"Some say they see poetry in my paintings; I see only science." -Georges Seurat



Monday, November 19, 2012

Does Bigfoot Exist? Statistical Evidence Clearly Says No

Most people, including scientists, think that pseudo-scientific claims are essentially unfalsifiable, since a skeptic can't typically provide direct physical evidence proving the non-existence of something. But, luckily, pseudo-science proponents sometimes make scientifically testable or falsifiable claims, or at least making claims we can analyze with some serious scientific confidence. Here's an example involving 'squatches that I noticed while watching Animal Planet's Finding Bigfoot television series.

Bigfoot ("sasquatch") believers have estimated that there are between 2,000 and 6,000 bigfoots in North America. Surely, a hairy lumbering 7 to 9-foot-tall giant primate living in modern North America, snooping around at all hours of the day and night in both rural and suburban areas, would have been struck by a motor vehicle by now, right? Not so, according to the Bigfoot Field Researchers Organization (BFRO) FAQ, which claims that this painful lack of evidence is explained away by simply saying "Around humans their typical behavior is to flee or hide. They try to stay out of view or at least in the shadows when near people or moving vehicles." But that behavior describes many, many species of real animals that actually exist and nonetheless end up as highway spaghetti. If bigfoots really exist, where are all the bigfoot roadkills? Let's find out...


Humans are the most intelligent of all animals, and generally of the understanding that it's not a good idea to get in the path of a huge metal box going at a high rate of speed, like a motor vehicle. Nonetheless, human pedestrians manage to get themselves into motor vehicle accidents tens of thousands of times per year, with many accidents resulting in deaths. According to a report by the National Highway Traffic Safety Administration, in 2006 there were 67,573 car accidents involving pedestrians, resulting in 4,784 pedestrian deaths. With the U.S. population right at 300 million people in 2006, this means that 0.0225% of Americans managed to get hit by cars, and 0.0016% died from the incident. These are the annual pedestrian hit rates and the kill rates by vehicles for the human population in the United States.

Next, we can use these to estimate what should be the annual hit and kill rates for the hypothetical bigfoot population in North America. This will tell us how many bigfoot roadstrikes and roadkills we should expect to observe if bigfoots actually exist.

Non-human animals are surely more likely to be killed by vehicles, statistically-speaking, than humans who are intelligent, who build cross-walks, who are taught to look both ways before crossing streets, who typically don't cross freeways on foot, and who build pedestrian pathways to minimize the dangers of automobiles to people. Let's very generously assume that bigfoots are only twice as likely as humans to get hit by cars, since bigfoots are certainly less intelligent and presumably unlikely to be using crosswalks or sidewalks or wearing reflective vests in low-light conditions. Most good-sized animals probably have considerably higher hit and kill rates over humans, likely much much higher than only 2 times the rate of humans (possibly up into single digit percentages or more for some species). But we'll be generous for now and see where the numbers take us. So now we have an estimated annual hit rate for bigfoots of 0.0450% and a kill rate of 0.0032%.

Those rates don't initially sound high, being well under 1%, but let's apply them to the BFRO's population estimates (from 2,000 to 6,000) over the last 60 years and see what we find. If the population of bigfoot has been around 2,000 over the last half century or more, our estimated annual hit rate suggests that there should have occurred at a bare minimum 54 car accidents involving a bigfoot, with at least four involving a bigfoot fatality. If the population is instead 6,000 then there should have been 162 bigfoot vehicle strikes resulting in a dozen bigfoot deaths. Remember these are extremely conservative estimates.

How many bigfoot roadway fatalities have there been? Obviously, none. Zero, zilch, nada. OK, so how many times has a bigfoot been hit by a vehicle that did not result in its death? And what about near-misses? Let's go to the BFRO's FAQ for their own answers to these questions: "Only a very small fraction of the thousands of credible sighting reports describe near-misses with vehicles. No substantiated reports describe a collision with a bigfoot."

But wait! Our extremely conservative estimates based on comparably-sized primates living in similar areas (i.e., humans) over the last 60 years suggest that at the very least, at an absolute minimum, there should have been several bigfoots killed by vehicles, and dozens upon dozens of non-fatal bigfoot vehicle strikes leaving many maimed or wounded bigfoots to be recovered. By the BFRO's own admission, again, no carcasses have been recovered, no crippled or wounded bigfoots have been found, not a single report describes a collision, and there are almost no near-misses reported. Something is wrong here.

