CHAPTER 15.0 VEHICLE INTERCEPT AND EXTERNAL STATION SURVEYS

        15.3.1.2 Selection of the Survey Site Within a Chosen Roadway Segment

In cases where a tape recorder is used, a microphone which reduces background noise (directionality, and range) should be chosen.

Some surveyors have tried to eliminate the need for office data entry by having fieldworkers enter the data directly into portable personal computers. The technique can reduce survey costs, but it has the same disadvantage of most computer-assisted survey techniques – there are no raw source documents. The only record the survey team has is the data file itself, and perhaps a printed record of the file once the data are brought to the office. With the other techniques, there are handwritten, audio, or video records that can be referenced, if necessary. It should also be noted that lack of familiarity with the portable computers may cause slower data entry relative to written or recorded data entry. Small inefficiencies can be a problem at high volume locations.

An increasingly popular observation technique is the video data collection approach. In this method, a video camera is aimed at one or more traffic lanes and is used to simply tape all vehicles going by. A fieldworker usually monitors the video taping, but is not needed for any active data collection. The technique can be highly effective for multi-lane high speed traffic locations, because fieldworkers recording information at these sites can become overwhelmed.

The technique is usually cost-effective for sites that would require more than one fieldworker to observe license plates directly, because a single person can monitor more than one video recorder. Videotaping has the following additional advantages (Baker, Deardorf and Cathy Strombom, 1995): 

It can be used at sites where it would be impossible to station fieldworkers for extended periods;

· It is more accurate than the other methods if taping is carefully monitored;

· It can be used to provide additional data about the survey site, including the classification of vehicles and vehicle occupancy; and

· The use of videotape data collection is expanding in a number of fields, and so it is possible to locate specialist companies that are quite experienced with the required procedures.

The technique does have some problems, however, such as:

· It requires renting high quality video recorders (typically, top-of-the-line equipment is needed to capture the license plates of vehicles traveling at 60 miles per hour or more);

· Transcribing license plate information from the videotapes can be a tedious process (survey office workers need to fast-forward the tape to just the right moment, record the license plate, and then repeat the process again and again); and[1]

· The survey site is less flexible, and harder to adapt if special unforeseen circumstances arise (such as an unexpected lane closure or a rain shower).

Some researchers are now using automated collection and transcription of license plate data from videotapes.  This development will not only reduce the cost of the survey since office screening and transcribing of license numbers would not be needed, but potentially would increase the accuracy of the data and provide a larger sample (i.e., more license plate “matches”).

15.3.3.2 Data Collection Techniques for Roadside Interviews 

Roadside interviews can be performed using conventional paper and pencil interviewing (PAPI) procedures or by using computer-assisted personal interviewing (CAPI) procedures. A number of recent studies have taken advantage of CAPI techniques for roadside surveys, including efforts in Florida, Georgia, Texas, New Hampshire and in Canada since 1995 for a national commercial vehicle survey program. Muntean lists the following advantages of CAPI for roadside interviews (Muntean, 1995):

1. Real-Time data entry and coding

- Creates a time/date stamp tied to the survey record with site direction required for hourly data expansion 

- Direct data entry to data file

- Creates an electronic record of surveyor productivity with a benchmark for the site team

- Trained interviewers can actually complete the survey faster than with PAPI methods; improves consistency in coding

2. On-Site Editing

-  Instantaneous validity check

- Mapping utilities are commonly used to verify routing and gain immediate driver input at the time of survey to visually validate the surveyor interpretation of the survey site, direction of travel for this trip, confirm definition of a trip, assign practical route, verify trip start time and estimated arrival time at destination, highways used and intermediate stops.  These systems can be preloaded with zone systems and generate travel miles (VMT) in each traffic zone and confirm travel times    

- Sip routine

- Interviewer and supervisor can review record upon completion of interview, supervisor input is critical in the early stages of the survey to detect surveyor issues and standardize data collection methods of the team 

3. Automatic Geocoding

- Computer provides interviewers with place name look-up tables, when these systems have encrypted vehicle registration files pre-loaded, but with minimal information, access to home base trip information can be verified and speed data collection and improve data integrity 

- Interviewer can scroll through list or type first several letters to find the place name

- Database can record zone number, as well as place name

These advantages have been realized in other studies, as well. As described in other chapters of this manual, CAPI also provides other advantages. However, because roadside interviews are by necessity relatively short, survey teams often do not require advanced capabilities.

The primary disadvantage of the CAPI technique is the expense of obtaining usable laptop or palmtop computers, or personal digital assistants (PDAs). Although the unit costs of these machines have dropped precipitously, most survey efforts require the purchase or rental of several machines. The computers used in roadside interviews need to (Anderson, 1994): 

Be small and lightweight;

· Have sufficient battery life for daily survey needs;

· Have a sunlight readable display;

· Have means of easily transferring data to office computers;

· Have time-of-day stamping capabilities;

· Have sufficient memory for the survey program; and

· Have sufficient storage space for a day or more of surveys.

