Authored by Luiza Quinn* & Linda K. Kaye
*Luiza Quinn is currently a BSc (Hons) Psychology Student at Edge Hill University
Digital traces are a fascinating field which can bring on many opportunities for psychological science. That is, to what extent can we understand people based on their digital footprints? Here is an example of a study in this field that helps us understand people’s personality based on photo behaviour on Instagram.
Ferwerda et al. (2015) aimed to show how Instagram picture features could relate to personality traits. These traits often hold valuable information that can be used by algorithmic systems to tailor services to individuals. However usually collecting personality data can be time-consuming and extensive for psychologists. In this article, the researchers wanted to illuminate a new approach by demonstrating how personality might be inferred from online behaviour instead, in the form of photo-sharing characteristics.
How Was This Done?
Like many studies in this field, the researchers used the Big Five Inventory which incorporates personality traits of openness, conscientiousness, extraversion, agreeableness and neuroticism. An online survey was issued to 113 American Instagram users, aged between 18 and 64 (median 30, 54 male and 59 female). The participant’s Instagram account pictures were also collected, totalling at 22,398. Distinct features of each photo were measured including:
· Saturation
· Brightness
· Hue
· Content-based features
Combining saturation and brightness also allowed the researchers to determine emotional tone of the image through Pleasure-Arousal-Dominance dimensions (see below visual for example of these dimensions)
So, What Did They Find?
According to their findings, there are specific markers we can use to detect people’s personality from users' photo behaviour!
Those scoring highly in openness uploaded pictures with cold colours, such as blue and green, containing few people and faces while expressing more dominance and arousal than pleasure.
Those scoring highly in conscientiousness had a mix of dull and vivid coloured pictures, meaning high saturation variance.
Those scoring highly in extraversion showed more dominance than pleasure and, similarly to openness, uploaded pictures in green and blue hues with low brightness.
Those scoring highly in agreeableness were found to upload pictures that were neither too dark nor bright, meaning balanced brightness.
Those scoring highly in neuroticism were found to use variations in brightness.
Looking further into this, the researchers next trained a machine-learning-model to predict future data patterns using the software tool Weka. They found the most successful predictions related to the personality traits of openness, conscientiousness, and agreeableness, whereas the more difficult traits to accurately predict were extraversion and neuroticism.
Overall, the findings imply that there are ways of understanding personality traits from the way Instagram users filter their photo content.
Article citation
Ferwerda, B., Schedl, M., & Tkalcic, M. (2015). Predicting personality traits with Instagram pictures. EMPIRE '15: Proceedings of the 3rd Workshop on Emotions and Personality in Personalized Systems (pp. 7–10). https://doi.org/10.1145/2809643.2809644
Comments