Now consider these numbers. There are roughly 400 million animals struck and killed by motor vehicles each year in the U.S. alone. That means over the last 50 years, in the U.S., there were something like 20 billion animal roadkills. And not a single one of those 20 billion carcasses were a bigfoot? How is that possible? Seriously, what are the odds that not a single bigfoot has ever been maimed or incapacitated by a vehicle strike, and there are almost no near-misses reported, and not one bigfoot has been found among the 20 billion animals killed along U.S. roadways in the last half century? Surely any one of these events should have happened by now, but none have. With these probabilities considered together, it's spectacularly and astronomically unlikely that not one of these events has happened yet, if indeed bigfoot truly exists. It's effectively impossible.

This same sort of analysis applies not just to the lack of bigfoot roadkill, but also to the observations that no bigfoot has ever accidentally (or purposefully) been shot by a hunter or caught in a trap; no bigfoot has ever accidentally drowned or slipped on a mountainside or died of natural causes in such a way that their carcass could be discovered by humans; and of course that not a single bigfoot has ever had a clear non-grainy photograph or video taken of it. All of these things, or at least some of them, or even just one of them, should have happened by now. The statistics are clear. The fact that not one of these events has happened yet tells us something important. The inescapable conclusion is that bigfoots don't exist, and we can say this with high scientific confidence.

14 comments:

  1. Garbage in, garbage out. Your knowledge of the history of bigfoot and auto accidents is abysmal.Non human animals are surely more likely to be hit by cars? Im sure when you were getting your doctor, like I did as well, your profs would have questioned you using an assumption like that!The variables themselves are unbelievably large.Best wishes.

    ReplyDelete
    Replies
    1. There is no history of bigfoot, since they have not been demonstrably proven to even exist. My knowledge of auto-accidents is abysmal? I only cited US government reports, what is the problem exactly?

      And yes, non-human animals are much more likely to get hit by cars. For instance, there are about 80 to 100 Florida panthers, but 15 of these were killed by vehicles in 2007. That's a 15% annual road kill rate for that population.

      And yes, some of the numbers I cite are large, that's sort of the point.

      Delete
    2. The issue, here, is that you assume that Sasquatch are necessarily less intelligent than humans. I assume that this is based on the fact that they are less advanced - except that there are many human beings still stuck in a tribal state of living. Not that I necessarily believe that Bigfoot exists - but there's always the chance that they're more intelligent than us and avoid our society and technology because of the proven damage that they cause.

      Delete
    3. So wait, you are accusing me of assuming that Sasquatch are less intelligent than humans? I am assuming they don't exist at all, so their intelligence is basically an irrelevant question. Also, I would never equate "un-intelligent" with "living in a tribal state of life".

      Delete
  2. Awesome article! You complete me. Yes I'm sure moose, elephants, rhinos, etc. have been hit. A good thing to keep hammering home is you are just talking about the USA, but your points also totally hold, per logic and reason, when looking at the situation globally i.e. Russia, Canada, China, Brazil, Australia, etc.

    ReplyDelete
  3. Thanks Link, I would prefer to use comparable non-human animal data for better estimates but have been having trouble tracking some down. This definitely should hold for all locations as long as the statistics exist.

    ReplyDelete
  4. At first I found the analysis slightly compelling, but as I scrutinized it more, I don't think it works at all.

    First of all, the collision rate is going to be directly proportional to the distance between the animals natural habitat, and the vehicle density in that area. Humans, are most likely to be found in the areas where there is the largest number of people. Deer and Bear don't frequent Urban areas but it in not uncommon to have them show up in urban or sub-urban settings. Without a doubt racoons are better at avoiding getting hit by cars than deer, but the fatality rate for racoons is probably higher. Exactly 0 porpoises are ever hit by cars, because the habitat is exclusive to cars and trucks.
    Of course we don't know the habitat of the creature, but dense forest seems to be negatively correlated with motor vehicle density.

