Beyond “Doing Better”: Ordinal Rating Scales to Monitor Behavioural Indicators of Well-Being in Cats

Abstract

Safeguarding the well-being of cats is essential to the mission of any responsible animal shelter. Environmental enrichment and behaviour modification are often key to this goal. Measuring response to these interventions is essential to ensure strategies are successful. There are often many staff and volunteers involved in these efforts, and a lack of standardised language can make monitoring progress difficult. Ordinal rating scales of key behaviours can be a useful way to summarise observations and ensure that common language is used. However, it is crucial that these scales have good interobserver agreement and reliability, so operational definitions and training systems are important. This paper presents a method for evaluating four different expressions of cat welfare on a 0–5 scale: modified Fear, Anxiety, and Stress score; Response to Petting score; Participation in Play score; and Food Intake Summary score. All scales showed almost perfect average interobserver agreement (linear weighted κ) and excellent average interobserver reliability (interclass correlation coefficient). These scales may prove useful to other shelters, or any other environment where evaluating response to interventions is important to the welfare of cats, such as research facilities or home environments. The exceptional interobserver agreement and reliability of this study compared with some others highlights the importance of standardised training programs.

Charting Ethical Tensions in Multispecies Technology Research through Beneficiary-Epistemology Space

“…The artists felt this to be an ambiguously provocative proposition for the audience; the computer scientists developing the technology were looking for something that could be readily implemented; but animal researchers worried about the scientific validity of an overarching and simplistic idea of happiness. Ultimately the decision was taken to adopt the Participation in Play scale [16] with the rationale that it would best represent the AI's own view of happiness -i.e., that its task was to play with cats to make them happy.…”

“…Outside the enclosure, the artists monitored the cats from behind one way glass mirrors, manually rating their engagement with each game using the Participation in Play scale [16], a tool used in Veterinary Science to measure behavioural indicators of wellbeing in cats through ordinal scales that rate aspects of play, such as playing enthusiastically or tentatively, playing while moving or stationary, watching toys passively without engaging, appearing disinterested, or retreating. The results were fed into an AI 'decision engine' that attempted to learn which toys and treats the cats preferred, and that would then recommend the next toy or treat to be offered, which the artists could accept or reject (recording the reasons for their decision).…”

Designing Multispecies Worlds for Robots, Cats, and Humans

“…The robot performed more than 500 activities over the course of the 12 days, ranging from providing small cat treats 2 , to simple toys like a cardbox or a string, all the way to more elaborate toys such as an orange bird toy or a battery powered wiggling fish toy. Following each activity, each cat's engagement-as a proxy for positive attitude towards the robot's provided entertainment-was ranked using the Participation in Play scale [18]. The task executed (i.e., 'present a prey game with a toy bird'), the targeted cat, and the impact on happiness where subsequently fed back into the robot's machine learning system (i.e., the decision engine) informing future proposed activities.…”

“…8 Following each performed robot activity, the system needed information regarding the cats' happiness score, i.e., the impact of the given activity on each individual cat's happiness score. To achieve this, the lead artists-who closely monitored the environment, focusing on the cats' engagement-used the Participation in Play scale [18] to rank each cat's engagement, which acted as a proxy for happiness on a scale from 0 to 5 (e.g., 'No interest in playing or retreats from attempts at play' to 'Enthusiastically playing, moving around the environment'). Furthermore, the cats' stress score [28] for each cat was recorded using instantaneous sampling (i.e., at predetermined time intervals of 15 minutes) by the animal welfare officer.…”

“…In the same fashion, operator interference with toys' trajectories changed the activity. This had the potential to alter the cats' engagement with the activity and ultimately the activity's impact on their happiness score (using the PIP scale [18]), whether positively or negatively. This, just as the artists' interference, had implications for the data fed back to the decision engine, thereby potentially affecting future activities.…”