Bus Stops & Perceived Safety Through a Gender Lens

A good transit system is one people can rely on for both the essential and non essential trips that are part of everyday life, but most importantly, one which people see as safe. The goal of the project is to find whether gender plays a role in how safe people feel waiting for the bus and whether safety perceptions are different in the daytime vs nighttime. This project is made up of research as well as personal anecdotes and observations in an effort to make this kind of work more relatable to folks. The issues discussed here are near and dear to my heart, so I'd like to share them as plainly as I can so that others may also see the importance of these topics or perhaps see themselves in this work.

This is also an effort to build on the work of others and expand the definition of accessibility. Accessibility goes beyond physical factors and includes "invisible" social factors as well, such as how safe people feel waiting for their bus. The hope is to normalize having gender related data in transit planning and as a topic of discussion in advocacy circles so that we can gain more insight into the rider experience and continue to improve it and make it accessible for more people.

There’s been many studies in the Boston area that have identified the differences in bus ridership by race, such as  Free Buses Advance Equity,  particularly in light of the pandemic when it became unmistakably clear many essential workers are people of color who rely on the bus as their main mode of transportation. It’s important that we don’t stop there as there are further inequalities even within this group.

In an effort to achieve a more nuanced view of the role of public transportation in people's lives, more research could include qualitative research elements such as surveys and interviews. In  Evaluation of Bus Reliability Measures and Development of a New Composite Indicator , the authors state indicators that rely on arbitrary thresholds to define reliable service work well to classify lines but have little relation to the travellers perspective. It's important to pair both the technical elements of transit planning with the experiences of riders and in this way help test the fidelity of the models to the real experiences of riders. 

This project takes a micro scale approach and focuses on bus stops along a single bus route- Massachusetts Bay Transportation Agency, MBTA, bus route 1 which goes from the Harvard Yard in Cambridge, Massachusetts to Nubian Station in the Boston neighborhood of Roxbury. It’s important to note the stark difference in socioeconomic status of both of these places. The  median household income  for the zip code where Harvard Yard is located, 02138, is $75,446, while that of the Nubian Square zip code, 02121 is $28,238. The route goes through important corridors such as Massachusetts Avenue for most of the way and passes by important sites such as Boston Medical Center and Northeastern University as it makes its way to Back Bay, crosses the Charles River, leaves Boston and enters the Massachusetts Institute of Technology campus in Cambridge and finally ends at the Harvard Yard. The route covers a one way distance of about 4.45 miles and serves a total of 46 bus stops.  

From personal experience, I can say that the surrounding environment can play a large role in how comfortable and safe I feel waiting for the bus. As a student in a public transportation class at Northeastern University, I learned about the concept of Potential Waiting Time developed by Professor Peter Furth. It is the "name given to the difference between budgeted and mean waiting time." -Furth. This means that if a bus is unreliable, people will "budget" in extra time for their trip to have some wiggle room and ensure they don't miss their bus.

Part of the concept is adding a weight to the time spent waiting for the bus. However, after having lived without a car in Boston, where the winters are long and the sun sets at 5pm, I believe it’s worth exploring if perhaps different weights could be applied for waiting time during the daytime vs nighttime. Does a place that feels safe to wait at during the daytime also feel safe during the nighttime? How much of a role do amenities such as trees, lights and bike lanes play in how safe we feel?  

The MBTA is currently undergoing a  Bus Network Redesign  Process and I believe it’s important to include the needs of folks who identify as women and expand our equity lense into one that includes both race and gender. As a woman of color who uses transit, I want a better experience not only for myself, but for all women. I want our experiences to be recorded and made visible so that we can have a baseline to work from and ultimately create better outcomes that are more in tune with our needs. Transit is a public good which will only improve when it meets the needs of its users and for that we must better understand who the user is.  

I conducted a bus stop safety audit via a text survey and coupled the responses with the Plan for Accessible Transit Infrastructure (PATI) dataset from the MBTA which is an inventory of the amenities located at each of the bus stops in the MBTA's service area. "Each bus stop has information about the location of the stop, sidewalk condition, if shelters or benches are present.."   mbta-massdot.opendata.arcgis.com  

How the safety audit was conducted

I used the platform  SlickText  to conduct a perceived safety audit via text at bus stops along MBTA Route 1. I also created an email where folks could contact me about the project with any questions,  bus.stop.audit@gmail.com . The purpose of the survey was to ask riders how safe they felt waiting at a specific bus stop during the daytime vs nighttime and roughly about how many minutes they spent waiting for the bus.

