Societal views and animal welfare science: understanding why the modified cage may fail and other stories

Abstract

The innovations developed by scientists working on animal welfare are often not adopted in practice. In this paper, we argue that one important reason for this failure is that the solutions proposed do not adequately address the societal concerns that motivated the original research. Some solutions also fail because they do not adequately address perceived constraints within the industry. Using examples from our own recent work, we show how research methods from the social sciences can address both of these limitations. For example, those who persist in tail-docking cattle (despite an abundance of evidence showing that the practice has no benefits) often justify their position by citing concern for cow cleanliness. This result informs the nature of new extension efforts directed at farmers that continue to tail dock, suggesting that these efforts will be more effective if they focus on providing producers with methods (of proven efficacy) for keeping cows clean. Work on pain mitigation for dehorning shows that some participants reluctant to provide pain relief believe that the pain from this procedure is short lasting and has little impact on the calf. This result informs the direction of new biological research efforts to understand both the magnitude and duration of any suffering that result from this type of procedure. These, and other examples, illustrate how social science methodologies can document the shared and divergent values of different stakeholders (to ensure that proposed solutions align with mainstream values), beliefs regarding the available evidence (to help target new scientific research that meets the perceived gaps), and barriers in implementing changes (to ease adoption of ideas by addressing these barriers).

Implications

Many of the innovations developed by animal welfare science have failed to see widespread adoption in practice. One reason for this failure may be that science-based ‘solutions’ fail to resonate with the values and concerns of other important stakeholders. Here we use examples from our past work with dairy cattle to argue that new scientific research should occur in tandem with social science research that provides a better understanding of the welfare concerns of the general public and perceived barriers of livestock producers.

Introduction

Here is a dirty secret. Careers have been frittered away, and with them millions of dollars in investment, on ideas of animal welfare scientists that have never gained traction in the real world. Consider, for example, the modified cage for laying hens. The modified cage is an elegant scientific solution, taking into account our best research on social group size, risks of feather pecking and cannibalism, the benefits of perches, dust bathing opportunities and a secluded nest site (see Lay et al., 2011; Mench et al., 2011). This research, combined with years of engineering and marketing efforts, to create another cage for laying hens. This happened without, it seems, any serious investment in finding out what the general public ¹ (and critics of existing cage systems for layers) wanted and what elements they would consider essential in housing systems for birds. Arguably this leaves us with a solution that appeals to a more narrow conception of welfare held by the scientists, but fails to resonate with the concerns of others who from the outset wanted cage-free systems and are thus generally less willing to support a production system based on cages, even if these are labelled ‘enriched’, ‘modified’, etc.

As explained by Fraser (2008), animal welfare science is mandated to provide solutions to the problems brought to us by the society that funds our positions and the institutions in which we serve. However, our own values bias the work we do, including what we choose to measure and how we interpret the results. Consider a hypothetical researcher who sets out to study the welfare of laying hens kept in modified cages v. a well-designed free-range system. She might choose to measure pathogen exposure or risk of feather pecking, and on this basis conclude that welfare is superior in the modified cages. But this choice of measures (and many other choices including the specific conditions under which the different systems are studied) reflects her values as an individual. Other scientists might have made different but equally rational choices, for example, studying the hen’s ability to fully stretch her wings, or aspects of foraging behaviour (both behaviours might be restricted in cages relative to cage-free systems). The problem is not that such biases occur; it is our blindness to these biases. This failing is arguably more problematic if the biases we bring to our work do not correspond with the values of the society that mandated our work in the first place. In the current example, people who buy eggs of course expect that animals will be kept healthy (Vanhonacker et al., 2010; Ingenbleek and Immink, 2011), but public concerns regarding how farm animals are reared also focus on long periods of close confinement and how this reduces opportunities for animals to live a more natural life (e.g. Boogaard et al., 2011; Vanhonacker et al., 2008).