    Second, the (supposed) creature is (supposedly) nocturnal. This in and of itself wouldn't effect the hit rate, since the decrease in vehicle density would be countered by the increase in difficulty of seeing the animal. However, we would expect daylight to increase the effect of density. Meaning a very low density road is going to have even fewer deaths at night, and a high density road will have more, because the roads will have a proportional reduction based on time of day.

    Three, the population of cars has not remained static over 60 year, but has been increasing. Assuming the number of cars on the road has tripled, you would expect the first year to have a 1/3 the likelihood of a hit as right now.

    Four, are Gorilla's ever hit by cars? What about Bonobo's? Do we know of a case ever of a non-human ape being hit by a car anywhere?

    Five, I find the 400 million animals number hard to believe. there are only 250 million registered vehicles in North America, that statistic would indicate each vehicle in America killing an average of almost 2 animals per year.

    I understand how your line of reasoning took you to your conclusion, but I'm overwhelmed by the possibilities in alternative hypothesis.

    ReplyDelete
    Replies
    1. Thanks very much for your comments Shea, I can tell you put a lot of thought into my analysis, and critiquing it, and I appreciate it. I think you bring up some very good points, especially that humans are probably most likely to be struck by cars in urban areas (where bigfoots are unlikely to be treading). And I'm sure like in most vehicle accidents, alcohol is a primary culprit for both drivers and pedestrians, which probably has little bearing on bigfoot strikes. This could make human pedestrian accident data potentially questionable as applied to bigfoots.

      I used human data for my estimate, despite these drawbacks, for three main reasons: [1] the statistics are very clear (regarding accident rates and population rates); [2] humans are comparably sized primates living in similar though not strictly the same areas; and [3] this was likely to provide a very conservative estimate in comparison to animals.

      Indeed, if you fiddle with the estimates a bit, tweaking this way or that, you almost always get a non-zero estimate for the number of bigfoots that should have been struck or killed by now in auto accidents. But I'm happy to grant you your points and use animal data instead. I already pointed out that Florida panthers had a road kill rate as high as 15% in a single year. Let's look at black bears next, some good data seems to exist on this animal:

      Black bears are comparably sized mammals, are purported to feed on much the same things, and have almost the exact same habitat as bigfoots (indeed, an analysis of sighting locations suggest that bigfoots and black bears have nearly identical habitats, see: http://onlinelibrary.wiley.com/doi/10.1111/j.1365-2699.2009.02152.x/full. One might suspect from that analysis that most bigfoot sightings as mis-identifications of black bears, but that's not the issue I want to dwell on here. Instead let's look at the black bear car accident rates to apply to bigfoot.

      Again, in Floria, there are about 1500 black bears in existence (http://www.swfwmd.state.fl.us/education/interactive/springscoast/blackbear.shtml). Recently, over a five year period, 125-175 black bears were killed each year in Florida (http://www.defenders.org/black-bear/threats). This is a 10% annual road kill rate. If bigfoots had only half this rate (5%) then there should be 100 to 300 bigfoots killed by car accidents EVERY SINGLE YEAR in North America, if the BFRO estimates of 2000 to 6000 are accurate.

      Consider also that up to 90% of black bears die of non-natural causes, including car accidents, shootings, trappings, and related human activity (http://www.obergatlinburg.com/activities/wildlife-encounter/american-black-bears/). Being a mammal of similar size living in the same areas, bigfoots should be dying in similar magnitudes from similar causes. But none are. Bigfoot believers can come up with all kind of theoretical and mostly non-sensical explanations for why not, but they are missing the obvious, more pragmatic explanation: bigfoots don't exist.

      Lastly, I'll add that even supposing the 400 million animal roadkill statistic is an overestimate by a factor of 10, then that still leaves several BILLION dead animals over the last 60 years on US roadways of which ZERO were a bigfoot. Exact numbers are less important than the general magnitudes we are working with, so don't get too bogged down with estimate accuracies for numbers so large.

      Thanks again for your comments.

      Delete
    2. That's a good response.
      I'm typically ambivalent in my positions on things like this. Unless a claim is preposterous (i.e. homeopathy), or confirmed (i.e. Giant Squid), I tend to take a "perhaps" point of view. I take a claim to be preposterous if and only if there are no alternative explanations worth actual merit that would allow it's possibility.
      So, I'll definitely not defend the claim that sasquatch is real, only that there exists not vanishingly small possibility.
      Or at least I'm going to try right now...