15.3.4 Survey Station Design

15.3.4.1 The Design of Roadside Survey Stations 

The final element of the survey design that is critically important is the design of the survey station. Roadside handout and interview surveys require survey fieldworkers to work in and around moving traffic, and, because most respondents will have never been asked to participate in such a survey effort, they are not likely to fully understand what they are supposed to do. Therefore, it is essential that the survey team carefully design the survey station to be as safe and clearly understood as possible.

The survey station design is affected by:

•  The survey method   license plate, roadside handout, roadside interview, combined method (in particular, the amount of time vehicles need to be stopped)

•  The volume of traffic on the highway segment;

• The roadway geometry (in particular, the roadway and shoulder width);

•  The availability of trained fieldworkers; and

• The percentage of vehicles passing the station that are needed for the survey.

The survey team needs to work closely with highway engineers and local law enforcement agencies to develop station designs and survey procedures that are acceptable in terms of safety and the potential effect on traffic flow.

Figures 15.2 and 15.3 show some example survey station designs, but the survey team will probably want to customize their sites to match local conditions and specific survey needs. For instance, if the roadway traffic volume will support it, many survey teams use several interviewers or questionnaire distributors. Instead of intercepting a single vehicle at a time, an entire platoon of vehicles is intercepted together. In addition, some sites will allow for vehicles traveling in the survey direction to pass the survey station in a second lane while interviews are taking place, while other sites simply will not be wide enough.

Figure 15.2 Sample Vehicle Intercept Mailback and Interview Traffic Control Plans from Vermont

Figure 15.3 Sample Vehicle Intercept Survey Site Design for Interstate Highways

At a minimum, all roadside interview and roadside handout stations require the following:

1. Extensive signage, warning motorists of the survey (some survey efforts have used variable message signs);

2. A large number of traffic cones for channelling traffic;

3. Safety equipment for fieldworkers (orange vests, hard hats, etc.);

4. Supervisors and flaggers who have experience working in traffic; and

5. A police detail.

Even though police officers are not always required at the site, having officers on hand is usually worth the additional expense. Generally, police departments provide off-duty officers working overtime for efforts such as vehicle intercept surveys. The officers often have police cars available to them which can help attract motorists’ attention to the fieldworkers on the roadway. In some cases, the police officer at the survey site plays an active role in directing vehicles into the survey station. In other cases, the officer simply oversees the safety procedures at the site. The police presence reduces the risk of accidents or incidents, because the officers are trained in safety procedures. In addition, the officers are much better able to deal with irate respondents than most survey crews.

15.3.4.2 The Design of License Plate Survey Stations

The survey station for license plate surveys is much less complicated. The survey team simply needs to locate camera positions which can record license plates from all lanes of passing traffic and that allow fieldworkers to monitor the cameras. The stations are commonly established on highway overpasses. Fieldworkers for license plate surveys will also be near moving traffic as they work, so it is important that personal safety precautions similar to those used on roadside surveys be used.

Figure 15.4 shows an example survey station design for a videotape license plate survey. Like the roadside survey station designs, the license plate survey station designs need to be customized to the particular survey sites.

As discussed below, for both roadside surveys and license plate surveys, it is usually advantageous (or necessary) to conduct vehicle counts at or near the survey station. Usually tube counts are collected a few hundred feet beyond the survey station, and manual classification counts are made throughout the day. The count information is used for expansion of the data.

The organization of the vehicle intercept survey includes the following elements:

· Management of the survey effort;

· Staffing analysis;

· Hiring methods;

· Coordination with other agencies; and

· Advance publicity.

Figure 15.4 Example Vehicle Intercept Survey Design for a License Plate Survey

15.4 Organizing the Vehicle Intercept Survey

15.4.1 Management of the Survey Effort

Chapter 4.0 summarizes the three primary management tasks of travel surveys:

· Overall management and leadership of the effort;

· Day-to-day management of survey fieldwork; and

· Analysis of the survey effort from the data user perspective.

Vehicle intercept surveys present unique management challenges for each task, but perhaps the most important aspect of these surveys from a management perspective is the need to effectively adapt general survey procedures to specific survey sites, each of which is different from the others. Often survey efforts rely on more than one data collection method or technique, and even when the same general procedures are used, they must still be adapted to each site.

High quality field supervision is a key to the success of vehicle intercept surveys. The survey team should ensure that the supervisors understand the goals of the survey effort, and that they are equipped to make on-site decisions about how to solve unforeseen problems without endangering the overall data quality.

15.4.2 Staffing Analysis

The fieldworker staffing needs for vehicle intercept surveys are related to the survey method and the site being surveyed. The staffing analysis should consider attributes of each selected roadway, such as the number of travel lanes, average daily, morning, and afternoon peak-period traffic flows, time periods to be analyzed, and logical and safe survey operation locations. The basic steps presented below are recommended:

15.4.2.1 License Plate Survey 

1. Identify the number of roadways, lanes, and the survey times. This information will establish the coverage of roadways to be surveyed. It will also identify specific videotape crew, license identification crew, and camera specifications per survey location.

2. Identify the estimated daily peak-traffic volumes using the roadways to be surveyed. Much of this information may already be available from existing traffic counts or volume estimates. Based on these volumes, define the analysis needs of the survey in terms of sampling (see Section 15.5) and travel demand modeling. This information, in conjunction with the various time periods identified in Step 1, should provide the appropriate level of survey coverage to represent typical peak and daily roadway travel on the selected roadways.