This type of mismatch may be better avoided if efforts to describe and understand social concerns (e.g. Krystallis et al., 2009; Boogaard et al., 2011) are used to inform the direction of scientific work. In the current example, we suggest that if a sustained research effort to understand societal values around cage and non-cage rearing for laying hens had pre-dated or at least accompanied the scientific work, research would have instead focused on the development of high-welfare non-caged systems that are more likely to see widespread adoption in practice.

Understanding the values and beliefs of farmers and other industry professionals (e.g. veterinarians) is equally critical. These individuals often serve as the ‘gate keepers’ in deciding if and how to implement innovations (Kauppinen et al., 2010; Driessen, 2012). Understanding the perspectives of farmers is now increasingly recognized as a critical component in resolving challenges in food animal production (Feola and Binder, 2010; Kaupinnen et al., 2010; Driessen, 2012; Hansson and Lagerkvist, 2014).

In this paper, we make the case that some solutions developed by animal welfare scientists fail to gain traction because (1) they do not adequately address the societal concerns that motivated the original research and (2) they do not adequately address the perceived constraints within the industry. We argue that scientists, policy makers and funders working on animal welfare must increasingly invest in social science research that addresses these limitations. We provide examples from our own research on dairy cattle to illustrate how social science methodologies can document the shared and divergent values of different stakeholders, the beliefs regarding the available evidence and the barriers in implementing changes. We argue that the results of these studies can help identify solutions that better align with mainstream values, help target new scientific research that meets the perceived gaps in knowledge, and help ease the adoption of innovations by specifically addressing perceived barriers.

A short primer on theory and methods

The Theory of Planned Behaviour (Fishbein and Ajzen, 1975; Ajzen and Madden, 1986) is one of the prevailing paradigms in social psychology and communication. This framework posits a causal chain of influence in which attitudes, influenced partially by one’s beliefs and values, affect human behaviour.

According to the Theory of Planned Behaviour, one’s intention to perform a particular behaviour is the most powerful predictor of that behaviour. Intention in turn is then influenced by both attitudes and subjective norms, where subjective norms involve how a person perceives social pressure from peers to perform or eschew the behaviour in question (Ajzen and Fishbein, 1980). An attitude is a disposition to respond favourably or unfavourably to an attitude object, which can be anything concrete or abstract that a person is able to hold in mind (definition adapted from Cross, 2005). For example, the general idea of a ‘factory farm’, the act of using analgesics for dehorning, or even the dairy cow herself can all be considered attitude objects towards which someone may hold positive or negative attitudes (N.B. ‘view’ and ‘opinion’ are often used as proxies for ‘attitude’).

Attitudes in turn are a partial function of the evaluation of the beliefs held about the attitude object along with the strengths of those beliefs. Apart from helping to predict behaviour, the study of attitudes also becomes useful because it can help reveal the values and beliefs that underlie them. Beliefs represent the composite of information someone holds about an attitude object and link the object with certain attributes (Fishbein and Ajzen, 1975). For example, the belief that ‘dehorning is painful for the calf’ links the psychological object ‘dehorning’ to the attribute ‘is painful for the calf’, and would likely contribute to forming a negative attitude to the practice. Beliefs are often evaluative in nature, such that they apply a value judgment.

Values root both belief and attitude formation (Bem, 1970; Seligman et al., 1996; Ajzen, 2001). Values can be thought of as ‘desirable, trans-situational goals…that serve as guiding principles in people’s lives…’ (Seligman et al., 1996). Values form a foundational platform through which an individual views the world. This is why people’s values about animal welfare are relatively impervious, regardless of the level of information that they might have about a specific practice (see Hansen et al., 2003). It also explains why educational efforts towards the goal of, say, heading off criticisms about husbandry practices in the livestock industries often fall short: they fail to address the deeply held values that people hold around animal care.

