Welfare scientists investigate the well-being of animals managed or impacted by humans – for example living as pets, on farms, in research labs, in zoos etc. Their research involves collecting behavioural and physiological data to make careful, objective inferences about how animals feel, with the goal of improving animals’ lives in an evidence-based way. Typical aims are to assess the relative impact of practices like different handling methods, or different types of housing, to identify those best for welfare and a good quality of life. Another key aim is validating indicators (e.g. investigating whether particular vocalisations reliably signal pain or hunger). As well as its practical value, welfare science touches on some fascinating, fundamental biological problems (e.g. the nature of sentience).
Welfare science has come a long way since the early days of CCSAW. There are now dedicated peer-reviewed academic journals such as Applied Animal Behaviour Science and Animal Welfare, as well as professional academic societies that run international conferences (especially the International Society for Applied Ethology and Universities Federation for Animal Welfare). There are also now many excellent welfare science groups around the world, including those at University of British Columbia – UBC, University of California Davis – UCDavis (in the US), Scotland’s Rural College – SRUC, University of Bristol, and several more. And there are growing connections with other disciplines too, especially social science and veterinary medicine, because solving animal welfare problems in practice involves much more than just understanding animals.
The slideshows below (best seen on a computer) provide more detail on the types of research question that welfare researchers aim to tackle, and how we do so.
Do cats and dogs smile? When animals look happy, are they? Without being anthropomorphic, can certain signals and expressions indicate well-being? 
When farm or lab animals are separated from mothers earlier than natural, what are the best ways to reduce weaning stress? And do they quickly recover? Or are there lasting effects on well-being? 
When animals (like these broiler chickens) grow up in indoor systems well-designed for efficiency but lacking natural stimuli, does this reduce well-being? Or do animals not miss what they have never seen? (Image: Food & Farm Care) 
Laying hens in aviary systems can damage their bones in collisions. How can we improve their agility, bone strength and housing design to eliminate this? 
These rat cages contain woodchip bedding, paper nesting, and tubes for hiding in. What other ‘enrichments’ would enhance welfare? 
Some practices essential for welfare, like veterinary visits, cause stress and fear. Can we make these experiences less aversive, or even positive? 
Like ‘Uncle Fatty’ (a macaque who was fed too many snacks by tourists in Thailand), pet, laboratory and zoo animals can become overweight. When do lives of leisure and ad lib food (which animals might want) backfire for welfare? 
This cow’s right hock is swollen, but does this cause her pain? If she does not limp, nor flinch when touched, is she stoically hiding discomfort, or actually pain-free? 
Hens are social, but in barns of many thousands, some cause harm by plucking each other’s feathers out. How can we best prevent this behaviour? 
Tropical animals in Northern zoos, beef cattle in shadeless feedlots, laboratory mice in ‘ventilated cages’ … some practices may cause thermal stress. How can we quantify and resolve such problems? 
Why do barren conditions matter to animals? And faced with commercial realities, which improvements will most effectively enhance animal welfare for the smallest economic cost? 
For captive wild animals, which aspects of natural life should be mimicked to ensure good welfare? For example, do polar bears need lots of space? Cold and ice? Even opportunities to hunt? 
When is aggression a sign of poor welfare like fear or pain? And how can it best be prevented or controlled, to improve well-being and the human-animal bond? 
Is playing a sign of good welfare, even joy? Does it help create resilient animals? (Or does it just make us humans happy to watch it?) 
Does stereotypic behaviour indicate frustration? Abnormal brain development? Both? And how can we most effectively prevent it? 
Which species and developmental stages have any feelings at all? Do fish have welfare? Do chicks in the egg? Mosquito larvae? Where are the boundaries of sentience?
The methodologies used range from controlled experiments to epidemiology and meta-analysis, and involve assessing diverse indicators of animal state (endocrine responses, behavioural changes, animal cognition, and more). The slideshow below gives examples.
Controlled experiments in the lab let us manipulate specific aspects of an animal’s experience, and test hypotheses in careful (sometimes creative) ways. Here, factors affecting hen agility are assessed. 
Such experiments include offering animals choices, to find out what they value. These hens are being tested to see what type of pen design they prefer. 
We can also experimentally manipulate early experience (e.g. rearing pen size, or opportunities to perch) to see how this affects welfare in adulthood. 
Some welfare science is conducted in real world situations, collaborating with transport companies, slaughterhouses, shelters and so on. This can be challenging but is a great way to obtain practically relevant data and to connect with industry groups. 
‘Number-crunching’ pre-existing data (e.g farm or breeding company records, or zoo studbooks), and running meta-analyses, can be extremely powerful approaches, revealing new (sometimes unexpected) predictors of welfare impact. 
Surveys are great tools for issues that matter to people as well as animals (e.g. feather-pecking in poultry, separation anxiety in dogs). They can glean responses from potentially thousands of people on welfare problems, and their animals’ living conditions. Epidemiological statistics can then identify practices linked to better welfare. 
Behavioural changes often indicate welfare states. For example, what animals choose to eat, and how much, can warn of sickness or infection. Methods to assess these include live observation, and video recording. 
Behavioural data can be collected automatically and be valuable too. For example, data from robotic milking systems can be processed via machine learning to spot patterns predicting illness. 
Behavioural tests in the lab are often great tools. In this apparatus, a force plate measured the startle reflexes of chickens exposed to a camera flash. Birds from highly enriched pens were found to be calmer and less ‘jumpy’! 
Cognitive tests can also be revealing. Humans in low moods often interpret ambiguous stimuli in negative, ‘pessimistic’ ways. Judgement Bias tasks can detect similar effects in animals. Here, picking the white stimulus leads to a food treat. Picking the black one triggers an aversive blast of air. Faced with intermediate stimuli, what to do? ‘Optimistic’ animals treat these as if predicting snacks. Just as in humans, good welfare can boost this cognitive ‘optimism’. 
Some physiological and immunological indicators of emotions and moods come from studies of humans (e.g. changes seen in people under pressure or in distress). Measured in samples from animals, they allow us to infer the presence or absence of similar states. 
Some of these physiological measures can be taken non-invasively. Here, thermography quantifies hens’ reactions being handled. Subtle falls in comb temperature (left) occur as blood is redirected to the body (a ‘fight or flight’ response). Minutes later, hens recover and their combs rewarm (right). 
Well-being can also be revealed by signals: how animals’ facial expressions, postures and vocalizations change in rewarding or aversive situations. We can quantify and validate these using video and other technological tools. 
Drugs can be used validate new indicators. To illustrate, when dairy calves are ‘disbudded’ to stop horn-growth, they react to this tissue damage (e.g. struggling, ear flicking). Researchers can use analgesics to pinpoint the most accurate indicators of pain. These can then be used to identify better ways to disbud. 
Can animals experience specific states like contentment? Or boredom or depression? This requires carefully ‘operationalizing’ such states in humans, before assessing whether animals show homologous behavioral and physiological responses. 
Much welfare research is multi-disciplinary. We often reach out to geneticists, immunologists, economists and others for collaborative solutions to welfare problems. For example, working with nutrition experts is essential for working out how to improve bone strength in hens. 
The overlap between welfare science and neuroscience is also highly collaborative. Neurological data can help elucidate links between housing conditions, abnormal behaviour, and cognitive changes. 
Interdisciplinary work can also assess the practical value of improved animal welfare. For example, how does better animal welfare affect meat quality, or farm-level productivity, or the usefulness of data from lab animals?
