Research Explainer · Social Effects & Equity

Why do people really use pooled rides? The answer is about safety as much as speed.

The biggest study yet on what drives people to use app-based pooled rides in Mexico City surfaces a finding that goes far beyond transport planning: for many women, pooling a ride is a rational response to a genuinely unsafe public transit system.

Abouelela · Tirachini · Chaniotakis · Antoniou
TU Munich · U. Twente · U. Chile · UCL London
N = 2,484 users · Mexico City · 96,317 trip records · Sentiment analysis
User voices — open survey responses
"The service is excellent and convenient. The drivers are always attentive and provide excellent service." Positive respondent · sentiment score: +8
"Safe and reliable. Please expand the routes — PT coverage in the north is very limited." Regular user · route expansion request
"The biggest issue is the unclear app interface. However, once you are in the car, the service is very good." Critical respondent · sentiment score: −2
"Given the cost, convenience, and comfort traveling between home and work, I plan to continue even if my company withdraws support." Loyal user · willingness to pay own cost
Scroll to explore

What actually makes someone choose a pooled ride over all the alternatives?

The standard transport economics answer to "why do people use a mode?" is: because it is faster, cheaper, or more convenient than the alternative. These factors matter for pooled rides too. But this paper — the broadest analysis yet of pooled-ride users in a developing megacity — finds that the picture is considerably richer.

Mexico City's public transit system carries half the city's 34 million daily trips. It is also, by independent assessments, one of the most dangerous urban transit environments in the world. Ninety percent of PT users report feeling unsafe. Twenty-three percent of women avoid PT entirely for personal security reasons. Sexual harassment is frequent across all modes, from metro to minibuses to BRT.

Into this environment, Jetty offers something qualitatively different: a seat you book in advance, a vehicle you can track in real time, a driver with a work contract. For many users — especially women — this is not merely a convenience upgrade. It is the difference between a trip that feels manageable and one that feels threatening.

Women respondents reported using the service for security against harassment at six times the rate of men — and they still pay the same price, despite earning on average 21% less.

§

What users say — and what the models confirm

Users could cite up to six reasons from a list of 14. The top four — chosen by more than half of all respondents — tell a coherent story about a city where formal transit has quality and safety deficits.

Reasons to use Jetty (% of 2,484 respondents who cited each reason)
Booking of seat
69%
Security against theft
67%
Travel time saving
66%
Travel time reliability
53%

Access/Egress time
42%
Vehicle quality
36%
Fare
36%
Security vs. harassment
16%
Avoid parking problems
11%

The 16% who cite harassment security represent an absolute signal: women reported this reason at 6× the male rate. The structural gender gap this reflects is the paper's central finding.

§

Gender, demographics, and what drives each choice

The models (binary logit and hybrid choice) reveal how demographic and trip characteristics shape why users choose Jetty, and what they do on board. Explore the findings by selecting a factor below.

