Behavioural Design Fellowship At First Principles Lab

Reducing Train Deaths in Southern India's Railways

Applying behavioral design principles + human-centered design to understand what causes deaths on Indian railways in one division, and how to prevent future deaths through systemic and individual level interventions.


Research team: Kashish Java, Kamakshi Arora

Some details are confidential due to a non-disclosure agreement. Some of the findings below describe patterns and scale in relative terms to be compliant. Click here to learn more about First Principles Lab.

Background

Indian Railways carries over 23 million passengers daily, making it one of the world's largest transportation networks. But alongside this scale comes a devastating human cost: tens of thousands of railway-related deaths occur nationally each year, with the majority classified as human-run-over (HRO) incidents.

As part of the Behavioral Design Fellowship at First Principles Lab, we partnered with Southern Railways to examine HRO incidents in one division in Kerala. Our research aimed to move beyond simplistic "trespassing" narratives to understand the behavioral, systemic, and infrastructural factors driving these preventable deaths.

I was particularly drawn to this problem because of its tangible nature. Unlike designing an experience for an app or an interface where stakes might be low, this problem exists within a system that has real implications-- people are dying, the causes are observable, physical, systemic. Indian Railways is also an interesting representative ecosystem of broader challenges India faces in the areas of equity, infrastructure, labor rights, and modernization.

This case study documents our research phase: problem definition, data analysis methodology, stakeholder interviews, and the behavioral insights that will inform future intervention design.

So my partner Kashish Java, a clinical psychology graduate, and I got to work.

How we worked

This project required adapting to constraints I had not faced before. Kashish was based in India; I was in Philadelphia, working full-time. The problem itself - deaths on railway tracks - is deeply contextual to Kerala's geography, culture, and infrastructure. We could not just apply frameworks from a distance.

The Challenge

At the outset of this initiative, the term we frequently encountered was "trespassing," suggesting personal negligence or violations of rules. However, as we explored further, our investigation revealed more complex dimensions to this issue, highlighting how the framework was structured in a manner that "promoted" risky decisions as reasonable actions. Moreover, trespassing was often employed as a general, overarching term to classify a range of troubling behaviors that extended beyond mere thoughtless incursions.


Our analysis of five years of division-wide incident data revealed a crisis that is accelerating, with deaths increasing nearly threefold over the study period. But the "trespassing" label in official records obscured what was actually happening: people making decisions shaped by infrastructure gaps, time pressure, normalized risk, and detection failures.
These people, also known as "intruders," aren't always acting carelessly; they are simply individuals maneuvering through a system that frequently presents unsafe actions as the most viable option.
stakeholder map
Approach
Data Analysis
We analyzed five years of incident data across the division, mapping patterns across time, location, and circumstance. The analysis surfaced clear signals:
Stakeholder Research
Our research combined quantitative analysis with qualitative research, mainly in the form of interviews, and observations. We conducted interviews with Railway Protection Force (RPF) officials at multiple levels, including staff from control room operators to senior leadership uncharge of managing entire divisions, and gathered insights from passengers, trackside community members, and railway workers.

These conversations surfaced the systemic pressures invisible in incident reports: the RPF officer managing over a thousand commuters alone, the track worker relying on sound cues that electric trains no longer provide, the commuter who boards a moving train because missing it means losing a day's wages.
Understanding the people
To move from statistics to actionable insights, we developed four personas representing distinct risk profiles. Each emerged from our research and embodies different behavioral drivers.
  • Rajan
    The Calculated Risk-Taker
    Rajan is a daily commuter who boards moving trains because missing his connection means lost wages. Additionally, buses take too long and the roads are always congested. Trains are his best option, even though he has to navigate growing crowds.

    What does he represent?
    Overcrowding creates competitive boarding conditions, and research shows that increased crowd density perception elevates risk-taking behaviors while reducing feelings of safety. Repeated safe boardings also normalize the risk, leading to desensitization of the associated dangers.
  • Shri
    The Trackside Resident
    Shri has lived near the tracks since childhood. Crossing them and walking along them has always felt routine; the tracks feel like an extension of his backyard. His house is just a few meters away from the tracks-- he has even sat on the tracks and hung out with his friends.

    What does he represent?
    This familiarity has made the danger invisible; it is a psychological phenomenon known as the familiarity heuristic.
  • Shyam
    The Overlooked Trackworker
    Shyam maintains tracks while trains are running--his job requires proximity to danger. He often works in extreme heat, as track temperature (Railway temperature) runs 3-5°C higher than the surrounding air. Currently, he relies solely on vision and sound to detect approaching trains, but track curves and the proliferation of increasingly silent electric trains have made quick detection far more difficult.