I hung up large posters with the project and survey details at 33 out of the 46 bus stops along the route. The posters were quite heavy since they were made out of durable weatherproof material and I used a grocery cart to carry them which made transporting them a bit of a hassle. To hang them up I used double sided weatherproof tape for bus stops with a shelter. For stops with just a pole, I used a hole puncher to add holes at the top and bottom of the posters and zip ties to tie them to the poles.  

I chose not to hang up a poster at certain stops since there were many people around and I didn’t feel comfortable. This was the case with the stops near Central station and a couple near Nubian Station. 

The posters were up for roughly a 2 week period- June 29th - July 16th. During this time there were heavy rains in the Boston area and I believe the winds blew away the posters that were hanging on the poles since when I went to check on them the posters were gone, but the zip ties were still present, leading me to believe no one cut the posters, but rather they flew away.   

The text survey was available in the following languages:

• English
• Spanish
• Haitian Creole
• Portuguese

Survey questions: 

Message 1. Hi! Welcome to the bus stop safety audit. 

• You must be at least 18 years old to be in this Northeastern University research project (IRB#21-04-25) 
• This survey is voluntary and will take less than 5 minutes. 
• Your part in this study will be handled in a confidential manner. Any reports or publications based on this research will use only group data and will not identify you as being of this project.
• You will be entered to win one of ten $50 Visa gift cards. Winners will be notified on July 17th. 
• Standard messaging and data rates apply.
• If you have any questions about this study, please email the researchers: bus.stop.audit@gmail.com
• If you have any questions about your rights as a research participant, please email Northeastern University's Office of Human Subject Research Protection: irb@neu.edu
• Let’s get started! What is the name of the bus stop you’re texting from? (you can find the name on the flyer).   

Message 2. Do you identify as a woman?

• Yes
• no
• prefer not to answer

Message 3. How safe do you feel waiting at this bus stop when the sun is out?

1. very safe 
2. safe
3. neutral
4. not safe
5. not safe at all
6. I don't use this bus stop when the sun is out
7. Prefer not to answer

Message 4: How safe do you feel waiting at this bus stop when it's night time?

1. very safe 
2. safe
3. neutral
4. not safe
5. not safe at all
6. I don't use this bus stop at night time
7. prefer not to answer

Message 5: About how many minutes do you usually wait at this stop before the bus gets here? (just text the number of minutes or text NO if you don't usually use this bus stop)

Steps taken - cleaning survey data

Here are the steps I took to organize and clean the dataset

• Downloaded English and Spanish CSV from the Slicktext platform
• Took out date, birthday, email columns (these were empty)
• Added column for daytime and nightitme score, and convert people’s responses to a 1-5 scale
• For rows with an “I don't use this stop at night” I left the safety score blank 
• For “average waiting time (perceived)”, one person,857-492-7089, gave a range, so i split the difference.
• Changed the NOs to blanks
• Removed incomplete entries, 9 from English and 1 from Spanish 
• Changed Nubian stop ID from 25 to 64 (Originally I couldn’t find the ID for Nubian square so I chose #25 and later discovered the ID is 64)
• Added lat lon coordinates to corresponding bus stop IDs, use same ones from PATI dataset
• Split into inbound and outbound 
• Combined english and spanish CSVs
• When I was making the graphs and comparing safety feeling for night and day for women identifying folks and non women identifying folks, some bus stop IDs only had one response so I decided to exclude these and only count stops with at least 3 or more responses.   

Steps taken- Analyzing and visualizing survey data

I used the free programming software, R, to filter the survey dataset and create the different boxplot and bar graphs. I then used  Canva  to add icons next to some of the bar graphs and add additional informational text to the graphs.

Steps taken- making the map

Steps I took to add the PATI bus stop amenities dataset to ArcGIS

• Contacted the MBTA’s OPMI office to obtain the PATI dataset 
• Used Excel to separate the different types of amenities into their own tabs. It's important to note I did not map out all of the physical amenities provided to me by the OPMI office. I left out traffic signs, hydrants, newspaper boxes, and mailboxes, I originally planned to include planters, but it turned out non of the stops had one listed. I also excluded measurements such as height of seating, length of shelter, slsope of shelter pad, etc.
• Used google maps to map out the amenities next to the corresponding bus stop ID and gave them enough space so they could all be visible when zoomed out and could be seen at a glance
• Used QGIS to convert the KML google maps files to shapefiles 
• Uploaded shapefiles containing the amenities’ locations to ArcGIS 
• Used the  Flaticon  to find free icons to represent each different kind of amenity on ArcGIS

116 people took the complete English survey

2 people took the complete Spanish survey.

