Research plan – Københavns Universitet

Research plan

A large part of the study will be done by means of a questionnaire distributed and filled out by a group of approximately 200 experts (here understood in a wide sense including practising veterinarians, farm advisors, animal welfare assessors, and researchers). The different expert questionnaires mentioned below are sub-elements of this large questionnaire. Since this expert group is relatively small, and since it members are in general very busy, and the questionnaire requires that they take time out of their busy schedule, they will be given a voucher as an incentive. The experts will be recruited from Scandinavia, UK, the Netherlands, Finland, Germany and possibly other Western European countries.

The project consists of three parts:

1. Improving the scoring scheme to increase inter-observer reliability

This will be studied based on a questionnaire. As mentioned earlier, the type of scale used may influence the reliability of the scoring as well as the information that can be gained from the measure. Traditionally measures used in assessment schemes, whether specifically for lameness or in general animal welfare assessment are divided into, at the most, a few categories. The argument is that fewer categories makes it easier to score and leads to higher inter-observer reliability (IOR).  A few recent studies, however, have shown that using continuous scales, i.e. a Visual Analogue Scale, improves the IOR1,2. The assessors using the scales also reported preferring the VAS in contrast with the classical categories. The suggestion for this part of the project is to examine whether there is potential for converting a number of indicators from the Danish index project from their current categorical nature3 to continuous scales and by comparing the IOR of continuous scales as and the categories presently used.

Practically speaking, the participating experts will in the questionnaire be presented with a number of situations on farm, and they will be asked to score different severity levels for a number of indicators by means of images, video-clips and textual information describing the situations. A split-design will be employed where one-half of the experts are asked to score the indicators on a scale with only a few categories, and the other half will be asked to score it on a continuous scale. As appropriate images or footages that show different levels of severity on the studied indicators do not exist, it will be necessary to go out in the field and collect relevant images/footage). Some important indicators suitable for continuous recordings will be selected for this analysis. This will first be done using video recordings or pictures, and based on the finding further studies will at a later stage be done on farm to test the feasibility and ease of use under practical conditions.

2. Aggregation of welfare indicators

In the second part of the project focus will be on how best to aggregate the different indicator values into a total score for the welfare at farm level, which in turn will serve as input to the total welfare index for the country for each species of farm animals.

In the current version of the index, all indicator values with a lower welfare status compared to the theoretically best score contribute with a certain weight to the total index value at farm level. These weights have been calculated by means of input from a group of experts. By means of a questionnaire, it will be studied to what degree there is agreement across experts about these weights and how the actual disagreements would translate into different outcomes on the total index values. For practical reasons, a limited number of indicators and related weightings will be chosen.

The focus will also be on the assumptions underlying the whole index model, i.e. the hedonist definition of animal welfare and the additive (“utilitarian”) view on how to add up welfare problems to a total score. The questionnaire will address the degree to which the experts agree with these animal welfare assumptions, and whether the expert’s animal welfare assumptions may have an impact on their scoring behaviour. In parallel, there will be a questionnaire aimed at a representative sample of the Danish public (approximately N=500) to be able to compare to which degree animal welfare assumptions among experts are in line with the views of the public. Lack of agreement will not per se be a problem for the index, but it is certainly of relevance for the way the results of the index are communicated.

3. Comparing aggregated welfare scores with holistic view of experts

In the third part of the sub-project the link between the aggregation of indicator scores and more “holistic” scores will be examined. The idea is that experts will not only have views about the individual indicators but also about welfare at the farm as a whole (a “holistic” view). Ideally, this can be studied by having a group of expert both doing on-farm scores of all indicators and stating their holistic view of the total welfare at the farm both prior to and after the scoring.

It is not feasible at this stage to examine the correspondence between farm score levels that are based on indicator scores and the holistic view. Instead, we will do two other things:

Firstly, for at least one set of criteria the questionnaire will present experts with the task of scoring indicators and via other stimuli (e.g. videos give them an impression of the farm in question) request them to make a more holistic score relating to these criteria.

Secondly, we will develop a small questionnaire to be used on farm to provide holistic scores which can be used when the index is applied in 2018 and onwards. Thereby there will for the future use of the index be the opportunity to compare the aggregated welfare scores with the holistic views of the assessors.

The findings of this part of the project may together with findings from the second part of the study (‘Aggregation of welfare indicators’) feed into possible revisions of the aggregation procedures for the index.

1. Miljø- og Fødevareministeriet (2017). Udvikling af nationale dyrevelfærds-indekser for kvæg og svin – Afsluttende projektrapport. Dyrevelfærd og Veterinær-medicin 2015-15-19-00145

2. Tuttyens F, Sprenger M, Van Nuffel A, Maertens W, Van Dongen S (2009). Reliability of categorical versus continuous scoring of welfare indicators: lameness in cows as a case study. Animal Welfare 18: 399-405

3. Nalon E., Maes D., Van Dongen S., van Riet M.M.J. Janssens G.P.J., Millet S., Tuyttens F.A.M. (2014). Comparison of the inter- and intra-observer repeatability of three gait-scoring scales for sows. Animal 8 (4): 650-659