2.2. Assessment of the Behavioural Indicators Using Scan Sampling

A single experienced observer in equine ethology walked regularly in front of the individual boxes, at a distance of at least 1.5 m from the door making as little noise as possible. She observed the horse for 3 s and then recorded whether the animal expressed one of the behavioural indicators.

Each horse was observed for five scans per day on 50 non-consecutive days distributed over a nine-month period. Observations were equally distributed across times of the day (9:00 to 10:30, 10:30 to 12:00, 12:00 to 13:30, 13:30 to 15:00, and 15:00 to 16:30). The average number of total scans analysed per subject was 200 ± 18 (mean ± SD); variations in the number of scans resulted from the absence of the horse or the presence of the caretaker in the loose box at the time of the observation. Descriptive statistics regarding the four behavioural indicators are presented in Table 2.

Description of the behavioural indicators of the compromised welfare state in the sample. Stereotypies were divided into two sub-groups (oral and locomotion) because they could be influenced by different environmental factors. Descriptive statistics are presented as the percentage of horses expressing stereotypies and aggressive behaviours and the percentage of scans during which the “withdrawn” and alert postures were observed (mean ± SEM, Min–Max).