The Relationship Between Transit Asset Condition and Service Quality (2018)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

14 C H A P T E R 3 This chapter presents a framework for evaluating service quality and a proposed set of relation­ ships formalizing the link between asset condition and service quality. The previous chapter summarized the information from various materials reviewed as part of this task. Service Quality Framework This section describes the basic attributes of transit service quality. It is based on U.S. and international research described in a range of publications, including TCRP Report 165: Transit Capacity and Quality of Service Manual (3rd Edition) (KAI et al. 2013) and other references. Service quality itself may be expressed through a range of transit system performance and customer satisfaction attributes: • Reliability and customer satisfaction. In terms of performance, service quality is often con­ sidered to be about reliability, meaning that customers expect the transit system to consis­ tently deliver on­time service. If it does not, then reliability declines, customer perceptions are affected, and customer satisfaction is likely to decline. If customers perceive that service performance is unpredictable because of the uncertainty in travel time, then they will regard the transit trip as risky and perhaps not worth taking. • Travel time expectation. Another consideration is that the travel time expectation is not nec­ essarily an absolute—but, rather, that customers will readily tolerate a threshold that is likely to vary from customer to customer. Uncertainty is associated with travel time performance that exceeds a certain threshold above the expected travel time (which itself may include some baseline delay over best­expected­travel time). • Performance of the “system.” Reliability is a function of the performance of the transit “sys­ tem” which, in turn, depends on various factors, including operating policies, occurrence of asset failures that may depend upon asset condition, and external events (such as traffic con­ gestion, adverse weather, or police incidents). Where an asset is to blame for a service delay, it may be a vehicle mechanical breakdown or a defect in fixed infrastructure such as track, signals, or the traction system that has failed and resulted in delays to service. These are the performance issues central to this research. • Technical obsolescence. A complicating factor in considering how specific assets affect service quality is technical obsolescence, particularly regarding areas such as ticketing and communi­ cating travel information with customer­facing electronic components. Over time technical standards have changed, and older assets—even if well­maintained—may be obsolescent. Assets of lower technical standards may affect customer perceptions and, in some cases, may Framework for Relating Transit Asset Condition and Service Quality

Framework for Relating Transit Asset Condition and Service Quality 15 contribute to transit system unreliability compared to systems where obsolescent components are replaced. In formulating a relationship relating asset condition and quality of service, it is important to first recognize the different factors incorporated in defining service quality. The following paragraphs summarize the different service quality attributes and performance measures, based on the literature reviewed in Chapter 2. Definition of Service Quality TCRP Report 165, Chapter 4, which draws from a wide body of research and literature, defines “Quality of Service” (QoS) as follows in relation to those aspects of transit service that directly influence how passengers perceive the quality of a particular transit trip: The overall measured or perceived performance of transit service from the passenger’s point of view In addition, the reference discusses how expectations of QoS vary among different types of passengers and the importance of maintaining high QoS for attracting and retaining riders. QoS is both a reflection on the service a transit agency provides and an indicator of potential changes in ridership. Attributes of Transit Service Quality Based on the review of resources, the research team developed a comprehensive set of QoS attri­ butes and evaluated the extent to which asset condition may affect these attributes. Table 3­1 pro­ vides an overview of identified attributes of transit service quality affected by asset condition. The table contains the superset of service quality attributes identified in the references reviewed, as well. The first part of the table identifies the attributes that have a direct relationship to asset condition and the second part of the table identifies other attributes of service quality referenced in the litera­ ture but not directly related to asset condition. Key conclusions from this exercise are as follows: • Asset condition is related to the quality attributes of Frequency, Reliability, and Travel Speeds. As assets decline in condition, failures become more likely, reducing frequency, reli­ ability, and average speeds. • Asset condition is related to Appearance/Aesthetics in that customer perceptions of this factor are likely to be worse when assets are in poor condition and/or are technically obsolescent. • Asset condition is related to other service quality attributes, including Comfort, Ease of Access, Environmental Impact, Information, Safety, and Security. For these attributes, increasing asset failures may affect quality, although the effects may be difficult to model. For instance, reduced service frequency resulting from increased failures can lead to greater crowding and less comfort. A particularly challenging area is safety. Given that operating severely deteriorated assets could compromise safety, in theory and in general practice, transit agencies establish thresholds for safe operations and remove an asset from service, rather than operating unsafely. Thus, potential safety issues tend to become operational issues, rather than actual safety issues. Regardless, customer perceptions are likely to be worse when assets are in poor condition for these factors, as in the case of Appearance/Aesthetics.

