Bus-Riders of the World Unite!

Remarkable facts from a new paper by James Habyarimana and William Jack of the Center for Global Development:

The World Health Organization (2004) reported that 1.2 million people died from road traffic injuries in 2002, 90 percent in low- and middle-income countries, about the same number as die of malaria. In addition, between 20 and 50 million people are estimated to be injured or disabled each year.

Road traffic accidents constitute the largest share, 23 percent, of deaths due to injury, nearly twice as many as the 14 percent due to war and violence combined. Traffic accidents were ranked as the 10th leading cause of death in 2001, and are projected to be the third or fourth most important contributor to the global disease burden in 2030 (Lopez et. al., 2006). By that date, road accidents are projected to account for 3.7 percent of deaths worldwide — twice the projected share due to malaria (Mathers and Loncar, 2006).

It’s interesting how we hear all the time about malaria and mosquito nets (it was even World Malaria Day last week), but we never hear about third-world traffic accidents.

Habyarimana and Jack report the results of a fascinating field experiment they carried out, putting posters in over 1,000 randomly chosen Kenyan mini-buses. The posters told passengers to speak up if the driver drove dangerously.


And it really seems to have worked. Using data on insurance claims, the authors find that the buses that got these posters saw large declines in crashes relative to the control group, and the accident reduction appears to persist, as long as the signs remain posted.

I love it when I stumble onto a paper that A) teaches me some important facts, B) has a clever idea, C) is believable, and D) makes the world a better place.

Doug B

I wonder if trucking companies in the US have seen a similar effect from the ubiquitous "How's My Driving?" signs on the backs of trailers?

I also what would happen if we had road signs that encouraged drivers to report other bad / illegal drivers. Would we have enough law enforcement to respond to all of the complaints?

Maybe require bumper stickers ("How's my driving? If I am driving recklessly, call 911!") for minors and convicted DUI offenders?

Eric M. Jones

Most amazing public safety sign ever posted on a bus:

My wife and I took a bus to the beach on Guadeloupe (Fr.). Prominently displayed in tthe front was a sign that said in French, "Women Must Wear Tops".


Doug B: look up "How's my driving? for everyone". I think it's been referenced on this blog.


what's the incidence of seatbelt use in the third world?


#4: Buses don't have seatbelts in the first world, so probably not that high.

Utku Suleymanoglu

It might also be interesting to see how the stickers affect the number of fights started in the minibuses. In my country (more developed than Kenya), bus or minibus drivers usually don't take it kindly if some passenger interferes with their driving habits.


1.2 million people died from road traffic injuries in 2002 and about the same number die of malaria and about 1 million abortions (killed babies) in the USA each year.

What an opportunity we have to make a difference! Now, if only we can get someone to pay attention...


good point #5, should also be noted that in the US many people stand on buses, so even if there were seat belts on buses, at best half of a full bus's passengers would be wearing them.. Being from a developing country, I can confidently say there are more people standing on buses than sitting and I agree with #6, even here in the US (at least in Chicago), drivers don't react well to being criticized, so increase in bus fights would be an interesting statistic to have.


Having survived being driven around perpetually congested Hong Kong and the arena style experience that is the Shanghai freeways, I would say that US drivers(and most developed countries) are far safer than its residents give them credit for. That or everything you hear people say about Chinese drivers is true.

I never could see the difference when I heard people complain how drivers from were such terrible drivers. Now I'm thinking these people lack a sense of perspective.

Paul Farrell

Do How's My Driving? programs work? Yes, they do. Insurance carriers in the US and large private companies have repeatedly validated that the feedback process works. (it's called behavior safety by insiders).

I was one of the skeptical study sponsors back in 1996 that worked with our actuarial department to study a group of 30,000 power units (cars, pickups, vans, and larger trucks) that had used the program for up to three and a half years. We compared their crashes prior to using the program and after. We also checked to see if they had implemented any other new safety programs during the test period. Finally we compared their results as a group against all the other fleets who had not used the program.

We documented a 22% reduction in collisions and that more of the "serious/costly" claims had been avoided than the simple stuff like backing into a mailbox, etc.

Other insurance carriers have done similar studies with similar results over the years since then. One of the most recent was completed about two years ago showing an average 30% reduction in crashes.

