Caretaker Score Reliability for Personality Assessment of Bottlenose Dolphin (Tursiops truncatus)

SIMPLE SUMMARY: The assessment of animals’ personalities can help manage decisions concerning zoo animals more appropriately; for example, a proper personality evaluation helps create stable social groups or increase the chances of breeding success if compatible breeding pairs are chosen. In zoos, the animals’ personality is often evaluated by the caretakers due to their familiarity with the animals and their behaviours. In this study, we aimed to evaluate how reliable caretakers’ ratings are when assessing dolphins’ personalities. With this aim in mind, we asked 24 caretakers to score a variety of personality traits of bottlenose dolphins under their care through a questionnaire in two periods. Our findings showed fair to good degrees of agreement within scores of the same rater and across raters within the same centre. We were also able to identify which raters and centres showed significant score mean differences systematically. We believe the study of raters’ outcomes reliability is crucial to make appropriate management decisions based on the animals’ personalities. ABSTRACT: The evaluation of zoo animals’ personalities can likely lead to a range of benefits, including improving breeding success, creating stable social groups, and designing and developing environmental enrichment programmes. The goal of this study was to use caretakers scores to evaluate personality in bottlenose dolphins and to assess the reliability of scores within each rater and among raters from each centre. To this end, 24 caretakers from 3 countries (Spain, France, and Argentina), including a total of 5 dolphinariums and 6 groups of dolphins, used a questionnaire based on the Five-Factor Model of Personality to score bottlenose dolphins on a number of personality traits in three different contexts. Each caretaker evaluated the animals under their care twice, ensuring that raters did not share thoughts nor impressions with other raters. Our findings showed a good degree of agreement between each rater’s scores and a fair degree of agreement among scores of raters from the same centre. We also identified which raters and centres had significant mean score differences and detected that 4 out of 24 raters from two different centres showed such differences systematically. The evaluation of raters’ reliability and the identification of particular inconsistent raters and centres is critical to make more appropriate and realistic management decisions that, in turn, directly impact animals’ welfare.