A New Way to Look at MBTA Demographics
In 2022, the MBTA launched our rolling “rider census” – a.k.a. the System-Wide Passenger Survey – to better understand who rides the system, and how that shifts over time. While this “census” is not a full technical accounting of every rider, it is our most comprehensive survey of ridership and our official tool for determining the socioeconomic and demographic characteristics of MBTA riders across each mode: bus, subway or light rail, Commuter Rail, and ferry.
Traditionally, we use this survey to assess of the fairness of our system in serving low-income riders and riders of color, as required by the Federal Transit Administration (FTA). However, this survey also allows us to consider both intersectional demographics and the time of day that different riders are on the system. In short, the “rider census” allows us to understand the ways that different people use the system, providing contextual information that is not available when we publish daily ridership figures.
More specifically, identifying the demographic profiles of public transit riders throughout the day allows the MBTA to understand how external factors like age and household incomes interact and relate to the use of our services. As a result, we can make better policy recommendations across capital planning, service planning, fare policy, and public outreach planning that incorporate more intersectional, granular demographics.
Our new Rider Demographics Interactive Tool highlights the percentage of riders in specific demographic categories who are on different modes at different times in the day. For example, white men make up 14% of rapid transit ridership at 7AM-9AM, while Black, Hispanic, Asian, or Native American women make up 37% of rapid transit ridership in the same time window.
Better Understanding the Data
Data for our rider census is collected from spring to fall every year, which means we are in the middle of our 2023 survey efforts. Surveyors intercept riders at stops and stations throughout the MBTA system, who complete a brief questionnaire about their most recent trip using public transit. For our analysis, we create and apply survey weights to ensure that our results accurately reflect how many riders are on each line or route in a given year.
To allow for analysis of multiple intersecting identities (i.e., gender, income, race/ethnicity, and age) and avoid dividing the data into insignificantly small sample sizes, we created binary groupings for this dashboard. In other words, riders fall into one of two categories for each demographic variable. While we lose some of the granular details about riders by using just two categories, this improves our ability to compare groups across mode and time of day.
Of note, this data currently limits our ability to disaggregate by line or route. But, as we collect more survey responses this year and in subsequent years, our ability assess ridership in greater detail will improve as our sample size increases. We also do not survey during early morning, late evening, or weekend hours. This further restricts our ability to make inferences about these time periods. For this reason, 6:30PM-6AM is a single time period on the dashboard.
For the other time of day buckets, we set these time periods based on the number of riders using our services throughout the day. Because different volumes of riders are on the system at different times in the day, the 6AM-7AM bucket captures a similar number of riders to the 9AM-1:30PM bucket.
Applying the Tool to Policy
How can the MBTA use this data to improve our service planning? Since a key goal of the organization is to ensure that our service options reduce the transit burdens of riders who have lower incomes and identify with racial and ethnic groups that have been historically disadvantaged, we can begin to assess how service throughout the day is associated with the different groups riding the system.
For example, we can better understand who was affected by a 2021 service change to Commuter Rail. On the recommendation of the Rail Vision Advisory Committee, the MBTA changed Commuter Rail schedules to more closely reflect clockface timetables and better accommodate travel outside the morning and evening rush hours. Using this tool, we can look closer at what proportion of rider categories are actually traveling during these off-peak hours.
Riders who have household incomes below $75,000 and are Black, Hispanic, Asian, or Native American make up a small proportion of Commuter Rail ridership, but this share rises during the midday hours. From 6AM-9AM, these riders make up around 22% of Commuter Rail ridership, while they constitute 32% of riders from 9AM-4PM. A group like women and gender non-conforming riders ages 25 years or younger also make up 24% of ridership 9AM-1:30PM which drops to 12% from 4PM-6:30PM.
These groups of riders likely benefited more from the T shifting service to not only accommodate the office commuters who are most often traveling at peak hours. The relative size of rider groups who have incomes below $75,000 or are Black, Hispanic, Asian, or Native American also grows after 6:30PM, which suggests that providing improved regular Commuter Rail schedules in the evening should be considered, especially if we want to lower the transit burdens for these groups.
In looking closely at bus ridership, a large majority of riders during the day have household incomes below $75,000 and are Black, Hispanic, Asian, or Native American. However, their share of ridership is nearly 10 percentage-points higher during the afternoon rush hour (4PM-6:30PM) than the morning (7AM-9AM). While we do not show route level data, this initial picture suggests that service cuts in the morning as opposed to the afternoon have an outsized impact on this group of riders.
Plans for Future Rider Analysis
This type of intersectional analysis is essential to better conceptualizing equity for MBTA riders. Our demographics tool presents a small slice of the trip data we have collected, and we have many additional questions that we plan to answer. How far are different groups traveling on each mode? Is identity interconnected with transport mode shift? How do station level demographics relate to the accessibility of those stations?
As the T undergoes innumerable challenges in service delivery, getting answers to these questions is even more important for benchmarking the behaviors of our ridership. Year two of the System-Wide Passenger Survey will have a featured role in understanding which riders have been most impacted and how the composition of ridership is changing in the face of these issues. OPMI is currently analyzing this next round of data, and plans to release the next report and interactive data access web tool later this spring.