The key is that the reporting feedback should be used to spur re-training, not "bashing" or "blaming".

The program is in place on literally hundreds of thousands of commercial vehicles, but many still refuse to use the program believing that the feedback is useless and delivered by "pranksters".

For the record, while I had no interest in promoting the How's My Driving concept during the study period (as an employee of an insurance carrier who was paying for the program -- I was actually interested in making sure we weren't wasting our time and money), but now I actually work for the leading, national provider of this service. It doesn't change the statistics, but I want to be open and fully disclose my current "bias" towards the program. Thanks.



@4, @5 and @8:
it is not always the bus passengers that die or are injured when a bus is involved in an accident. In countries like India, two- and three-wheelers are very common, and are often the "victims" in an accident.

The trouble with the whole proposal is that if it becomes universally adopted, it just becomes "background noise" and people stop paying attention to it. Part of the reason why people paid attention to it was the fact that they came out of nowhere, and aren't part of what one normally sees in a bus. How many of us actually believe that the presence of "How's my driving" stickers makes any difference in the driving of trucks in the U.S?

Maybe, maybe, if there's a "Driver Number" posted on every bus very prominently and people can submit a quick complaint by a text message...

David Silverman

We tend to ignore real killers be it road accidents or cigarette smoking (kills 1100 a day in the US).

Paul Farrell

If anyone would like to dig deeper into the tragic road safety results of the world, there are several resources:

A recent international road safety conference was held in Washington DC (in February).

The World Health Organization has done some super work in understanding regional differences in traffic satety issues:


Presumably the greater fear of Malaria can be attributed to it being, to an extent, out of human hands, unlike automobiles.
Similar to "since I control the car, I am the one keeping myself safe; since I have no control of the airplane, I am at the mercy of myriad external factors." Except that in this case it's human control being trusted rather than individuals.
Also interesting is that it is estimated that 8% of deaths of children under 5 years were caused by malaria but only 3% caused by injuries according to the WHO. http://www.cdc.gov/malaria/impact/index.htm
Also 60% of cases of Malaria occur in Africa South of the Sahara and 80% of fatal cases occur there.


Have any of you all ever been to India. Talk about mayhem. Basically the only rule is biggest vehicle rules the way. Anything else need get out of the way. There are no rules here, and if they have any they are not enforced. Police usually stand on the wrong corner to ticket anybody, and if your in a car, how is a policeman on foot suppose to stop you? Basically there is no common courtesy what so ever, there is no common sense as most that drive here are illiterate, so maybe that is why India has the highest per capita accident death rate in the world. It is basically a no brainer, no pun intended!


#5 and #8:

I reviewed a cost-benefit analysis of seatbelts on buses a few years, which determined that it was not a cost-effective road safety measure to reduce road fatalities/injuries. Two reasons for this: first, buses are much less likely than other vehicles to be involved in a crash in the first place; second, buses are designed so that passengers are extremely unlikely to be flung through windows or around the bus in the event of a crash.


I'm a LITTLE skeptical. Is comparing the buses with signs to the control group necessarily show that the signs made a difference? Wouldn't you need to track a group of buses/drivers without signs for a given period, then track the same group of buses/drivers with signs to see if the difference was made? There is no guarantee that the control group and test group demonstrated the same risk before the signs were implemented.


@17: No, your suggestion is not a good idea, because it is possible that the stickers have long term effects on driving habits on the treatment group drivers. That is, they may not be the same any more. Maybe, if you wait a long time for the second round of data collection. But then, it is also possible that they will mature/older and become better or worse drivers. Some of them might also opt out of driving and worse this decision might be related their driving skills. That is, over time, you might lose good or bad drivers, which would bias your treatment effect estimates.

The idea behind the simple design implemented is that when you pick 1000 minibuses at random, the average driver has the same habits as the average driver in the who population and will respond to the policy, on average, identically.


I live in South Africa. Here, it would be necessary to make an allowance for the number of passengers thown out of the taxi (with appropriate subdivisions for the speed the taxi was traveling at the time) or shot by the driver. Further allowances would also need to be made for the drivers who speeded up or drove even more recklessly.
Besides, no one would be able to afford to pay the drivers enough to allow such notices to stay in the taxi.