Improving Language Access on NYC Transit

By Edwin Jeng, Research Assistant, NYU Rudin Center for Transportation

 

Language access (the provision of equal services for people with Limited English Proficiency) is an under-researched aspect of public transportation in New York City, even as the city is considered one of the most linguistically diverse urban areas in the world. This report considers the need for language access on NYC transit systems and the current state of service for Limited English Proficiency (LEP) riders. Overall, language access in the subway can and should be improved, and likely presents a barrier for LEP New Yorkers accessing equal economic and social opportunities. This work considers select language access improvements that would serve a substantial number and proportion of riders without excessively burdening the MTA, and best practices for language access from elsewhere that could apply to NYC’s uniquely challenging transit environment. 

 

Background

LEP residents account for nearly 25% of the New York City (NYC) population, or 1.8 million people (NYC Department of City Planning). LEP New Yorkers are highly concentrated in specific areas: almost half live in just one-fifth of NYC’s residential neighborhoods (Jeng, 2022a). By contrast, those 40 neighborhoods contain only 27% of the total NYC population.

 

The 40 highlighted neighborhoods contain nearly half of the city’s LEP population. Just the 10 neighborhoods with the most LEP residents, labeled, contain nearly 20% of all LEP residents in NYC. Source: American Community Survey 2016-2020 data by neighborhood tabulation area (US Census, NYC Department of City Planning).

 

Top 10 Neighborhoods by Number of Limited English Proficiency (LEP) Residents

Neighborhood Tabulation Area (NTA)	Borough	Residents (5+) with Limited English Proficiency (LEP)	Total Population	Percentage of Residents (5+) with LEP
Elmhurst	Queens	45,724	98,560	49.10%
Bensonhurst	Brooklyn	42,633	96,331	47.80%
Flushing-Willets Point	Queens	38,468	59,237	69.30%
Jackson Heights	Queens	36,047	89,628	42.80%
Corona	Queens	31,971	69,333	49.90%
Washington Heights (South)	Manhattan	28,505	81,271	36.60%
Murray Hill-Broadway Flushing	Queens	28,424	54,362	54.90%
Sunset Park (Central)	Brooklyn	26,432	44,187	64.70%
Washington Heights (North)	Manhattan	25,667	80,831	33.40%
Gravesend (West)	Brooklyn	25,070	56,148	48.10%

 

The top 10 neighborhoods in terms of LEP population. See Appendix for top 40 neighborhoods. Data Source: American Community Survey 2016-2020 data by neighborhood tabulation area (US Census, NYC Department of City Planning).

 

Public transportation riders with limited English proficiency are insufficiently served in terms of language access on the MTA’s transit services (namely subways and buses, but potentially also Access-A-Ride, which is not covered in this analysis). Most signage and announcements are in English only, though there are some signs in other languages. The MTA also places Braille text on signs that are placed on station pillars. Planned service change notices, both within stations and sent over email and text, are commonly translated, and the languages chosen for display seem to be tailored to the demographics of the local community. However, there is currently no publicized plan to extend the translation process to announcements, permanent signage, or other more constant forms of wayfinding, despite calls from legislators to do so (Kaufman). 

 

The MTA has made efforts to make on-demand language assistance more accessible. Riders can call 511 to speak to a customer service representative who can request an interpreter from a third-party vendor. Riders can also request information about service changes or help with trip planning in over 100 languages using WhatsApp or iMessage. These text-based services, like the MTA’s website, make use of Google Translate to allow the MTA’s staff to assist in languages other than English. WhatsApp in general is popular among speakers of Spanish and other languages, so it is likely a good strategy for reaching LEP riders. According to an MTA representative, the agency received 15,586 WhatsApp messages in November 2022, of which approximately 3% were in a language other than English.

 

Still, transit users often need to make quick, context-specific decisions that are better served by wayfinding infrastructure than through time-consuming calls or texts, which may be unable to help with location-specific inquiries such as, “Am I standing on the right platform?” That may explain why asking fellow riders seems to be the most common means of obtaining assistance, anecdotally speaking. Nonetheless, these forms of on-demand language assistance are important to continue improving, especially for users of less-commonly spoken languages that are less likely to receive planned translations for physical or digital materials. 

