Below we present three high level insights about accessibility concepts and measures in practice that surfaced in interviews, followed by a discussion of findings about how accessibility is used, barriers to accessibility measurement, and opportunities to expand and improve accessibility measurement.
Accessibility Concepts And Measures In Practice
Three key insights about accessibility concepts and measures in practice were apparent from our analysis of interview and survey data.
Accessibility uniquely captures the vital connection between transportation and land use
Many respondents indicated that accessibility is a fundamental objective of transportation systems, and that accessibility measures are designed to capture this critical benefit. Advanced, standard, and non-users all pointed out that accessibility is affected by land use and transportation systems. They indicated that accessibility measures add value to performance assessments because they capture the important influence of land use on transportation system goals and related objectives centering on multimodal travel options.
Consequently, some respondents indicated that accessibility measures serve a critical function supporting planning projects that otherwise would be dismissed. For example, an advanced user from a large MPO said:
…In a lot of planning studies…it would be a complete nonstarter to talk about taking that level of capacity away from a current high-volume corridor…Having the accessibility measures to … round out the comparison... allowed that [multimodal improvement] alternative to survive to the end of the project when it probably wouldn’t have otherwise.
At the same time, many respondents qualified their response by noting that they lacked power to affect land uses and development patterns. As one nonuser said:
…These are all great things that you can measure, but are we …able to effect a change? And that’s the problem… if you're not able to effect a change based on what you’re measuring then why are you measuring it?
Overall, while some users find accessibility’s ability to capture land use to be powerful and useful, others find it less useful for the same reason.
Accessibility applications can vary substantially based on local needs, contexts, and perspectives
Multiple interviewees indicated that the desirability, design, and use of accessibility measures depends on the context. Measures that work well in urban areas may not work well in suburban or rural areas due to different transportation and land use conditions and performance objectives. For example, congestion may not be a concern in some rural areas and accessibility may be undesirable for residents seeking remoteness. At the same time, in some highly congested urban areas, accessibility via land use change may be easier to affect than congestion.
Several respondents highlighted the unique objectives of rural areas and how they relate to accessibility. One DOT respondent in a largely rural state noted that the notion of accessibility is contrary to the reason that some people choose to live in rural areas:
There’s a desire to remove themselves from accessibility because… while it’s great that you can get to the store and the grocery store and doctor and the pharmacy and to all these different things easily, everybody else can get to you easily as well. So there’s some of that mentality that people want to be away from it all.
Rural areas also pose unique considerations in terms of how to measure accessibility. An advanced MPO user reflected on challenges to applying commonly used employment access measures in rural areas, noting that some communities have high accessibility for non-employment destinations:
Most of these [rural] places would come across as fairly accessible if someone lives in [a rural] county... They have to have a car because there’s no transit there but it’s the rolling up to the grocery store, running to my church, running to CVS. Those are all they would say they do locally within their community. I guess it just depends on the scale of accessibility. Dad drives 45 minutes to get to work Monday through Friday, but he’s going to little league games down the street...those sort of internal trips are still being contained locally within those [rural] areas.
For this MPO, measuring accessibility to non-work destinations was a way to show that urbanized areas are not the only areas with high accessibility.
On the other hand, a standard DOT user indicated that access to employment opportunities is important to measure in rural areas that may feel left out from the economic growth experienced in more urbanized areas because they capture connections to urban job centers:
I think the intent behind the highway accessibility/connectivity criteria being in the law was this whole idea of getting people from the rural areas to the urban centers for jobs…since there is growth in urban areas…from a political side it seems reasonable to think: ‘hey, if we improve the access from those areas that are kind of feeling left behind to those areas that are growing, then it's kind of a win-win… [Without the accessibility measure] our rural projects would have a little bit harder time competing on some of the other criteria.
Tailored variations of accessibility measures may also be needed to account for differences across a region in areas that includes a mix of urban, suburban, and rural communities.
Measures must be easy to interpret and communicate to the public and decision makers
As public agencies and stewards of public funding, nearly all interviewees from state departments of transportation, metropolitan planning organizations, and transit authorities expressed that accessibility measures must be easy to communicate and interpret. Multiple respondents suggested that elected officials, board members, and the general public must be able to understand calculated measures. Concerns about communicability and interpretability often emerged when discussing the potential to use some of the more technically demanding measures.
