Quick Response Freight Manual II

10.0 Freight Data Collection

This section provides a detailed discussion on data collection for freight planning and forecasting. The subsections in this section include a discussion of the need for freight data collection, the common types of data collection supporting freight planning and forecasting, and key issues associated with collecting freight data pertaining to costs, sample sizes, and implementation processes.

10.1 Need for Freight Data Collection

Section 9.0 of the QRFM provided a detailed description of the various existing freight data sources available at national and regional levels for freight planning and forecasting applications. These included commodity origin-destination databases (such as FAF2 and TRANSEARCH), modal flow databases (such as the Carload Waybill Sample), vehicle data (for example VIUS), and employment/industry data (for example County Business Patterns). Although these data sources provide comprehensive information on base year and forecast freight demand, transportation supply, and economic characteristics in a region, there are often a host of other critical data needed for freight planning/forecasting that are beyond the scope and coverage of these standard data sources. For example, truck volume data on the highway network is a critical need for MPOs and other regional planning agencies for the validation of the regional truck models. Understanding time-of-day characteristics of truck traffic is another important need for planning agencies to understand peak-hour interactions between passenger and freight traffic, and plan for congestion alleviation measures during peak hours. These data elements can only be compiled from a local freight data collection effort. Also, in many cases, the data available from standard freight data sources may not be representative of the actual freight traffic characteristics in the planning region under consideration. For example, truck payload factors derived from the VIUS database are only available at the state level of detail and cannot always be applied to an urban area. Clearly, local data collection efforts can provide more representative and accurate data in such cases to support the freight demand analysis and planning process.

Although local freight data collection efforts require additional resources in terms of time and costs, they provide much needed data for a planning agency to conduct a comprehensive analysis of freight traffic flows in a region and develop more accurate freight forecasts for planning applications. Some critical factors that impact the time, costs, and level of effort associated with local freight data collection programs include the following:

The level of availability and comprehensiveness of existing data;
Type and volume of data/information needed;
Time needed to conduct data collection;
Desired level of accuracy and detail in the collected data; and
Types of equipment and resources (manual, etc.) required to perform data collection.

10.2 Local Freight Data Collection

Though there are a host of local freight data collection methods, this section covers the most important methods from a freight planning and forecasting application perspective. In addition to presenting the essential concepts associated with each data collection method, some key issues pertaining to costs, sample sizes, and implementation for each type of data collection, also are discussed.

The following types of local freight data collection methods are covered in this section:

Vehicle classification counts;
Roadside intercept surveys;
Establishment surveys; and
Travel diary surveys.

10.2.1 Vehicle Classification Counts

Introduction

Collecting vehicle classification counts is a common local freight data collection method, which involves counting traffic for each vehicle class (based on a particular vehicle classification system) for a certain duration of time at key locations on the highway network. Typically, the counts are collected during weekdays and may be collected for more than one day to get average weekday traffic volumes at the count location. Collecting counts by vehicle class is important in order to differentiate between automobile and truck traffic volumes, as well as analyze truck traffic volumes by truck type (the applications of vehicle classification counts are discussed in detail in a subsequent section).

The four primary methods of collecting vehicle classification counts are:

Manual Observation – Manual counting procedure involves a trained observer collecting vehicle classification counts at a location based on direct observation of vehicles. This procedure is generally used for short durations of count data collection (for example, peak hour), and in cases where available resources do not justify the use of automated counting equipment. Typical equipment used in manual counting for recording observed traffic include tally sheets, mechanical count boards, and electronic count boards.
Pneumatic Tubes – This data collection approach involves placing pneumatic tubes across travel lanes for automatic recording of vehicles. These tubes use pressure changes to record the number of axle movements to a counter placed on the side of the road. They can record count data for 24-hour periods or more and are easily portable.
Loop Detectors – This data collection approach involves embedding one or more loops of wire in the pavement, and connecting to a control box, excited by a signal (typically ranging in frequency from 10 KHz to 200 KHz). When a vehicle passes over the loop, the inductance of the loop is reduced, indicating the presence of a vehicle. One of the main benefits of this approach is the reliability of count data under all weather conditions. Load detectors are mainly used as permanent recorders, at locations where counts are required for a longer-time duration.
Videography – Videography involves collecting vehicle classification counts using video tape recorders and tallying them manually by observing vehicles on the video. Similar equipment, as described under the manual observation data collection approach above, can be used for tallying the data. A primary advantage of videography is the ability to stop time and review data, if necessary.