 

Example of multilingual planned work and service change notice at East Broadway. Information is available in English, Spanish, and Simplified Chinese, likely based on the ridership demographics of the station. Source: Sonia Wu. 

 

As a recipient of federal financial assistance, the MTA is required to adhere to Title VI of the Civil Rights Act of 1964, or in other words to “take responsible steps to ensure meaningful access to the benefits, services, information and other important portions of their programs and activities for Limited English Proficient ("LEP") persons” (MTA, 2019). The on-demand language assistance options discussed above, and outreach in multiple languages for projects like its bus network redesigns, represent steps the MTA already takes. It is possible that the agency considers changes like additional translated signage or multilingual announcements to fall short of being “reasonable steps”—because they would not serve a substantial number of riders or would use too many resources, for example. 

 

Ample evidence supports the importance and necessity of improving language access on MTA services. Ridership on NYC public transportation remains below pre-pandemic levels as a whole, but has disproportionately rebounded in working-class neighborhoods, often those with large immigrant populations (Martinez and Bhat). While immigrant or foreign-born status does not necessarily equate to language access needs, subway ridership data also shows a correlation between neighborhoods with more LEP residents and stations that have recovered (or kept) a greater share of their 2019 ridership (Jeng, 2022b). Moreover, stations in these neighborhoods have increased their share of total subway ridership since 2019, meaning they are now even more important in terms of farebox revenue than before the pandemic. These neighborhoods’ residents tend to be more dependent on public transportation to commute to work than residents in the median NYC neighborhood, according to Census data (Jeng, 2022a). 

 

Looking at ridership totals instead of percentages, the top quintile of subway stations in terms of neighborhood LEP population saw more than 375,000 entries on an average day from 10/8/22 to 11/05/22 (Jeng, 2022b). The “post-COVID” daily ridership of just those stations is, for comparison, greater than even the pre-pandemic weekday ridership of the entire LIRR network. Clearly, a significant number of riders would likely benefit from improvements to language access. Many of these improvements could be made as low-cost additions in tandem with planned ADA improvements.

Subway stations color-coded by their neighborhood’s population of LEP residents. The 20% of stations with the most LEP residents in their neighborhoods (those in the darkest shade of blue) saw more entries on an average day between 10/8/22 and 11/05/22 than the entire LIRR system on an average pre-pandemic day. Created using MTA and American Community Survey data. 

 

For an individual LEP resident, poor language access could mean getting on the wrong train, missing the intended stop, getting lost, or being late—scenarios which cause stress and a tangible loss of time (or potentially even wages or employment). On a citywide scale, this lack of language access is likely causing losses in economic productivity from LEP residents. As New York rebounds from the pandemic, the city and the MTA would likely also benefit economically from improving language access for non-residents, especially foreign tourists, who are potentially choosing other forms of transportation due to inability to navigate transit. 

 

Besides economic considerations, improving language access on public transportation is important for achieving more equitable health and social outcomes for LEP New Yorkers. Given the importance of public transportation for most LEP residents, difficulty navigating the transit system could translate into fewer trips, increased isolation, and a smaller range of movement within the city. According to the NYC Health Department, language barriers can lead to a higher risk of loneliness and other mental health issues. An LEP resident may eventually memorize a trip to work and back home, but visiting new or unfamiliar parts of the city is more difficult, and may result in a reluctance to travel, visit medical specialists, or explore new job opportunities. In addition, an LEP resident may turn to more expensive car services, which will not be sustainable for every trip. Traveling to see family or friends or even just for leisure should be just as much of a possibility for LEP New Yorkers as it is for English speakers.

 

Considering Language Access Improvements:

Implementing Multilingual Announcements

Neighborhoods with the most LEP residents in NYC contain speakers of numerous languages. However, in many of these neighborhoods, the majority of LEP residents speak either Spanish or Chinese (including Mandarin and other dialects) as their primary language. This is the case for the top 10 neighborhoods with the most LEP residents in the city (Jeng 2022a). A citywide approach to language access should begin with those two languages. But even more tailored solutions are possible by utilizing knowledge about individual stations’ demographics (as the MTA already does when translating service change notices). Stations located within the most “LEP-dense” neighborhoods could be a good starting point for implementing language access improvements.