In general, respondents expressed a desire to use measures that are easy to explain and noted their reluctance to use measures that are difficult to explain to ordinary residents due to “black box” analyses or software used to produce the results. Both advanced and standard users reported that proximity, trip characteristics, and access to opportunities measures are the most easily communicable metrics that staff, residents, and decision makers could relate to easily. Logsums were cited multiple times for communication challenges. For example, an advanced MPO interviewee working within a large region said:
I would never want to try to communicate to policymakers or the public what logsums are. I think that cumulative opportunities measures are really appealing because they translate directly into things that policymakers and business leaders tend to care about, which is the size of markets...Anything I could ever do to avoid talking about logarithms in public to a non-technical audience, I would do.
When the audience differs, considerations of communication and interpretation also shift, and measures can be more sophisticated. Advanced users that computed logsums often did so for internal analyses to validate travel demand model outputs or provide another perspective on public-facing proximity or access to opportunities measures.
Multiple respondents also described tradeoffs between technical completeness and communicability when deciding which accessibility measures to use. Comparing logsums to access to opportunities measures, a planner working for an advanced MPO user indicated that communicability “is more important than technical correctness.” The same planner also emphasized an agency’s responsibility to clearly explain to stakeholders why measures were chosen and used. The tension between usefulness and user-friendliness echoes tradeoffs described in prior work in the US and Europe (Silva et al. 2017; Siddiq and Taylor 2021).
Overall accessibility measures have different strengths in terms of their ability to capture specific objectives and their transparency and technical accuracy. Thoughtful consideration of the particular needs of a specific context can help practitioners determine which measure(s) are useful, as discussed further below.
Accessibility Measure Use And Evaluation
Accessibility measures are used by transportation planning agencies of various sizes, in urban and rural areas, and with systems ranging from auto-centric to multimodal. Not all measures are used at similar rates. Our respondents most commonly reported using proximity and access to opportunities measures. We identified no respondents using competitive measures or potential path areas, although multiple agencies employ logsum measures in public-facing documents. Logsums are also sometimes used internally during exploration or to validate modeling. Although trip characteristics are used by some agencies, they did not generally seem to be interpreted as accessibility measures.
Interview and survey results point to challenges and opportunities posed by different accessibility measures in a number of different areas. Figure 1 summarizes the median survey response for each of the six accessibility measures discussed in interviews and surveys in terms of three categories: usefulness, ease of application, and appropriateness for application areas. Median values were calculated for each measure using two different subsets of respondents: those that report moderate, very, or extreme familiarity with all six applicable measures (six respondents), and those that report slight familiarity or no familiarity at all with at least one measure (36 respondents).
A respondent who is less familiar with a measure may rate it negatively based on a perception of poor performance, which might change if they learned more about the measure or used it more in their work. Less familiarity may also be related to an agency’s needs, which may affect their rating of usefulness or appropriateness. An agency’s resources or staff capacity can also relate to their familiarity and can also affect their ability to easily apply measures. Not surprisingly, most respondents reported lower levels of familiarity with more complex measures (competitive measures, potential path areas, and logsums) when compared with simpler measures (proximity, access to opportunities, trip characteristics).
Respondents with both high and low measure familiarity indicated that measures of proximity, access to opportunities, and trip characteristics were very useful for most modes, except in cases where familiar users indicated that access to opportunities and trip characteristics are extremely useful. Logsums and potential path areas were rated moderately useful for most modes by both familiar and less familiar respondents, while competitive measures were rated moderately useful by less familiar respondents and very useful by familiar respondents. Lower usefulness scores indicated by less familiar respondents may reflect less knowledge of the measures, or they may reflect differences in the needs of their agencies.
In terms of ease of applying the measures (assembling data, computing measures, and interpreting results), familiar and less familiar respondents both rated proximity measures as somewhat easy to apply, placing them among the easiest to apply measures. Familiar respondents indicated that access to opportunities measures were as easy to apply as proximity measures. This contrasts with the assessment of less familiar respondents, who indicated that access to opportunities measures were somewhat difficult to estimate and neither easy nor difficult to interpret. Similarly, familiar respondents rated the remaining measures neither easy nor difficult to apply in most cases, in contrast to less familiar respondents who often rated them somewhat difficult to apply. This may reflect the latter group’s lower level of knowledge of the measures or lessor technical resources or capacity. These survey results align with interview findings that reveal barriers and limitations to applying accessibility measures in practice. Acquiring and assembling the necessary data for accessibility analyses was described as one of the most challenging steps in the process according to both survey and interview data, particularly for interview subjects with less experience and capacity. Both familiar and unfamiliar survey respondents indicated that proximity, access to opportunities, and trip characteristics measures are easier to interpret when compared with competitive measures, potential path areas, and logsums. This finding was echoed in interview responses related to each measures’ communicability.