A variety of methods have been applied to inquiries into stakeholder attitudes on farm animal welfare. Much of the work to date in North America has been based on quantitative surveys (e.g. Heleski et al., 2004, 2005 and 2006; Heleski and Zanella, 2006; Levine et al., 2005; Prickett et al., 2010). Survey studies have real benefits: they can be deployed on a broad scale and can be designed to capture representative samples that offer the opportunity to generalize to broader populations. But people’s perceptions of farm animal welfare are complex; uncovering the basis of concerns likely requires a more in-depth approach than can be achieved by quantitative survey work alone. Indeed, ‘social scientists tend to be sceptical about the notion that lay attitudes to distinct issues…can be studied in abstraction within a quantitative or statistical framework. In daily life, these issues rarely arise other than in an interconnected or merged way […] experience is less likely to be distorted if it is studied discursively within a less rigid, more qualitative framework’ ( Hansen et al., 2003, p. 117).

European researchers have pioneered work on qualitative approaches to understanding stakeholder values around animal welfare (e.g. Borgen and Skarstad, 2007; Kauppinen et al., 2010; Vanhonacker et al., 2010; Benard and de Cock Buning, 2013). Qualitative methods of inquiry can vary, but typically include interviews, focus groups and field research (ethnography). A common thread is the motivation to understand the meanings and experiences of these stakeholders, and to uncover the social and psychological phenomena behind the patterns described in quantitative approaches (Warren and Karner, 2010).

Qualitative research can involve either in-person or online interactions; both have their advantages. For example, online forms of inquiry provide an efficient and cost-effective way to reach large groups of people in a short period of time (Danielson, 2010), but lack the flexibility that would allow clarification or pursuit of a relevant thread of conversation. In-person approaches such as interview and focus group studies can be difficult to organize and expensive to deploy, but allow the researcher to probe for more information if a respondent is unclear. Focus groups also have the added benefit of eliciting dynamic conversation in a social context (Pivetti, 2007).

A mixed method approach will aid in capturing both the breadth and depth of the various psychological constructs that inform how people relate to the welfare of farm animals. Quantitative results help to contextualize qualitative findings, and qualitative results help to understand the reasons behind any quantitative differences observed. Qualitative methods such as interviews or focus groups can also be used to identify the range of relevant issues and used to design better-targeted quantitative studies (Brannen, 2005).

Case studies of our work

One approach that we have found useful is an interactive online platform, designed in concert with the University of British Columbia’s Centre for Applied Ethics (Danielson, 2010). Our intent has been to foster informed discussion between farmers, critics, consumers and other stakeholders on controversial issues relating to dairy cow welfare. Our Cow Views website created online forums designed to mimic town hall meetings, in that people from anywhere could access the site, access some background material on the issue (to help ground people’s comments with a minimum level of basic knowledge), and indicate whether they supported, opposed or were neutral to the practice in question. Participants were also asked to explain their position or to select reasons written by previous participants (providing a type of social context). This mixed approach allowed us to not only gauge levels of support (the quantitative aspect), but also the reasons behind it (the qualitative aspect).

Below we examine four case studies using this method. These examples illustrate how participant responses can identify contentious issues and approaches to resolving them. Our surveys were available on the Internet so anyone with Internet access could participate, but we encouraged participation of people in the North American dairy industry by publishing the link with brief articles in producer magazines (Progressive Dairyman and Ontario Farmer) that invited readers to participate. We also recruited dairy industry participants at conferences (e.g. the Western Canadian Dairy Seminar) and via e-mail (e.g. to members of the American Association of Bovine Practitioners). In addition, we recruited more demographically diverse participants using a commercially available recruitment tool (Amazon’s crowd-sourcing service, Mechanical-Turk). Mechanical-Turk provides access to respondents known to be more diverse than standard Internet samples and college students (Casler et al., 2013). These recruitment methods were used to elicit a wide range of perspectives and were not intended to be statistically representative of any particular population. Indeed, our sample had more cattle industry participants than would be expected in a random sample; participants also tended to be younger (under 30 years old) and female.

