top of page
  • 執筆者の写真David Creelman

The End of Personality Assessments?



 

Personality assessments have had a long and important role in the history of psychology. However, new machine-learning tools may make the assessments obsolete. We might be entering an era where personality assessments seem as quaint as doing calculations with a slide rule.


We know personality assessments have limited predictive power. We use them because they are the best tool we have. They help us answer questions such as:


· Will this person be a good fit for the job?

· How should I behave with this person so that we work well together?

· What are my traits that I ought to be aware of in self-management?


The alternative to an analysis of personality traits would be a model of someone’s personality. Rather than think “They are high on extraversion so I guess they’ll be good for a sales role” we would run a model of their personality against the particular sales role and let it predict if they would be a good fit or not.


We can imagine what this predictive model would be like if we think in terms of what an experienced manager, who knows a person well, might say. They might predict a person would be good for one particular sales role but not for a different one. The manager would go on to make a nuanced explanation of what it was about the person’s personality that led them to this prediction. The nuanced explanation would go well beyond just citing five personality traits.


So this is what we aspire to, a model built by machine learning that could give the same kind of reasonably accurate, nuanced predictions about a person that an experienced manager would give. How would we create this machine-learning model? We would do it the same way a manager does, by observing the individual and building a mental model that predicts how the person will behave.


The technology to do such a thing isn’t mature. The machine would need to be able to listen to the person’s conversations, watch their body language, and read their emails all the while understanding enough about the context to make sense of the behaviour. None of this is science fiction, but it’s not totally science fact yet either.


Why speculating about the end of assessments matter

Why would we bother speculating about technology that is probably 10 years away from being implemented? I care about this because psychologists have been using personality traits as a core idea for nearly 100 years. The notion that this well-established tool might just evaporate in the face of superior technology gives us a sense of the drama that might be ahead of us in HR.


It’s useful to think of the big five personality traits not as dials that exist in someone’s head but as a simplistic mathematical model that predicts behaviour in a limited way. Seen in that light it’s astonishing in today’s day and age we would still be using a predictive model of something as complex as a human being just based on measures of five factors. Vastly better models are bound to be developed and the next generation of HR pros will shake their heads and wonder how we ever managed without more sophisticated personality models.


Big change is coming. Things we thought fundamental will be replaced. Be ready for a brave new world.

Comments


bottom of page