5.0 I-35W TRANSIT ON-BOARD SURVEYS

5.1 Purpose and Approach

Metro Transit conducts on-board customer satisfaction surveys every two years. The two most recent on-board customer satisfaction surveys were conducted in 2006 and 2008. The next on-board customer satisfaction survey will be conducted in the fall of 2010. The national UPA evaluation will use the results from these on-board surveys for the routes in the I-35W corridor as part of the national evaluation. This information will allow for comparisons pre- and post-deployment. In addition, on-board ridership surveys targeted specifically to the Minnesota UPA projects will be conducted on MVTA routes in the I-35W corridor. These surveys will provide more detailed information on the prior mode of travel, frequency of use, reasons for use, influence of the Minnesota UPA project on use, and related information. As discussed in Section 4.0, the surveys will also obtain information on perceptions related to the real-time transit and traffic information DMS and use of transit based on this information as part of Metro Transit's request to experiment submitted to FHWA.

Additional Metro Transit and MVTA On-Board Surveys. The transit on-board ridership surveys will be conducted by Metro Transit and MVTA. On-board ridership surveys will be administered on Metro Transit routes serving the new and expanded park-and-ride lots along I-35W north of downtown Minneapolis in May-June 2010, approximately six months after opening of the new and expanded park-and-ride lots and implementation of the new transit services and the MARQ2 bus lanes in downtown Minneapolis. On-board ridership surveys will be conducted on Metro Transit and MVTA routes serving the new and expanded park-and-ride lots along I-35W and Cedar Avenue south of downtown Minneapolis in September 2011, approximately one year after the opening of the final park-and-ride facilities and the HOT lanes in the Crosstown Commons section. This schedule allows for conducting the surveys at times other than September 2010, which is when the ongoing Metro Transit customer satisfaction surveys are conducted.

Metro Transit On-Board Customer Satisfaction Surveys. The Metro Transit Bus Ridership Survey is intended to meet a number of objectives. The objectives include determining why customers ride the bus, identifying the primary sources of bus information, assessing how Metro Transit communicates with customers, and evaluating riders' overall level of satisfaction with Metro Transit. Other objectives are to determine riders' level of satisfaction with specific service components, to identify which components influence overall satisfaction the most, and to identify which service elements are of greatest importance to customers. Two final objectives are to evaluate the importance of possible service improvements to riders and to assess the demographic characteristics of current riders. The results will also be compared with those of previous years to monitor attitudinal shifts over time.

The following questions from the 2008 Bus Ridership Survey will be examined for routes in the I-35W corridor to help establish pre-deployment rider characteristics.

• How many days per week do you ride the bus?
• How long have you used Metro Transit service?
• What influenced your decision to first try transit?
• From which zip code did you begin your commute today?
• To which zip code are you commuting today?
• Which days of the week do you usually ride the bus?
• When do you usually ride the bus (rush hour, non-rush hour, special events)?
• How did you pay your fare today?
• Does your employer offer transit passes?
• What is the primary purpose of your trip today?
• If transit were not available, how would you have made this trip today?
• Which best describes your racial or ethnic background?
• What is your age?
• Are you female/male?
• Approximately what was your family's total income last year?
• What is the one main reason you use transit?
• How many working automobiles do you have available for your use?
• Please indicate your level of agreement with the following statements (strongly agree, agree, disagree, strongly disagree, don't know)
  • Park-and-ride lots are conveniently located.
  • You feel your car if safe in the park-and-ride lots.
  • Hours of operation for express service are sufficient.
  • Morning rush hour service runs on schedule.
  • Afternoon rush hour service runs on schedule.
  • There are enough express routes.

As available, data from the 2006 survey will also be examined to develop trend lines for rider attributes. Finally, the results from the 2010 Customer Satisfaction survey will be analyzed.

5.2 On-Board Ridership Surveys Protocol

These surveys are targeted at obtaining information from passengers on Metro Transit and MVTA buses operating in the I-35W corridor. Metro Transit and MVTA have extensive experience conducting on-board ridership surveys, including the on-board customer satisfaction survey, which is completed every other year. The same process used on other surveys will be followed on the UPA on-board ridership surveys. The actual process includes individuals handing out surveys and pencils to riders as they board buses and collecting the surveys as the passengers exit the buses. The routes and runs to be included in the surveys will be determined by Metro Transit and MVTA, with input from the Battelle team.