3. Identify the estimated number of license plates to be observed by time period and roadway lane. This information will be required to identify the data collection periods. In most cases, observation can be conducted in hourly segments to facilitate license plate identification and transfer of information for merging with DMV databases.

4. Estimate license plate transcription and data entry rates to determine office staff requirements.

15.4.2.2 Roadside Handout and Interview Surveys 

1. Identify the number of roadways and survey times. This information will establish the extent of coverage of the roadways to be surveyed.

2. Analyze the traffic control plan developed during the survey design task for each roadway to be surveyed. This plan includes the setup and control of the roadway survey station. It should include a given number of surveyors, roadway flaggers (for each direction) to slow traffic, a given number of relief surveyors, and one or more supervisors. Each station should have a law enforcement officer on duty to ensure safety and motorist cooperation. This plan also provides information on the effort required for site setup, including coning roadway travel lanes and stationing of advance warning signs.

3. Identify the estimated daily and peak-traffic volumes using the roadways to be surveyed. Based on these volumes, the analysis needs of the survey are defined related to sampling and travel demand modeling.

4. Use the traffic control plan and the traffic volume information to estimate site productivity rates.

15.4.3 Hiring Methods

Based on the staffing analysis, the next task will be to identify the methods for hiring surveyors. The hiring options for the vehicle intercept surveys include the following:

1. Contract with a local or, in the case of the License Plate Survey, a specialized data collection firm to conduct the survey. The data collection firm will be responsible for hiring surveyors, and conducting, administering, and supervising the survey in the field.

2. Contract with a local or specialized data collection firm to conduct the survey with supervision provided by the project sponsor. The data collection firm will be responsible for hiring surveyors and conducting the survey in the field while the project sponsor will be responsible for overall management and supervising the survey in the field.

3. The project sponsor will be responsible for recruiting and hiring surveyors, and conducting, administering, and supervising the survey in the field

In most cases, the license plate survey will be conducted by private consulting firms specializing in this type of data collection. If consultant assistance is obtained for the roadside surveys, most survey teams usually hire traffic engineering firms, as opposed to survey research firms. This is because of the need for experienced traffic personnel. The engineering firms either use in-house staff or hire temporary employees for the work.

If the sponsoring agency chooses to manage the effort themselves, they should schedule interviews with prospective surveyors from various organizations, including college and university employment agencies, planning, and engineering departments; state and local government employment agencies; and private temporary employment organizations. Based on these interviews, the sponsor(s) will select the best qualified personnel.

In most cases, hiring surveyors familiar with the region to be surveyed is important. It may also prove useful for the project sponsor(s) to be involved in the initial administration and supervision of the survey (possibly during the pretest stage) effort to ensure that the surveyors (or local contractor) are conducting the survey properly.  In some cases, survey firms have offered financial incentives to temporary survey crew employees to increase retention by increasing the pay for each subsequent week on the job.  Loss of surveyors not only requires additional resources to train replacements, but delays the schedule as well.

15.4.4 Coordination

Coordination and communication between the various participating agencies should be established early on in the development and implementation of the vehicle intercept and external station survey. In many vehicle intercept surveys, State Departments of Transportation (State DOTs), Departments of Motor Vehicles (DMVs) from the survey state and perhaps from adjoining states, state and local Highway Patrols, and local Departments of Transportation and Public Works may potentially be involved in the survey effort.

The survey team needs to identify the special concerns of the different agencies early in the survey design. In some cases, agencies, such as State DOTs, require the use of standardized and accepted Traffic Control Plans to be used for vehicle intercept surveying. State DOTs may not allow vehicles to be stopped for surveying. There are also legal implications to stopping vehicles for surveys, including insurance liability issues for data collection firms. This initial coordination between the surveyor and participating agencies is critical to the successful implementation of the survey, and typically helps to identify the acceptable and appropriate survey method early on in the effort.

15.4.5 Advance Publicity

For vehicle intercept surveys, it is highly desirable to publicize the overall effort well in advance of the fieldwork dates. Potential respondents are more likely to believe that the survey effort is legitimate and important if they have heard that the survey will be taking place. This is especially true for the license plate method, because respondents will receive the survey instrument without any notification. The publicity effort is the only effort to alert potential respondents of the effort prior to the actual survey.

However, the survey team should exercise care in providing publicity about the specific sites to be surveyed. If motorists know that roadside surveys are being conducted at a particular location on a certain day, they may change their travel patterns to avoid (or to include) that location. In this case, the usefulness of the survey data is jeopardized, particularly if the collection of origin-destination data is the primary survey objective.

15.5 Sampling

As noted above, the vehicle intercept and external station survey should be developed and conducted simultaneously with either an existing or proposed Traffic Count Program. The purpose of the Traffic Count Program is to identify the level of traffic volumes expected at each vehicle intercept survey location. These traffic volumes can help to define the sample size requirements of the total respondents required for given survey time period.