As a whole, participants reported feeling less safe waiting for the bus at nighttime.

115 out of the 118 total respondents shared how they self identify. The number of people who identified as women was 78 vs 37 who did not identify as women (78+37=115). They each gave responses for daytime and nighttime with some stating they did not use the bus stop at nighttime. This accounts for the differences in the number of daytime vs nighttime responses. While this survey sample is small, it is a start in beginning to understand how different groups of people experience public transit and also beginning to document and track people's perception of safety so that we may begin taking steps to close the gap between how safe different groups feel while waiting for their bus.

Folks who identify as women reported feeling the safest in the daytime at stops 74 and 67. The each tied with a score of 4.67. In the safety feeling scale 4 means feeling safe and 5 means feeling very safe. Folks who do not identify as women reported feeling the safest at stop 74, with a score of 4.67 as well.

What amenities are here? Both stops are located in Cambridge near major university campuses and have a shelter with seating available and a bike lane running past it. Stop 67 is located near the Harvard Yard and also has two trash cans, a tree and a bench as amenities. Stop 74 is located just outside the MIT campus. From anecdotal evidence, both stops have high foot traffic, likely due to the number of restaurants, businesses and universities nearby.

Folks who identify as women reported feeling the safest waiting at stop 73 at nighttime, score of 3.71 out of 5, which falls between feeling safe and very safe on the safety feeling score scale. Stop 67 had the highest safety feeling score, 3.67, from folks who do not identify as women.

Both of stops with the highest safety feelings scores are located in Cambridge and have a shelter with seating and a bike lane. In addition, Stop 73 also offers two bike racks, three trees and a trash can. Stop 67 has a bench, two trash cans and a tree.

Bus stop 10590 had the lowest safety feeling score from both groups during the daytime. It has a seven eleven behind it and what looks to be an abandoned building across the street. As for amenitites, it offers two trash cans, a bench, a light and a bike lane. The sidewalk is standard width. By just looking at the physical space it’s hard to say what factors are having an influence on people’s safety perception. More research such as direct interviews and or/ethnographic research would have to be done to really get a good understanding of what’s affecting how safe people feel waiting for the bus here.

It's worth noting that this bus stop has the same safety feeling score during the daytime as the lowest night score from both women and non women identifying folks, a score of 2.

In the daytime, there is a greater point difference between the safest feeling bus stop for folks who do not identify as women than there is folks who do identify as women- 1.99 vs 1.52. Bus stop 74 scores the highest among both groups even though it offers less amenities than the other stops, except for 89. What else could be going on here? It's important to note stop 74 is located near a university campus, MIT, which could mean higher foot traffic at various hours of the day since college students often have what could be considered irregular schedules, not the 9-5pm workday.

The five stops with the highest scores are all located in Cambridge, while the bottom six stops are located in Boston. The stops with the lowest safety scores, 10590, 58, and 57, are all right next to each other. This mini corridor running along Boston Medical Center, could perhaps be considered a hotspot or point of interest that calls for further research and resources.

At nighttime we see a greater point difference than we did in the daytime for both groups (women and non women identifying folks) - 1.71 points. This could mean a greater variation in the environments the bus stops are located in or a greater variation at night in factors folks consider to induce feelings of safety, or perhaps a combination of the two. Both 73 and 67 are located near university campuses, MIT and Harvard, respectively, while 58 and 6 could be said to near Northeastern University, they are not as near to a campus as 73 and 67. Perhaps during the nighttime, there may be a greater range of factors that induce feelings of safety in people. While we see that the stop with the most amenities, 73, also has the highest safety score for women identifying folks, this doesn't necessarily translate to more amenities equaling greater feelings of safety across the board. For example, stop 58 has two shelters, as opposed to the single one stop 73 has, and yet it's tied with stop 6 for the least safe feeling bus stop on the scale. And we see that stop 74, with just two amenities is near the top of the scale.

At this point it's important to look beyond the physical infrastructure like amenities, and look to the social environment as well. The top four scoring bus stops are all located in Cambridge, while the lowest four scoring bus stops are all located in Boston. Of most concern is that the lowest three are located near the Roxbury neighborhood. While bus stop 6 has a shelter with seating as well as trees, it's located along a five lane road and the sidewalk is located behind the stop. When I visited the stop to hang up the poster trash was scattered all over the ground. When I again went to the stop to remove the poster, the stop was still littered with trash and old clothes.