Table 5-1 presents the Metro Transit and MVTA routes influenced by the Minnesota UPA projects that will be included in the transit analysis. The route number, route description, and service type (express or local) are included in the table. The number of a.m. inbound bus trips are also presented to provide an indication of the current service level. A total of 30 routes have been identified in the I-35W corridor, including Cedar Avenue, to be included in the analysis. As shown in Table 2-1, the majority of Metro Transit and MVTA routes in the corridor provide express service to downtown Minneapolis. The MVTA also provides local service oriented to the Mall of America in Bloomington.

Additional runs may be added to some of these routes to accommodate the anticipated ridership growth from the new park-and-ride lot spaces and the travel-time savings provided by the HOT lanes, PDSL, and MARQ2 lanes. Other routes may have available capacity for new riders. Also, new routes may be added to serve new park-and-ride lots. The service plan for the new park-and-ride lot on I-35W in Roseville is still being finalized, some changes are known at this time. A new route, Route 252, will be implemented in September 2009 from the 95th Avenue/I-35W Park-and-Ride Lot to the University of Minnesota. Three a.m. and three p.m. trips are planned initially. A new route from the new park-and-ride lot in Roseville with six a.m. and six p.m. peak period trips is planned for implementation in December 2009. As discussed in Sections 3.0 and 4.0, implementing new routes, adding trips to existing routes, and other service changes will be documented in the national evaluation, along with ridership and travel-time data.

The sample for the on-board ridership survey will be drawn for the routes presented in Table 5-1. The sample plan methodology outlined next focuses on providing a statistically valid sample of boarding on these routes. The sample plan is based on the ridership on these routes and the overall response rates from the recent Metro Council Customer Satisfaction Survey. It also assumes geocoding is not needed for the origin and destination information. If address geocoding is needed, the assumption for the response rate drops and the sample size needs to be increased. The sampling plan will be finalized based on additional discussions with personnel from Metro Transit, MVTA, the U.S. DOT, and the survey contractor.

The sampling plan also considers a number of other factors. As highlighted below, these factors include type of route, primary destination, direction of travel, and service area for origins.

• Type of route – express and local.
• Primary destination – downtown Minneapolis, University of Minnesota, and Mall of America/I-494 corridor.
• Direction of travel – peak direction and reverse commute (off-peak direction).
• Service area for origins – by corridor for Metro Transit routes (I-35W North and I-35W South) and by route group for MVTA (460 routes and 470 routes).

Based on these factors, the sampling plan should provide a statistically valid sample for each of the following groups.

• Express routes in the I-35W North corridor to downtown Minneapolis in the peak direction.
• Express routes in the I-35W South corridor from to downtown Minneapolis in the peak direction.
• MVTA express routes in the 460 series to downtown Minneapolis in the peak direction.
• MVTA express routes in the 470 series to downtown Minneapolis in the peak direction.
• Local MVTA routes to the Mall of America.

Because analyses are expected to be conducted for each of these groups, we recommend that a stratified cluster sampling approach be utilized to ensure the representativeness of the sample. Each of the five route groups listed above will form a sampling stratum while individual buses/bus trips will form the sampling stratum. The sample size will be finalized based on updated ridership information. Following the Metro Transit procedures, every rider on the selected bus trips will be requested to complete a survey. Based on the preliminary analysis of ridership and response rate to previous surveys, a reasonable goal is 400 valid surveys of riders in each of the route groups noted above. Assuming a minimum design effect of 1.2 (i.e., responses by riders on the same bus are assumed not to be very correlated) this sample size would enable the ability to estimate percentages of riders with a particular attribute or opinion within plus-or-minus 5.4 percent with 95 percent confidence. The exact number of buses that will need to be sampled will depend upon ridership levels per route category and response rate. For example, if one assumes that roughly 10 surveys can be completed on each bus, a sample size of 400 completed surveys would require 40 bus trips to obtain the required sample size. A one-day survey will be conducted for each of the identified bus trips. It is anticipated that the one-day survey will be conducted on either a Tuesday, Wednesday, or Thursday. A 100 percent sample is recommended for the following two route groups stratum.

• Express bus trips to the University of Minnesota in both the I-35W North and South corridors.
• Reverse commute bus trips.

Table 5-2 provides additional information on the sample size. It presents the four major route categories, the number of bus trips for the express routes to downtown Minneapolis and the sample size for routes within the categories.