16 The Relationship Between Transit Asset Condition and Service Quality Safety Vehicle safety as measured by accident numbers by distance. Passenger safety as measured by injury or fatality rates. If deteriorated assets are left in operation they may compromise safety, but potential safety issues are generally operationalized. Regardless, deteriorated asset conditions may contribute to customer perceptions of reduced safety. Security Security for the whole journey – relates to personal safety and security rather than injury prevention. Deteriorated asset conditions may contribute to customer perceptions of reduced security. Attribute Service Quality Measure/Description Relationship with Condition Service Quality Attributes Related to Asset Condition Appearance / Aesthetics Visual appearance from a customer perspective. Deteriorated conditions may affect customer perceptions of service quality. Comfort Includes availability of seats, standing room/degree of crowding and ride comfort. May include a range of other service elements designed to make the service more comfortable (e.g., temperature control). Increased headways or reduced reliability may result in greater crowding, indirectly reducing service quality. Also, assets in poor condition may not provide the same level of comfort. Ease of Access Access and egress to/from the system including the interface with other modes and the ease of reaching facilities, particularly for passengers with disabilities. Generally this attribute relates more to design issues than condition. Asset failures such as failed wheelchair lifts or elevators may limit accessibility, but intermittent failures can be modeled as causing significant increased delay for passengers dependent on them. Environmental Impact Externality effects such as wayside noise and pollution. Energy efficiency may drop as vehicles/systems age, resulting in increased emissions. Frequency How often service runs, inverse of headway. Asset failures may increase vehicle headways, reducing frequency. Reliability Asset downtime. Standard deviation of headways. Perceived and/or actual wait times. MDBF. Asset failures become more likely as assets age and decline in condition. Failures result in increased delay and reduced system reliability. Ride Quality Quietness and smoothness of the ride. Deteriorated asset condition may result in reduced ride quality, but this is difficult to model. Further, conditions may contribute to customer perceptions of comfort. Travel Speeds Journey time. May also address extent of speed restrictions. Asset failures create delay, reducing travel speed and increasing journey time. Table 3-1. Summary of transit service quality attributes.

Framework for Relating Transit Asset Condition and Service Quality 17 Asset Condition and Service Quality Relationships The measure, Effective Journey Time (EJT), is proposed to integrate consideration of dif­ ferent aspects of service quality into a single overall measure and demonstrate the effects of changes in asset condition. EJT combines actual journey time with adjustment factors for dif­ ferent components of the journey, as well as for customer perceptions. Changes in asset condi­ tion affect EJT in two basic ways: they affect actual journey time and they may affect customer perceptions, resulting in revisions to the adjustment factors for one or more components of the journey. EJT can be converted into Effective Journey Cost by multiplying EJT by the personal value of time. Effective Journey Time Formulation The equation for EJT is shown below. The first term in the equation represents average travel time. For each segment of the trip, the average travel time is multiplied by a factor (f ) that reflects the conditions experienced by the passenger. The second term in the equation represents the travel time variability. This includes the standard deviation of travel time (s) and a constant (k). EJT is calculated as follows: = Σ + Σ σ (3-1)2EJT f t ki i i i i where • ti is the average travel time for segment i. A trip would typically include at least two segments (waiting and IVT) but could include more. • fi is the factor applied to time on segment i to reflect the conditions experienced by passen­ gers on that segment. By convention it is assumed that fi is 1.0 for travel on a vehicle in good condition. • si is the standard deviation of travel time for segment i. • k is a constant that reflects the relative importance of travel time variability (expressed as the standard deviation of travel time) in relation to average travel time. The formulation of EJT is detailed further in Appendix B. Affordability User costs. Availability Defined in terms of service extent – does a service exist, and if so where and over what time period? Note: as defined here, availability relates to whether service is provided – frequency and delays are addressed through other attributes. Cleanliness Customer perception of transit vehicle/system cleanliness. Customer Care How staff interact with customers. Information Availability and visibility of information to help customers plan and execute their journeys. Transit Amenities Whether particular facilities are present or not (e.g., extra features and services to enhance user experience). Other Service Quality Attributes from Literature Attribute Service Quality Measure/Description Table 3-1. (Continued).