Tail docking

Why do some dairy farmers continue to dock the tails of their cows? The procedure was once believed to reduce the risk of the zoonotic disease leptospirosis but this idea has long been debunked (Mackintosh et al., 1982). Others have suggested that the cows with docked tails are cleaner, and thus at lower risk of udder infection, but a series of studies have shown that there is no protective effect of docking on cleanliness or mastitis (Eicher et al., 2001; Tucker et al., 2001; Schreiner and Ruegg, 2002; Fulwider et al., 2008; Lombard et al., 2010). Why then would some dairy farmers continue to spend time and money to conduct a procedure that harms their cows (by removing their natural fly-swatter) and harms the reputation of their industry?

The answer can be found, at least in part, in the responses obtained from social science work (Barnett et al., 1999), including our own online engagement exercise on the question of tail docking (Weary et al., 2011). The reasons provided by participants in support of docking showed a strong concern for cow cleanliness and udder health, combined with a belief in the efficacy of docking in addressing these problems. For example, one participant argued that docking is beneficial because… ‘on some farms it is an effective way to keep cows clean and prevent them from splashing manure everywhere’. Another argued that ‘cattle in tie stalls and free stalls tend to have their tail lying in urine and manure, thus swishing urine and manure all over themselves and their stable mates. Now picture this, you have to go wash udders and attach milkers all while tails are swishing back and forth’.

Thus, one reason why some farmers continue to dock their cows (even in Canada where the practice is not allowed under the producers’ own Code of Practice for the Care and Handling of Dairy Cattle, National Farm Animal Care Council, 2009; see also Lerner and Algers, 2013 for the related case of poor compliance within Europe concerning tail-docking for piglets) is that they place high value on health and hygiene and thus are highly motivated to do something that they believe might help. This motivation to address problems with cow cleanliness may explain why farms that dock dairy cows actually have dirtier cows (Lombard et al., 2010); these farmers may recognize that they have a problem with cow hygiene and use docking in a misguided attempt at improvement. Similarly, lying cubicles for cows are often configured with the aim of keeping the lying surface clean (e.g. by using an aggressive neck rail placement that prevents cows from standing in the stall; Bernardi et al., 2009), even though prolonged standing outside of the stall increases the risk of lameness (Bell et al., 2009).

One productive approach may be to focus education events for dairy farmers on effective methods for improving cow cleanliness. For example, one important cause of faecal soiling is that cows sometimes defecate onto the lying surface. The resulting soiling may be addressed in several ways, such as more frequent bedding management, including routine raking of the stall surface to ensure that this is clean and dry. Another way of keeping cows clean is to allow them improved opportunities to groom (e.g. by providing access to a mechanical brush; see DeVries et al., 2007). Grooming not only restores hygiene, it also provides cows the opportunity to engage in a valued natural behaviour. Farmers especially concerned about the role of the tail as a source of faecal soiling may wish to consider switch trimming as a less invasive alternative. This combined approach targets the concern and provides positive alternatives that address farmers’ values around cleanliness without creating the welfare issues associated with tail docking.

Research is also required to better understand the effects of cow housing and management on cow cleanliness. The natural behaviour of cattle is to defecate while standing, but much of the soiling of the stall surface occurs when cows defecate while lying down (Tucker et al., 2005). Understanding why cows fail to stand before defecating is key for the design and management of lying stalls that keep cows clean. We hypothesize that this failure to stand before defecation is a learned behaviour resulting from the pain and injuries that result from standing up and lying down in poorly designed stalls, but new research is required to understand the development of this behaviour and to determine how this understanding could be used to develop housing and management that improves cow cleanliness.

In summary, docking is not motivated by the desire to harm cows but rather to keep cows clean. Increasing our understanding of the views of producers, including those in favour of docking, has helped to identify new approaches to achieving change. Specifically, we recommend extension efforts directed at producers that provide effective management strategies for keeping cows clean, in addition to showing the evidence that docking harms cows and is not helpful for improving cleanliness or udder health.