Table 5-2. Recommended Sampling Plan by Route Category

Route Category	Number of AM Bus Trips per Route	Sample Size for Each Route
Express Routes to Downtown Minneapolis	20 or More Trips	50% of AM Bus Trips
	10-19 Trips	60% of AM Bus Trips
	1-9 Trips	100% of AM Bus Trips
Express Routes to the University of Minnesota		100% of AM Bus Trips to the University
Express Routes – Reverse Commute		100% of all Bus Trips in Reverse Direction
Local Routes to Mall of America		100% of AM Bus Trips in Peak Direction to Mall of America

5.3 On-Board Ridership Survey Questions

The following questions are recommended for inclusion in the on-board ridership surveys. The questions are modeled after those used in previous Metro Transit and MVTA surveys, those used on the MnPASS and carpool surveys, and those used in conjunction with the Miami UPA transit projects. The questions presented focus on the I-35W south on-board surveys. The questions will be modified for the I-35W north on-board surveys to remove references and questions relating to the I-35W HOT lanes and use of the MnPASS toll lanes. The final wording, sequencing of questions, and format for the surveys will be determined by Metro Transit, MVTA, and their contractors.

10. How would you rate each of the following aspects of the I-35W bus service?

11. How does the I-35W bus service today compare to the same service before December 2009?

Please indicate if you strongly agree, agree, neither agree or disagree, disagree, or strongly disagree with the following statements.

5.4 Analysis Methods

The results from the Metro Transit and MVTA on-board ridership surveys will be used primarily in the transit analysis. The survey results will be used to identify individuals changing from driving alone or carpooling to riding transit, as well as individuals making new transit trips. The survey results will be analyzed by members of the Battelle team in a number of ways. In addition to examining the responses to each question, cross tabulations will be run to explore the interaction of different variables, such as income and bus use. Examples of the analyses to be conducted using the survey data are highlighted below.

• Prior mode of travel and mode change to transit. The survey results will be used to assess the prior mode of travel for bus riders. This analysis will examine possible mode change to transit resulting from the Minnesota UPA projects. The analysis will identify if current bus riders changed from driving alone in the general-purpose freeway lanes or carpooling as a result of the UPA projects. The surveys will also identify new transit trips in the corridor which may be made by new residents in the area, individuals entering or re-entering the work force, and incoming students at area universities and colleges. The on-board surveys provide the key source for information on mode change to transit and the factors influencing this mode change.
• Frequency of use and use of other modes. The survey results will identify frequency of bus use by bus riders in the corridor, as well as use of other modes, including carpooling or paying a toll to use the HOT lanes and the PDSL.
• Perceptions of the bus services on I-35W. The questions relating to the reasons for riding the bus, value of the lanes, overall satisfaction with the bus service, safety, and other attributes provide insights into the perceptions of transit riders. Responses to these questions will be used in the congestion, tolling, and other analyses.
• Ridership by income levels, gender, zip code. The responses related to frequency of bus use, factors influencing use, and benefits of use will be examined by income levels, gender, and zip code zones as part of the equity analysis.

Although the on-board surveys will be conducted at multiple time-points and does include the possibility that a particular survey respondent may participate in multiple surveys, the national evaluation team assumes that this will not be tracked as part of the survey. As such, the resulting statistical analysis will be conducted under the assumption that each survey is conducted on an “independent,” though representative sample of riders. The national evaluation team anticipates largely relying upon descriptive statistics, such as estimating means, percentages, ranges, etc. as well as associated tests such as t-tests, likelihood ratio F-tests, and Chi-Square tests to determine if there are significant differences among rider groups, time points, etc. Table 5-3 summarizes the statistical precision (the width of 95 percent confidence intervals for statistical estimates) that is anticipated, based upon various sample sizes of riders.

Table 5-3. Anticipated Width of 95 Percent Confidence Intervals for Statistical Estimates

