Dummy Diversity

I’m going to kick things off with a case study that makes me angry.

It also illustrates to me one of the most important lessons of gender innovations: gender/sex analysis isn’t just nice, it’s a matter of life and death.

In 2018, there were 6.7 million car crashes in the US. As a result of these crashes, over 36,000 people died and 2.7 million were injured. Motor vehicle crashes are a major cause of death in the US, and the leading cause for people under the age of 30.

Yet car accidents are far more dangerous for some people than others. National data from 1998 to 2008 show that female drivers were 73% more likely to sustain an injury and 47% more likely to sustain a serious injury in a car crash than male drivers. That, by the way, takes into account factors like age, body mass, and crash type and comes from drivers wearing seatbelts, which means that cars’ safety equipment is failing female drivers. Men get in more crashes than women, but when women do, the results are far worse.

Black and white image of first anthropomorphic crash test dummy.

Why is driving so much more dangerous for women than men? The answer has to do with the design of crash test dummies.

Crash test dummies were first used by the Air Force to test ejection seats. The first dummy, “Sierra Sam,” was born in 1949, modeled on the 95th percentile of height and weight for adult men in the US (meaning “he” was taller and heavier than 95% of the adult male population).

The crash test dummy commonly in use today is the descendant of a series of dummies produced by General Motors in the 1970s, based on then-average male data. When the National Highway Traffic Safety Administration—the US government agency in charge of vehicle safety—began its crash test program in 1978, it adopted GM’s dummy (called the Hybrid III). The private Insurance Institute for Highway Safety also uses the Hybrid III, as does the European New Car Assessment Programme.

In fairness to the safety testers, there has been some attempt to represent more than just average-sized males in safety tests. What about the picture of that nice 1950s-style crash test dummy “family”—with the big man, the little women, and the three assorted children?

Five crash test dummies of different sizes arranged like a family sitting for a portrait, with two "parents" and three "children" of different ages.

Ms. Crash Test Dummy is designed to represent the 5th percentile “female” by height and weight (meaning “she” is very small). Her design, however, is not based on female data: she’s just a scaled-down male dummy. What’s worse, she is rarely a driver—in most mandatory tests, the small dummy sits in the passenger seat. It’s as though the organizations in charge of vehicle safety assume women don’t drive. (A very large 95th percentile male is also used, though not pictured in this “family” shoot).

Does it really matter that the “female” dummy is based on male data? After all, if very small and very large people are represented, won’t cars be safe for everyone? Unfortunately, no. Muscle strength, spinal alignment, geometry, and mass distribution—not just height and weight—affect how a body responds in a car crash.

No dummies currently required in safety testing anywhere—not even the 5th percentile “female” dummy—are built using data from female people.

Of course, female and male bodies overlap in many ways—they don’t fall into two neat categories and any individual might have a mix of male-typical and more female-typical features. It’s clear from the data, however, that female-bodied people are suffering the consequences of under-representation. (I’ve been using female and woman interchangeably, because that’s how the research we have available does it, but not all women have female-typical bodies and not all people with female-typical bodies are women).

And it’s not just people with female bodies who are being poorly served: older drivers, drivers classified as overweight, and pregnant people all have worse outcomes in crashes because they are not represented in safety testing.

And the Hybrid III—designed with population data in the 1970s—doesn’t even represent the “norm” anymore. Average body weight has increased dramatically, especially in the US, and populations are aging worldwide.

In short: the use of the “normal” male in vehicle safety tests leads to dangerous conditions for anyone whose body doesn’t match that norm.  

One of the other lessons I take from this case study is that it doesn’t take deliberate sexism (or ageism and sizeism) to create biased design. Biased crash test dummies are a holdover from days when most military pilots (but not all!) were men. But I have hard time imagining that if the statistics were reversed—that if men were 79% more likely to be injured in a crash than women—we wouldn’t have made dummy diversity a priority by now.

Computer simulation of a pregnant crash test dummy with fetus,

Fortunately, there are researchers working to represent a greater diversity of human bodies in crash tests, including properly modeled females, pregnant people, larger people, older people, and children. Most of this work uses computer modeling. Swedish researcher Astrid Linder, for example, has developed a computer model based on 50th percentile female data and a matching physical prototype. So, it’s not all bad news. Now that these models exist, there is nothing to stop commercial dummy producers or national automotive testing agencies in Europe or the US from using them (though they haven’t yet).

And they do have to be required by automotive testing agencies if we want to change crash statistics. We can’t rely on voluntary testing. As an individual driver whose body doesn’t fit the norm, I don’t have a lot of choice other than to drive less (good for many reasons, but hardly a realistic choice for many given our car-centric infrastructure) or to choose one of the few manufacturers that have committed to an expanded set of vehicle safety tests—if I can afford it. Only when diverse bodies are mandated in vehicle safety testing will cars be designed to keep them safe.

See the full case study here.

Published by Hannah LeBlanc

Researcher and writer with Gendered Innovations.

4 thoughts on “Dummy Diversity

    1. They do use them! But they’re all built by scaling down data from men in the 1970s. So it’s more like a family of different-sized men?


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create your website at WordPress.com
Get started
%d bloggers like this: