Validity and feasibility of human-animal relationship tests for on-farm welfare assessment in dairy goats

This study aims at establishing suitable tests to measure the quality of the Human-Animal Relationship (HAR) in dairy goats for on-farm welfare assessment protocols by evaluating the predictive validity of different categories of HAR tests and their feasibility in on-farm condition. Twelve commercial dairy farms in Northern Italy were selected and classified as "good" (n = 6) and "poor" (n = 6) HAR on the basis of reports from a technical advisor. Some variables were tested to measure the HAR: sneezing - the number of alert sounds; voluntary approach (VA) test - in a stationary situation, the latency to the first contact by goat and the% of goats that entered in contact with the test person and within a 1.5 m radius around her were recorded, both continuously and at scan sample intervals; avoidance distance (AD) test the avoidance distance from a moving and approaching person and the% of tested goats, of contacts with the test person followed by withdrawal or of acceptances of gentle stroke were recorded. The feasibility of each test was evaluated based on costs, time consumption, safety and training requirement. Sneezes were rarely and random expressed, so they do not seem suitable to be included in a protocol. Both Principal Component Analysis and One-way ANOVA confirmed the predictive validity of most of the variables, which were able to discriminate between "good" and "poor" HAR farms. Latency to first contact resulted valid (P = 0.01) and a high feasibility was reported. The indicator was easy to be recorded and the VA test could be stopped immediately after the first contact, saving time. Variables from the AD test resulted valid (% tested goats: P = 0.006; AD mean: P = 0.016; % contacts: P = 0.006; % acceptance: P = 0.003), although they were more time consuming or required a more specific training compared to latency to first contact. The correlation among variables seemed to support also a convergent validity of the tests used. The investigation pinpointed promising behavioural tests to be included into on-farm welfare assessment protocols in dairy goats. Taking into account species, test results and feasibility considerations, we suggest the inclusion of latency to the first contact with the test person into on-farm welfare assessment protocols. However, these results should be further tested in a larger number of farms of different dimensions, to overcome the limitations of this study due to the small sample size and to check the effect of farm size