      In order to do that I'm going to need to make a compelling case that brings your numbers to something more feasible.

      Your analogy with black bears is pretty good. Though the alternative explanation was actually embedded in the evidence. Specifically, bigfoot seems to share the same habitat because a large number of bigfoot sightings are misidentifications of black bears. I will of course note that once you grant that, the entire basis for the population estimate seems more absurd than it already was...

      There is also a chance that sasquatch have been hit by cars, shot, etc... and that it was just never reported. Certainly we know that there are people who have been afraid to report encounters before...

      So lets take your hit rate:
      162 bigfoot
      If we assume that due to them being nocturnal and shy that we could reduce your hit rate in half, that would bring us down to 81 hits.

      Assuming a linear trend in motor vehicle growth, we could assume base on 1/3 the number of cars 60 years ago, that that would knock off about another 30ish. We'd be down to 51.

      The last possibility for bigfoot is that the habitat estimates are wrong. That could reduce the hit rate down to 5 (though it would call into serious doubt to concept that they are migratory).

      So how many accidents could we expect before we had could be reasonably certain a report would be made. I'd say that number is probably 4. (Note, I think 1 would be at least 50%, but reasonable certainty 95%)

      Given that 5 is really close to 4, I'd say that sasquatch has a few more years till we can say No Way...

      Though these are napkin numbers. The fact is that we don't have enough data. I thought of doing a Bayesian analysis to figure out if you were right, but I couldn't find enough data to fill in the blanks. Since I'm skeptical of the numbers we're bringing up I have doubt about your conclusion, and doubt about my own doubt. :)

      What I'd really like to find is a single case of a Gorrilla getting hit by a car (I googled it, and couldn't turn up anything, but I think a report about that is likely to be in another language). If a Gorilla has been hit by a car (even once), it would certainly seem to be a strong validation of your argument (Gorillas share many of the same features, and the population of gorillas relative to nearby humans is perhaps equitable).

      Anyway, I think you make a strong case, but I don't think it's a death blow...
      I'll look at your hunting one a bit later, as you may be able to tell from this response I'm kind of tired.

      P.S. 90% of black bears are killed by human activity? If that is true, it is rather depressing.

      Delete
    3. I kept thinking about this, and I decided to do some research. Again, the ideal situation here would be a confirmation of either the hypothesis that the creature is real, or a denial of the possibility. You are making, a rather good, argument for the matter.

      So redoing your work with available data...
      I had some other numbers here, but then I looked back at your source, and noticed you got the numbers wrong.

      Page 6: "Nationwide, nearly two pedestrians died in vehicle crashes per 100,000 population, and a pedestrian crash death occurred every 70 million miles walked."

      That means if bigfoot has a population of 2000, you'd expect .04 bigfoot deaths per year (assuming a hit rate equivalent to humans). Which still would mean 2.4 expected over the course of 60 years assuming the rate was constant.

      Page 16: 3 times as likely to be hit in an urban area.

      .8 expected bigfoot hits over 60 years.

      I tried looking at season, but couldn't find any reasonable information linking sightings to season.

      Though the majority of fatal crashes do happen between 6pm and 9 pm, the 6am to 6pm are low enough, that I think the data suggests that I was wrong about time of day, it either increases the likelihood of bigfoot hits, or has no effect.

      So, assuming your thesis, and given your data, I'm getting .8 bigfoot deaths over 60 year. I'm open to the possibility that it was me that did something wrong...

      I also want to look at my Congo Gorilla counter argument for comparison. This link shows a 5.5 per 100000 in terms of all accidents. If we assume the same ratio of pedestrian involvement in crashes, that gives us .825 per year per 100,000. There are a little over 100,000 gorilla's. Assuming that we can apply the logic we used on bigfoot, we should expect almost .3 Gorilla hit per year in the Congo. Or 20 over 60 years, unfortunately I can't find any evidence to corroborate our mathematical hypothesis.

      Therefore, I can not reject the null nor the alternative hypothesis...

