Physiognomy, the practice of deciphering a person's personality based on appearance, dates back to 500 B.C. The question is, is it pseudo-science, or does it actually hold credibility? One study published by Scientific Reports may support the theory that physiognomy may actually be legit after all.
The research took place in Denmark and utilized machine learning techniques on thousands of faces to predict their political ideology. The study, led by Stig Hebbelstrup and his research team, explored if computational neural networks (CNN) can accurately determine a politician's political stance based on a single photograph of their face. Sounds dystopian, right? Surprisingly, the predictions were successful 61% of the time.
CNN aims to learn patterns and connections in data by adjusting neuron links. This learning process is referred to as "training" or "optimization" – so whenever there's an output error, the functions are modified in the previous nodes to be more accurate each time. The training process took some work. Hebbelstrup and his colleagues used photos of Danish politicians, with the first dataset consisting of 5,230 pictures. Those who were more moderate – or could not be identified as right-wing or left-wing – were excluded. They also excluded candidates of non-European origin because they were 2.5 times more likely to be representative of left-wing parties. Lastly, the researchers removed photos of politicians with beards because they "can impair some of the subsequent analyses."
In the end, they were left with 4,647 images, with 1,442 of them being female. The sample was divided by both genders, and the algorithm was applied to them separately. They found that masculinity and attractiveness weren't linked to ideology in men, but happy faces (both men and women) were likely to be representatives of right-wing parties. Meanwhile, politicians who had a neutral expression or showed contempt were more likely to represent left-wing ideologies. How interesting, but we're not surprised.
Three years ago, we reported on a study that revealed over half of white, liberal women under 30 have a mental health disorder. These findings were backed by more recent data that showed that liberal women are statistically the unhappiest and most mentally ill demographic in America. As we all know, most U.S. media leans to the left. So could the harmful narratives and advice be the reason for liberal women's unhappiness?
Believe it or not, this is not the first study to come out on attractiveness and political beliefs. In 2018, a study published in the Journal of Public Economics echoed these findings by stating that attractive people were likelier to be right-wing. Authors Rolfe Dause Pterson and Carl Palmer wrote, “Controlling for socioeconomic status, we found that more attractive individuals are more likely to report higher levels of political efficacy, identify as conservative and identify as Republican.”
Let's go back to the Danish study. 60% isn't high, but it's not exactly low either. Considering this seems like the first time Hebbelstrup's done this, that rate is actually pretty good, and there's a chance it'll only increase with more training. “Our results confirmed the threat to privacy posed by deep learning approaches. Using a pre-developed and readily available network that was trained and validated exclusively on publicly available data, we were able to predict the ideology of the pictured person roughly 60% of the time in two samples,” the authors conclude.
They continue, “We also provide the first demonstration that model-predicted ideology connects to independently classifiable features of the face. For females (though not males), high attractiveness scores were found among those the model identified as likely to be conservative. These results are credible given that previous research using human raters has also highlighted a link between attractiveness and conservatism.”
Evie deserves to be heard. Support our cause and help women reclaim their femininity by subscribing today.