Respondents indicated that all measures except for logsums were appropriate for most application areas. Familiar respondents indicated that proximity, access to opportunities, and trip characteristics were modestly more applicable than more complex measures in most cases. Logsums were rated as less applicable than other measures, particularly by respondents with more familiarity. This aligns with interviewer subjects’ insights about the difficulty communicating logsums to non-technical audiences and their use of logsums for internal deliberations.
Barriers To Accessibility Measurement In Practice
Several non-user agencies and standard users described barriers to using accessibility measurement or using more sophisticated measures. Resource limitations in terms of staff time and expertise as well as technical limitations on tailoring the measures correctly were commonly cited by interview respondents as a barrier to use or a consideration limiting the type of measures in use. This is consistent with the work of others that has highlighted planners’ lack of knowledge and data constraints as inhibiting the more widespread adoption of accessibility measures (Boisjoly and El-Geneidy 2017).
Resource limitations were described by a range of agencies. One non-user representing a predominantly rural area described the limitations faced by smaller agencies in terms of both capacity of staff to compute accessibility measures as well as expertise of decision-makers to interpret the measures even if they were calculated:
Small communities, they don’t have that tax revenue in order to hire those engineers who are high-skilled and usually high-wage employees so they have to rely on, you know, less-skilled employees and employees that wear many hats, rather than being specialized on one area …
Even advanced users indicated resource and time constraints in considering more robust accessibility measures. One advanced MPO user stated:
It would be awesome to be able to get into some of these accessibility measures and look at which projects could have the greatest impact on enhancing accessibility, but the kind of “keep it simple, stupid” mantra always has a limiting effect on how complex we get with project evaluation…we would love to do stuff like this [more advanced demand-aware measures], but it's a manpower issue.
At the same time, many standard and advanced users reported working with software products and vendors to automate accessibility measure calculation using travel demand models, GIS, or other software. Some agencies also purchased commercial datasets describing employment locations and/or origin-destination flows to aid in these calculations.
Technical limitations often limit evaluations in specific contexts. Several standard and advanced users indicated that they had had trouble obtaining data needed to understand walk, bike, and transit travel, including data that reflect real-world traveler counts, travel behavior (e.g., transit rider origins and destinations), and infrastructure presence or quality (e.g., sidewalks). Many respondents struggled to incorporate active travel modes into their travel demand models. Both advanced and standard users expressed concerns about the difficulty of calculating region-wide accessibility measures due to lack of consistent regional datasets. Aside from mode-specific data, maintaining consistent spatial data representing destinations (the opportunities side of accessibility) across an entire state (or region) can also be challenging.
Data limitations can have unintended consequences when they leave out particular services or populations. A planner at a standard MPO user revealed that their accessibility measurement falls short in terms of including all types of public transit services. Their modeling focuses on fixed route services and does not include demand responsive service such as paratransit. They indicated that “a big part of the population of people with disabilities aren’t really included in the accessibility measures.”
Even when adequate staff resources and data are available, other conceptual barriers can arise. Some respondents indicated that it was difficult to consistently estimate accessibility where interpretations would vary across geographic contexts. For example, a grocery store may be considered accessible in a rural area if it is within five miles, but in a more urban area that measure may be reduced to a ½ mile.
Some respondents also noted that accessibility is just one of many factors that aid in transportation decision making and performance evaluation. As one standard MPO user put it, “accessibility is probably one of the primary reasons for making a transportation decision, but it’s not the only reason that we do things.” In fact, accessibility is not an explicit part of the federally required MAP-21 performance measures (Federal Highway Administration, Department of Transportation 2013). There are federal goals to “reduce congestion” but none relating to “increasing accessibility,” so agencies seeking to incorporate accessibility analyses must go above and beyond federal reporting requirements to do so. Many users and non-users kept track of several other performance measures in accordance with federal MAP-21 reporting requirements but noted difficulty obtaining resources or permission to compute accessibility measures above and beyond these requirements. Ultimately accessibility is not a required performance measure, so it is difficult for some to direct resources towards it or view it as fundamental or distinct from other performance measures.