Each survey method – the License Plate Survey, Roadside Handout Survey, and Roadside Interview Survey – use the same sampling strategy. The typical sampling procedures for selecting sample time periods and vehicles are described in the following sections. A general description of sampling methods and statistical analysis appears in Chapter 5.0.

15.5.1 Selecting Time Periods and Days

The Staffing Analysis of the organizational plan describes the staffing requirements based on specific roadway location, time period, and number of days identified for surveying. Depending on the scale and scope of surveying, the following guidelines should be followed for selecting survey time periods and days:

1. Identify the schedule for surveying. This schedule should describe the start and end dates proposed for the survey. For example, the survey may be scheduled for one month, roughly starting April 1 and ending May 1. In most cases, vehicle intercept surveys are scheduled in the Spring (March through May) and Fall (September through November – before the Thanksgiving Holiday) of any given year.

2. Within the defined schedule, identify the number of potential roadway locations and weekdays for surveying. Vehicle intercept surveys are typically conducted on Tuesdays, Wednesdays, or Thursdays during the week to gather representative samples of weekday traffic.

3. Identify the time periods for analysis. Generally, a 12-hour time period, from 6:00 a.m. to 6:00 p.m., is selected for passenger, but the selected period should be based on the daylight hours. A 12-hour period will provide representative samples of weekday daily, and morning and afternoon peak-period traffic volume flows.  Commercial vehicle surveys are often conducted on a continuous 24*7 basis where possible, and repeated over multiple seasons to sample the variation in the commodity mix. 

At this point, a detailed schedule for the number of days and time periods for surveying should be identified. Adjustments to this sampling can be made to account for weekend surveying, peak-period surveying, and time extensions (from the spring to fall) because of large scale surveying.

15.5.2 Selecting Vehicles

The selection of the number of vehicles to be sampled is dependent on the number of vehicles using the particular roadway under study. The development of this survey sample should be made based on the most recent traffic volume information available. Typically, vehicle intercept surveys are conducted for 12-hour periods and include the following sampling procedures:

1. Identify the daily two-way directional volumes for each of the selected roadway locations to be surveyed. In some cases, previously collected traffic data can be used to estimate daily traffic volumes. Optimally, newly collected traffic volumes collected from the Traffic Count Program should be used.

2. Identify the proposed sample rate for each of the selected roadway locations to be surveyed. This proposed sample rate is commonly defined to achieve an accuracy of 15 percent (error) at a confidence of 95 percent for a 10 percent proportion of the surveyed roadway’s total traffic having a particular origin and destination. This sample rate equation will provide the expected number of driver responses required at each roadway location.

Table 15.5 shows the sample size estimation for a multi-site license plate survey recently performed by the Metropolitan Washington Council of Governments.

During the conduct of the survey, a detailed account should be identified of the expected survey response rate and predicted number of completed questionnaires required for each roadway location surveyed. Adjustments to this sampling can be made to account for low response rates by specific location. For example, roadway locations can be surveyed again to obtain the proposed sample rates.

While the actual response rates, of course, are unknown until the survey is completed, the estimated response rates are critical information in the determination of sample size. The expected response rate will depend on the survey method, the type of roadway being surveyed, and the characteristics of the travelers on the roadway. Surveys which require a mailback response will, naturally, have much lower response rates than in-person interviews.

If license plate matching is used, there is an additional factor in the response rate estimation: the plate match rate. The match rate is the percentage of observed license plates which are recorded and matched to a vehicle in the motor vehicle department file. There are two main reasons why some vehicles will not produce matches: 

· The recorded plate is not from a state to which the data file will be sent for matching (usually it will be efficient to exchange data with from only one to three states; it is seldom worthwhile to exchange data for states with only a handful of plates); and

· The plate cannot be read or is mis-recorded.

Table 15.5 Selected Survey Station Sites and Estimated Survey Samples

There are several reasons why license plates are mis-recorded or unread. These include:

· Lack of light, or glare from sunlight;

· Obstructions, such as tow bars, bicycles, and dirt;

· Bad viewing angles, especially for vehicles which are changing lanes or have overhangs above the license plates; and

· Simple recording or keypunching errors.

Although a greater percentage of license plates may be read using video technology, the match rate may be even lower than for plates that are manually recorded. This is because the camera is in a fixed position and cannot “look around” obstructions or at plates that are outside the field of view. Automated plate matching programs can also have high error rates as similar characters (e.g. “D” and “O” “B” and “8”) may be mistaken for one another, and plates that appear visible to the naked eye on a video display are too dirty or unlighted for the program to match. Still the potential savings in labor, both for license plate transcription and data entry, makes the use of video technology attractive and can more than offset the costs of having to collect a greater sample.

Table 15.6 shows response rates for a vehicle intercept survey in New Hampshire which used three different survey methods. In this survey, the response rate was 17% for the postcard handout/mailback survey and 21% for the license plate recording/mailout/mailback survey. With the match rate at 73%, this survey had an equivalent response rate of 15%. While the response for these two methods turned out to be similar, it should be noted that the license plate method was used only on two expressways on the Massachusetts border in the most heavily developed part of New Hampshire. The license plates were recorded from overpasses, which did not exist in most of the other locations. The same response and match rates might not have been achieved if the license survey were performed at other locations.