Languages spoken by LEP residents in the top 10 neighborhoods in terms of LEP population. Either Spanish or Chinese is the most common primary language of LEP residents in all of them. The two languages combine for more than half of LEP residents’ primary languages in all 10 as well. Data from American Community Survey 2016-2020 by neighborhood tabulation area (US Census and NYC Department of City Planning).

 

Upgrading automated announcements on subway cars and buses to feature additional languages based on the needs of local populations might be the most immediately feasible way to improve language access on the MTA’s transit network. On the subway, custom announcements on New Technology Trains (those with pre-recorded announcements) have been done before: on the 7 train, announcements were recorded by SNY broadcasters for the New York Mets and by the actress Awkwafina to promote the TV show Nora from Queens. In addition, the MTA tapped a broad range of NYC celebrities to record announcements for stations, subway cars, and buses during the COVID-19 pandemic. These efforts prove that announcements outside of the standard script are possible and can be broadcast at specific stations or transit services. Moreover, just as the aforementioned campaigns were temporary, multilingual announcements can be adjusted or swapped out based on feedback as needed.

 

The first challenge to the implementation of multilingual announcements is the need to establish consistent translations. As seen in the image of translated signage above, the MTA already has a process of translating information to fit a station’s demographics while keeping station names in English. The process could be adapted for recorded announcements, with the proper amount of vetting by trained translation professionals. Feedback on translations could also be gathered from riders through surveys within the system and neighborhood-level outreach. It is likely impractical to translate non-numbered station names, but even if those remain in English, the recording of translated directional information (such as “Manhattan-bound”) and numbered station names will already be an improvement.

 

The second challenge is the potential for longer announcements to increase dwell time at stations. This challenge is tied to the questions of how many languages to include and which types of announcements to translate. As discussed above, the majority of LEP speakers in neighborhoods with the most LEP residents speak a select few languages (Jeng, 2022a). Thus, adding announcements in one or two languages (namely Spanish or Chinese) could reach the majority of LEP residents at many stations—still a compromise, but one that would improve upon the status quo and could be supplemented by other forms of language access for the languages that aren’t included in announcements. The primary audio announcements that could be translated are those announcing the upcoming stop (including transfer and accessibility options) and those announcing a departing train’s service, direction, and next stop. The latter type presents more of an issue for dwell time, since it is played while a subway car is in the station and precedes the doors closing. Possible solutions include shortening and speeding up the relatively wordy English announcements and including only essential information in the translated announcements. As the often long-winded SNY announcements showed, adding more information is doable. Other types of announcements, such as public service announcements, can also be translated but they are lower priority.

 

The MTA can draw from elements of multilingual announcements in both United States and international transit systems, each with their own method for incorporating multiple languages. For example, Houston’s light rail and buses feature English and Spanish announcements. On the light rail, English and Spanish are combined into one statement for brevity: “next station, siguiente estacion, [station name in English].” A similar structure could be adopted for buses and subways in NYC when conciseness is needed and the station name is not translated. In Taipei, upcoming stations and transfers are announced on Metro trains in four languages as required by law, while information about trains arriving at a station are announced by the station’s PA system in Mandarin and English (Taipei Rapid Transit Corporation). The MTA could similarly feature more languages in announcements between stations, while limiting announcements that may affect dwell time to fewer languages. The São Paulo Metro announces information about the next station in Portuguese and English after the doors close, a timing choice that seems to be common in many systems outside of NYC. That would represent a more radical solution to the dwell time issue, but could pose an issue for the NYC subway because of the need for boarding riders to quickly determine if they are on the right service when there are multiple lines stopping at the same station. Some combination of these methods could be developed. For example, station platform audio and digital signage announcements could include information in multiple languages about upcoming trains that might be too lengthy if announced in full on board.

 

Updated announcements may only work on newer train models, such as the R188 and R179, but more could be rolled out over time as the fleet is updated. Given the specific nature of announcements, a pilot program could be run to test and gather feedback for multilingual announcements before rolling them out to other parts of the system. The 7 line, with its newer rolling stock, relatively simple service pattern, history of accommodating prior custom announcements, and the high linguistic diversity of its riders, could be a good option. Alternatively, individual stations or bus lines with higher numbers of LEP riders could be targeted for trial. Furthermore, with the MTA’s new program to place station booth attendants throughout stations, staff with multilingual capabilities can be prioritized at these stations.

