Professional Documents
Culture Documents
Roll no: 43
others’ or one’s own personality. In order to make these judgments, one may use quantitative
data obtained from personality tests or use subjective data such as accounts of routine behavior
and common patterns seen in them, or use a combination of both these methods. Typically,
various factors impact the accuracy of this assessment, especially the amount of visible
behavioral cues and one’s ability to integrate them into a singular judgment. The display of these
behavioral cues might vary according to factors like cultural upbringing and social desirability,
for e.g, even though one might be highly extraverted and jovial, it may be frowned by one’s
culture upon to behaviorally express the same in public. When it comes to being the recipient of
personality judgment, one’s reference group seems to hold importance. In a series of 5 studies,
Wood et al (2012) found that the way an individual’s behavior is ranked in a group and in
comparison to behavior shown by other group members affects his/her personality ratings.
Researchers today are also aiming to reduce self-other and subjective rating errors in personality
judgment with the help of machine learning and AI. Youyou et al (2015) conducted one such
study, comparing the accuracy of judgments as made by computer models and humans with
relation to self-ratings. Their results showed that computers’ accuracy increased with no. of likes
on the person’s profile, its judgment being as good as that of an average judge and spouse.
Prominent accuracy was seen for the dimension openness. Agreement amongst computer models
as well external validity of these models was also high. Overall, computer models were seen to
be on-par or even better in providing personality judgments. Despite robust findings, there are
First, facebook likes were used as a basis in computer models. It is a known fact that social
desirability is just as prominent in one’s online presence as in reality. Arguably the type of
content one puts “likes” on facebook might reflect real-life personality however we cannot deny
that targeted ad campaigns and content also influence behavior. For example, I might be buying
few novels every week because I like reading and thus I like every post of an online bookseller
but I also might like those posts because a popular influencer posted a picture of herself with
those type of novels which makes me want to buy them. Such cases have to be accounted for
when making accuracy predictions using data such as likes from a social media platform.
Secondly, handling and usage of digital footprint brings ethical issues; how much of footprint
gets accessed by computers is difficult to handle and sometimes the individual may not want
certain things to be known. Recruiters from top companies may choose to be biased towards
interviewees because of what the digital footprint predicts about their personality, regardless of
how relevant that factor is for job performance and other job aspects.
Thus, although machine learning and models can be useful in making more accurate personality
predictions, it doesn’t liberate the user from the responsibilities that come with usage of such
models.
References
Alex M. Wood, G. D. (2012). How Are Personality Judgments Made? A Cognitive Model of
6494.2012.00763.x
Wu Youyou, M. K. (2015). Computer-based personality judgments are more accurate than those
doi:https://doi.org/10.1073/pnas.1418680112