What do people THINK a serial killer’s smile looks like?

As an experimental and informal survey, I posted four photos of different people smiling with only their lower faces exposed. Two of the four were smiles of convicted serial killers. The other two were actors. Survey participants were asked to vote on which two smiles they believed to be the smiles of serial killers. The main purpose of this survey was to explore potential biases in what people THINK a serial killer’s smile may look like.

^{See NOTES 1-3 at the bottom of the page for survey limitations.}

Results revealed near the end of this post.

beautiful vs. creepy

jeepers creepers 

Pre-serial killer considerations, I simply intended to study what causes people to perceive certain smiles as creepy. From observing trends in art, social media, ranking systems, and pop culture, I identified two main types of “creepy smiles”: Type I, which I coined The Grinch Pinch and Type II, which I coined The Muted Shark. Types I and II typically contain all or many of the following features:

Type I – The Grinch Pinch

• • wide, high-intensity lip-corner pulling (mouth smiling from zygomaticus major muscle)
  • sharp features (can be related to cheeks, nose, eyes, teeth, etc.)
  • exaggerated orbicularis oculi contraction (smizing) 
  • low facial fat content- especially around the infraorbital triangle (area between eyes and laugh lines)
  • multiple pronounced facial wrinkles due to exaggerated orbicularis oculi contraction, high intensity lip corner pulling, and/or low facial fat content

Type II – The Muted Shark

• • increased eye openness due to lack of orbicularis oculi contraction or to presence of upper lid raiser (widened eyes)
  • muted lip corner pulling due to either:
    — low intensity zygomaticus major muscle
    — co-activation of muscles associated with emotions such as anger, contempt, fear, and disgust

It should be pointed out that Type I and Type II represent exaggerated versions of a trend; however, “creepy smiles” exist on a spectrum and features from Types I & II can be blended to form a combo type. Observe below:

maybe it’s Maybelline 

One interesting finding was that there seems to be a fine line between what defines a “beautiful smile” and what defines a “creepy smile.” 

A few months ago, while researching anatomical variations of the zygomaticus major muscle, I got pretty deep into the aesthetics of a smile. Considerations for “beautiful smiles” – taken from trends in plastic surgery requests, celebrity smile rankings, art, beauty concepts – often featured celebrities with a specific (but not uncommon) muscle variation known as the bifid zygomaticus major. Bifid smiles tend to be wider and more exaggerated with sharper lip corners. Interestingly, exaggerated mouth smiles is one of the cornerstone features of Type I creepy smiles: The Grinch Pinch. 

how to draw the line

While Grinch Pinch smiles and “beautiful smiles” both feature wide, high intensity lip corner pulling, perceptual differences regarding pleasantness seem to be tied to a few different things:

• • sharpness of features (consistent with theories in shape psychology)
  • pronouncedness of wrinkles around the eyes and mouth
  • a marginal difference in intensity balance

Sharpness of features and pronouncedness of wrinkles can be linked to facial fat content. With more facial fat comes rounder features and fewer wrinkles. With less facial fat comes sharper features and more wrinkles.

So, perhaps the line between creepy and beautiful is not one line at all – but many lines (and some edges). 

back to the serial killer stuff

Again, here is the original survey I posted on LinkedIn.

Here are the identities behind the smiles and the results from the 37 valid answers collected. 

theory vs. practice

Reviewing the collective characteristics of Type I and II creepy smiles, do the % results align with the theories of what makes a smile “creepy”?

Let’s see.

Type I & II features:
(Remember, we must refer to the lower face only, since that is what voters based their assumptions from.)

• • wide, high-intensity lip-corner pulling (mouth smiling from zygomaticus major muscle)
  • sharp features (can be related to cheeks, nose, eyes, teeth, etc.)
  • exaggerated orbicularis oculi contraction (smizing)
    While the eyes are not shown, most effects of orbicularis oculi can still be inferred from cheek area.
  • low facial fat content – especially around the infraorbital triangle (area between eyes and laugh lines)
  • multiple pronounced facial wrinkles due exaggerated orbicularis oculi contraction, high intensity lip corner pulling, and/or low facial fat content
  • muted lip corner pulling due to either 1) low intensity zygomaticus major muscle 2) co-activation of muscles associated with emotions such as anger, contempt, fear, and disgust

Did the features associated with The Grinch Pinch and The Muted Shark predict the outcome of the serial killer survey?

