Observations versus assessments of personality: A five-method multi-species study reveals numerous biases in ratings and methodological limitations of standardised assessments

Jana Uher, Elisabetta Visalberghi
  • Journal of Research in Personality, April 2016, Elsevier
  • DOI: 10.1016/j.jrp.2016.02.003

Observing or assessing personality? Five methods study shows the differences

What is it about?

This study investigated the theoretical and methodical differences between two major methods used to study personality: Assessments (e.g., rating scales, questionnaires) and observations of behaviour. Five methods were used. Individuals' behaviours were measured with computerised and video-based behavioural observations. Raters' impressions of the studied individuals' personality were studied with standardised rating scales. Finally, raters' interpretations of the two questionnaires used were studied with open-ended survey questions. A key feature: The individuals whose personality was under study were Capuchin monkeys, a primate species endemic to South America and famous for their cognitive and tool use abilities. The raters were persons well familiar with the studied monkeys (e.g., behavioural researchers, keepers). This enabled interesting comparisons because other species have social and behavioural systems that differ from ours. Therefore, attribution biases can become particularly apparent. The monkeys showed pronounced individual differences in a broad range of behaviours, which can be summarised in 20 broad personality categories. But interestingly, sex differences emerged only in two of these categories (aggressiveness and dominance). By contrast, the personality ratings for these monkeys showed sex differences in 12 personality categories! Hence, the raters tended to see far more sex differences than were actually observed in the monkeys’ behaviours. These attribution biases likely derive from the raters’ stereotypical beliefs about gender differences in humans.

Why is it important?

Gender differences in human personality are widely discussed. But most of these findings were obtained with personality questionnaires. This study showed that ratings on such questionnaires can be substantially biased. Another important finding was substantial variation in raters' interpretations of the questionnaires. This challenges the widespread belief that rating scales could enable researchers to assess individuals’ personality in standardised ways.


Dr Jana Uher (Author)
University of Greenwich

Personality questionnaires are popular in research and applied settings. But the ways in which questionnaire findings are actually related to people’s observable behaviours are still largely unknown. It is high time to explore these relations and to investigate how people actually understand and use questionnaires.

Read Publication


The following have contributed to this page: Dr Jana Uher