Pain control for dehorning and disbudding

It is clear that the common procedures of disbudding and dehorning cause pain (for a review, see Stafford and Mellor, 2005), but despite the ready availability of effective pain control methods many farms provide no pain mitigation for these procedures (USDA, 2009; even in Canada where the Code of Practice for the Care and Handling of Dairy Cattle requires the use of pain control, Vasseur et al., 2010).

Unlike tail docking, which simply requires producers to stop a management intervention, providing pain relief involves some extra labour and expense for supplies. These extras are sometimes viewed as barriers to change. Indeed, in our work examining views around this issue (Robbins et al., in press), those against providing pain relief sometimes cited concerns about costs for the producer and the time required for the procedure. Others opposed to providing pain relief argued that the pain experienced was minimal and short lasting, and that pain control methods had little effect.

These comments suggest at least a few different approaches for improving the implementation of pain relief. One is that extension efforts need to provide more information to farmers on the likely costs and to better contextualize these figures. Multimodal treatment (including a local block and a non-steroidal anti-inflammatory) can cost less than a cup of coffee. Moreover, providing pain relief actually reduces the growth check that otherwise occurs after dehorning (Faulkner and Weary, 2000; Baldridge et al., 2011; Coetzee et al., 2012), so even in strict financial terms the minimal cost may be a good investment. One way to contextualize the cost is to compare this expense with total costs of rearing a replacement heifer (≈2500 USD; Gabler et al., 2000). In this context, it is easier to judge if the extra expense (if any) meaningfully affects the bottom line.

Beliefs about the duration of the pain response and the efficacy of treatment can be targeted directly by reviewing the scientific evidence from a number of sources showing the more than 24 h of behavioural and physiological pain responses and the efficacy of various methods of pain mitigation, including local blocks to essentially eliminate the intra-operative pain and non-steroidal anti-inflammatory drugs to greatly reduce post-operative pain (for review, see Stafford and Mellor, 2011).

The belief that pain is relatively minor or somehow does not really matter to the calf must also be specifically targeted. This view may correspond to our own everyday experiences in which we often experience at least some physical pain that has little effect on our quality of life. Unfortunately, very little work to date has attempted to determine how much different types of pain matter to animals. In humans, pain associated with low mood is much more likely to be detrimental to the patient’s quality of life, and this thinking can also be applied to consider when pain is likely to result in suffering in animals (Weary, 2014). In recent work, we used cognitive bias testing (see Mendl et al., 2009) to determine if calves experience a low mood state (as indicated by a pessimistic response to an ambiguous visual stimulus) in the hours following hot-iron disbudding. Calves were first trained using a milk reward to discriminate between red and white colours on a video screen, and then tested with ambiguous screen colours (i.e. shades of pink) before and after disbudding. Before disbudding, the calves responded as expected, for example visiting the screen midway between red and white about half the time. However, in the 24 h following disbudding, calves showed a pessimistic bias, responding less frequently than expected to the ambiguous test screens (Neave et al., 2013; Daros et al., 2014). This type of evidence directly targets the beliefs of skeptics by showing that dehorning pain has a sustained effect on the calf’s mood.

In summary, responses on the issue of pain relief illustrate a number of beliefs (e.g. that pain is short or relatively trivial) and perceived barriers (e.g. that effective pain control methods are expensive or unavailable). These results provide direction on how to better target extension efforts for producers, specifically focusing on the nature and duration of the pain response, providing lists of approved and efficacious drugs, providing a breakdown of costs and benefits together with context to evaluate the impact on farm profitability, and reviewing the evidence that this pain does cause calves to suffer.

Pasture access

Access to pasture is an increasingly contentious issue in countries where total confinement systems have become common. For example, <30% of US cows have access to pasture at some time during their lactation (USDA, 2010), but when asked their views many people place considerable value on cattle having access to pasture (e.g. Boogaard et al., 2008; Ellis et al., 2009). Work using the Cow Views engagement tool described above found that almost no participants were opposed to providing pasture access, including producers and people unaffiliated with the dairy industry (Schuppli et al., 2014). However, one important theme to emerge from the qualitative analysis was the perceived lack of feasible methods for how to incorporate pasture into the management of large, modern farms. Perceived barriers included the lack of access to land, poor grazing conditions and concern that pasture access would lead to lower levels of milk production.