Bus stops 58 and 57 are both located near Boston Medical Center and a Homeless Services Intake facility. If I was just going off of amenities, I would have expected stop 58 to have a much higher safety score since "on paper" it looks to have all the ingredients of a great bus stop! But after visiting the area in person, I see there is more at play here. A high quality bus stop goes beyond amenities. Public transportation is a social service that is interconnected with many other social services and so is it's quality and accessibility. When we have heavy issues such as homelessness going unaddressed, it affects not only the quality of life of the people going through that hardship, but also the quality of life of others who may not be using services like public transit to access opportunities due to safety concerns. It's not enough to advocate for more amenities, like trees and bus shelters as these may just be drops in the bucket of what is a much bigger and nuanced issue.

Accessbility goes beyond physical infrastructure. According to the MBTA's Plan for Accessible Transit Infrastructure (PATI) database, bus stops 73, 57 and 58 all have a score of "Medium", yet there is a 1.6 point difference in terms of feelings of safety between these two stops. While this may not seem like a large lap, I believe it is an important one that could receive more attention, as it is currently not receiving any. If the goal is for transit infratructure to be accessible, we must look beyond the physical factors and begin looking at the entirely of people's transit experience.

Stop 67 scored highly for both daytime and nighttime across both identity groups. In person interviews could be done at this stop to get a better sense of what exactly about the environment here is having this effect on folks in an effort to replicate this at other stops if possible.

Deeper dive into the safety feeling scores of folks who identify as women

Not only did stop 6 score the lowest for nighttime (score: 2) but it also had the greatest difference between daytime and nighttime scores - a 1.67 point drop in how safe folks who identify as women feel waiting at this stop during the daytime vs the nighttime. Located behind this stop is what appears to be a private parking lot for a business of sorts.

It's interesting that stop 74, which scored the highest for daytime, has a drop of 1.33 points (about the middle among all the drops). This stop is located closer than stop 73 to the MIT campus, but does have less amenities than 73. Perhaps the parking lot located behind it could also have something to do with this shift in daytime vs nighttime perceived safety. Not many of the bus stops along route 1 have parking lots near them. Perhaps a future study could look at a route that does in an effort to test for correlation between parking lots and a sharp drop in daytime vs nighttime perceived safety.

The three stops with the greatest difference in daytime vs nighttime scores are located in Boston, 6, 89, 57, and the three stops with the least difference are located in Cambridge, 73, 68, 67.

Stop 89 stands out because it's located near the Northeastern University campus and on the corner of a busy pedestrian intersection, which I thought might mean more people at various hours of the day walking by. It could be worth doing in person interviews here as well to see what else is at play.

Stop 73, which scored the highest for nighttime among the folks who identify as women, also had the least difference, .59, in daytime vs nighttime scores in this group. However, it did not score the highest for daytime scores, coming in at fourth place

Deeper dive into the safety feeling scores of folks who do not identify as women

Only stops 67 and 57 had at least 3 or more responses from folks who do not identify as women for both daytime and nighttime and thus it is harder to get a clear picture of what factors could be contributing to the drop between daytime and nighttime scores at the micro level (individual bus stops). But we can at least compare the score differences for these stops between the two identity groups.

Stop 67- day vs night difference in perceived safety feeling scores

• identify as women: 1.07
• do not identify as women: .83

Stop 57- day vs night difference in perceived safety feeling scores

• identify as women: 1.42
• do not identify as women: 1

At both of these stops there is a drop in safety perception for both groups, with the women identifying group having the greatest difference between daytime and nightime scores. During the daytime folks who identify as women gave a higher safety score for both stops than the non women identifying group.

Stop 67 daytime score

• identify as women: 4.67
• do not identify as women: 4.5

Stop 57 daytime score

• identify as women: 3.67
• do not identify as women: 3

Both of these findings are in line with the general trend discussed at the beginning of the results portion stating that those who identify as women tend to feel safer than non women identifying folks during the daytime, but both identity groups experience a drop in feelings of safety between daytime and nighttime and this drop is greater in the group that identifies as women.

See what physical amenities are present at each stop along MBTA bus route 1 in the northbound direction as well as the daytime and nighttime safety scores for folks who identified as women.

This map only shows northbound stops.

Go to the full interactive map:  https://bit.ly/bus_stops_perceived_safety_map  to access the southbound stops, southbound amenities and the safety scores for non women identifying folks.

Overall both groups feel less safe at the bus stops during the nighttime, but the drop in safety perception is greater among the group who identified as women, While the numbers themselves may not seem large, a 1.16 difference among non women identifying folks vs a .79 point difference among folks who identify as women, the bigger picture here is that a difference does appear to exist and this is worth exploring further in an effort to better understand the rider experience and improve transit service for all people.

