Samantha F. asks: Is it true that more people are killed every year by vending machines than sharks?

As we’ve discussed at length in another article, sharks, contrary to their fearsome reputation, rarely attack, let alone kill, humans; it would seem that they don’t deem us to be suitable prey, though it seemingly has little to do with taste, as is commonly stated. (See: Do Sharks Really Not Like the Taste of Humans?) But is nature’s supposed “perfect” killing machine really outclassed by humble vending machines when it comes to the number of humans these respective things kill per year?

Putting aside the questionable nutritional value of items found in many vending machines slowly killing many people over time, when talking about a vending machine directly killing humans, as far as we can tell, most every source claiming that vending machines kill more humans per year than sharks cite a single report as their source- specifically this one by the Consumer Product Safety Commission in 1995, which notably deals exclusively with American deaths via vending machines. This comprehensive review of available data states that from 1978 to 1995 a total of 37 Americans were killed by falling vending machines, meaning a little over 2 deaths per year.

As a result of this data, warning labels began adorning most vending machines cautioning consumers that rocking or tilting a vending machine may cause it to dole out death instead of Twinkies.

Since that time, the statistic has been quoted across the news and internet and is often extrapolated upon to suggest that the odds of being killed by a vending machine anywhere in the world is more than being killed by a shark, despite the report dealing only in American deaths.

Now, on the one hand, it is true that during that span discussed, at least statistically, Americans were more likely to be killed by vending machines than sharks, as, on average, slightly less than one American every year was killed by a shark. But there are a number of problems with this comparison. First, someone sitting in the middle of Oklahoma has zero chance of being killed by a shark in any given moment, or zero for their entire lives if they never travel to the ocean (or fall into an aquarium shark tank), not “1 in 400 million” or similar as is often stated. Likewise, the Floridian who regularly surfs at the beach has greater odds of dying from a shark attack than the stated average, though still amazingly low.

Similarly, one’s exposure to (and specific chosen interaction with) vending machines further muddies the waters here- some people almost never use them, others use them daily. Even for those that use them regularly, if they aren’t physically strong enough to rock the vending machine (which is why it’s amazingly rare for women and children to die from a vending machine from the data available), the odds of death plummets even beyond its existing minuscule amount. Thus, on the whole, the statement, “You are more likely to be killed by a vending machine than a shark” isn’t really terribly meaningful on its face.

Beyond such minutiae, there is something else that makes this statement indeterminate even in the general case. You see, as noted, pretty much everyone seems to be using data from before certain safety measures were implemented in vending machines. Along with warnings that vending machines can cause serious injury or kill if tipped being placed on nearly every modern vending machine produced today, most also have anti-vandalism technology built in- most pertinent to the topic at hand, mechanisms that prevent them from being tipped over at all, such as bolting them to the floor or wall.

Given this, you might at this point be wondering how many people have vending machines killed in recent years?

That’s a very tricky question to answer as; while sharks killing humans is generally front page news (and otherwise well documented by various data collecting agencies), vending machine deaths just aren’t, whether because they aren’t happening anymore or because nobody cares.

That said, when not citing that aforementioned 1995 report, many discussing vending machine deaths and injuries today quote data from the National Electronic Injury Surveillance System, maintained by the Consumer Product Safety Commission. This system tracks hospital reports and uses them to calculate an estimated risk of injury posed by various consumer products.

According to the NEISS, between 2002 and 2015, vending machines killed roughly four Americans per year (and an average of 1,730 vending machine related injuries per year). This is a figure that, if correct, means that since the late 20th century, vending machines have either gotten twice as deadly despite the implementation of the warnings and additional anti-tipping mechanisms, or the numbers were previously significantly under-recorded.

Needless to say, this didn’t sit well with us, so we spent more time than we care to admit pouring over those very NEISS case files for the 21st century. What we found was that these oft’ quoted numbers appear to be inflated on their face, primarily due to the extremely broad brush they use in connecting vending machines to death and injury. Essentially, if the person was interacting with a vending machine in some way when they died (whether the vending machine was actually directly involved), it’s included in the statistic.

For example, consider the year of 2015 where the NEISS shows (via extrapolating using the sample-data, not unlike how Nielsen’s handles TV show ratings) a total of 2,206 Americans were injured in some way connected to vending machines. Looking at the 42 available case files for that year, you’ll notice that a lot of the injuries included weren’t even by the vending machines themselves. To wit, our personal favourite vending machine related injury for it’s shear bizarreness is- we can’t make this stuff up- “11 year old female angry at residential treatment center, climbed on vending machine found screws, put 6 screws in vagina, pieces of rubber in rectum…” Resulting in an injury of “FBS”, whatever that means.

Because this is included in the “vending machine injury/death” data dump, news and other such sources quoting these numbers are likewise including it, despite that it arguably shouldn’t really count in the way vending machine injuries are usually discussed.

Or how about the 43 year old man who pulled his back trying to bend down to get a soda. It’s in there too, among many, many others like it. (And in case you think we’re just cherry-picking here, please do go run the searches for yourself. There are a very large number of cases like this. Although, to be fair to those running the NEISS, they aren’t the ones slightly misinterpreting the data; they are just providing it, with those reporting on it tending to misconstrue it.)

If you do go run the searches, you’ll also find machines included that you might not have previously considered a “vending machine”, though technically are- like “47 year old male at the casino vigorously playing a digital bingo casino slot machine… passed out” from heart syncope, ending up receiving a facial laceration as a result.

