A Rodent on Mars? More like Pareidolia and Data Dredging

Human beings are good at detecting patterns around them. Sometimes these patterns are real, like a predator rustling in a bush or the point-origin of cholera in London in the 1850s. Other times, the human pattern recognition systems make a false positive identification, such as faces on the Moon, the virgin Mary on a peace of grilled cheese sandwich, satanic backmasking in songs and so on. Therefore, it is very important to make sure that you validate a putative pattern before you announce your discovery to the world. If you do not and you are wrong, then you might just end up being criticized on a blog somewhere.

A video posted on Youtube earlier this month has been making its way across the Internet. It was put together by self-identified UFOlogist Scott C. Waring and allegedly shows what Waring thinks is a rodent on Mars. As we will see, it is just a rock that, if you zoom in, has a couple of features that kind of remind you of a rodent of some sort. However, the original large image has been selected among 100+ images and the data has been made less fine-grained by the zooming procedure. Among thousands and thousands of rocks of different shapes and sizes, some rocks are bound to look like something we know especially when our natural tendency for pareidolia has been enhanced by grainy data. Not to mention the fact that a rodent could not possibly survive on Mars.

Let us go over all of the reasoning errors, in order of appearance, performed by Waring in this video.

Confirmation bias and unchecked pareidolia

Here is how Waring describes his approach (0:23 – 1:01):

Now, I went to the Internet to try and prove that this object looks like a squirrel and there you go [showing a photo of a squirrel – Emil Karlsson’s note], there is the photograph, right there. It is in the same position. You see the arms, the pinkish around the hand, the pinkish around the nose, and the eye. Now, what is really unusual about this is look around the eyeball. Around the eyeball around this animal there is light color around it. Now look up here [at the rock – Emil Karlsson’s note] around the eyeball, there is lighter color around it, it becomes lighter. This story is getting harder and harder for NASA to deny.

Notice how Waring says that his mission was to try and prove that this object was a squirrel. An unbiased investigator would rather attempt to establish whether or not this object was a squirrel. This shows that Waring starts with his preferred interpretation of the data, and then goes out to find evidence supporting such an interpretation. In addition, he focuses exclusively on things that appear to confirm this position, rather than things that count against it. This is clear from the fact that he only looks at one reference photograph of a squirrel and most discuss the ways in which the rock resembles a squirrel and not the ways in which it is different from a squirrel. To his credit, Waring does point out that the “squirrel” rock does not have a long tail, but instead of letting this come in the way of his belief, he dismisses it by saying that NASA probably just cut it off.

From his conclusion that the rock kinda looks like a squirrel, Waring concludes that it is, in fact, a squirrel on Mars. He does not even stop to consider that his assessment of the resemblance is due to pareidolia. Instead, he moves straight to discussing how this “squirrel” has gotten on Mars.

Follow Debunking Denialism on Facebook or Twitter for new updates.

Unscientific fantasies

His main “explanation” for how this “squirrel” (which is actually just a rock) ended up on Mars is that NASA took a squirrel from earth, placed it into a state of hibernation and transported it to Mars on the Mars Rover for animal experimentation. The rationale for this is supposedly that NASA wanted to see how long a squirrel would survive on the planet with the radiation, heat and lack of food and water etc.

However, Waring is forgetting one important thing: squirrels, like all animals, require oxygen or they die. The atmosphere of Mars consists of 95% carbon dioxide and only trace amounts of oxygen. This means that a hypothetical squirrel placed on Mars would not be able to survive long enough for radiation, heat or the lack of food and water to become an issue. Death due to suffocation would occur within minutes.

The notion that this “squirrel” is part of an animal experiment by NASA to test animal survival on Mars also has another flaw: it is possible to recreate the large-scale physical conditions of Mars in the lab. An agency like NASA would probably not waste a lot of money for something that (1) can be done in a lab setting faster and cheaper and (2) wont be of any scientific relevance since the squirrel will suffocate to death pretty fast (so it is not possible to study squirrel survival in a Mars-like environment using accurate atmospheric conditions anyways).

Waring later provides an even more astonishing hypotheses: the “squirrel” represents an actual rodent species native to Mars. This has the same problems as before: how can a mammal that is large compared with organisms that acquire oxygen though passive diffusion survive in an atmosphere that is 95-96% carbon dioxide and hardly any oxygen? Further, what kind of food does it eat? On earth, squirrels eat plant material such as seeds and nuts, but there are no such plants living on Mars. Although Waring does admit that this is a far-fetched idea, I am not sure he understands exactly how ludicrous it actually is.

Fun with zooming and data dredging

Waring then zooms out from the zoomed-in photograph of the squirrel-like rock to give viewers a larger view. From a larger view (although not large enough to fit the entire photograph), the squirrel-like rock appears to be nothing more than a common rock. I fact, it is barely noticeable. This means that it would have to have been identified either by chance or by scanning an enormous amount of zoomed in frames in order to find one that looked good. Thus, the result may be due to data dredging: looking at tens of thousands of zoomed in shots, one is bound to find some false positives. If you highlight these false positives and ignore all the other frames that just look like rock-covered surfaces, one is guilty of this kind of dishonest data mining.

He also goes through the process of how he got the zoomed in frame in question. He downloaded the full-size photograph from the NASA website, enlarged it, and went through repeated 80% resizes until the squirrel-like rock is visible. While such a process may prevent or reduce the risk of the image becoming pixelated, it does nothing to prevent what is known as interpolation. Basically, the moment you enlarge a picture, you create more pixels. But what are these pixels filled with? As it turns out, the image editing software makes an interpolation: it estimates an approximation based on the color, intensity etc. of surrounding pixels. Every time an image is resized, interpolation cause a loss in image quality. The less image quality a photograph has, the higher is the risk of pareidolia.

Conclusion

It is just a rock.

Almost every part of Waring’s inquiry is fraught with error: he starts from the position he wants to prove, selectively focus on evidence that supports his idea while dismissing evidence that does not, unchecked pareidolia, data dredging and a profound ignorance of biology and scientific research. In addition, he does not even take into account the very real problem of interpolation.

References and further reading:

Pareidolia: The Skeptic’s Dictionary: Pareidolia.

The atmospheric composition on Mars: NASA Quest: Mars Facts, Space.com Mars’ Atmosphere: Composition, Climate & Weather and Case Western Reserve University Department of Astronomy: Mars.

Interpolation: Cambridge in Colour: Understanding Digital Image Interpolation.