Table 15.6 Response Rates from A New Hampshire Vehicle Intercept Survey

15.6 Drafting and Constructing Vehicle Intercept Survey Instruments

 Like the other surveys described in this manual, the wording of questions must be carefully considered and consistent with other types of travel surveys recently performed or currently underway. For example, consistent questions about trip-making should be determined for household travel surveys and vehicle intercept and external station surveys if they are to be used together for model development. This ensures consistent information will be collected, especially regarding travel behavior characteristics of persons within a region. In addition, information collected in the vehicle intercept surveys should also maintain consistency with the current Census data specifications. The household, trip, and person information typically collected in the intercept survey should maintain the same categories as specified in the latest Census including breakdowns of income levels, occupation codes, and ethnic status.

Survey instruments and questionnaire designs will be different depending on the survey method selected. General survey instrument and questionnaire guidelines related to the types of questions, survey instrument considerations, and timeframes are provided below.

15.6.1 Data Elements

Table 15.7 shows the data elements typically collected in vehicle intercept surveys. Questions should be structured to obtain information about travel and trip-making behavior, households, and individuals. Travel and trip-making behavior questions should include origin and destination locations, vehicle type, trip purpose, and auto occupancy information. Questions about households should include number of available automobiles, household income levels, and number of persons per household. Demographic questions about individuals should include sex, occupation, and employment status information. Vehicle intercept surveys are also commonly used to collect attitudinal and stated-response information from motorists.

Because vehicle intercept instruments are generally very short, the number of data elements collected by these surveys is limited. Many vehicle intercept surveys collect only the barest essentials about the auto trip. Therefore, the range of data elements covered in these surveys is much smaller than for the household travel and activity surveys. The survey team should define essential data elements and other desirable elements that would be included if space or time permit.

Table 15.7 Vehicle Intercept Survey: Typical Data Elements Collected

15.6.2 Translating Data Elements Into Questions and Response Categories

As discussed in Chapter 6.0, a survey question should be included in a survey interview script or questionnaire if:

· The information obtained from the question is relevant to the models being developed or refined, or to other anticipated analytical efforts.

· The question and response categories are expected to be valid measures of the modeling variables.

· The responses can be coded meaningfully for modeling analyses.

· Analysts, interviewers (if any), and respondents agree unambiguously on the meaning of the question and response categories.

· The question and response categories have no wording problems. The wording of questions and responses is the same or equivalent to any measure from other surveys that will be used in the modeling work.

· Response categories exhaust all meaningful answers that can be anticipated.

· Response categories are meaningful and understandable to respondents.

· (For interviews) the questions and response categories are easily learned by interviewers.

· The benefits in the survey analysis from the question outweigh its costs in terms of survey length, respondent burden, and increased non-response.

· The information gained from the question is more useful than the information that would be gained from other questions that will not be on the survey.

· The question does not provoke respondents to be hostile to the survey effort or to question the goals of the surveying agency.

Chapters 2.0 and 6.0 provide detailed guidance on drafting survey questions.

15.6.3 Survey Instrument Considerations

For the License Plate Survey, the questionnaire layout parameters should be identified and implemented based on the types of questions selected, appropriate card stock, page layout formats, and return of address envelopes with postage. This layout should include a brief project description, a request for respondent participation, and a logical organization of the travel behavior, household, and demographic questions. For example, the “to trip” (inbound, to work) and “from trip” (outbound, from work) travel behavior questions should be separated and asked at the start of the questionnaire. Household and demographic characteristics questions should follow the travel behavior questions. The layout should also consider double-sided questionnaires to reduce printing requirements and costs. In addition, an upper limit on the number of questions should be identified to increase the survey response rate. A cover letter, signed by the project sponsor(s) and/or elected officials should also be provided along with detailed instructions for filling out the questionnaire.

Additional instrument considerations must be addressed in the License Plate Survey design to account for unreadable license plates on the videotape, license plate and DMV mismatches, and coding errors when transcribing license plates from the videotape. In addition, the DMV will not be able to identify out-of-state vehicle license plates that could be videotaped on a surveyed roadway. The surveyors must identify the appropriate strategies to account for these potential problems to maintain the established sampling strategy and response rates.

The Roadside Handout Survey can follow the same format as the License Plate Survey. Often, handout surveys include fewer questions, and utilize a postcard style format. This questionnaire form may be limited to questions specifically related to travel and trip-making behavior.

Many survey teams feel that it is imperative to keep the survey instruments as short as possible either because longer instruments require additional expense (in terms of higher postage, need for envelopes, etc.) or because of the concern that response rates will drop precipitously if the questionnaires are too long. Unfortunately, this policy is sometimes enforced to the point where analyses of the data are harmed, because key data elements are left out. The survey team should indeed keep the questionnaires short, but not at the expense of the analysis.

The appropriate card stock, page layout formats, and multi-language formats should also be developed. Questionnaires for this type of survey tend to be multi-colored card stock the size of a typical postcard, or slightly larger. In most cases, the questions are provided on one side of the questionnaire while the return of address and postage are provided on the other side. Figures 15.5 and 15.6 show two recent roadside handout survey forms.