Public concerns about animal agriculture are often related to restrictions of movement (Benard and de Cock Buning, 2013), suggesting that indoor housing may still be accepted if less confining. Pasture access also may be valued for different reasons by different people, for example because grazing is seen to be as natural, or because access to pasture comes with other attributes such as more space per animal and access to fresh air.

Work by our group at UBC (Legrand et al., 2009; Falk et al., 2012) has shown that cows have a strong preference for access to pasture during the night but remain indoors for the majority of the day, when temperature and humidity are high. Others have also shown that cows preferentially use indoor housing for protection from environmental conditions (Charlton et al., 2011a and 2011b) and spend more time indoors as the season progresses from summer to winter (Krohn et al., 1992) and on days when it rains (Charlton et al., 2011a). Moreover, high yielding dairy cows spend more time indoors when provided a choice between indoor cubicle housing with access to a total mixed ration (TMR) v. access to pasture where the TMR was also provided (Charlton et al., 2011b). Collectively, this work suggests that integrating the notion of ‘choice’ may be one worthy line of research to pursue when designing housing systems for dairy cattle, given that this enables the cow to decide when she wants to access the outdoors. This type of hybrid system allows for the high levels of milk production. In one study, we provided cows either exclusively indoor housing or indoor housing during the day and pasture at night. Cows allowed pasture access at night maintained the same high levels of milk production and TMR intake as the cows that were kept exclusively indoors (Chapinal et al., 2010).

We recognize that the issue of access to pasture is complex, particularly when one considers the potential impacts of housing systems on the environment (see von Keyserlingk et al., 2013). However, work to date has shown a high level of consensus on the desirability for some type of pasture access. We encourage more social science work to better understand what aspects of ‘pasture’ are considered most important, together with more biological and production work to design dairy systems that address the perceived constraints of producers and resonate with the broadly held public values.

Early separation of cow and calf

The calf is separated from the dam within the first hours after birth in many modern dairy operations, but this practice is the most contentious of any topic we have studied date (Ventura et al., 2013). Opposition to early separation was high among participants with no involvement in the dairy industry, while most people working within the dairy industry (including farmers) were in support of early separation, sometimes because they perceived that there were no practical alternatives.

Despite this diversity in views, participants often expressed similar types of values (including key components of animal welfare like affective state and health) when justifying their position, suggesting some room for consensus. For example, those who supported early separation argued that the practice minimizes emotional distress for cow and calf and that it promotes calf health. Opponents argued that early separation reduced opportunities for cow and calf to experience positive states and that it compromised cow health. Interestingly, both perspectives have at least some scientific support. For example, research has shown that separation distress is greater if cow and calf are allowed to remain together for longer periods before they are eventually separated (Weary and Chua, 2000; Flower and Weary, 2001). On the other hand, allowing the cow and calf to remain together does appear to confer benefits for cow health (Krohn et al., 1990 and 1999) without compromising production (Metz, 1987; Flower and Weary, 2001). Thus, while the science provides an unambiguous resolution to issues like tail-docking, the science in support (or opposition) to early separation is less clear.

Clearly, there is a need for further health-based research to more fully document the health risks and benefits associated with cow–calf contact. In particular, we call for research that is not overly simplistic (e.g. focusing on just a single type of infection); evidence-based practice requires a comprehensive assessment of the various risks and benefits. We also advocate demonstration projects showing a variety of systems for allowing contact in practice and thus targeting the perceived lack of realistic methods.