18 The Relationship Between Transit Asset Condition and Service Quality Journey Time Components Table 3­2 summarizes the components of journey time implied by Equation 3­1. Actual journey time is the sum of the components listed in the table. For journeys consisting of multiple segments (for example, a bus trip followed by a connecting train trip), all segments need to be included to determine the total actual travel time. Although all of the components notionally exist, In­Station Conveyance Time is relevant only in cases where one is considering passenger movement through stations. It is included in the comprehensive EJT tool described in Chapter 6, but as detailed in this chapter, consideration of this factor requires quantifying additional data items. Adjustment Factors Each journey time component in Table 3­2 can be adjusted to account for customer percep­ tions of the type of time and the asset condition. An adjustment factor of 1.0 is assumed for travel while seated in a vehicle in good condition. This factor adjusts primarily for the customer experience of comfort while at a station or on a vehicle. Other attributes that affect customer experience, such as reliability, are incorporated elsewhere in the model. Table 3­3 summarizes the adjustment factors recommended for different journey time components and conditions relative to the assumed value of 1.0. Effects of Asset Condition Changing asset conditions may affect both actual journey time and passenger perceptions of that time. Major condition effects include the following: • Changing vehicle failure rates. As vehicles age, their MDBF is projected to gradually increase. Likewise, replacing a deteriorated vehicle fleet may result in an increase in MDBF. TCRP Report 157 (Spy Pond Partners et al. 2012) provides models for predicting changes in MDBF as a function of changes in vehicle age or mileage and includes default values for different vehicle types. In models detailed in this report, a vehicle failure is assumed to create a delay equal to one headway, but, particularly on bus rapid transit or rail systems, a single vehicle failure may create cascading delays for other buses or trains. The likelihood of a failure can be multiplied by the average delay duration to calculate the effect of vehicle failures on headways and the standard deviation of journey time. • Changing failure rates for fixed assets. As in the case of vehicles, fixed assets such as guideway, elevators, and escalators, are more likely to fail as they age and may affect journey times when they do. Guideway failures may increase headways, travel time, and the variation of travel time and may affect multiple buses or trains. Deteriorated guideway conditions may result in Component Description Buffer Time Time passengers build into their travel time in anticipation of delays or unreliable service In-Station Conveyance Time Time required to navigate a bus or rail station, including time walking, handling fare payment, traveling up or down stairs, and taking elevators or escalators Wait Time Time passengers wait at a stop or station In-Vehicle Time Time passengers spend on a vehicle to reach their destination Table 3-2. Journey time components.