The Roadside Interview Survey requires a similar, but slightly different, survey instrument as the Roadside Handout Survey because an interview between the surveyor and respondent is conducted. The questionnaire should include a questionnaire formatted on a single-sided, 8 1/2 by 11 inch paper. This format allows for quick interviewer (surveyor) tallying of driver responses and should be limited to the same types of travel and trip-making behavior questions asked for the Roadside Handout Survey. It is imperative to limit the interview time to one to two  minutes to facilitate the conduct of the survey. Each interviewer will be given a clipboard and several questionnaires to conduct the survey. Figures 15.7, 15.8 and 15.9 show examples of roadside interview recording forms. Figure 15.10 shows a typical interviewer script used in a PAPI vehicle intercept survey. A computerized questionnaire would be similar, but would include error checking, look-up tables, etc.

Figure 15.5 Sample Vehicle Intercept Mailback Survey Form from Vermont

Figure 15.6 Example Vehicle Intercept Mailback Questionnaire from Upstate New York

Figure 15.7 Sample Vehicle Intercept Interview Form from Upstate New York

Figure 15.8 Sample Vehicle Intercept Interview Form from Northern Colorado

Figure 15.9 Sample Vehicle Intercept Interview Form from Colorado

Figure 15.10 Sample Vehicle Intercept Interviewer Script

15.7 Pretesting

The pretesting requirements of the vehicle intercept and external station surveys will be dependent on the survey method selected. The survey logistics, procedures, and instruments should be thoroughly pretested before the full survey is implemented. Depending on the survey method chosen, the number of roadways (and in the case of the License Plate Survey, the number of lanes per roadway) should be identified to test the survey instrument, surveyor organization, and survey response rates under different conditions. These conditions may include surveying for peak-periods and daily conditions on different roadways to identify any problems that may arise through the conduct of the survey.

Fully implemented survey procedures should be pretested on at least two roadways with different characteristics, such as a freeways, highways, and arterials. For example, the entire survey method should be tested for an all-day period, generally from 6:00 a.m. to 6:00 p.m. The pretest staff should include the designers of the survey procedures and forms, survey supervisors, and key surveyors hired to provide support in the fully implemented survey. The surveyors conducting the pretest will use the same procedures and techniques identified for use during the survey training sessions.

The pretest requirements for the License Plate Survey will be more rigorous than the other survey methods. This pretest should include filming one lane in each direction for a selected multi-lane-divided highway and a two-lane state route. In this case, the fully implemented survey procedures should be followed from videotaping, license plate identification and processing, transfer of license plate information to the local DMV, transfer of address information to the project sponsor(s), mailout and collection of returned questionnaires, and questionnaire processing.

Pretesting for the Roadside Handout and Interview Surveys will be very similar to one another regarding the number and time periods of locations selected for pretesting and should be the same as conducted for the License Plate Survey pretest (6:00 a.m. to 6:00 p.m. for two roadways). However, slightly different pretesting needs are required for the Roadside Interview Survey to refine surveyor interviewing techniques. The pretest should be used as an opportunity to help surveyors refine their interview style and interview times before the fully implemented survey begins.

Because of the high cost of setting up a survey station, pretests are usually conducted at one of the sample locations. If the survey effort works well, the pretest data are used. To do this, the survey team should schedule one or two survey locations several weeks ahead of the rest. The results of the early survey can be evaluated, and procedures may be modified, as needed, before the general data collection occurs.

After completing the pretest, a debriefing meeting should be held to identify any discrepancies in the survey procedures. The survey control sheets (which vary by the selected survey method) should be rigorously examined after the pretest for discrepancies and uncontrolled information. The information contained in the returned questionnaires should also be processed immediately to identify any potential problems associated with respondents filling out the survey properly. Frequency distributions should be conducted on each question to determine the validity of the survey instrument.

15.8 Training and Briefing

Training methods will vary according to the survey method chosen. License Plate Survey training may require license plate identification and geocoding while the two roadside survey methods will require additional training. In general, survey staff training should consist of a project briefing and demonstration of the surveyor responsibilities. In addition, surveyors should be provided with the necessary materials and procedural notes to conduct the survey. In cases in which local data collection firms are hired, it proves highly useful for the project sponsor(s) to train survey administration and surveyor staff. Typical survey training procedures include:

1. Project Briefing – The project sponsor(s) should hold a project briefing for all hired survey staff (or local contractor survey staff). This meeting should be used to describe the background and purpose of the survey, and the administrative procedures to be followed during the course of the survey. At this point, all surveyors, including the “break” or replacement surveyors, should be given their survey assignments and work schedules.

2. Survey Demonstrations – The surveyors should be given individual demonstrations on the procedural conduct of the surveys. This includes describing the tasks of the surveyor regarding responsibilities for distributing (handing out) survey materials to drivers and/or collecting travel behavior information (interviewing or license plate data collection). The traffic control setup and breakdown requirements for each of the roadside surveys will also be demonstrated during this step.

