Skip to content

Project Meeting 2023.06.01

Michelle Bina edited this page Jun 2, 2023 · 10 revisions

Agenda

  • Admin Items
    • Doodle poll results for proposed regular meeting time change (CS)
  • Household Travel Surveys (MWCOG)
    • Discussion on survey frequency

Action Items

  • Meetings to change time (11am PST/2pm EST) starting July 1. Tentatively planning time change for a month or two and then reassessing which time works best for folks.

Meeting Notes

Meeting Time

  • Planning to move meetings after July 1, try it out and see how things go.

Frequency of Data Collection

Questions posed to the group?

  • What impact does smaller samples but more frequent surveys have on model development?
  • Do you wait until you have several years of data before estimating/calibrating?
  • Do you have economies of scale, resulting in more samples for lower effort?
  • You could conceivably do calibration more frequently - whether or not that’s a good thing, depends.
  • What about transit onboard surveys? Do you coordinate those with household travel surveys? Is it coordinated across operators?
  • Is anyone doing NHTS add-on instead of their own survey?

SANDAG

  • Conducted a small sample survey in 2022 (2800 households)
  • Model base year will be moved to 2022
  • There is another follow-up survey currently in the field and near the end of data collection (targeting about 2000 households)
  • Planning to do smaller but more frequency surveys once every two years
  • Commercial vehicle and establishment surveys were conducted last year.
  • Transit onboard survey got delayed, but is now in the field. They usually do this about once every four or five years. There aren’t as many operators as other regions, so it’s easier to coordinate.
    • They are currently doing a transit onboard tour survey to get trip diary data to complement household survey data.
  • Regional plan is every four years, so they might do model updates with that frequency.

CMAP

  • Have traditionally conducted surveys about once a decade. The last one was spring 2019, providing a good pre-COVID snapshot, with about 12k households.
  • Starting next spring, they will be testing 3 small wave surveys every 2 years but haven’t set target sample numbers yet.
  • At this point, recalibrating the model is not intended after the initial round of surveys comes in - it will probably be too small of a sample. It’s also bumping against the next LRTP and calibration wouldn’t be done in time. They might consider a larger model update effort after the second wave.
  • Two largest transit operators currently have an approx. 5 year cycle of onboard surveys but hasn’t been coordinated process at this point.

MWCOG

  • Their last survey was 2017-2018 and has been a once-a-decade survey.
  • They are currently looking at state of the process; interviewing 6 MPOs. About half have moved to more frequent smaller sample.
  • Particularly for ABMs, they would like to know what is the sample size needed to update those components?
  • They have memos on this topic that they would be willing to share with the group.

MetCouncil

  • They used to do a 10 year survey with 10k households (last one was 2010).
  • Starting 2019, they switched to a smaller more frequent survey approach (7500 households in 2019).
  • They also did a quick contact of 2019 participants to ask COVID-related questions during the pandemic. This was not planned but quick thing they wanted to do.
  • 2021 sample included 3500 households.
  • 2023 sample is currently going out in the field now and target is 3500 households.
  • They are not sure if management will continue this effort after the third wave. Funding is limited to these three waves.
  • Surveys are cross-sectional sample, not panel.
  • In 2019, the plan was to combine samples for calibration, but first sample was pre-COVID.
  • Advantage of this approach is that you can use different waves to try different methods to reach low participation rate populations.
  • Recent systemwide on/offs and customer surveys available spring of 2023.

PSRC

  • They are moving more towards frequent and smaller surveys.
  • They have surveys from 2014, 2017, 2019, and 2021 – in field again this year 2023 (targeting about 3k samples, 2400 for the region plus add-ons for cities).
  • Surveys are a popular source of data for the agency, beyond modeling, and has been the impetus to keep it more frequent.
  • They combined 2017+2019 for a bigger sample.
  • Region has actively changing transit system with major light rail changes, so it’s difficult to reconcile network levels of service and travel behavior changes.
  • MPO doesn’t coordinate transit onboard surveys, not as common in the region.
  • Not doing panel surveys anymore, as far as they know.

MTC/SFCTA

  • MTC is leading travel survey effort for Bay Area.
  • 2019 was an all app-based travel diary, testing person-based vs household-based survey. The effort was meant to be one off, also capturing unique travels, but has been used to re-estimate many components except for one with kids since only surveyed 18+ residents.
    • Leveraged the app trace data in numerous ways. It was great for QAQC on what people are reporting.
    • Caveats about requirement and participate rates
  • Staff (30-40 people) tested out multiple apps for user experience and data produced
  • Current procurement is for 6 years, with 3 cycles every 2 years
  • They are currently testing different recruitment approaches to quantify tradeoffs
  • Reasons behind every 2 year sampling:
    • Travel behavior is changing rapidly and 5 years would be too infrequent
    • Even more true from uncertainty of COVID impacts – telecommuting and transit ridership
    • Questions from planners don’t need all the details and can use aggregate stats (without needing large sample sizes)
    • Administrative advantages with procurements, survey instruments, etc. Easier to budget for smaller more frequent survey than a very large one infrequently.
  • Transit onboard surveys
    • Used to be operator by operator, going 2 or 3 at a time
    • This effort was paused during COVID
    • Now moving to a “snapshot” survey, doing all operators at the same time. Less detailed sampling – more breadth than depth – to be done within the next year. Valuable because transit ridership has changed.
  • Any other surveys?
    • Comparisons of survey with big data sources. Trying to work out what that means but will be exploring possibility of getting more information via big data. MTC is mostly interest in specific facilities. They will oversample users of specific facilities and compare to data sources to see what those data sources can contribute to that conversation.
  • Zephyr session is planned for what people are doing with big data in modeling.

SEMCOG

  • Partnered with state DOT for next survey.
  • Thinking of surveying more frequently than every 10 years but not there yet.
  • Didn’t discuss implications on estimation or other challenges that may come along with collected data this way.
Clone this wiki locally