The vehicle classification system used for the count program can vary depending on the need, as well as the type of method used for counting vehicles. Classifying vehicles by the number of axles is the most basic vehicle classification scheme. However, this has some limitations with respect to applications for freight planning; for example, the inability to differentiate between automobiles and two-axle trucks, which is an important piece of information for freight planning applications in urban areas. The FHWA 13-group vehicle classification system is a common and efficient scheme for classifying vehicles (trucks are classified based on number of axles and number of units). This system is described in detail at the following web site: http://www.fhwa.dot.gov/ohim/tmguide/tmg4.htm#chap1. However, some data collection methods such as pneumatic tubes are only based on counting the number of axles and cannot classify vehicles based on the FHWA 13-group system. The key to classifying vehicles in count programs, and using them for freight planning applications, is to understand how the different classification schemes relate to one another. For example, translating length-based classification from loop detectors to axle-based classification and vice versa.

Applications of Vehicle Classification Counts

Vehicle classification counts are useful in freight planning and forecasting. Some applications are described below.

Model Calibration and Validation

One of the most important applications of vehicle classification counts is in performing model calibration and validation. Truck counts by truck type can be used to calibrate input origin-destination (O-D) trip tables of regional truck models using an Origin-Destination Matrix Estimation (ODME) process, if the collected counts provide a good geographic coverage of key truck traffic locations in the region. The ODME process iteratively updates the input O-D trip table of the model so that model truck volume results match with observed truck counts. Truck counts also are used for validation of regional truck models by comparing model results with observed truck traffic volumes at screenline locations. By collecting classification counts, this validation process can be performed by truck classes in the truck model. However, model calibration and validation processes cannot be conducted simultaneously because if an ODME process is conducted for model calibration, the model results are matched with observed truck counts, resulting in redundancy of a model validation process.

Time-of-Day Analyses

Another important application of vehicle classification counts is in performing time-of-day analyses of truck traffic volumes. Hourly counts collected over a 24-hour period can be used to develop time-of-day distributions of truck traffic volumes to analyze peaking periods for truck traffic. Classification counts also allow for the simultaneous comparisons of time-of-day distributions between automobile and truck traffic to understand peak-period interactions between passenger and freight traffic, and to plan for the efficient movement of traffic during the peak period. Classification counts also can be used to analyze time-of-day truck traffic characteristics by truck class, as well as by highway facility type (freight access routes versus major freight corridors).

Trip Generation

Vehicle classification counts also are used to develop truck trip generation models. For example, counts by truck class on access routes to major freight facilities provide inputs for developing regression models by truck class for truck trip generation. Truck counts also can be used to develop truck trip generation rates for freight facilities as a function of economic variables such as employment. Directional counts on access routes around freight facilities can be used to develop separate trip generation rates for production and attraction trips. However, the application of counts for trip generation analysis entails the availability of freight facility economic or land use data.

Identification of Major Freight Corridors and Access Routes

If the vehicle classification count program provides a good geographic coverage of sites on the highway network, it then can be used to identify major freight corridors and freight access routes in the region, based on an analysis of locations with high truck traffic volumes. This information serves as an essential input for defining the regional highway freight system of a region, which can be used for highway freight planning purposes.

Implementation Issues

Site Selection

The success of a vehicle classification count program in terms of its applicability for freight planning in a region is to a large extent determined by the selection of sites for collecting counts. The best approach to site selection is an initial assessment of the truck count data needs in the region and selecting sites based on a prioritization of needs. This approach not only ensures that the most critical data needs in the region are satisfied by the count program, but also is useful for the optimal utilization of resources available for conducting the count program. Some examples of critical freight planning data needs that feed into the site selection process include the following:

Truck volumes on truck model screenline locations;
Truck volumes on major freight corridors;
Truck volumes on major freight access routes; and
Truck volumes to meet specific jurisdiction truck traffic data needs.

An important consideration in the site selection process is the geographic coverage of the region, particularly if a primary application of the count data is for performing an ODME process for input truck trip table calibration.

Costs

The costs of collecting vehicle classification counts are primarily governed by the type of method used for collecting counts, as well as the number of sites selected for the count program. To reduce the overall costs of compiling traffic volume information on the regional highway network, the planning agency must consider availability of count data from existing count programs, in order to avoid duplication of count data collection efforts. For example, vehicle classification counts are collected by state departments of transportation at key locations on the highway network as part of their requirement to report traffic data to the FHWA for the Highway Performance Monitoring System (HPMS). Similarly, existing count programs of regional jurisdictions (for example, counties) and authorities (for example, sea and air ports) can provide traffic volume information by vehicle classes.

For each type of data collection, the actual costs will vary significantly depending on the efficiency of operation of the data collection firm, the accuracy sought from the data collection effort, as well as the characteristics of the site for ease of count data collection. Based on a review of previous count data collection efforts, the unit costs (per site) for conducting 24-hour vehicle classification counts by manual and video counting methods are approximately $650 and $500, respectively. In addition to these costs, there are typically costs associated with data synthesis and analysis that vary depending on the extent of the count data collected, as well as the tabulations associated with the analysis.