 

Implementing Multilingual and Language-Agnostic Signage

Increasing the amount of multilingual signage is an equally important yet arguably more difficult aspect of improving language access. Currently, there is virtually no translation of permanent signage within stations, which share information like off-peak service hours, platform stoppage locations, and signs pointing to police stations. Avenues for improving wayfinding signage include adding physical signage, utilizing digital screens, adding QR codes that point to audio translation of signs, and enhancing the entire wayfinding system to be more language-agnostic.

 

Adding multilingual signage indicating the next stop and/or direction of a platform’s trains would help address a commonly asked question—”where is this train going?” Platforms with multiple services present logistical challenges, but these can be solved; the same platforms already display text-based, often wordy explanations of service patterns. A reworking of wayfinding information for language access purposes would thus improve navigability for English speakers as well.

 

Example of existing NYC subway platform signage explaining the service pattern of trains stopping there. Wordy and English-only, it is confusing for those unfamiliar with the system.

 

The proliferation of digital screens in the subway and bus system provides an opportunity to offer wayfinding information (such as line maps reflecting current service patterns) in languages other than English without the need to install new physical signage. These screens could rotate between languages, just as some screens already rotate between PSAs in different languages. While translating written station names is likely impractical, translated directional information (“Manhattan-bound” or “North-bound”, for example) could also be implemented on newer buses and trains with the capability to display custom text digitally. Using the flexibility of digital signage is especially important in the context of changing demographics—a neighborhood could shift to a different set of languages over time, so having the option of quickly and inexpensively changing signs is a major benefit over physical signage, which can be costly to install or replace.

 

Enhancing the MTA’s wayfinding system with language-agnostic information could potentially improve language access more universally than simply translating wayfinding signage piecemeal. As with announcements, the MTA could learn from international best practices, such as the use of station numbers in the Tokyo Metro (see image below). Station numbers already used internally at the MTA could be adapted for this purpose. 

Example of bilingual signage in the Tokyo Metro indicating the current stop, previous stop, and next stop on a platform, with language-agnostic station numbers. Original image source: Andrew Alexander Price. Annotations added.

 

Other possibilities for language-agnostic wayfinding include the expanded use of pictographs or symbology, as is common in some other systems such as Mexico City’s (see image below). While such comprehensive changes would require substantial internal advocacy, technical research and resources, and external outreach, they have the potential to solve issues that translation on its own cannot. They would also address the sheer range of languages spoken in the city that makes it impossible to serve all LEP riders using pre-planned translations. Moreover, universal, language-agnostic wayfinding can last even if the specific languages used change over time.

Example of pictographic signage in the Mexico City Metro, with pictographs providing an alternate means of understanding both Salida (exit) and Politécnico station. Image source: David Ostrowski. Annotations added.

 

Station numbers or other forms of language-agnostic information combined with next/previous stop information could allow someone without English proficiency or knowledge of relative station locations to find the right platform. Multilingual strip maps—an expansion of the digital strip maps that exist already—within stations like those in London, could further complement those improvements and help riders find the right direction to get to their destination station without already knowing where it is relative to their current station. These navigational aides would benefit LEP riders and English-proficient New Yorkers alike. While such systems would take time and resources to implement, and would need to be adapted to fit the NYC transit system, possibilities for language-agnostic wayfinding should not be dismissed simply because they are somewhat radical departures from the status quo. The NYC transit system has made major wayfinding improvements before, such as font standardization and changes to line colors in the decades following subway unification (New York Transit Museum). 

 

New Roles for Multilingual Ambassadors

 

Transit employees who speak the languages riders speak should be considered for situations when talking to another person is the best option. At stations or stops with high LEP ridership and/or complex service patterns, there could be a more coordinated effort to have trained staff fluent in the languages spoken in the area available for assistance. The MTA likely has numerous employees in various roles who speak multiple languages. The agency recently announced plans to open Customer Service Centers at select stations and to have station agents leave their booths to become more public-facing station ambassadors who can help with wayfinding or other customer experience needs. If there are agents fluent in multiple languages, a formal language access plan could allow them to assist in multiple languages either themselves or by contacting an interpreter. This plan could incorporate relatively low-cost communications materials like posters with language options that customers can point to or pins for agents that share what languages they are able to speak. The new Customer Service Centers could be ideal locations to pilot an enhanced language access plan. 