Opportunities For Expanding The Use Of Accessibility Measures
Several respondents indicated their intention or desire to improve or expand their use of accessibility measures. In some cases, these improvements would apply accessibility measures in ways that are relatively well established in research and practice. For example, some respondents expressed interest in applying access to opportunities measures, using accessibility measures to highlight the transportation-land use connection, or using accessibility to better understand the needs of vulnerable populations. In other cases, respondents indicated a desire to apply emerging data sources or methods in their accessibility measurement. For example, a few respondents expressed interest in using less commonly applied measures such as potential path areas, logsums, or competitive measures. Some expressed a desire to account for travel time reliability, while others were interested in obtaining and using data that reflect the quality of networks and destinations in their accessibility measures.
One area of interest mentioned by several respondents is the use of repeated accessibility analyses to support scenario evaluation. This approach would allow agencies to use accessibility measures to plan more proactively. As one standard MPO user put it:
We haven’t used [accessibility analysis] in scenario planning yet. I think that's something that our MPO board has expressed interest in … Because it is done … after the projects have been chosen. I think they’re interested in using it in a more proactive way… I think that's something that they'd like to see and hopefully that would … at the very least give them better understanding of … what impact their decisions are likely to have.
Some expressed a similar desire but noted that they had yet to find a tool to make this approach tractable. As an advanced MPO user stated:
The idea of doing …scenarios for system planning is probably one of the more emerging areas … I’d like to see it … streamlined, easier, able to be used to evaluate more iterations so we can think about it more as a scenario planning tool.
Others expressed a desire to use accessibility to guide investments proactively. One standard MPO user in a mid-sized region had the following to say:
Going forward, I'd like [accessibility] analysis to help us shape the future more than just analyzing what we're doing…to help us really spend the billions of dollars we have control over better going forward to actually change things for the better, to again create better access.
Another frequently mentioned area of interest is evaluation of nonmotorized travel options. Respondents noted that evaluation of non-auto travel has been hampered in the past by a lack of consistent and large-scale data for pedestrian and cyclist conditions and an inability to model their effects on travel choices. Several users mentioned that data and modeling improvements that would support more robust accessibility measurement of nonmotorized travel have recently occurred or are on the horizon, for example sidewalk/cyclist infrastructure data or the ability to model bike/pedestrian trip assignment. Respondents highlighted the growing potential of accessibility measurement to better capture nonmotorized travel as data and modeling methods advance.
In response to data challenges, a number of respondents also encouraged potential accessibility users to be pragmatic by using the data they already have. Others emphasized the value of simple solutions and publicly available data.
While respondents expressed a desire to expand or improve their use of accessibility, many also showed interest in assistance with learning how to do so. Several respondents asked interviewers questions about how to design or interpret accessibility measures in specific contexts, or about specific methods or data sources. Some respondents indicated that a more standardized way of measuring accessibility would improve overall understanding of accessibility. Others indicated that accessibility measurement should be tailored to a specific context and its particular needs.
Many respondents expressed interest in guidance about accessibility measurement to explain how they should be used and when to use different measures. A few respondents expressed a desire for guidance that spells out the concept of accessibility, noting that a standard understanding of accessibility is missing between agencies, stakeholders, and the public and that there are misconceptions about what accessibility means
Some respondents emphasized the promise of using the concept of accessibility more broadly to advance peoples’ understanding of transportation systems. An engineer working at a large MPO highlighted the value of accessibility for demonstrating fundamental transportation system benefits, indicating that
The concept [i.e., accessibility] really helps people understand what transportation is trying to do…Don’t focus on getting the analysis perfect … and nitpicking it. It’s take this … concept and allow it to be a learning tool.
In fact, a key lesson from our interviews is the power of using the concept of accessibility to think deeply about transportation-land use performance can be valuable regardless of the measure used. As one of our interviewees put it,
The huge opportunity [with accessibility measurement] is to get people … thinking about the destinations, rather than the travel times.
Similar arguments have been made in the research literature. For example, Handy (2020) notes,
the intense focus on accessibility [performance] measures might be distracting us from the power of the concept of accessibility itself. Why I am – and perhaps others are – so enamored with accessibility is its vast potential as a way to understand the urban environment. Using the concept of accessibility as a way to simply think about the urban environment can itself be helpful to planners. (p. 4).