Data Variability Issues

Data variability is an important concern that needs to be addressed by any vehicle classification count program. In addition to time-of-day variations, truck volumes can have significant day-of-week and seasonal variations, which have not been as well established as time-of-day truck traffic distributions. For example, how seasonal changes in industrial shipment characteristics translate into seasonal variations in truck traffic volumes on the highway network. Thus, truck counts that are typically collected on a specific day of the year cannot be representative of average annual truck traffic volumes at the location and need to be adjusted to account for seasonal variations. These seasonality factors are typically derived from permanent traffic recorders that collect continuous counts. However, these locations are not typically distributed across the region with sufficient coverage of all relevant areas (for example, there is usually very little coverage on state highways and no coverage on arterials). Thus, vehicle classification count programs designed to capture seasonal variations especially on these road types can significantly increase the understanding of temporal variability in the region.

Advantages

Vehicle classification counts have many advantages, which are presented below for each method of data collection.

Manual Observation

Some advantages of manual counts are presented below:

There is no disruption of traffic during data collection.
There is minimum risk to individual observers collecting vehicle classification counts, as they do not have to interact with the traffic flows.
They may be more accurate than automatic vehicle classification counting methods and can count vehicles based on both axle group and number of units, thus enabling vehicle classification by the FHWA 13-group classification system.

Automated or Electronic Data Collection

Some advantages of automated or electronic counts are presented below:

There is no disruption of traffic during data collection, even though automatic vehicle recording equipment are placed on the pavement to count vehicles.
They are particularly useful when classification counts are needed at many sites, due to the higher efficiency in data collection with minimum labor requirement.

Video Surveillance

Some advantages of video surveillance-based counts are presented below:

There is no disruption of traffic during data collection.
They offer the ability to stop time and review data for quality checking.
They can provide better information on type of truck (and consequently, the type of commodity hauled) compared to automated counting methods.

Limitations

Some key limitations of vehicle classification counts are presented below for each method of data collection.

Manual Observation

Following are some key limitations of collecting vehicle classification counts by manual observation:

There is a high personnel requirement, as well as training, for conducting manual counts.
Manual vehicle classification counts have the potential for human error, especially under heavy traffic conditions.
They are not a good approach for counting vehicles during the nighttime period, as visibility can cause a problem in effective counting of vehicles by vehicle classes.

Automated or Electronic Data Collection

Following are some key limitations of automated/electronic vehicle classification counts:

There is a potential for equipment failure, which will impact the accuracy of the collected counts.
They are relatively more expensive compared to manual counting methods, especially for a larger geographic coverage area.
They can count vehicles only based on a particular classification system (for example, number of axles), and consequently, there is a potential for error when converting counts from one classification system to the other.

Video Surveillance

Following are some key limitations of collecting vehicle classification counts based on video surveillance:

They are associated with high equipment costs, especially for larger geographic coverage areas.
Weather can have a serious impact on video count programs, due to the potential for equipment failure or reduced visibility.

10.2.2 Roadside Intercept Surveys

Introduction

Roadside intercept surveys belong to the category of primary truck trip data collection, which involve intercepting trucks on the road and interviewing truck drivers to get information on their truck trip characteristics. The surveys are administered through the use of survey questionnaires that are completed by the interviewer based on information provided by the driver from the personal interviews. Typically, the interviewer makes visual observation of the vehicle to gather information about configuration and number of axles. There are many key steps involved in developing and implementing a roadside intercept survey, which include preparation of the survey questionnaire, site selection, site preparation, recruiting and training of survey personnel, sampling frames, survey administration, and survey data synthesis and analyses.

Depending on the types of freight modeling and planning applications, roadside intercept surveys can gather comprehensive information about truck travel characteristics in a region. The key data attributes that can be collected through roadside interviews include O-D locations (state, city, zip code), routing patterns, type of commodity, vehicle and cargo weight, shipper and receiver information (business name, contact, type of facility, etc.), trucking company information (business name, contact, type of carrier – truckload, LTL, or private, etc.), and type of truck (number of axles and number of units).

The locations for conducting roadside intercept surveys depend on the O-D truck travel patterns that are being analyzed. To gather truck trip characteristics of internal-external, external-internal, and external-external (through) trips, the most common approach is to conduct surveys at external cordon locations. External cordons are the external highway gateways that are used by trucks to enter and exit the study area. Collecting roadside intercept surveys within concentrated urban areas for internal-internal trips can be prohibitive because of the need to conduct surveys at many locations (due to the complex internal distribution patterns of trucks), as well as traffic congestion and/or limited space availability at survey sites. Some common locations for conducting roadside intercept surveys include weigh stations, toll plazas, and border crossing locations.