Framework for Relating Transit Asset Condition and Service Quality 19 speed restrictions, thus increasing travel time. Failures of other customer­facing assets, such as elevators or escalators in stations, may result in increased in­station conveyance time and the standard deviation of journey time. TCRP Report 157 also includes models for predicting the likelihood of failure for these assets as a function of asset age or condition. • Changing adjustment factors to account for passenger perceptions. An adjustment factor of 1.2 is recommended for old vehicles (vehicles that have reached their useful life) relative to new vehicles, based on the literature review discussion in Chapter 2. However, at least one study reviewed recommends a much lower value, based on a revealed preference study of rail passengers in the United Kingdom (Wardman and Whelan 2001). Regardless of the specific factor used, the effect of the adjustment factor is to cause a gradual increase in EJT as vehicles age and a sudden improvement when a fleet is replaced. Adjustments for passenger percep­ tions can be made for stations and other customer­facing assets, as well, but the review did not yield any examples of studies relating passenger perceptions to facility age. Component Factor Apply to Reference Notes Buffer Time 1.3 All (Small et al. 1999) Factor is applied to standard deviation of travel time. Same factor is recommended, regardless of condition. In-Station Conveyance Time 2.0 Walking on a level surface (TfL 2014) Value used for London Underground (LU) Journey Time Metric (JTM). 2.5 Walking down stairs (TfL 2014) Value used for LU JTM. 4.0 Walking up stairs (TfL 2014) Value used for LU JTM. 1.5 Riding an elevator or escalator (TfL 2014) Value used for LU JTM. Wait Time 1.9 All Various Factors for excess waiting and/or standing time range from 1.7 to 2.6 (KAI et al. 2013, Littman 2015, TfL 2014, Balcombe 2004). In-Vehicle Time 1.0 Seated in an uncrowded, new vehicle 1.2 Seated in an uncrowded, old vehicle (Steer Davies Gleave 1991) Assumed to apply to a vehicle that has reached its useful life (estimated lifespan). This factor varies and may be as low as 1.02 based on Wardman and Whelan (2001). 1.0-2.3 Degree of crowding (KAI et al. 2013) Includes an equation for predicting a perceived travel time adjustment factor based on degree of crowding. This is multiplied by the vehicle condition factor. Table 3-3. Recommended journey time adjustment factors.

20 The Relationship Between Transit Asset Condition and Service Quality • EJT changes to account for changes to service attributes. The EEM (NZTA 2016) describes a range of user benefits and provides a way to calculate the economic benefits of making improvements to transit service. The values associated with these benefits can be used in the EJT model in appraising investments in transit asset maintenance, refurbishment, or replacement programs. Factors from the EEM and their EJT in­vehicle equivalent values on a per­trip basis are listed in Appendix B. The EEM defines a wide range of factors, including boarding, driver behavior, ride quality, facilities on the bus, seating, information, shelters/ stops, ticketing, security, and stations. Although many of these are not related to condition, they may still be useful for transit agencies considering the merits of other programs or ser­ vice designs. Example Applications Figure 3­1 provides an example EJT calculation showing the unadjusted and adjusted journey time, both before and after SGR investment, for a hypothetical rail line. Although the calculation is not based on actual data, it is intended to show that, after a SGR investment, both the adjusted and unadjusted journey times decrease. Before investment, the actual journey time for all com­ ponents is 19.1 minutes, increasing to 23.4 minutes when adjustment factors are included. Fol­ lowing an investment in improved vehicles and track, these journey times decrease to 13.3 and 16.6 minutes. Looking closer at the components of time that change in this example, the decrease in buffer time after SGR investment indicates that travel times are more reliable and passengers do not have to build in as much buffer when planning their trips. Overall in this example, SGR investments reduce EJT an average of 6.8 minutes per trip. This savings can be multiplied by an average value of time and number of trips per year to yield a prediction of the annual user benefit resulting from the investment. Figure 3-1. EJT example calculation.

Framework for Relating Transit Asset Condition and Service Quality 21 By changing the different adjustment factors and considering the three types of time, EJT can be used in several ways to analyze asset condition and service quality. For example, EJT can be used to 1. Calculate a baseline value for EJT per passenger or per day. This requires calculations be done for a route segment or an origin­destination (O­D) pair for a representative route. The result can be expressed in minutes or dollars. 2. Create an alternative scenario. EJT can be used to simulate different conditions. For example, the effects of higher vehicle or fixed asset failure rates, greater IVT, or increased headways could be assessed by changing those parameters. 3. Calculate the difference to determine excess journey time or cost. EJT can be converted to a journey cost by multiplying it by value of time. The difference between the journey cost for a future or worst case value and that of a baseline scenario can be used to demonstrate the effects of deteriorated asset conditions. 4. Use to supplement or replace cost calculations in other models. Existing modeling approaches for transit assets typically lack a user cost component. TAPT includes user costs, but has a simplified cost prediction relative to that described here. Model developers may wish to incorporate the EJT calculation approach described here as an addition to models that lack a prediction of user costs or as a refined approach for predicting these costs.