Comparison of in-person qualitative behaviour assessment (QBA) and video qualitative behaviour assessments (V-QBA) of beef cattle

The objective of the study was to compare Qualitative Behaviour Assessments (QBA) conducted in person/live with those conducted by video (V-QBA). The livestock used in this assessment were two groups of 30 finishing suckler beef cattle. Forty live QBA events were carried out between 20th November 2019 and 18th March 2020, with 2 assessors performing 20 each. The same assessor watched video footage of the time period of the live assessment a minimum of one month later, and repeated the QBA.

The cattle from this dataset were reared on the North Wyke Farm Platform, a National Bioscience Research Infrastructure. The North Wyke Farm Platform is data rich and highly documented, and details about the farmlets and livestock can be found in the related information. There is also a lot of related data available on the North Wyke Farm Platform data portal (https://nwfp.rothamsted.ac.uk/), which is free to use after registration.

The objective of the study was to compare Qualitative Behaviour Assessments (QBA) conducted in person/live with those conducted by video (V-QBA). The livestock used in this assessment were two groups of 30 finishing suckler beef cattle, weaned a week before the start of the observation period. Comparing the two groups was not an objective of the study and the groups are not identified in the data. Cattle were housed in a barn for the winter, fitted with 4 CCTV cameras placed 4.3 m above the ground and on 3 different walls of the barn. One of the cameras gave a similar view to that gained by the in-person QBA assessor. The barn dimensions and the in-person and camera observation points can be seen in Figure 1 of https://doi.org/10.3389/fvets.2022.832239.

Forty live QBA events were carried out between 20th November 2019 and 18th March 2020, with 2 assessors performing 20 each. The assessors had received the same training and had similar experience around cattle; inter-observer reliability was high (0.779 - 0.871), details of which can be found in Supplementary Material C of https://doi.org/10.3389/fvets.2022.832239. During the live QBA events, assessors made the minimum of noise or movement so as not to disturb the cattle. Once any behavioural effect resulting from the assessors’ entrance had subsided, a 10-minute observation period began. As soon as this period finished, assessors filled out observation forms. A minimum of one month later, assessors were given video footage of the cattle covering the same time period as the initial assessment, with the 4 videos of the different angles of the barn formed into a collage, so that the assessor could view them all simultaneously. The assessor then carried out the QBA in the same way as before.

The assessors rated the cattle behaviour as a group for 20 characteristics, both positive and negative, using definitions derived and adapted from the Welfare Quality protocol for cattle (2009; Lelystad, Netherlands: Welfare Quality Consortium). Scoring resulted in a numerical value between zero (complete absence of that characteristic) to 125 (observed to the greatest realistic extent possible). Definitions of the characteristics, as used by the assessors, can be found in Column_units_and_definitions.csv or Supplementary Material A of https://doi.org/10.3389/fvets.2022.832239

Three of the measured characteristics were removed from the dataset as these were not used. “Distressed” and “Fearful” both had a score of zero in all instances, and “Frustrated” had a score of zero most of the time (65/80 assessments).