Terminal gateway surveys are a special class of roadside intercept surveys, wherein trucks entering and exiting terminal gateway facilities (seaports, airports, and intermodal rail yards) are intercepted and surveyed to get information on O-D locations, routing, commodities, payloads, truck types, types of carriers, O-D facilities used, etc. for trucks using terminal gateways.

Applications

Data collected from roadside intercept surveys and useful for freight modeling and planning applications, which are discussed in the following sections.

Origin-Destination (O-D) Freight Flow Matrix

A primary application of roadside intercept surveys is in the development of O-D freight flow matrices for a region. Depending on the extent of data available and the level of accuracy of the geocoding process for the O-D locations, a TAZ-level O-D freight flow matrix can be developed from the survey data. However, only the external gateway surveys offer the ability to develop accurate O-D matrices, since surveys conducted at internal locations are typically inadequate for developing a comprehensive O-D matrix incorporating all the possible O-D flow combinations in a region. The O-D matrix developed from external gateway surveys contains truck freight flows between external cordons and internal regions (TAZs or districts), and can be in terms of commodity flows (in tons) by trucking submodes (truckload, LTL, and private), or truck trips by truck class. These O-D matrices serve as key inputs in the development of external truck models for urban areas. Commodity-specific flows from these matrices also can be used to validate the TAZ-level disaggregation procedures in existing urban truck models for production and attraction trips.

Trip Distribution

Truck O-D information collected from terminal gateway surveys are essential inputs for developing truck trip distribution tables for terminal facilities. These tables can be developed by type of commodity and/or truck classes to understand the variations in terminal gateway truck O-D distributions as a function of these parameters.

Payload Factors

Roadside intercept surveys collect information on the type of commodity, weight of cargo, and type of truck that can be used to develop weighted average payload factors by commodity group and truck classes. These factors can be used in the development of commodity-based urban truck models (which involve conversion of commodity flows to equivalent truck trips by each truck class), or validation of payload factors in existing truck models to improve the accuracy of predicted truck trips.

Commodity Tonnage Distribution to Truck Classes

The type of commodity, weight of cargo, and type of truck information collected from the surveys can be used to develop tonnage distributions for each commodity group carried by each truck class, at each external cordon location. This information is a key input in commodity-based urban truck models to distribute total commodity flows to each truck class, in order to predict truck trips by truck classes.

Empty and Through Truck Factors

Empty truck trip fractions at external cordons are key inputs in commodity-based urban truck models in order to account for empty truck trips. Collecting through truck traffic information is a key requirement for developing robust truck models. Some of the key inputs for which include the fraction of total trips that are through trips at each external cordon, and through truck traffic distributions (the distribution of through trips at each external cordon through all the other external cordon locations).

Market Research

Roadside intercept surveys can be used for market research and have been applied successfully in many studies, particularly related to cross-border movements. Using intercept surveys at border crossing locations, information can be collected on major shippers and receivers involved in cross-border trade, as well as major carriers performing cross-border shipping operations.

Advantages

Following are some key advantages of performing roadside intercept surveys for gathering truck travel information:

They offer the best statistical control and reliability, since sample is from known traffic population.
They have high response rates compared to mail or telephone surveys, due to direct one-on-one interview with the driver.
Surveys at external stations provide a good statistical representation of trucks entering, exiting, and passing through the study area.
They have low investment costs, if managed and administered properly.
They offer the ability to gather comprehensive truck trip information in a single interview pertaining to O-D, routing patterns, commodities, shipment sizes, truck types, and facilities used.

Limitations

Following are key limitations of performing roadside intercept surveys:

There are only a limited number of locations where intercept surveys may be implemented in a region. This can lead to sampling bias in the truck travel characteristics determined from the survey.
There is potential disruption of traffic, especially when surveys are conducted by roadside pull-offs.
There is potential risk for survey personnel, related to safety risks from traffic and security risks from direct contact with interviewees.
They can only capture truck traffic characteristics of trucks passing through survey sites. They are not particularly effective for collecting information on internal-internal truck traffic characteristics because of the limitations in the number of sites, and the complexities in distribution patterns of internal-internal trips.

Roadside intercept surveys generally focus on “last stop-next stop” origins and destinations since questions involving multiple stops (trip-chaining activity) can be confusing to the driver and may yield less reliable data. This can be a potential limitation if last/next stops of the surveyed trip involve activities that are not related to goods movement such as fueling and rest areas.

