The prediction of human sociability in the domestic cat

Temperament consists of stability in behavioural tendency across both contexts and time. Variability in temperament may affect how suited cats are to reside in certain types of environment. Over the past three decades many studies have been performed on the domestic cat in order to assess their temperament, often with a focus on social aspects of their behaviour, particularly how amenable they are towards humans. Whilst such tests could be used to evaluate or infer an individual’s suitability for domestic living, research in this area frequently lacks sufficient demonstration of reliability and validity. In addition, little consideration is often given to the practicality of the measures and tests proposed. Thus the ability to accurately measure key traits of relevance to human-sociability in a practical way remains largely unknown, as does the predictive validity of such tests in relation to future behavioural tendencies in other contexts. The aims of this PhD were to address these issues by developing robust measures with demonstrable reliability, validity and practicality, which could be used to evaluate the sociability of cats towards humans, and in a predictive capacity in relation to post-rehoming behaviour. A neurobiological framework based on relevant affective systems (a balance between FEAR, SEEKING and RAGE sensu Panksepp) was used to define ‘Human-sociability’ and the ‘aggressive response’. The deconstruction of these traits allowed the operational definition of psychobiologically based behavioural tendencies, which were hypothesised to be useful predictors of sociability towards humans. Four standardised behavioural tests and 65 basic behavioural measures were then used to provide information on these traits. These measures were subsequently refined based on their reliability, validity and practicality. Many measures were discarded because they were found to be influenced by (short-term) temporal and/or social factors (i.e. familiarity of a person or their interaction style), or were less practical to perform (without appearing to explain much additional variation within the data). At this point only nine individual measures were retained, from only one of the four initial behavioural tests. However, further analysis indicated that the majority of these remaining measures were either influenced by longer-term temporal factors or by the environment (i.e. individual rehoming centre). These findings suggest many behavioural measures currently used in the assessment of temperament may be invalid. In light of these results, other methods of trait assessment were explored. A series of questionnaire items were developed and put through a similar process to determine their reliability, validity and 4 practicality, based on the same neurobiological framework as the previous behavioural experiments. From an initial twenty-eight items, ten demonstrated sufficient practicality, inter and intra-rater reliability and temporal stability within the rehoming centre and were thus retained within a final model that could be used to generate behavioural ‘profile scores’ for cats within the rehoming environment. This refined questionnaire model demonstrated good content and face validity, containing three clusters of measures that were hypothesised to represent social aspects of all core processes of interest (FEAR, SEEKING and RAGE). In addition, good construct validity of measures was also demonstrated via the convergence of individual items hypothesised to share similar emotional underpinnings and via the discrimination between those that were not. A subset of measures from this model were found to have good predictive validity in relation to future post-rehoming behaviour, and were also associated with owner post-adoption ‘satisfaction’ scores. However, this subset of items no longer included RAGE reactivity and thus the criterion validity of the questionnaire in relation to the behavioural manifestation of this core process post-adoption is limited. Based on these findings, an initial framework process for the behavioural assessment, management and matching of cats to suitable owners for use within the rehoming environment has been proposed (the Lincoln Rehoming centre Cat Assessment Tool (L-RCAT)). This research has resulted in the provision of a first-of-its-kind tool that can be used to generate practical, non-invasive and valid information about the behavioural tendencies of cats in relation to aspects of human-sociability, not only in-situ (i.e. within the rehoming environment), but also in a predictive capacity in relation to future ex-situ behaviour (i.e. post adoption within the home), as well as to aspects of owner-satisfaction and thus the potential ‘rehoming success’ of individuals