3. Survey Materials – Each surveyor should be given the appropriate survey materials and supplies at the start of his/her survey shift. The materials provided are dependent on the scale and method of the survey to be undertaken. The materials for conducting the License Plate Survey will be provided by the contractor hired by the project sponsor(s). The materials identified below are intended to provide basic guidelines for each roadway to be surveyed using either of the roadside survey methods:

- Individual packages containing questionnaires to either hand out or tally responses, pencils for tallying responses, and large pencil erasers;

- Survey assignment sheets;

- One or more envelopes per specific time period and direction;

- Traffic cones and advance warning signs; and

- Surveyor badges, hard hats, and (highly visible) safety vests.

4. Survey Procedures – Each surveyor should be provided with an outline (in writing) of the procedures described in the previous training sessions. In some cases, several copies of a letter describing the reasons for the survey may be enclosed with this package. The letter should be signed by the project sponsor(s) and should provide interested drivers with information about the survey. In addition, this will ensure that the surveyor will not be distracted from his/her responsibilities by answering questions about the survey that may be served through this letter. At a minimum, the procedures outline should contain the following information:

- Where and when survey crews are to meet at the start of the survey shift;

- The set up of the survey traffic control plan at the specified roadway location;

- How to determine which questionnaires are to be used for specific roadways, directions, and time periods;

- How to hand out questionnaires and/or conduct the interview and how to complete tally sheets for those drivers agreeing to participate in the survey;

     - (Roadside Interviews) What to do with completed questionnaires;

- What to do with the survey control sheets that track the number of completed, incomplete, and refused interviews; and

- Removal of the traffic control equipment at the end of the survey shift.

- (License plate surveys) Procedures for transferring raw license plate data to the office, procedures for matching the license plate database to DMV databases, and procedures for mailing surveys.

- Office staff procedures to follow when completed survey forms are mailed back.

15.9 Interviewing and Questionnaire Distribution

This section describes the typical techniques used in Roadside Handout and Interview Surveys. For each survey method, the surveyor should be trained to solicit agreement of participation from drivers traveling on the selected roadways. Descriptions of surveyor interviewing techniques are described below for each survey method chosen.

15.9.1 Roadside Handout Survey

1. Set up the traffic control plan for purposes of providing advance warning to drivers of the upcoming survey.

2. Once questionnaires are handed out, surveyors should be trained to briefly describe the survey to solicit passenger participation. Topics covered include the purpose of the survey and description of the length of response time. Politeness and conciseness are required for each surveyor to obtain representative survey samples.

3. If passengers refuse to participate, surveyors should politely and quickly thank them for their time. If passengers agree to participate, the surveyor should distribute the questionnaires to potential respondents for mailback.

15.9.2 Roadside Interview Survey

1. Set up the traffic control plan for purposes of providing advance warning to drivers of the upcoming survey.

2. Surveyors should be trained to briefly describe the survey to solicit driver participation. Topics covered include the purpose of the survey and description of the length of interview time. Politeness and conciseness are required for each surveyor to obtain representative survey samples.

3. If drivers refuse to participate, surveyors should politely and quickly thank them for their time and move on to the next driver. If passengers agree to participate, the surveyor should ask the related travel behavior questions and tally the passenger responses. The surveyor will be equipped with several survey questionnaires, a clipboard, and pencils to tally passengers’ responses. The surveyor should limit individual interview time to one to two minutes. A general rule for commercial vehicle surveys is to complete the survey within 7 minutes when axle weights are collected. 

4. Surveyors should record all collected information neatly and accurately on the questionnaire forms. During the slack periods, the surveyors should use the time to code each questionnaire for all information other than the origin and destination data obtained during the interview.

5. Surveyors will be responsible for organizing questionnaires collected from drivers by uniquely coded time period and one-way roadway manila envelopes for data processing/checking purposes.

The surveying and interviewing techniques for each survey method described above are very similar. Variations of these techniques can be used by the survey team to respond to unique situations likely to be encountered in different areas throughout the country.

15.9.3 Additional Administration Issues

The organization plan outlined previously in this chapter identifies the staffing analysis, hiring methods, and supervision required for each survey method. Additional administrative support issues should be considered before the survey is fully implemented. The requirements for this administrative support include:

· Developing the necessary forms such as the questionnaires, control sheets, and surveyor assignment sheets;

· Providing schedules to survey crews (make sure surveyors and film crews will show up at their scheduled times);

· Developing a manual of survey instructions;

· Developing and conducting the surveyor training program;

· Specifying and assembling survey materials including surveyor badges, vests, hard-hats, pencils, clipboards, envelopes, traffic cones, and advance signs;

· Conducting the pretest; and

· Ensuring the appropriate level of police enforcement;

· Establishing a criteria for cancellations due to weather, and a plan to notify crew members.

This material must either be compiled and/or purchased at least one month before the implementation of the full survey. Additional materials, especially cones, and surveyor vests, and hard hats may be required for purchase as a contingency.

15.10 Coding

Vehicle intercept and external station surveys are coded using similar procedures as for other surveys described in this manual. In the case of the vehicle intercept, similar techniques are used to code the mailback/self-administered questionnaires obtained from the License Plate and Roadside Handout Surveys. The survey questionnaires are typically designed to be self-coding (except for the origin-destination information), where each survey response can be coded to correspond to its answer check box number.