Not really.

But why? 

parts vs. whole

Aside from the truth that most things depend on things that depend on other things, we process faces holistically, and isolating features may not be helpful in this scenario. 

Additionally there were just about a billion variables that could easily confound the survey results, including but not limited to:

• • face direction
  • head position
  • lighting
  • photo quality
  • conscious identity recognition
  • subconscious identity recognition 
  • groupthink due to visibility of answers

but still – most people chose A and left out B

Variables aside, what interested me the most was the disproportionate tendency to assign serial killerness to option A, Rodney Alcala, and the disproportionate tendency to assign serial killerness to option B, James Franco. 

Before we dig into those stats here are some “fun” facts about Rodney Alcala:

• • Rodney Alcala was a serial killer active around the same time as Ted Bundy.
  • He is known as The Dating Game Killer. 
    • In 1978, Alcala was a winning contestant to a popular ABC show, The Dating Game. 
    • Despite his victory, the prized bachelorette refused to go through with the date, because she considered Alcala “creepy.”
    • The photo displayed in the survey was taken from the film footage during his appearance on The Dating Game.
  • He studied film under Roman Polanski at NYU.
    • Alcala had been described as a successful photographer.
    • He frequently posed as a fashion photographer to attract his victims.

Both James Franco and Rodney Alcala attended UCLA.

back to the stats

So, 78.4% of voters chose Rodney Alcala to be the most likely serial killer. Followed by Ted Bundy at 59.56%, Willem Dafoe at 45.94%, and James Franco at 16.22%. Because voters were asked to choose two of the four options, each identity had a 50% chance of being selected. If we take chance into account, Ted Bundy and Willem Dafoe do not deviate too far from 50%. Alcala and Franco, however, both seem to defy chance – in opposite ways.

But wait . . . There’s more!

There is one additional interesting finding of the Franco-Alcala stats. When looking at how people chose answers in group form (remember – participants were required to guess two answers): the Alcala-Franco combo was never chosen. 

There were six possible combinations, each with a 16.67% chance of being chosen.

The chances of 0 people out of 37 people choosing AB is pretty low – especially when taking into account how many people chose A. That probability combined with the pro-Alcala and anti-Franco voting trends is intriguing. 

Let’s go back to some variable considerations.

In this very specific set of lower-face only snapshots, Alcala’s photo features the:

• • least wide and most muted smile
  • worst quality photo 
  • darkest skin pigmentation
  • only non-forward facing photo

Whereas James Franco’s photo features the:

• • widest and most exaggerated smile
  • good lighting
  • best quality photo

Which of these characteristics could have contributed to the disproportionate votes? Realistically, it is extremely unlikely to gather useful insights from this micro study; however, just to be safe, make sure your LinkedIn headshot is:

• • high quality
  • well lit
  • mostly front-facing
  • showing off a sturdy smile

Again, if any researchers would like to collaborate on a related (but more  legitimate) study, please reach out to facetheFACS@melindaozel.com

NOTES & CAVEATS

NOTE 1: I do not directly believe it is possible to correctly identify a serial killer from a single photo. From a series of photos or from video footage? More likely – but still a maybe. Even in such instances, the ability to determine character would be deeply affected by various obstacles such as individual assessment skills, individual biases, cultural biases, relationship to said killer, perceived attractiveness of said killer, etc.

NOTE 2: Due to sample size, potential recognition of actors or famous killers, and research methods, survey results should not be considered a true representation. Please be aware of the many factors that can influence one’s perception of a facial expression such as: cultural biases, racial biases, mood biases, photo quality, head angle, lighting, teeth, etc. That being said: If any researchers with access to more resources would like to collaborate on a real study, please reach out to me at facetheFACS@melindaozel.com. I have more ideas for similar – but more robust – studies.

NOTE 3: Most of these statements may apply primarily to Western culture. I cannot generalize globally.