Why Eye Witness Accounts Are Not Good Scientific Evidence

People think that because there are thousands of eye witness reports of Bigfoot sightings (or UFO's, or angels, or Loch Ness monsters, or unicorns, or whatever) there must be some truth to the matter. They admirably want to remain open-minded about things. Surely all those eye witnesses can't be wrong, right?

Let's use some simple math to show where this line of thinking leads people astray. Suppose we are saying to ourselves: "There were something like 3,500 'credible' Bigfoot sightings in the U.S. over the last several decades. Even if 90% of these eye witness reports are wrong or inaccurate, that still leaves 10% or around 350 accurate reports. Therefore, Bigfoot likely exists, and there are probably hundreds or maybe even thousands of them! Yay!!"

The problem here is that people are working with the wrong set of numbers. What we should be looking at instead as our starting number is the number of events in which a Bigfoot sighting is at least theoretically possible, and from that starting point figure out how many might be likely to be credible or reliable (and how many not). As an example using real data, we can check how many black bears there are, use this to estimate the number of black bear sightings by humans, and then use an estimated misidentification rate to see how many mistaken eye witness reports we might expect to see in which 'squatches are mistaken from black bears. We will try to use conservative numbers when estimating. Then we will compare our results with the actual count of eye witness reports...

There are at least 735,000 black bears in North America (excluding Alaska, Texas, and a few other states due to lack of data). If we assume that on average each of these bears gets sighted once per year (some get sighted multiple times, others don't get seen at all), then there are 735,000 events per year with the potential for a mis-identification by a human viewer. Over the last 60 years, this means that there were roughly 44.1 million sighting events of black bears by humans with the potential for a mis-identification.

Let's be generous to witnesses and assume that 99% of the time a black bear is seen by a person, it is correctly perceived as a black bear. But 1% of the time, a misperception occurs, perhaps due to environmental factors like low light levels, fog or rain, far viewing distance, tree coverage, "eyes playing tricks", or whatever else. Sometimes, the black bear is misperceived as a brown bear, other times as a deer, still other times as other mammals, maybe as people, maybe as an alien. Occasionally, though, a black bear is misperceived as a non-human primate (let's say 1% of these misperceptions).

So taking 1% of 1% of 44.1 million viewing events leaves us with the prediction that there should be about 4,410 reported instances over the last six decades of someone seeing a non-human primate when in fact what they really saw was a black bear. It's an interesting coincidence that there are just about this many "credible" Bigfoot sightings reported; while the numbers aren't identical they are certainly close and of similar magnitude. And it's also a strange coincidence that the hypothetical Bigfoot habitat, which has been geographically-determined based on the spatial locations of eye witness reports, coincides almost exactly with the habitat regions for black bears, to a degree that strains credibility.

The nice thing about this analysis is that it makes absolutely no assumptions regarding whether Bigfoot exists or not. It simply shows that there are more than enough expected mis-identification possibilities to account for all the 'credible' Bigfoot sighting reports, without any of them actually being true. And it seems likely that black bears are the likely culprit driving many (if not most) mis-identifications.

...........................

I'm an experimental psychologist by training, specializing in perception. If there is anything close to a scientific law in psychology, it's that what people report seeing, hearing, feeling, remembering, or experiencing is just not reliable. Perception is not a reflection of reality, it's a model, a mental representation constructed by our brains. Usually this model is fantastically accurate, but sometimes it's not. If we bring people into the lab, sit them in a dark room, and have them press a button whenever a dim light is flashed in front of them, their performance will almost never be perfect over repeated trials. Accuracy may be high, but it's rarely perfect.

So even on the simplest possible perceptual task we can imagine (seeing a flash of light in a dark room), people still make mistakes and misperceptions about what was experienced (sometimes seeing the light when it is not there, sometimes missing the light when it really is there). They might even be absolutely certain and 100% confident that their perceptions were accurate, and still be wrong, I see it all the time just in the lab. Now imagine how many more perceptual errors we might expect when identifying an animal through trees, from a distance, in possible low light levels, with maybe poor eye glass prescriptions; it could be foggy or rainy, plus a million other intervening variables that complicate visual processing ability when identifying a moving blob.

The legal and law enforcement professions know all about this eye witness reliability problem. Eye witness inaccuracies are the single greatest cause of wrongful convictions, contributing to something like 75% of the legal cases that are later overturned through DNA testing. Unfortunately, you just can't totally trust what people say, or report seeing. You can certainly believe them when they say they saw something; I don't think most people are lying or are stupid or anything like that. I think they literally saw what they say, in the sense that their perceptual system reported to their consciousness that a visual identification was made, a match was found in their heads, even if it happened to be the wrong one. But again, perception is not reality, it's just our brain's best attempt at a representation of reality given limited, noisy, and sometimes poor-quality data about what's out there in the world. This is well-understood in the psychological sciences, neurosciences, legal, and law enforcement professions. By now it should be common knowledge, but often isn't.

Sadly, we must acknowledge that eyewitness reports are not good or reliable evidence, and they certainly don't count as good scientific evidence for proving the existence of mythological creatures for which no other good quality scientific proof can be fostered.