Model-Based Factors Explorer

BINARY LOGIT & ICLV MODELS — ABOUELELA ET AL. 2025, TABLES 5–7

Women vs men
↑ More likely to cite seat booking
Bachelor/professional degree vs others
↓ Less likely to cite seat booking
Masters or PhD vs others
↓ Less likely to cite seat booking
Full-time employed vs other
↑ More likely — schedule rigidity increases value of guaranteed seat
Access by walking vs other modes
↓ Less likely — walkers may have shorter, easier trips
Complexity of replaced trip (more modes)
↑ More likely — complex trips make a guaranteed seat more valuable
Women vs men
↑ Significantly more likely to cite theft security
Age 36–45 vs 18–25
↑ Older users prioritise personal security more
Age 46+ vs 18–25
↑ Strongest age effect for security concern
Higher education (bachelor+)
↑ More likely to cite theft security
Driving licence availability
↓ Having a licence reduces probability — car option reduces exposure to PT crime
Access by walking vs other modes
↓ Less likely — possibly because walking access implies shorter exposure to PT environments
Complexity of replaced trip (more modes)
↑ Multi-modal trips mean more crime exposure, so security valued more
Latent variable: frequent PT user
↑ PT users value time-saving more — Jetty is faster than their complex PT routes
Age 36–45 vs 18–25
↓ Older users less likely to cite time saving
Household size 5+ vs 1–2
↑ Large households — more constraints on travel time
Access by walking
↓ Walkers less likely to cite time — short access means the overall trip is already simpler
Access time duration
↓ Longer walk to pickup reduces perceived time benefit
Complexity of replaced trip (more modes)
↑ Biggest effect — replacing 3-mode trips makes Jetty time savings enormous
Women vs men
↑ Women significantly more likely to sleep — reflects sense of safety not available in PT
Full-time employed vs other
↑ Fatigue from early-morning commutes drives sleeping behaviour
Age 36–45 vs 18–25
↓ Older riders less likely to sleep
Age 46+ vs 18–25
↓ Strongest age reduction in sleeping probability
Driving licence (yes vs no)
↓ Drivers less inclined to sleep — different comfort baseline
Latent variable: frequent PT user
↑ PT users more likely to sleep — safety contrast is most acute for them
Complexity of replaced trip (more modes)
↑ Bigger relief from trip complexity → more likely to relax and sleep
Age 36–45 vs 18–25
↓ Older riders less likely to use smartphone
Age 46+ vs 18–25
↓ Strongest age reduction — generational smartphone usage difference
Driving licence (yes vs no)
↓ Drivers less likely to use phone — possibly use cars instead sometimes
Latent variable: frequent PT user
↓ Frequent PT users less likely to use smartphone (safety concern in PT may carry over)
Complexity of replaced trip (more modes)
↑ More modes replaced → more convenience gain → more smartphone use
§

What people do while riding — a window into perceived comfort and safety

What passengers do on board is not just a curiosity — it is evidence of their perceived comfort and security level. In a vehicle you feel unsafe in, you do not close your eyes. You do not read. You stay alert.

😴 Sleeping 74%

The most revealing finding. Women are significantly more likely to sleep than men — the model links this directly to the safety of the service relative to CDMX public transit. People do not sleep in places they feel threatened.

📱 Using smartphone 71%

Productivity and entertainment on a 46-minute average journey. Using a phone requires physical safety (risk of theft) and cognitive ease. That 71% do this comfortably in a shared vehicle in Mexico City is itself a data point about perceived security.

🪟 Looking out of window 34%

A restful, passive activity consistent with the early-morning, long-distance commute profile. 75% of trips are under 49 minutes; these are dedicated commute journeys where decompression is valued.

📖 Reading for pleasure 23%

Requires concentration and a stable, comfortable environment. The fact that nearly 1 in 4 riders reads for pleasure contrasts sharply with the crowded, standing-room-only reality of many colectivos.

💼 Working 14%

Reclaiming commute time as productive work time. Combined with studying (8%), approximately 1 in 5 riders turns the journey into useful working or learning time — something impossible while driving.

💬 Talking to passengers 6%

The lowest-scored activity. Pooled rides are not primarily a social experience. People value the convenience and safety — not the company. This finding aligns with similar results from Europe and North America on ride-pooling socialization.

§

What 864 users said in their own words

An optional open-ended question generated 864 Spanish-language responses. Using N-gram analysis and AFINN lexicon scoring, the researchers extracted patterns invisible to the closed survey questions.

AFINN lexicon-based sentiment scores (scale: −5 to +5 per word)
+2.87 Overall avg.
All users
+2.87
Women
+3.51
Men
+2.14
Bus users
+3.52
Van users
+3.03

Women's scores are statistically significantly higher than men's (p = 0.006). Bus users score higher than van users — possibly because buses have more personal space at lower cost (40% seat utilisation vs. 60% for vans). The gender gap in sentiment mirrors the gender gap in security reasons: women have more to gain from Jetty, because they have more to lose in the absence of it.

The N-gram word network surfaced key themes that closed survey questions could not capture. The most frequent word associations:

excellent service safe · reliable service · convenient expand · routes friendly drivers lower · cost uncomfortable seats bad experience

Requests to expand routes dominated the geographic comments — most concentrated in the north of the city, where PT coverage is worst. This represents both a satisfaction signal (people want more of it) and an equity signal (the areas with greatest need are where coverage is thinnest).