    What does he represent?
    Occupational habituation has normalized the danger, while chronic understaffing (41 of 138 maintainers diverted to office duties) and inadequate safety equipment leave workers reliant on human vigilance alone. Failed safety initiatives like the Rakshak pilot have eroded trust, and many contract workers lack basic safety awareness. For this low-paid, neglected workforce, systemic failures compound individual risk. And on top of all this, railway workers are often left out from data, leading to them being often underrepresentated or missing in prevalent statistics that might lead to safety planning.
  • Sayil
    The Overworked Railway Protection Force (RPF) Officer
    Sayil is an RPF sub-inspector assigned to a station where passenger traffic has increased dramatically in recent years. He has to manage 22 different RPF-mandated operations on his watch—from preventing child trafficking and assisting railway children to maintaining women's safety—all while manning a station with two platforms and an impossible monitoring job.

    What does he represent?
    Accountability is placed on officers like Sayil when incidents occur, obscuring the systemic failures that make their mandated duties unachievable
Our analysis revealed several interconnected behavioral and systemic factors:

Normalized Risk
When people cross tracks daily without incident, the perceived danger fades. And social proof reinforces this--if everyone around you crosses, it feels safe. On top of that, each successful crossing strengthens the belief that risk can be managed through personal vigilance.

Detection Failures
People crossing tracks rely on sight and sound. But the division has numerous sharp curves creating visibility problems, dense vegetation that limits sightlines, and increasingly silent electric trains. The shift to electrification has been particularly consequential as newer trains are quieter and faster, eliminating the loud engine sounds and strong vibrations that people historically used as warning cues.

Behavioral Insights

They get run over because often they don't hear or see the train coming. Sometimes, they move to the adjacent track to avoid a train but end up getting hit by a train coming from the opposite direction.
Infrastructure- Forced Behavior
In several stations in the division we looked at, crossing the tracks is the only practical option. The absence of safe crossing infrastructure does not eliminate the need to cross, it just makes the crossing unsafe. Even where infrastructure exists, lack of crowd control or perceived time barriers make the more immediate choice of trespassing appealing. When the alternative is a multi-kilometer detour, the 'unsafe' choice becomes the rational one.

Occupational Exposure
Track workers face compounding risks across three levels: environmental (extreme heat impairing cognition, proposed speed increases reducing reaction time), institutional (reliance on untrained contract workers, lack of safety enforcement, invisibility in data and policy), and technological (no train-tracking mechanisms beyond vision and sound).
What gets measured gets managed. The data gaps around worker fatalities mean this populations's risks remain invisible to policy and key decision-makers.

Balancing Perspectives

Designing for railway safety is not just about understanding users - it requires navigating organizational realities and political sensitivities, which as researchers we admittedly did not anticipate right away, but picked up on early in the process as we dug deeper.

Stakeholder Access
Railway officials are busy people with demanding roles. Scheduling interviews took persistence, and responses rarely came quickly. We learned to work with the access we could get, triangulating across sources rather than waiting for perfect information. It forced us also to be creative; we reached out to labor historians, non-profits working on worker safety and rights, journalists, and an ethnographer who l tried understanding the experience of railway children.

Sensitive Territory
Some of our findings touched on politically charged issues: labor conditions, staffing shortages, and death classification. When a track maintainer dies on duty, is it classified as 'trespassing' or an occupational hazard? Who counts as staff versus contract labor? These questions go beyond just semantic ruminations, they determine what gets reported, what gets measured, who receives safety protections and recognition, and ultimately what problems get addressed.

Equity and Investment
The demographic patterns we found - working-age men, migrant laborers, residents of trackside communities - pointed to an uncomfortable truth: those most at risk are often those with the least political leverage to demand safer infrastructure. General compartments, used disproportionately by lower-income passengers, see less investment than premium classes (A/C coaches).
Our research surfaced tensions between how incidents are officially classified and what the data actually revealed. Navigating what could be shared publicly, and how to frame systemic issues constructively, became part of the design challenge itself.
ecosystem map

Synthesis: A Systems View

As researchers and designers, we acknowledged that the curb-cut principle is relevant in this situation: when we create solutions for the most at-risk, specifically track workers and nearby communities, the outcomes can advantage everyone. And with that, we worked on our guiding problem statement/question, that would inform the next phase of this project, and guide us in our prioritization.

During our synthesis phase we worked on prioritizing our findings to help us condense our problem statement that would define the next phase of the project.
Prioritization matrix for problem definition
Our redefined problem statement
How might we create equitable railway safety that serves all users and nearby communities, while addressing track misuse and improving risk awareness in low-detectability zones?

Impact and Next Steps

Our research findings were presented to Southern Railways stakeholders who commissioned this study. Through systematic analysis of incidents, stakeholder interviews, and behavioral research, we identified high-risk corridors, temporal patterns, and user insights that revealed the systemic nature of railway safety challenges in the Trivandrum Division.

These findings are now actively informing operational decisions and shaping the next phase of work: intervention design. The research foundation we built - from data-driven corridor identification to behavioral personas - is being utilized to develop targeted safety interventions.
~
Made on
Tilda