Implementation Issues

Sampling Rates

Because of the impracticality of intercepting all the trucks passing through the survey site, sampling rates are typically developed to select a sample of the total truck traffic for the surveys. These rates can vary based on the total truck traffic volumes at the location, as well as the type of truck. The sampling rates also can depend on the rate of processing of surveyed trucks at the site, which is a function of the number of interviewers, as well as slot space available at the site for the surveys. Typically, roadside surveys at the site are accompanied by vehicle classification counts in order to determine total trucks passing through the location for expanding the survey sample data.

Three questions need to be answered when performing sampling analyses for roadside intercept surveys, which are as follows:

Where to sample (which sites to be selected for performing surveys)? Key parameters that help answer this question include the major locations for entry and exit of truck traffic in a region and locations of existing truck stop sites such as weigh stations, rest areas, toll plazas, and border crossing.
Who to sample (which trucks to be selected for surveys, and how many)? Key parameters that help answer this question include the types of trucks passing through the site and the volume of traffic by each truck type.
When to sample (which day of week and seasons to be selected to account for weekly and seasonal variations in truck traffic patterns)? Key parameters that help answer this question include the volume of truck traffic in the region by day of week and seasonal truck traffic volumes. These data can be typically collected from Weigh-in-Motion (WIM) sites and permanent traffic recorders.

There are no specific guidelines for arriving at sampling frames for the surveys, since each region has unique truck traffic characteristics in terms of total traffic volumes, types of truck, site characteristics, time-of-day truck traffic distributions, and weekly and seasonal traffic variations.

Costs

The cost of conducting roadside intercept surveys for a region depends on many factors, including the number of survey sites, time period of data collection, site preparation, costs of equipments such as cones, signs, etc., as well as the efficiency of the data collection firm, and the quantity and quality of data collection desired. Based on an analysis of previous roadside intercept studies, the average cost of conducting a 24-hour intercept survey is estimated to be around $5,000 per site. However, actual costs of data collection can vary significantly based on the characteristics of the sites, the quantity and quality of data collected, and the data collection firm employed for conducting the surveys.

Personnel Training and Other Operational Issues

Recruiting and training personnel to conduct interviews of truck drivers is a critical component in the design and implementation of a roadside intercept survey program. There are, however, many data collection firms specializing in roadside intercept surveys that can be hired to conduct roadside interviews, but this approach can be potentially more expensive. An alternate and less expensive approach is to recruit personnel from local organizations and/or volunteer groups (community service clubs), comprised of individuals with good knowledge of local roads, and understanding of general traffic patterns in the region. Typical components of personnel training for roadside intercept surveys include instruction in personal interviewing techniques, accurate identification of different truck and tractor-trailer combinations (along with number of axles), and procedures and requirements for ensuring personal and third-party safety at the survey site. Other operational elements to be considered for the survey include the provision of accessories such as clipboards and pens, as well as reflective safety vests, headlamps, and hats to survey personnel and equipment of each site with survey crew signs and traffic cones. Additionally, it is advisable to deploy a Commercial Vehicle Enforcement officer at the site to ensure safety of survey personnel, as well as effective direction of selected trucks to the survey site, in order to ensure a high degree of compliance, which leads to high response rates.

10.2.3 Establishment Surveys

Introduction

Surveying establishments engaged in freight activity is an important element of a local freight data collection effort for a region, since they generate a large fraction of local, and long-haul (internal-external and external-internal) freight flows. This data collection method involves surveying owners, operators, or fleet managers of key establishments, which may include manufacturing facilities, warehouses, retail distribution centers, truck terminals, and transload facilities. These surveys may include terminal gateway facilities like seaports, airports, and intermodal yards. However, the utility of establishment surveys for terminal gateways is generally limited to getting information on economic characteristics of the facility (such as number of employees), since the extent of truck traffic volumes and patterns associated with terminal facilities make terminal gateway intercept surveys more optimal compared to establishment surveys for collecting information on truck traffic characteristics. The use of business directories, such as Dun & Bradstreet, may be useful in identifying personnel contacts who can provide the required information.

The primary methods of conducting establishment surveys include telephone interviews, mail-out/mail-back surveys, and combined telephone and mail surveys. Establishment surveys can be used to collect comprehensive information regarding economic, land use, and modal freight (trucking, rail, etc.) activity characteristics of freight facilities, which may provide key inputs for freight modeling and planning applications. Specific data attributes that can be collected include facility hours of operation, number of employees, facility land area, fleet size, fleet ownership, types of trucks in fleet (straight, tractor-trailers), commodities handled, average payloads by commodity and type of truck, types and share of trucking services used (parcel, truckload, and LTL), average daily inbound and outbound truck shipments, average trip lengths, truck trip-chaining activity, truck O-D distribution patterns, types of facilities used, etc. In addition, establishment surveys also can be used to understand how key transportation performance variables such as transportation costs, travel times, reliability, highway regulations, and roadway closures impact shipment decisions.