Data Element	Measures of Effectiveness	Statistical Methods and Assumptions	Anticipated Precision for 95% Confidence Intervals
4.1 Prior mode of transit riders	Actual and percent change in drivers and carpooler switching to transit	One-sample estimation of percentage Percentage of current bus riders that changed from driving alone or carpooling as a result of UPA	Width of 95% Confidence Interval for the estimated percentage of current bus riders that switched due to UPA n=200 - ± 6.9% n=300 - ± 5.7% n=400 - ± 4.9% n=500 - ± 4.4%
4.2 Reasons for using transit		One-sample estimation of percentage Percentage of riders citing a particular reason	Width of 95% Confidence Interval for the estimated percentage n=200 - ± 6.9% n=300 - ± 5.7% n=400 - ± 4.9% n=500 - ± 4.4%
4.3 Length of commute in time and distance	Calculation of change in VMT	Average commute distance prior to UPA assumed to be 10 miles with relative standard error of 25% Average trip commute time is 35 minutes with a relative standard error of 25%	Width of 95% Confidence Interval for the estimated trip length (miles) n=200 - ± 0.35 miles n=300 - ± 0.28 miles n=400 - ± 0.25 miles n=500 - ± 0.22 miles Width of 95% Confidence Interval for the estimated trip duration (minutes) n=200 - ± 1.0 min. n=300 - ± 51 sec. n=400 - ± 44 sec n=500 - ± 40 sec.
4.4 Perception of UPA transit improvements (need list, e.g. park and ride, travel time DMS, more frequent bus service)	Percentage of respondents citing a reduction in travel time Percentage of respondents citing an improvement in travel reliability	One-sample estimation of percentage Percentage of riders citing a particular reason	Width of 95% Confidence Interval for the estimated percentage n=200 - ± 6.9% n=300 - ± 5.7% n=400 - ± 4.9% n=500 - ± 4.4%
4.5 Perception of tolling equity and level of congestion on I-35W		One-sample estimation of percentage Percentage of riders citing a particular reason	Width of 95% Confidence Interval for the estimated percentage n=200 ? 6.9% n=300 ? 5.7% n=400 ? 4.9% n=500 ? 4.4%
4.6 Perception of Safety using HOT lanes, MARQ2 lanes, and guided bus	Changes in the perception of safety by travelers	One-sample estimation of percentage Percentage of riders citing a particular reason	Width of 95% Confidence Interval for the estimated percentage n=200 - ± 6.9% n=300 - ± 5.7% n=400 - ± 4.9% n=500 - ± 4.4%
4.7 Socio-demographic descriptors	Change in travel time and distance by user groups Change in total transportation cost by user group	Average commute distance prior to UPA assumed to be 10 miles with relative standard error of 25% Average trip commute time is 35 minutes with relative standard error of 25%	Width of 95% Confidence Interval for the estimated trip length (miles) n=200 - ± 0.35 miles n=300 - ± 0.28 miles n=400 - ± 0.25 miles n=500 - ± 0.22 miles Width of 95% Confidence Interval for the estimated trip duration (minutes) n=200 - ± 1.0 min. n=300 - ± 51 sec. n=400 - ± 44 sec n=500 - ± 40 sec.

Differences between rider groups or between survey waves will be identified through two-sample t-tests and/or Chi-square tests. Table 5-4 summarizes the anticipated statistical power that can be achieved for these tests based upon different levels of respondents participating in the on-board surveys. Overall, there should be sufficient numbers of samples for the national evaluation to estimate measures of effectiveness with reasonable precision and to conduct statistical hypothesis testing for differences between groups and/or time points, provided that these on-board surveys include at least 400 riders in each group.

5.5 Schedule and Responsibilities

Table 5-5 presents the recommended schedule for conducting the on-board surveys of Metro Transit and MVTA passengers. It is recommended that the on-board surveys of Metro Transit buses operating on I-35W north of downtown Minneapolis be conducted in April-May 2010, approximately six months after implementation of the new and expanded park-and-ride lots, new transit services, and the MARQ2 bus lanes in downtown Minneapolis. Conducting the on-board ridership surveys for Metro Transit and MVTA buses operating on I-35W and Cedar Avenue south of downtown Minneapolis in September 2011 is recommended. This schedule will allow for completion of all the park-and-ride lots in the corridor and the HOT lanes in the Crosstown Commons section. With the completion of these UPA projects, bus riders will have convenient and fast service into downtown Minneapolis using the HOT lanes, PDSL, and MARQ2 lanes. As noted previously, some questions on the surveys will be different to reflect the different UPA projects on I-35W north and south of downtown Minneapolis. The regular Metro Transit customer satisfaction survey is scheduled for September 2010.

Metro Transit and MVTA will conduct the on-board surveys of riders on buses on I-35W and Cedar Avenue following established protocols. Metro Transit and MVTA will finalize the questions to be included in the on-board surveys, in cooperation with the Battelle team. Metro Transit and MVTA will administer the on-board surveys and will compile the results in electronic format. Metro Transit and MVTA will provide the survey results to the Battelle team electronically. Members of the Battelle team will analyze the on-board survey results and will incorporate the results into the interim and final reports.