Roadside Interview Survey coding can use the same techniques as the other survey methods. For example, the interviewers typically code the motorist responses into the check boxes on the questionnaire form, either by tallying motorist responses directly onto the forms or by directly entering responses into the survey databases on hand-held computers. Data coding can be completed manually or entered directly (using hand-held computers) into the coding database, by motorists completing the self-administered mailback survey, or by CAPI.

The survey data are typically punched into a numerical ASCII data block for a specified width and length as determined by the number of questions/responses and sample size of the survey. Individual survey questionnaire responses are typically given an identification number to track the responses for each vehicle surveyed. Survey origins and destinations must be geocoded to identify the geographic locations of the vehicles surveyed. Chapter 14.0 provides a detailed discussion of survey geocoding techniques.

15.11 Cleaning and Editing

The requirements for data editing the returned questionnaires is very similar for each selected survey method. However, the requirements for the License Plate Survey also require the additional task of editing and cleaning license plate information gathered from the videotapes and addresses obtained from the motor vehicle departments. Since turnaround time is critical, electronic data file transfer is becoming increasingly popular. Requirements for the Roadside Interview Survey consist of ensuring that all responses are tallied legibly and accurately by the surveyors. The questionnaires for each completed one-way roadway trip should be edited as soon after collection as possible to ensure that the questionnaire represents a valid response. At the end of each day, the trip envelopes should be opened and the control sheets should be checked for completeness.

For each survey method, the number of completed questionnaires and blanks should be tallied for each trip. Once the editor is satisfied that the trip information is complete, the questionnaires returned for that trip are sorted. Blanks can be discarded while the completed questionnaires should be sent for data entry. Each completed questionnaire may not have all questions answered. Therefore, the rules to identify a completed questionnaire should be given to the editors. Generally, a questionnaire is considered usable if the origin and destination of the trip, trip purpose, vehicle occupancy and time are filled in and codable.  In addition, some models may also require demographic information such as home location, income, household size and auto ownership.

Cleaning the data once it is entered begins with range checks. For example, if the possible answers to a question are numbered 1-4 and the non-response is coded as 9, then all answers between 4 and 9 must be erroneous. In addition, certain cross checks must be performed to verify the accuracy of the data. These types of checks are called logic and consistency checks. 

15.12 Survey Expansion

Survey record expansion and weighting depends upon an accurate, 100% count of traffic at the survey site during the day of the survey.  Ideally, this count should be stratified by hour, and type of vehicle (auto and truck).  The survey records are then assigned a combined expansion/weighting factor by hour and type of vehicle.  Quality assurance checks should be made to ensure that the sum of the final valid survey records' expansion factors equal the observed total vehicle trips on the day of the survey.  A final adjustment factor may be added to adjust for the day of week and seasonal variations in traffic.  This can be done by obtaining the average annual daily traffic volume from the jurisdiction's highway department and comparing it with the survey day counted traffic. 

Roadside interview and handout surveys are usually conducted at multiple data collection sites.  When surveys take place at multiple the simplified record expansion outlined above is not satisfactory.  The reason being that the population of vehicles could have travelled through any combination of two or more sites.  Accordingly, if a calculation for the treatment of the site weights are not introduced, multiplicity of surveys results and double counting quickly contaminates the population.  Without the proper statistical treatment, bias is introduced, particularly for the longer distance trips where their likelihood of multiplicity increases.  Unfortunately, this issue is not commonly addressed and analysis becomes flawed (Sureshan S., 2008).

Associated Documents

Collecting External Data Using Bluetooth Technology 

Field Practices in External Survey Data Collection

ALPR Camera Surveys in Ohio

Estimating Through Trips Using a Two-Phased Logit Model 

Synthesized Through-Trip Models & Economy-Based External Trip Models for Small and Medium-Sized Urban Areas 

REFERENCES 

Adler, Thomas, Stephen Lane, Nicholas Brand, and Harold Wilson. A Quick Response Screening Model for Planning Statewide Origin-Destination Surveys, presented to the 5th National Conference of Transportation Planning Methods Applications, Seattle, May 1995.

Anderson, Marsha. Development of Notebook Computer Software, 1994.

Baker, T. Brent, Raymond Deardorf, and Cathy Strombom. Sampling Travel Survey Participants Using Video Technology on the Tacoma Narrows Bridge, presented at the 74th Annual Meeting of the Transportation Research Board, January 1995. 

Muntean, David, Jr. Origin-Destination Travel Surveys in the 1990s Using Microcomputers, ITE Journal (January 1995) pp. 39-43.

Sureshan, S. Commercial Vehicle Survey Data Expansion Process, Ontario Ministry of  Transportation, Toronto Ontario Canada, September 2008.

U.S. Department of Transportation, Federal Highway Administration. Urban Origin-Destination Surveys, Washington, D.C., 1973 (reprinted 1975).

[1] Travel surveyors have developed a number of in-the-field approaches to improve the transcription process. For instance, for a Massachusetts survey, fieldworkers passed their clipboards in front of the camera right before a platoon of cars was about to pass. This would alert the office worker to switch from fast-forward to play. A Washington survey team deliberately over-exposed the videotape by a small amount so that the white license plates would stand out better (see Miller, et al. and Baker, et al.).