arcroad · evidence base · grades A and B only
Civic sense: what the evidence shows, and what it doesn't.
"No civic sense" is said often and measured rarely. What the studies show: much of what looks like bad behaviour is a response to bad design.
Only peer-reviewed studies (A) and official statistics (B).
Wrong-side driving and lane discipline
Where a median gap is closed and the legal U-turn is 1–2 km away, some people ride the wrong way instead. arcroad measures the length of that legal detour, so the design that invites the violation is on the record beside the violation.
The studies
| Finding | Grade | Source |
|---|---|---|
| In 2023 wrong-side driving and lane indiscipline were linked to about 5.5% of police-recorded road deaths, the third-largest cause. Over-speeding was about 68%. | B | MoRTH, Road Accidents in India 2023 |
| On national highways in 2022, wrong-side driving was the second-largest cause of deaths (about 5.8%). The authors call for enforcement and design changes, not only new laws. | A | IIT Delhi TRIPP (Jain et al.), World Congress on Injury Prevention |
| Wrong-way driving is common enough in India that countermeasures have to be designed for Indian roads; findings from countries where it is rare do not transfer. | A | Field investigation of crash sites, TRB TRID 01994083 |
| In mixed traffic, the side-to-side movement of vehicles and the gaps they keep differ measurably by vehicle type. | A | Chauhan, Dhamaniya, Arkatkar (SVNIT), Springer 2024 |
Headlight glare on the oncoming driver
Glare is paid for by the person facing the light, not the one who switched it on. Roads without median screens or street lights push drivers to high beam, so arcroad records lighting beside glare and never claims glare caused a particular number of crashes.
The studies
| Finding | Grade | Source |
|---|---|---|
| The law requires headlamps that dip so they do not dazzle the driver of an approaching vehicle. Fines fall under the Motor Vehicles Act. | B | CMVR 1989 rule 106; MVA 1988 s.177 |
| No separate national data on glare crashes is kept. About half of 2022's road crashes happened between 6 pm and 6 am. | B | MoRTH reply in Parliament |
| Glare that blinds is different from glare that only annoys. After a bright oncoming lamp the eye needs time to recover, longer for older drivers and for objects at the edge of view, such as a pedestrian. | A | Skinner & Bullough, SAE 2009-01-0546; NHTSA DOT HS 811 043 |
| A US pilot found about two in three headlights aimed wrongly. | B | NHTSA report to Congress on glare |
Pedestrians and signals
Long waits, missing crossings and badly placed bridges produce the violation. That is arcroad's whole point: measure what the design costs people, and the "civic sense" argument changes.
The studies
| Finding | Grade | Source |
|---|---|---|
| About 45% of pedestrians who arrive on red cross against it, mostly in the second half of the red phase. | A | Kumar & Ghosh, Safety Science 2021 |
| Across 55 Kolkata junctions, those with more red-light crossing also had more fatal pedestrian crashes. | A | Safety Science 124:104610 (2020) |
| Longer red phases go with more violations. Of 2,360 pedestrians observed, about 29% were distracted, and those texting violated more. | A | Raoniar & Maurya (IIT Guwahati), 2022 and 2024 |
| Where a foot overbridge stands beside a street-level crossing, only about a quarter of pedestrians used the bridge. The choice followed the time it cost, not attitude. | A | Mumbai video survey of 1,134 pedestrians (Patra et al.) |
| Design walking speeds in Indian mixed traffic are about 1.12 m/s for adults and 0.95 m/s for older people. | A | Marisamynathan & Perumal, 2014 |
Safety gear
Behaviour responds to enforcement people can see, and to design. It is not fixed by "culture".
The studies
| Finding | Grade | Source |
|---|---|---|
| In 2023, 16,025 people killed were not wearing seatbelts. Two-wheeler riders were about 45% of road deaths and pedestrians about 20%. | B | MoRTH 2023 |
How arcroad uses this evidence
- A study's rate is compared with arcroad's own only when the place type and the definition match.
- Nothing is said about any group of people, community, faith, region or gender. Where studies break results down that way, the breakdown is not carried.
- Official figures are police-recorded and undercount; the site says "police-recorded".
- Behaviour arcroad measures is aggregate, by stretch and time, never by person or vehicle. No plates or faces.
- Every figure is re-checked against the primary report before a release.