Applications

Following are some key freight forecasting and planning applications of the data collected from establishment surveys.

Trip Generation

Data collected from establishment surveys on number of employees, land area, and average daily truck trip productions and attractions can be used to develop truck trip generation estimates. These data elements can serve as inputs to the two common approaches for trip generation, which include trip generation rates, and regression equations. Establishment surveys may be more feasible compared to collecting traffic counts for trip generation since daily trucking activity information for the facility can be collected at a fairly reasonable level of accuracy using limited resources, compared to conducting traffic counts that might prove to be more expensive. In addition to providing data for the estimation of trip generation rates and regression equations, establishment surveys can collect forecast economic data (future employment and labor productivity) for the facility, which are key inputs for facility freight forecasting and planning.

Truck Trip-Chaining Analysis

Establishment surveys generally provide better data for understanding truck trip-chaining activity compared to other types of data collection, such as terminal/facility gate surveys. Gate surveys of truck drivers are most effective when collecting only the last stop-next stop activity information, since drivers tend to get confused about questions related to multi-stop trip-chaining activity and may not provide reliable information. In the case of establishment surveys, however, the interviewee can provide information for each commodity group handled by the facility on the fraction of total truck trips performing multi-stop tours. This information, combined with the types and locations of facilities used by truck trips of each commodity group, is key to understanding and modeling truck trip-chaining activity associated with specific freight facilities. Establishment surveys of motor carrier terminal facilities are useful for understanding truck trip-chaining behavior based on the type of carrier (truckload, LTL, or private), and the type of commodity hauled.

Payload Factors

Establishment surveys offer a resource efficient and optimal approach for collecting payload data for truck shipments by commodity and type of truck. An alternative approach is through gate surveys, but they are not only more cost- and time-intensive to implement but only capture a sample of the truck shipments that can potentially lead to statistical bias in the estimates. A facility/fleet operator can provide more reliable information, with relatively lower data collection effort, on average payload factors by commodity and truck classes, using records/logs of daily truck shipment activity at the facility. Data collected from facility/fleet operators on average payload factors for different trip length categories (long- versus short-haul/local distribution), also can be used to understand the impacts of market area on payload factors for different commodity groups.

Other Applications

Other key applications of the data collected from establishment surveys include the following:

Time-of-day analysis to understand variations in trucking activity at a facility by time of day – This is useful for site/facility planning, to understand time-of-day interactions between trucks and automobiles, and to plan for the efficient movement of freight during peak periods.
Analysis of the types of facilities used by trucks generated by a facility for different commodity groups – This can be useful for developing trip distribution models (for example, truck traffic disaggregation models), as well as land use planning associated with large freight generators such as seaports. Establishment surveys of trucking terminals also can yield useful data on the types of facilities used by type of carrier (truckload, LTL, or private) to validate trip distribution patterns based on truck trips by carrier type.

Implementation Issues

Type of Data Collection

Deciding on the type of data collection (telephone, mail-out/mail-back, or combined telephone and mail) is a primary issue in the implementation of establishment surveys. Each of these methods has advantages and limitations associated with the type and volume of data collected, and the time and costs associated with the data collection effort. Generally, mail surveys have been the commonly used method for establishment surveys, particularly due to the relative ease of implementation compared to telephone or combined telephone and mail surveys. The investment costs and personnel requirements associated with mail surveys also are typically the lowest. However, mail-out/mail-back surveys have many limitations, most notable being their low response rates, as well as the inability to clarify responses to specific questions. Telephone surveys have relatively higher response rates compared to mail-out/mail-back surveys; however, they may be less effective in getting comprehensive trucking activity information, since identifying and reporting specific trip detail about all shipment types can be prohibitive in a telephone conversation. Telephone interviews also require the availability of accurate data on telephone numbers and interviewees (owners, operators, fleet managers, etc.), and compiling that data can be a time-consuming and costly undertaking. Combined telephone and mail surveys offer high response rates, since the establishments are notified beforehand through telephone contact about the mail survey. However, this survey approach typically has the highest cost of implementation. Table 10.1 presents the advantages and limitations associated with each type of data collection, pertaining to implementation, investment, statistical reliability, data attributes, and geographic coverage.

