Defra - Department for Environment, Food and Rural Affairs.

Science Search

Science and Research Projects

Return to Science Search homepage   Return to Project List

Validation and Differentiation of Welfare Indicators in Laying Hens - AW1146

Description
Valid measures of animal welfare are required to underpin societal and legislative judgements about how to keep and treat
animals. Both scientists conducting experiments to assess factors that might influence farm animal welfare (e.g. stocking
density or feed type), and assessors visiting farms on behalf of farm assurance programmes, need practical measures of
welfare. It is now accepted that, wherever possible, direct measures of animals should be taken (e.g. health, body condition, locomotion, behaviour) rather than simply measures of the environment animals are in (e.g. space allowance,
type of bedding) as it is the impact of the environment on the animal that matters for welfare. Because there is no one
measure of good or poor welfare, scientists and assessors often take a wide range of measurements. This can be
expensive and time-consuming so it would be useful to develop more refined sets of measurements that could be used in a
practical context. Of even greater importance, the validity of many currently-used measures is not well-established. In
chickens for example, we may measure the plumage condition, keel-bone deformation, prevalence of hair-line fractures, or
even walking abnormalities without knowing whether these are perceived as painful or aversive; and there are similar
difficulties in identifying indicators of positive welfare. Thus, to identify practical welfare indicators that can be used in a
wide range of situations, we need to establish which are associated with the animals' own perceptions and emotions i.e.
which indicators really matter to the animals.
In our previous work we validated indicators against hens' own expressed short-term preferences and aversions. This
confirmed some useful indicators, threw doubt on the validity of a minority, and resulted in the identification of new
indicators. For example, it was not previously known that head-shaking was a reliable indicator of an aversive environment,
or that faecal dryness was reliably associated with being in a preferred environment. This information is already being used
to improve farm assurance audits. Our original approach works well in the short-term, but cannot be used to validate
indicators of long-term welfare, which better reflects lifetime quality of life, because of limitations in how birds express longterm
preferences. Our proposed research is thus built on the knowledge that long-term exposure to preferred or nonpreferred
events or environments can change long-term emotional mood state. Thus, if we cannot validate long-term
indicators directly against birds' preferences, we can validate such indicators against the changes in birds' moods that
result from long-term exposure to preferred or unpreferred environments.
In the first part of our work we will house chickens in environments that are known to be strongly preferred (positive) or
unpreferred (negative)in the short-term. We will assess how long-term exposure to these environments influences mood by
performing 'cognitive bias' tests, based on our previous work in other species: animals in a positive mood judge situations
differently from animals in a negative mood. We will then use statistical methods to describe welfare indicators (measures
of behaviour, physiology, health) that are associated with housing in a positive environment and with measures of positive
mood, and to describe welfare indicators associated with housing in a negative environment and measures of negative
mood. The second part of the programme will be to expose birds to environments that are positive or negative in different
ways (e.g. rich in rewards or sparse in punishments) and to see which indicators are good 'general' indicators of positive or
negative experience, and which indicators are 'specific' indicators of particular types of positive or negative experience.
These results will produce a new generation of practical indicators for use in refined protocols for assessing laying hen
welfare.

Objective
The overall objective of the proposed research is to validate welfare indicators for laying hens against clearly defined
'ground truth' measures of welfare.
In the short-term, welfare indicators can be validated against the animals' own expressed preferences and aversions (as in
our previous experimental work). This approach is difficult to pursue when trying to identify indicators of long-term welfare,
as there are limits on animals' abilities to express their long-term preferences. However, over the longer-term, repeated or
continuous exposure to preferred or non-preferred events or environments is likely to alter affective mood state. Thus, we
propose to validate indicators of long-term welfare for laying hens against measures of affective mood state. The required
objectives to be met in Programme 1 are:
(i) To use hens' own expressed preferences and aversions to design events and environments that can be empirically
identified as (rather than assumed to be) positive or negative.
(ii) To expose hens long-term to these positive or negative events and environments and evaluate the extent of their
affective mood change using cognitive bias tests
(iii) To take a wide range of behavioural and physiological indicators at defined points in the long-term housing period
(including post-mortem indicators at the end) and use multilevel statistical modelling to determine: (A) indicators associated
with long-term housing in positive environments and (B) indicators associated with cognitive biases indicating positive
mood. A perfect validation would result in an identical set of indicators being identified in analyses A and B. Similarly,
statistical modelling will determine: (C) indicators associated with long-term housing in negative environments and (D)
indicators associated with cognitive biases indicating negative mood. A perfect validation would result in an identical set of
indicators being identified in analyses C and D, and no overlap between A/B indicators and C/D indicators.
We will then extend the work by designing events and environments designed to generate different 'types' of positive or
negative mood. The required objectives to be met in Programme 2 are:
(iv) To use hens' own expressed preferences and aversions to design events and environments that can be empirically
identified as positive (reward rich), positive (punishment sparse), and negative (reward sparse) or negative (punishment
rich).
(v) To expose hens to these events and environments and evaluate the extent and type of their affective mood change
using a wider range of cognitive bias tests
(vi) To take a wide range of behavioural and physiological indicators at defined points in the housing periods and use
multilevel statistical modelling to determine: indicators that are generally associated with positive housing and cognitive
biases indicating positive mood, indicators that are generally associated with negative housing and cognitive biases
indicating negative mood, and indicators that are more specifically associated with the type of positive or negative exposure
given (as detailed in objective iv) or the type of cognitive bias shown.
Time-Scale and Cost
From: 2013

To: 2016

Cost: £72,460
Contractor / Funded Organisations
B B S R C (BBSRC)
Keywords
              
Fields of Study
Animal Welfare