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



Monday, March 25, 2013

Bigfoot Footprints: The Problem of the Distribution Shape

A scientific study on Bigfoot footprints and other such data was published in the late 1990's:

Fahrenbach, W. H. (1997/1998). "Sasquatch: Size, Scaling, & Statistics," Cryptozoology, 13, pp. 47-75).

This work admirably attempted to tabulate and present 40 years worth of Bigfoot evidence data. I won't even get into the issue of what the researcher is counting as "good" or "acceptable" data to include in this collection of samples. Instead I want to focus here on the footprint data, because it appears to create an especially problematic issue for Bigfoot believers.


The paper considers over 700 footprint measurements purported to be left by the ever-elusive and mysterious Sasquatch creature. Here's the problem: The foot size distribution should be bimodal (with two 'peaks' or 'spikes' in the histograms) due to different sizes between males and females. But the paper (Figure 1) clearly shows only one peak in the data. The author interprets this finding as suggesting NO sexual dimorphism between males and females, but this is incredibly unlikely for the following reasons:

In primates, males are almost always larger than females. And the larger a primate is, the bigger the difference between males and females (they display a larger sexual dimorphism with increasing size). Maybe you didn't notice, but Bigfoot is supposedly 8 to 10 feet tall and weighing in at something like 400 lbs or some such nonsense (yet it's able to run with incredible speed, take bounding leaps, and elude all attempts at detection like a stealthy ninja in the woods? Right.).

The three largest primate species ever to exist (Gigantopithecus blacki, bilaspurensis, and giganteus) were truly massive, with some potentially approaching 10 feet in height and 1200 lbs in weight. All three of these species are now extinct but are thought to have possibly coexisted with very early human ancestors. Some wild and implausible theories suggest that Bigfoots are remnants of one of these Gigantopithecus species that have survived until modern times. The oversized height and weight certainly seem to match Bigfoot descriptions.

The problem is that these species had huge differences in size between males and females, exactly as we would expect, given real primate data on body size features. For instance, in the G. blacki species, males were twice as large as females, and so their footprint distributions would have been unmistakenly bimodal. And so should any of their remaining relatives, such as Bigfoots. Any primate that is as large as Bigfoot is claimed to be should have obvious sexual dimorphisms and differing footprint sizes. But the Bigfoot footprint distribution is unimodal! Even if Gigantopithecus were not Bigfoot's ancestor, the expected male-female size differential for primates should result in a bimodal histogram of footprint sizes.

The unimodal distribution of supposed Bigfoot footprints simply does not support the existence of the mythical creature, as the cited research paper suggests. Where we would expect to see two peaks in the histograms, we see only one. The shape of the distribution observed in the Bigfoot footprint data is more suggestive of hoaxers randomly selecting larger-than-human footprint sizes, or completely understandable misidentifications or misinterpretations of human/animal footprints or of just random patterns that aren't any form of footprints at all. It's another poor attempt at "proof" of an undiscovered but oft-sighted primate species secretly living in modern North America. Sorry Squatchers, but again, your supposed Bigfoot evidence just does not stand up to any serious scientific scrutiny.

5 comments:

  1. Dividing a normal distribution into males and females and expecting different peaks is kind of arbitrary. Why not assert that the distribution should be tri-modal (adults, teenagers, and children)? Virtually every population consists of sub-groups that have different means, medians, and modes, but that doesn't stop the overall distribution of everyone combined from being roughly normal. There are huge IQ differences various occupations but the IQ distribution of Americans as a whole is normal.

    According to this pdf file, you only get a bimodal distribution if the difference between two sub-groups exceeds two standard deviations:

    http://www.biostat.jhsph.edu/bstcourse/bio751/papers/bimodalHeight.pdf

    Also keep in mind that sexual dimorphism typically is much less pronounced before adulthood, and especially before puberty, and so a bigfoot distribution that samples the entire population, not just the adults, would reveal smaller sex differences.

    ReplyDelete
    Replies
    1. Dividing a population into subgroups based on age (adults, teens, children) is, I agree, an arbitrary distinction based on an underlying factor (age) that is itself a gradient, so that when taken in totality, would probably appear to be one large unimodal distribution. But dividing a population on gender is not arbitrary, as there are clear and distinct biological differences between males and females, with the size of their footprints being one obvious example. With a good-sized sample, as was apparently taken in this case, there should be a clear bimodal distribution reflecting different gender sizes. Gender size is not an arbitrary distinction. This is especially true for large primates, where the differences are bigger in comparison to "average" primates.

      Delete
  2. Couldnt the 14-16 group be female (or male) and the 16-18 group be male (female)?

    ReplyDelete
    Replies
    1. No, there is no 14-16 group versus a 16-18 group, that's the whole point I was making. There is apparently just one group, with one large unimodal peak in the data. Check out the definition of "bimodal distribution" versus unimodal, you should be better able to understand what I am getting at here. Thanks for reading.

      Delete
  3. Whatmakes you think you reaaluu know everythung

    ReplyDelete