Table 10.1 Advantages and Limitations of Mail-Out/Mail-Back, Telephone, and Combined Telephone and Mail Surveys

Method	Advantages	Limitations
Mail-Out/Mail-Back Survey	Ease of implementation Low investment costs, and minimal personnel requirement Generally good information and data detail from survey respondents	Low response rates Limited ability to clarify responses to specific questions Difficulty finding the appropriate person at the establishment to complete the survey
Telephone Survey	Ease of implementation Quicker turnaround compared to mail-out/mail-back surveys Low investment costs Ability to clarify responses Better success rates for follow-up surveys compared to mail-out/mail-back surveys	Compiling phone numbers and contact person information can be difficult and time-consuming Surveys can only be conducted during normal business hours Higher personnel requirement compared to mail-out/mail-back surveys Inability to collect comprehensive trucking activity information during a telephone conversation
Combined Telephone and Mail Surveys	Quicker turnaround that mail-out/mail-back surveys Improved ability to clarify intent of data collection, and explain questions, leading to better detail and accuracy in collected data Relatively higher response rates compared to mail-out/mail-back surveys	Compiling phone numbers and contact person information can be difficult and time-consuming Higher personnel requirement compared to mail-out/mail-back surveys

Sample Selection

Sample selection is an important element in the design of an establishment survey data collection effort. The larger the sample size, the more reliable and comprehensive the data collected from the survey. However, it would be practically impossible and cost prohibitive to survey the universe of establishments located in a region. Thus, attention to developing appropriate sampling frames is critical not only for minimizing the overall cost of the data collection effort, but also for ensuring that the sample surveys provide unbiased and reliable data on the economic, land use, and freight activity characteristics of establishments in the region.

It is important to note that there is no definitive methodology for arriving at the sample size. In the case of establishment surveys, the primary factor impacting sample size is the method of data collection since each method is associated with different response rates. For example, in the case of a mail-out/mail-back survey, the generally low response rates would entail the selection of a larger sample size compared to a telephone interview survey with relatively higher response rates. Other factors which will impact the sample size are the costs of data collection (tied to the method used), as well as the reliability and accuracy of the available contact information.

In the case of establishment surveys of freight facilities such as manufacturing plants, warehouses, and distribution centers, the usual sampling approach involves selecting establishments based on their employment size or land area. Standard privately owned data sources such as Dunn & Bradstreet are available for purchase and provide the universe listing of establishments in a region for sampling, along with their economic (number of employees, etc.) and land use (floor acreage, etc.) characteristics. Additionally, there may be publicly available data, compiled by state economic development departments, MPOs, or other organizations (such as port authorities) on major freight establishments in a region, which can be used to develop sample sizes.

In the case of establishment surveys of trucking terminals, the sampling strategy would typically depend on the trucking characteristics in the region. For example, the predominance of local distribution activity in large metropolitan areas (captured in the VIUS database) would imply that the sampling approach should focus on capturing a larger fraction of motor carrier establishments involved in short-haul local distribution activity compared to long-haul trucking, to perform a statistically reliable and unbiased analysis of trucking activity in a metropolitan area.

Costs

The costs involved in conducting establishment surveys vary depending on the method of data collection. Based on historical information available on establishment surveys, the average costs of conducting mail-out/mail-back (with a 10 percent response rate) and telephone (with a 20 percent response rate) surveys are estimated to be $100 per survey and $250 per survey, respectively. Historical cost information on combined telephone and mail-out/mail-back surveys is not available since this survey method is not very common due to the relatively higher level of effort involved in data collection. The cost of conducting combined telephone and mail surveys is expected to be higher than telephone surveys.

10.2.4 Travel Diary Surveys

Introduction

Travel diary surveys are a useful method of data collection, particularly for understanding internal-internal (local) truck trip activity in an urban area. The basic approach of data collection involves selecting a representative sample of trucks operating in the region, and obtaining travel diaries from truck drivers for a certain time duration. The usual time period for data collection is 24 hours; however, depending on the willingness of truck drivers to complete trip diaries, the surveys can be conducted for time periods extending more than a day (typically, three days or a week). The most common approach to providing travel diaries is through forms completed manually by the driver listing the truck trip characteristics for the time period of the survey. Typically, drivers are asked to record information on the truck trip regarding origin, destination, trip mileage, routing, travel time, trip time of day, commodity-hauled and size of shipment, truck type, and land use and activity (pickup, delivery, refueling, rest area, etc.) at trip end. Additionally, they may be asked to report their type of carrier operation (truckload, LTL, or private), if this information cannot be deduced from the source data.

An alternative and more advanced approach of travel diary surveys is the use of Geographic Positioning Systems (GPS) receivers, which are fit in trucks to trace individual truck trip activity. However, GPS-based data collection in itself cannot provide key truck trip characteristics pertaining to commodity hauled, shipment size, and activity at trip end. The maximum utility of GPS-based data collection for a travel diary survey is realized when combined with other data sources and methods of data collection. For example, origin, destination, and routing information received from GPS receivers can be used to validate and improve the information provided by truck drivers in manually completed travel diaries. Also, combining GPS truck trip information with GIS-based land-use data, for example, can yield useful information on truck activity characteristics at trip ends.

