Inferring Personality from Digital Footprints

Speaker:

Youyou Wu - Postdoctoral Fellow, Kellogg School of Management and the Northwestern Institute for Complex Systems (NICO)

Talk Abstract:

Personality is a key driver behind people's behavior, cognitions, motivations and emotions. The judgments people make of each other's personalities are the basis of common and important decisions in personal and professional lives, such as whom to befriend, marry, trust, avoid, hire, or elect as president. Like many other psychological traits, personality is latent and cannot be measured directly. Traditional personality assessment methods rely on human judgements, asking individuals themselves, peers, or clinicians to rate personality using a list of statements on a questionnaire. Despite their reliability and validity in most applications, these methods have several limitations: expensive, slow to administer, and sometimes biased.

In this talk, I will present the latest personality assessment technique free of human judgments—inferring personality from digital footprints. Based on a large sample of participants, I show that machine learning models built using one type of digital footprint—Facebook Likes—can be accurate in assessing personality. I also explore the minimal conditions needed to build models with satisfactory accuracy, including using small training samples or a short personality questionnaire. This method can be potentially generalized to the assessment of other psychological traits besides personality.

Live Stream:

To join the Meeting: bluejeans.com/8474912527
To join via Browser: bluejeans.com/8474912527/browser

Related Info