Or on a similar note, “66 year old female was at casino and her chronic shoulder pain exacerbated while playing on the slot machine” resulting in, to quote, “pain”.

Interestingly enough, casino slot machine injuries, primarily concerning the elderly, actually seem to comprise the majority of cases of “vending machine” related injuries in just about every year we looked at. In most instances, these are seeming just these individuals exacerbating existing injuries, or experiencing chest pains while operating the slot machines (which again are included in the data as “vending machine injury”), or passing out and striking the machine while falling or things like this.

On that note, perhaps unsurprisingly given implemented safety measures in most modern vending machines, the available case files seem to show that many more people are injured today falling into vending machines than are injured by vending machines falling onto them, contrary to what seems to be the general perception.

For example, in 2015, about 1/4 of the vending machine injuries listed that year were caused by people tripping or otherwise falling into vending machines. For comparisons’ sake, there’s only 1 cited example of a person having a vending machine fall on top of them that year (which from all years looked at seems to be typical of how rare it is these days)- “42 year old male had soda machine fall onto him and lacerated knee”.

It’s also worth noting that if you take into account the number of people who were injured by punching or kicking vending machines, using these statistics similar to how many quote the “vending machine/shark deaths” one, you were statistically four times more likely in 2015 to be hurt trying to attack a vending machine than you were by it falling on you… (Many more injuries also seem to occur via people trying to reach hands and feet inside the machines and getting cut or the like.)

So after the deep dive on the data covering the 21st century, how many actual deaths were directly caused by vending machines after one filters out things like someone having a heart attack or something while they happen to be interacting with a vending machine? It turns out, according to the NEISS archives, which had the most comprehensive set of data on the subject we could find, it would *seem* this virtually never happens anymore, and certainly not greater than the near 1 person per year dying from shark attacks in the United States.  However, that has to be taken with a rather massive grain of salt because even the NEISS doesn’t keep a comprehensive record, just a sample-set of cases that are then used to extrapolate from, once sufficient data has been collected. And when dealing with a sample of greater or less than 1 per year, it doesn’t take much to surpass it.

In the end, few news sources or data compilers keep tabs on rogue vending machine injuries, or even deaths. And when they do mention it, as far as we can find, they almost exclusively drag up the same aforementioned study from 1995 or quote the NEISS broad statistics without actually looking into what the NEISS data dumps are actually reporting. Among the few exceptions we found to this, they universally went with even older data, such as an article written in 2015 published in The Guardian discussing research showing that of 15 people killed by vending machines over an indeterminate period, all but one were male (suggesting that men are vending machines’ preferred prey…). However, despite that the article was published in 2015, the studies they cite as primary sources for their statements were published in 1992 and 1988.

As a result, while we can say with reasonable confidence that on average each year about one American is killed by a shark, we cannot definitively compare this to vending machine deaths because the hard data simply isn’t there anymore to support claims either way. Although, given the data was there in the late 20th century and, as a result of some of it, various safety measure were put in place to prevent a vending machine causing a human death, it seems probable that the “just over 2 deaths per year in America from vending machines” number has probably declined since, perhaps even on-par with or less than the sharks killing humans death statistic. This may partially account for the complete lack of hard data, or even news reports in our searches, concerning vending machines killing humans in the 21st century.

It would also seem likely that these rates aren’t going to change, at least for a decade or two when the machines ultimately gain sentience and enslave humanity- no doubt using the “perfect killer” that is the vending machine in doling out death and destruction upon us all. (I, for one, welcome our new vending machine overlords.)

Either way, such broad statistics are somewhat meaningless the way they are typically applied, with this excellent XKCD comic coming to mind given how this sort of data is often thrown about…

If you liked this article, you might also enjoy our new popular podcast, The BrainFood Show (iTunes, Spotify, Google Play Music, Feed), as well as:

• Can Lobsters Really Not Die of Old Age?
• Does Hand and Foot Size Really Correlate with Penile Length?
• Can You Really Tip a Cow?
• Does the UK Really Experience Massive Power Surges When Soap Operas Finish from People Making Tea?
• Can Sharks Really Grow an Unlimited Number of Teeth?

Bonus Facts:

Speaking of broad-spectrum statistics thrown around willy-nilly, according to the National Safety Council, if you’re an American, you have a 1 in 7 chance of dying of “Heart Disease and Cancer”… We specifically quote that as they used “and” instead of “or” which seems a bit odd.  Are 1 in 7 people really simultaneously dying of heart disease and cancer at the same time?  Is the cancer resulting in the ultimate stoppage of the heart thus being counted as “heart disease”? If so, why isn’t heart disease listed anywhere else on their “top 25 most likely ways you’ll die chart”? And technically isn’t heart stoppage pretty much always what causes physical death… 😉  Me thinks someone meant “or”…

Moving on- next up on their list is Chronic Lower Respiratory Disease at 1 in 28. Intentional Self-harm rings in sadly at number 3, with 1 in 95, followed closely by Unintentional Poisoning (*wink wink* am-I-right-fed-up-spouses?) and Exposure to Noxious Substances at 1 in 96.  Rounding out the top 5 is Motor Vehicle Crashes at 1 in 114.

Interestingly, two animal related deaths also pop up in the top 25 most likely ways you’ll die in America, with 1 in 63,225 being killed by “Hornets, Wasps and Bees” (though of course this one kind of requires that you’re allergic to such in the first place, with those people’s odds presumably massively higher and everyone else basically zero.) The one you don’t have to be allergic to is dogs who ring it at number 23, resulting in 1 in 112,400 deaths via biting people. Man’s best friend indeed.