work zone mobility impacts, stratified by the conditions and times when they occur.	In areas already experiencing recurrent congestion, it is necessary to first determine the delays normally occurring on the roadway prior to the start of the work zone
Average delay per: Entering vehicle Queued vehicle Peak period vehicle	Motorists are more sensitive to individual vehicle delays than to total delays; these measures can be stratified to better define the impacts to various vehicle subsets.	Continuous travel time and volume data are desired to compute average and maximum delay measures, as well as the percentage of vehicles which experience delays more than deemed acceptable by the agency.
Maximum per-vehicle delay	Knowing the upper bound on maximum individual delay experienced during a project can be helpful in responding to public complaints about perceived level of mobility impacts.	External influences (crashes, vehicle stalls, weather conditions) can dramatically increase work zone delays. Delays during these events should be categorized separately from those due strictly to work activities.
% Vehicles experiencing delays greater than xx minutes	In addition to knowing the magnitude of delays, knowing what percentage of drivers are experiencing more than agency-defined tolerable delays is also important for program-level reviews.	Diversion from the work zone due to delays makes continuous travel time and volume monitoring highly desirable.

xx = threshold level as defined by the agency

Recognizing that individual delays can vary significantly over the course of a project or even hours of a particular work shift, multiple versions of these delay measures may be needed to capture both the extreme and "typical" impacts. Another measure, percent of work activity time when motorist delays are exceeding some threshold, will be useful to agencies that have identified a maximum tolerable level of motorist work zone delay.

In locations that are not experiencing recurrent congestion, any delays occurring are directly attributable to the work zone. In locations where recurrent congestion is already occurring, it will be necessary to establish baseline delays prior to the start of the work zone so that the additional impacts of the work zone can be appropriately determined.

Travel Time Reliability Performance Measures

In addition to the congestion and delays that can be created, work zones can also make travel times less predictable or reliable for motorists who regularly use a particular facility. Drivers want dependable travel times so that they can better plan their departure and arrive at a destination near a desired time (11).

Roadways with highly variable travel times require motorists to "buffer" in more time in their departure time decision to ensure that they are likely to arrive on time, even though there is a chance that they will arrive much earlier than necessary if travel conditions are favorable. One way to describe travel time reliability is through a buffer index performance measure, defined for a particular time period (peak, nighttime, etc.) as:

(Equation 1)

The use of the 95th percentile travel time as the upper limit implies that someone who allows that amount of time for their trip would arrive late no more than once every 20 trips (or days). Similarly, use of an 80th percentile travel time in the above computation would correspond to arriving late no more than once every five trips (or days).

Currently, most state DOTs do not establish separate work zone travel time reliability performance measures. However, efforts to evaluate the potential usefulness of work zone travel time reliability performance measures were positive (2) . For those agencies that already track travel time reliability performance measures on a regular basis, examining work zone mobility-related impacts with this type of measure should be considered.

Some agencies may use other measure to characterize travel time reliability on their facilities, such as a planning time index or the percent of time that congestion (i.e., delays or travel times) exceeds some defined threshold level. These same measures can continue to be monitored during work zone conditions as well, and the change attributable to the work zone can be calculated.

Other Types of Work Zone Mobility Performance Measures

Although queues and travel time delays are the most direct indicators of work zone mobility impacts, some state DOTs have tried and are using other types of performance measures that relate to work zone mobility. Table 5 summarizes several of these types of measures. Average speeds in the work zone can provide an indication of the quality of service being provided. Often, multiple data collection locations are needed to fully characterize operating conditions throughout the work zone. Some agencies use level-of-service (LOS) as a mobility-related performance measure for work zones within or near at-grade intersections. Work zone volumes or throughput are another type of measure that has been used. A few agencies track the percent of time a project is operating at or near free-flow speeds as a work zone mobility performance measure. Although less objective than other mobility-related performance measures, customer complaints or surveys are also used by many agencies to assess how well mobility is being maintained in work zones.

Table 5. Other Potential Mobility-Related Measures

Measure	Why It is Important	Considerations in Using It
Average speed	Speed is easy to measure, and can be obtained at multiple locations (if desired) to evaluate operating conditions in various parts of a work zone.	The location of data collection can have a significant effect on speeds obtained. If handheld devices are used, measurements must be taken from close to the actual travel lanes to reduce errors.
Intersection Level-of-Service (LOS) maintained during the project	Some agencies use impact assessment tools that provide intersection LOS, and so have based their policies and acceptable thresholds around that measure.	LOS cannot be collected directly in the field. It must be computed from volume, delay, and signal timing data.
% Work zones meeting agency expectations for traffic flow	This is a program-level measure that can be directly and quickly obtained by agency staff as part of project field reviews and inspections.	Agency personnel performing the assessments may not all have the same expectations. Also, agency expectations may not always align with motorist expectations.
Customer complaints	Complaints can help identify in real-time which work zones are creating major impacts, and allow agencies to implement corrective measures quickly. Most agencies already have procedures in place to track and respond to all complaints received from the public.	Information received in a complaint can be subjective and incomplete. Agencies can spend considerable time deciphering and then responding to each complaint received.
Customer survey ratings	Surveys about travel conditions in work zones can be done fairly quickly and can yield useful insights about problems and possible corrective actions.	Unless properly designed, survey results can be somewhat biased (positively or negatively), which can reduce their effectiveness.

previous | next