      Delete
    4. Thanks again for you comments Shea. The 162 estimate is the human equivalent data, which by my reckoning was a seriously conservative estimate (or very low-ball estimate). Probably more realistic is the animal data which suggests hit rates in the single digit percentages per population annually. Applying this to gorillas or other primates in the Congo is problematic because there is so little data, and the BFRO bigfoot population estimates apply only to North America as far as I know. Plus there are so few roads in the Congo as to be useless for this analysis. In fact, if Bigfoot believers were arguing that Bigfoot only exists in the Congo I might be more likely to be sympathetic to their view, but they are claiming thousands of Bigfoots in North America, which is just plain rubbish based on comparable human and animal data taken as a whole.

      As long as the hit rate in North America is anything above humans (and anywhere close to animal rates which seem to be in the single digit percentages or higher) there should have been dozens or perhaps even hundreds or more Bigfoots hit by now.

      Delete
  5. Unfortunately I can't follow your math. Your own link had 2 per 100,000, which would put the hit rate at between .04 and .12 per year, far less than the 2.7 you calculated. I showed the math in my second comment. And how when you adjust for the likely rural distribution of bigfoot.

    I really want your math to work, but I'm just not seeing it.

    My analysis with the Congo, was using identical reasoning. Again I showed the math I used there.
    We are making the assumptions that:
    an ape should get hit at least as often as a human
    that the population of apes times the expected hit rate represents number of deaths
    And I'm assuming the the rural/urban distribution effects the hit rate

    If those assumptions are correct, we should expect the listed number of Gorilla deaths. I know the Congo is more rural, but it's not like they don't have cities or logging roads.

    Am I missing an assumption? Can you show me the math in such a way that derives your 162? Because this "in 2006 there were 67,573 car accidents involving pedestrians, resulting in 4,784 pedestrian deaths. With the U.S. population right at 300 million people in 2006, this means that 0.0225% of Americans managed to get hit by cars, and 0.0016% died from the incident." is far less accurate than simply using the listed number in your cited source of "Nationwide, nearly two pedestrians died in vehicle crashes per 100,000 population."
    You are claiming a hit rate of .000225, and the number from your source claims a hit rate of less than .00002. That's an order of magnitude difference.

    ReplyDelete
  6. Sure let's go over your points. I was using 2006 data, the only reason I didn't use the 10 year period is it did not specifically say how the per capita rate was calculated. But looking at the cited report (Death Density section), over the 10 year period spanning 1997-2006, there were 1.7 fatalities per 100,000 , and 27.8 crashes involving pedestrians per 100,000. Notice there is about one order of magnitude difference in these rates (which are hit rates versus kill rates). Not everyone dies who gets hit, that explains the disparity between my calculation and yours (hit rate is higher than kill rate).

    Here is some more mammal data I was able to track down for comparison: Deer collision rate per deer population in Ohio is approximately 5%:
    http://www.culturechange.org/issue8/roadkill.htm

    So far, virtually all the large mammal data (panthers, bears, deer, etc.), have annual roadkill rates in or above the single digit percentages per population (ranging from 5 to 15%, apparently).

    Here is some limited primate data, I could not find any on gorillas, chimps, or orangutans (this data may exist, I have no idea, I just don't know how to find it):

    Colobus monkeys now use specially-built "bridges" across roadways that were built due to the problem of vehicle strikes:
    http://www.neprimatesanctuary.org/animals_monkeys_colobus.html
    http://itech.fgcu.edu/faculty/sstans/colobus2.htm
    http://www.primate.org/news2000.htm

    Vervet monekys are reported to often be shot and hit by cars:
    http://www.vervet.org/

    Vervet monkeys & yellow babboons
    http://www.bornfree.org.uk/animal-future/project-focus/

    News report of several proboscis monkeys being killed by vehicle strikes:
    http://www.bt.com.bn/news-national/2010/09/23/proboscis-monkey-killed-traffic-accident-near-cleared-mangrove-site

    Hope this helps, and please keep calculating away and sharing your results.

    ReplyDelete
  7. Thanks for the response. That certainly explains the difference in our numbers.
    "this data may exist, I have no idea, I just don't know how to find it" is the same problem I ran into. It is possible that ape's get hit less than humans, or more, without tracking data, we can't really know. Thailand is the only place I can think of with a comparable modernity to america that has a native ape population.
    Like I said, though I find the argument interesting, it isn't compelling enough to reject the idea that bigfoot exists. :)

    ReplyDelete