In this case, as with pasture, there may also be room for the development of intermediate solutions. One reason some people oppose early separation is the high levels of calf morbidity and mortality on many farms (Ventura et al., 2013), a concern echoed by industry stakeholders in more recent focus-group-based research (Ventura et al., 2015). Thus, opposition to conventional methods of calf rearing may wane if the dairy industry is able to adopt and maintain consistently high standards of calf care. We particularly encourage the development of systems that allow calves to drink more milk, more naturally (e.g. teat-based ad libitum systems) and systems that allow for some type of social interaction (e.g. housing in pairs or small groups). More complex social environments confer important benefits for the calf, for example, decreased weaning stress, lower reactivity in novel environments, improved social behaviour and enhanced performance when compared with calves housed individually (e.g. De Paula Vieira et al., 2010 and 2012, Meagher et al., 2015).

As various alternatives are developed and tested by researchers, it is important to maintain on-going engagement with industry (to ensure that the innovations meet their perceived constraints) and the public (to ensure that these options resonate with mainstream values). We suggest thinking of the process of innovation as a cycle starting with consultation to understand stakeholder views, leading to scientific innovation, and then continued research and consultation to determine if the innovation actually meets the original concerns or creates new concerns.

Looking forward

Above we introduced just a few examples for how we engaged participants within and outside the dairy industry on contentious issues around animal welfare. Both our online and focus group studies provide a social context for these discussions. This context may introduce biases (e.g. making individuals more likely to say what they think ‘should’ be said; see Podsakoff et al., 2003), but people’s views about contentious issues rarely exist in a vacuum. We believe it is important to explore the interpretations they bring to these issues while they are in a social context. An advantage of our online approach was that it allowed user anonymity and created a safe-space in which individuals could discuss difficult issues with less fear of social judgment; many consultation efforts utilize online techniques as a means of improving engagement (Rowe and Gammack, 2004). Diverse methodologies are available for public engagement (e.g. Swanson et al., 2011); as more social science research takes place on animal welfare issues we encourage the adoption of a range of methodologies that should help understand the root of conflicts around farming practices.

It is often argued that people’s criticisms of practices in agriculture might sometimes be based upon a poor understanding of the real conditions on farms (Heleski et al., 2006; Hubbard et al., 2007; Spooner et al., 2012; Benard and de Cock Buning, 2013). According to this ‘informational deficit model’ (Wynne and Irwin, 1996; Einsiedel, 2000), if lay stakeholders were better educated about practices on farms some of their criticisms might evaporate. Although knowledge is not the sole determinant in people’s attitudes towards a particular issue, it does contribute to people’s frame of reference (see Te Velde et al., 2002) and we agree that engagement via the provision of some information is critical in our ability to deliberate on complex issues (Castle and Culver, 2006). In the studies described above we provided some context for the participants, but this was limited. In some other work we have provided much greater access to information (e.g. Schuppli et al., 2015). Other research has gone further, in some cases taking lay participants onto working farms (Boogaard et al., 2008 and 2011; Krystallis et al., 2009) and gauging their responses in context. We encourage more work on the role of information and how it can be best provided to help inform debate.

We also see the need for work to address the polarization of views and lack of inter-stakeholder dialogue on farm animal welfare. The social aspect of the Cow Views site was intentionally created to bring diverse stakeholders into contact with the views of others, but we have not formally tested the effect of this and other forms of interaction on stakeholder views. Other research has shown the value of engaging multiple stakeholders in the process of designing improved housing for farm animals. For example, Groot Koerkamp and Bos (2008) showed how multi-stakeholder engagement can help in the development of more socially sustainable housing and management systems for laying hens.

One promising approach for new research is frame reflection, as described by Benard and de Cock Buning (2013). In this study, both citizens and farmers were organized into focus groups and asked to temporarily shift their frames of reference to consider the other group’s perspective. The results indicated that farmers were less able or willing to shift their frame of reference, but citizens were able to do so and this seemed to help in the development of a more differentiated and inclusive concept of animal welfare. Such methods have the potential to develop a ‘shared vision’ for the future of animal care, in part by helping to address certain barriers like the inability of some people within one sector to acknowledge the legitimacy of others’ views.