Applications

Some key freight forecasting and planning applications of the data collected from travel diary surveys are listed below.

Trip Chaining

As discussed earlier, travel diary surveys are particularly useful for understanding internal-internal truck trip activity in a region and perhaps the most important application in this regard is truck trip-chaining analysis to develop more robust and accurate urban truck travel demand models. Travel diaries capture the entire trip making activity of each individual truck over a 24-hour period, which can be used to trace the occurrence of trip chaining. For example, a trip diary entry for a trip starting from home base to a pickup location, proceeding to a drop-off location, and then proceeding to another drop-off location indicates the presence of trip chaining (such trips are common in urban areas, particularly local distribution trips related to retail activity). Trip-chaining activity from travel diaries, coupled with information on type of commodity, type of carrier, and land use and activity at trip ends, can be used to understand trip-chain distribution patterns, and as inputs to develop activity-based truck travel demand models.

Trip Generation

A key application of travel diary surveys is in the development of trip generation estimates. Travel diaries provide data for the sampled trucks on total trip ends by land use, which after expanding to account for the universe of truck trips, can be used to develop trip generation rates or regression models for trip generation. Trip generation rates are derived by dividing the total trip ends for each land use category by the independent variable impacting truck travel demand (for example, economic/land use data such as employment/acreage). Similarly, trip ends for each land use category can be used to develop regression models. If there are sufficient data points spread across the region for trip ends, it can also be used to develop a statistically reliable model (for example, if most of the trips associated with a land use are concentrated at a couple of locations in the region, then there are only two data points for the regression model, which would impact the statistical validity of the model). Some important considerations affecting the accuracy of trip generation estimates derived from travel diary surveys are the source data used to develop the sample of trucking companies, as well as the trucking activity characteristics of the region. For example, if only those trucks registered in the region participate in the survey, while there are a large fraction of out-of-region registered trucks operating in the area, then the trip generation analysis will underpredict the total truck trips generated in the region.

Traffic Routing

Travel diaries record the routes taken by trucks for each truck trip between O-D pairs, which can be used to understand truck traffic routing patterns in the region for the validation of traffic assignment procedures. GPS-based travel diaries provide accurate and real-time truck routing information, which serve as critical inputs for the analysis of routing pattern variations by time of day. For example, how congestion during peak hours might impact truck routing patterns during the day, compared to the nighttime.

Implementation Issues

Sampling Frames

Selection of appropriate sampling frames is an important element in the design of travel diary surveys. Vehicle registration databases are commonly used data sources for developing sampling frames that contain the listing of all the trucks registered in a region. These databases are typically maintained by each state’s Department of Motor Vehicles (DMV). The approach used to sample the population plays a critical role in determining the utility of the data gathered for planning and modeling applications. For example, in order for the survey to provide data to better understand truck trip-chaining activity in the region, the sampling approach should consider selecting a larger fraction of trucks primarily performing short-haul local distribution activity, compared to long-haul shipments. Thus, random or systematic sampling techniques are generally not optimal for selecting sampling frames for travel diary surveys because the sample tends to have the same distribution of trucks as in the population. Stratified sampling is the best approach, which involves stratifying trucks in the population and selecting samples from each stratum to develop the sampling frame. Vehicle registration databases may provide average trip length information for each individual truck record, which can be used as a parameter to stratify trucks based on short- and long-haul trucking activity. The sampling frame is developed by selecting a larger fraction of trucks performing short-haul trucking activity. The sampling fractions, depending on the desired sample size for the survey Annual VMT information for each truck record might be another potential parameter used for stratified sampling. However, annual VMT is not a very good indicator of short- versus long-haul trucking activity.

Costs

Cost is a major implementation issue only in the case of GPS-based travel diaries, owing to the high equipment costs associated with GPS receivers, and the costs of installation on trucks. However, limited data are available on the costs of conducting GPS-based travel diary surveys because of the relatively fewer applications of this survey methodology.

Data Limitations

Some key limitations associated with data collected from travel diary surveys include the following:

Sampling process can be difficult, especially in cases where there is lack of good information on points of contact and their addresses and telephone numbers for trucks operating in the region.
The use of vehicle registration databases for the surveys can produce biased results in cases where there is a significant fraction of trucking activity associated with trucks not registered within the region. In this case, the travel diary surveys also will potentially underpredict the total trucking activity in the region.
One of the biggest problems associated with travel diary surveys is low response rates. Truck owners, in many cases, are not willing to participate in the survey due to confidentiality issues pertaining to sharing travel and customer information, as well as interruptions caused by the survey to drivers’ normal work schedule.
Travel diary surveys using GPS receivers are relatively more expensive to implement. There also is the potential for equipment failure in these surveys.

Previous Section | Next Section | Top