Currently the  MBTA data dashboard  offers information on reliability segregated by mode as well as peak and off peak hours. "Peak hours are from 7:00-9:00am and from 4:00-6:30pm on weekdays" -  mbtabackontrack . While this information does offer a window into how reliable the bus is, it's hard to easily glean information about bus reliability for nighttime hours, which as we saw, is when people from the two indentity groups in this safety audit report feeling less safe at.

"Every month the MBTA surveys a panel of its customers about their experience on our services."  MBTA Customer Satisfaction . None of the questions on the website relate to safety perceptions. Perhaps the team could explore adding additional questions that relate to safety, which is a critical part of the rider experience.

Out of the  ~20 questions  in the last  MBTA Systemwide Passenger Survey  (2015-2017) one question is about gender and none are about safety perception. One of the goals listed is, "The MBTA will use the results of this census to plan future service and to make sure service is provided fairly." How can we plan future service when we are not truly getting to know our riders? We must go beyond asking about origins and destinations, and start asking about other parts of the rider experience, such as comfort, safety, ease of use, etc. It may seem radical, but perhaps it's time we embed more empathy into our transit planning and begin understanding both the travel and thought patterns of our riders. In  An Overview of Women's Transport Issues in Developing Countries , the authors call to "reduce the transport burden, particularly of women and children, by consulting the women and increasing their representation in transport planning and decision taking."

If we only look at the physical infrastructure and technical aspects of public transit, we will miss key insights into what our riders experience day to day. We can't just add amenities to a bus stop and assume we have now made it safe. We must also engage with our riders in a more humane way and get to know them as people with varying needs, abilities and concerns. As the father of Latino Urbanism, James Rojas, often says, "It's about relationships, the relationships we have with the people around us."

When it comes to public transit, I truly believe our relationships with our physical environment go hand in hand with our relationships with each other in how we experience our cities and move through them.

We can explore physical, relational and technical aspects together and in this way get a better handle on what riders are experiencing and what we can do to improve their experience.

The safety audit conducted for this project used a small sample size to explore the topic of safety at bus stops through a gender lens, but perhaps the results could be used to begin exploring if potential waiting time could also be looked at through a gender lens. For example, the survey results showed safety perceptions are different for night and day across both identity groups. This could mean that waiting for the bus at nighttime may be a more stressful experience for some people since they feel less safe waiting for the bus at night. Perhaps additional weights or penalties could be placed on waiting time that takes place at night and thus emphasizing the importance of reliable service for people in general, but particularly for women identifying folks, who experience a greater difference between day and night safety perception.

Let's look at the technical aspects in our planning processes and also the social implications of public transportation so that we may gain a deeper understading of its more "invisible" components, like safety perception, which often are not tracked, but which play a role in service quality and how much access we have to our cities, to opportunities and to the people that are important to us.

Acknowledgments

I'm thankful for the oppotunity to complete this project during my time as a graduate student at Northeastern University. I'm grateful to the New England Chapter of the Congress for New Urbanism for awarding me the Spatial Justice Fellowship that awarded my funds and allowed me to print beautiful waterproof posters as well as fund my expenses for the summer. Thank you to the program coordinator, Jacinda Berbehenn, for her support throughout the summer and my two program mentors, Jon Ford and Ken Livingston. Many thanks to Professor Peter Furth for introducing me to the nitty gritty details of public transportation and the concept of potential waiting time, as well as trusting me to do this project. Thank you to Andrea Goldstein, Coordinator in the Human Subject Research Protection office of Northeastern University for the long phone calls and helping me through my first IRB approval process. A sincere thank you to Joaquin Osio-Norgaard, senior policy analyst in the MBTA's Office of Performance Management and Innovation for providing me with the PATI data in a format that was usable for the project and overall encouraging demeanor. I'd especially like to thank Patricia Daniela Retamal, coordinator of planning and management for the office of gender equality at the University of Chile (Coordinadora de Planificacion y Control de Gestion |Direccion de Igualdad de Genero, Universidad de Chile) for her incredible work around safety perceptions at transit stops and making time with me to chat more about the details of her team's research and providing guidance for how to do a daytime vs nighttime safety audit. Finally, thank you to my dear friend Imesh Panditha-Waasala for proving the emotional support needed to make it through the final stages of this project and for sharing his economics minded input with me and helping me think through ideas in real time.

This project is the first research project I've designed fully on my own and I am truly grateful to have had the opportunity and freedom to complete it on a topic that is so important to me as I have experienced first hand what it's like to be fully reliant on a mix of public transit, bike and my feet while living in Boston.