 

Conclusion

Improvements to language access will require a combination of technical best practices, management buy-in, capital investment, and community outreach. There are certainly challenges to implementing each of these potential improvements, including cost, timing and physical space. In addition, performing regular multilingual community outreach (as it does for large capital projects) would be essential for gathering feedback from actual LEP riders. Yet given how NYC’s transit system and its LEP community are in many ways interdependent, especially post-pandemic, better language access would help make transit more equitable and more economically productive citywide. For many riders, it would simply mean a better experience and possibly even more trips beyond their regular travel patterns. With the benefits of increased ridership, safety and navigability of the transit system, the relatively low cost investment of enhancing language access (compared to other capital investments) seems well worthwhile.

 

Appendix

 

Top 40 NYC Neighborhoods by Number of Limited English Proficiency (LEP) Residents

Neighborhood Tabulation Area (NTA)	Borough	Residents (5+) with Limited English Proficiency (LEP)	Total Population	Percentage of Residents (5+) with LEP
Elmhurst	Queens	45,724	98,560	49.10%
Bensonhurst	Brooklyn	42,633	96,331	47.80%
Flushing-Willets Point	Queens	38,468	59,237	69.30%
Jackson Heights	Queens	36,047	89,628	42.80%
Corona	Queens	31,971	69,333	49.90%
Washington Heights (South)	Manhattan	28,505	81,271	36.60%
Murray Hill-Broadway Flushing	Queens	28,424	54,362	54.90%
Sunset Park (Central)	Brooklyn	26,432	44,187	64.70%
Washington Heights (North)	Manhattan	25,667	80,831	33.40%
Gravesend (West)	Brooklyn	25,070	56,148	48.10%
Sheepshead Bay-Manhattan Beach-Gerritsen Beach	Brooklyn	22,832	63,752	37.80%
Borough Park	Brooklyn	22,146	78,836	33.60%
Concourse-Concourse Village	Bronx	20,990	66,082	33.80%
North Corona	Queens	20,936	39,263	58.50%
Chinatown-Two Bridges	Manhattan	20,316	41,236	51.00%
Coney Island-Sea Gate	Brooklyn	20,081	48,091	44.00%
Jamaica	Queens	19,870	55,891	38.60%
Bedford Park	Bronx	19,016	55,702	37.20%
Ridgewood	Queens	18,897	70,329	28.50%
Forest Hills	Queens	18,577	87,466	22.70%
Woodside	Queens	18,460	45,417	43.00%
Mount Eden-Claremont (West)	Bronx	18,226	50,073	39.80%
Soundview-Bruckner-Bronx River	Bronx	18,219	68,432	28.90%
Sunset Park (West)	Brooklyn	18,065	51,056	38.00%
Brighton Beach	Brooklyn	17,539	29,819	63.20%
University Heights (North)-Fordham	Bronx	17,158	45,235	41.00%
East Flushing	Queens	16,721	29,714	59.10%
Gravesend (East)-Homecrest	Brooklyn	16,533	51,665	35.40%
University Heights (South)-Morris Heights	Bronx	16,001	53,859	32.40%
Dyker Heights	Brooklyn	15,863	41,656	40.50%
Sunset Park (East)-Borough Park (West)	Brooklyn	15,413	29,563	56.40%
Sunnyside	Queens	15,399	48,609	33.50%
Mount Hope	Bronx	15,086	46,487	35.50%
Bay Ridge	Brooklyn	15,025	80,183	19.90%
Mott Haven-Port Morris	Bronx	13,889	53,469	28.20%
South Williamsburg	Brooklyn	13,526	45,664	36.20%
Midwood	Brooklyn	12,973	50,616	28.50%
Norwood	Bronx	12,862	39,290	35.20%
Madison	Brooklyn	12,502	40,227	33.40%
Inwood	Manhattan	12,464	41,450	32.20%

Data Source: American Community Survey 2016-2020 data by neighborhood tabulation area (US Census, NYC Department of City Planning).