Towards compatibility between artificial and psychometric personality models

Abstract

Purpose

The purpose of this paper is to investigate the statistical link between an artificial personality model and the leading psychometric model.

Design/methodology/approach

An online survey was conducted made of two parts: a 40‐sentence humor appreciation survey corresponding to the artificial personality model, and a 50‐sentence Big 5 psychometric survey. The cross‐correlation between the scales of the two parts was computed, and exploratory factor analysis performed on the Big 5 scores using three different sample age spreads.

Findings

The cross‐correlation coefficients between the artificial and psychometric personality scores supported the suggestion that there is compatibility between the two, albeit their absolute values were not as high as other studies due to the small sample size. Also, when computing factor analysis on Big 5 scores it was found that the loading of two factors identified as motivational went down systematically with the size of the sample, which empirically supports the suggestion that motivational and cognitive factors are distinct.

Research limitations/implications

The size of sample was not sufficient to reach a conclusive decision but the evidence was supportive and promising for additional research.

Practical implications

The compatibility between the artificial and psychometric personality models means that psychometric scores of real persons can be uploaded into artificial personalities in order to mimic real conversation and behaviour.

Social implications

Improvement of man‐machine interface, facilitating education.

Originality/value

Correlating the scores of different personality scores is not new, but correlating with artificial personality dimensions defined by humor appreciation scores is new. The suggestion that there is qualitative difference between factors of well established psychometric model is new and could have far reaching implications.

Citation

Kadri, F.L. (2013), "Towards compatibility between artificial and psychometric personality models", Kybernetes, Vol. 42 No. 3, pp. 497-505. https://doi.org/10.1108/03684921311323716