Before I get into why the study is flawed on the face of it, I want to make it clear that I believe that hybrids, when running in all electric mode, can be more dangerous than cars that are at least making a little bit of noise. Having said that, cars that are traveling at low rate of speeds these days are not necessarily all that noisy in the first place.
On to the NHTSA study:
According to the executive summary (emphasis is mine),
This study found that pedestrian and bicyclist crashes involving both HEVs and ICE vehicles commonly occurred on roadways, in zones with low speed limits, during daytime and in clear weather, with higher incidence rates for HEVs when compared to ICE vehicles.
Incidence rate of pedestrian crashes in scenarios when a vehicle makes a turn was
significantly higher for HEVs when compared to ICE vehicles.
bicyclist crashes involving HEVs at intersections or interchanges were
significantly higher when compared to ICE vehicles
Now, again, I want to point out that the reasons for worry are valid. Hybrid cars are quiet when they are running at low speeds. There is some cause for concern. Having said that, the analysis and conclusions in this study must be taken with a big grain of salt.
Issues that Cloud
First, the sample size is absurdly low. The researchers themselves point out the data is limited and further data is needed to draw any real conclusions.
This analysis was conducted on a total of 8,387 HEVs and 559,703 ICE vehicles that met the selection criteria. A total of 77 and 3,578 pedestrians were involved in crashes with HEVs and ICE vehicles, respectively. A total of 48 and 1,862 bicyclists were involved in crashes with HEVs and ICE vehicles, respectively.77 and 48 accidents involving hybrids is a small number to draw on. The less data you have the less likely you can make any strong statements about the analysis.
Second, related to the first point, there are only 12 states represented. I cannot say whether driving habits vary from state to state, but anecdotally, they certainly do. That's not really a big point, but it should be noted.
Third, the comparison was made with similar vehicle types, so it's not really a comparison of hybrid electric vehicles (HEVs) to internal combustion engine (ICE) vehicles.
The HEVs (case group) selected were the Toyota Corolla, Toyota Camry,I can't help but point out the Corolla, which has no hybrid version, is in the list of hybrids. It's a typo, I'm sure. But the conclusions may not be generalizable to HEVs vs ICE. But there's nothing really wrong with selecting the controls in this way (there is, but I'll get to that in a minute). It's just the distinction isn't necessarily being made about what the control group really is at other sites and it's an important one.
Toyota Prius, Honda Civic, and Honda Accord. The ICE vehicles (control group)
selected are the Toyota Corolla, Toyota Camry, Honda Civic, and Honda Accord. The
analysis is limited to vehicles of model year 2000 and later.
But the reasons above are just really incidental and not really important, other than to point out some of the limitations that may not be clear when you read the headlines.
The Real Reason This Study Should Not be Used
No, the big reason why this study is a problem is the lack of other confounding information. You may be comparing similar cars, but you may not be comparing similar drivers.
Hybrid car owners, up to a certain point, are a different type of driver, and are self selected individuals. People who buy a Prius are not necessarily going to choose a Corolla instead. Hybrid car owners are wealthier and are older. Drivers who are older and wealthier, if there were no such things as hybrids, aren't necessarily going to buy the ICE models shown.
It should also be noted that this is where the limitation of the number of states used is important. The reason why groups of buyers choose hybrids can vary widely from state to state. It's very easy to say, anecdotally, that hybrid buyers in Massachusetts are buying more for the environmental concerns, than buyers in Texas.
And an argument can and should be made that it's the people that cause accidents, not the cars.
But even if you don't buy that (and I believe you should), there's also no mention of how much the cars are driven? Incident rates for accidents should be calculated using miles driven.
Follow me on this. Hybrid car owners probably bought their car because they are more concerned with their fuel economy. They are most likely concerned with fuel efficiency because they drive farther than other drivers. More miles on the road, or more time on the road, means you have more of a chance to have an accident.
John, who only drives two miles to work every day is a lot less likely than Mary, who lives down the block but drives 40 miles to work every day, to have an accident. There's no mention of miles driven in this report, because the data isn't there.
If that's true, then it may be that hybrid car owners, if they drive more than others, may end up being safer drivers in most situations, since the incident rates for hybrid car owners in this study would end up being a lot lower. But that's just speculation and I have no way of backing that up.
Again, I do believe there is a reason to be concerned when it comes to hybrids at low speeds. They do run quieter and the likelihood of surprising pedestrians who aren't looking is most likely higher. But gas engines tend to be pretty quiet these days. I don't buy the 2x as dangerous results for certain situations shown in this study, and neither should you.