§

How Jetty differs from the global pooled-ride profile — and what that teaches us

Most shared mobility studies worldwide find men as the majority or most frequent users. In this sample, women use Jetty more frequently than men: 6.88 trips/month versus 5.82 for men, a statistically significant difference (p = 0.003). The authors suggest this is primarily driven by the safety premium the service offers relative to CDMX public transit, where women face disproportionate risks. This finding has a direct parallel in London, where 77% of an Alternative Transit Service studied were women — also explained by safety and operational structure.
74% of the trips replaced by Jetty would have involved at least two modes without the service; on average 2.2 modes per trip. One in five of those replaced trips would have involved Metro plus Microbus. This is a genuinely underserved need: the formal transit network connects the city's central areas well but provides poor direct connections from the residential north to the major job clusters at Santa Fe and Polanco. Jetty substitutes not just a mode but an entire multi-stage journey — which explains both the time savings users value and the complexity gains that show up in the model.
95% of users cited their latest trip purpose as work commuting. In European cities (Berlin, Munich, Hamburg) the majority of pooled-ride trips are for leisure, often with an evening peak. Jetty has a morning peak at 6 a.m. — nearly double the evening demand. This is consistent with the geography (connecting suburban north to downtown job clusters) and the strong "value of time before reaching work" effect visible in the frequency data. The commute focus also means users have fixed, repeated trip patterns: 22% of survey respondents made every trip on exactly the same route.
80% of survey respondents have at least one car in their household — nearly 2.5x the city average. At first glance this seems odd for shared mobility. But this is exactly who Jetty is designed for: people with a car who are choosing not to drive it. The service competes with private car use, not with metro. Users cite parking problems, time use while traveling, and trip cost as reasons — all car-relevant factors. These users are the sustainability target. Getting someone out of their car and into a van with eight other people is a far bigger VKT and emissions win than getting someone off the metro.
A methodological contribution of the paper is demonstrating how open-text sentiment analysis surfaces dimensions that closed-question models miss. The word-network analysis revealed "friendly drivers" as a significant positive association — a service attribute never asked about explicitly. It surfaced requests to expand routes as a key unmet need. It produced a valid quantitative sentiment score that showed statistically significant gender differences (p = 0.006). The authors conclude these two methods should be treated as complementary tools in transport research, not alternatives.
§

What this means for city planning and shared mobility policy

01
Subsidies for low-income users and women are not optional — they are structural

The service is currently affordable only for middle and high-income users. Women who use it for safety reasons earn on average 21% less than men — and pay the same fare. If pooled rides are to function as equitable infrastructure rather than a premium service, subsidy mechanisms targeting these groups are essential. France has piloted carpooling subsidies with some success; Mexico City needs an equivalent policy conversation.

02
Improving PT safety is a transport policy imperative — and a shared mobility policy lever

This paper documents in detail how unsafe formal transit drives women toward premium alternatives they can barely afford. Addressing sexual harassment and theft in the metro and colectivos is not separable from the question of shared mobility sustainability. It is, in fact, a precondition for it.

03
Route expansion in transit-poor northern CDMX is both a demand and equity priority

User comments overwhelmingly requested route expansion in the north — exactly where PT accessibility is lowest and the urban marginalisation index is highest. Planning pooled-ride routes to match unmet transit demand, rather than duplicating well-served corridors, maximises both impact and equity.

04
Integration with formal transit requires addressing the digital exclusion problem

Using Jetty requires a smartphone, a bank account, and a credit card. These exclude significant segments of CDMX's population — particularly low-income, elderly, and unbanked residents. Any policy push to integrate pooled rides into the formal transport system must include physical access points and alternative payment mechanisms.

05
Scheduled fixed-route pooling outperforms real-time ride-splitting for reliability

Unlike UberPOOL, Jetty operates on fixed routes and schedules — users know their pickup time and route in advance, guaranteed seating is possible, and detour uncertainty is eliminated. The paper's findings suggest these operational characteristics are a key driver of user satisfaction, particularly for time-sensitive work commuters.

Read the complete study

This explainer covers the paper's main findings. The full paper includes complete model tables, the exploratory factor analysis results, the N-gram word network, a detailed taxonomy of the pooled-ride ecosystem, and an extensive literature synthesis spanning 17 countries.

Read the Full Paper →

Abouelela et al. (2025) · Travel Behaviour and Society Vol. 41 · DOI: 10.1016/j.tbs.2025.101072 · Open Access CC BY 4.0