ArticlePDF Available

Assessing the human-animal relationship in farmed species: A critical review

Authors:

Abstract and Figures

The present paper focuses on six main issues. First, we briefly explain why an increased understanding of the human¿animal relationship (HAR) is an essential component of any strategy intended to improve the welfare of farmed animals and their stockpersons. Second, we list the main internal and external factors that can influence the nature of the relationship and the interactions between human beings and farm animals. Third, we argue that the numerous tests that have been used to assess the HAR fall into three main categories (stationary human, moving human, handling/restraint), according to the degree of human involvement. Fourth, the requirements that any test of HAR must fulfil before it can be considered effective, and the ways in which the tests can be validated are discussed. Fifth, the various types of test procedures that have been used to assess the HAR in a range of farmed species are reviewed and critically discussed. Finally, some research perspectives that merit further attention are shown. The present review embraces a range of farmed animals. Our primary reasons for including a particular species were: whether or not general interest has been expressed in its welfare and its relationship with humans, whether relevant literature was available, and whether it is farmed in at least some European countries. Therefore, we include large and small ruminants (cattle, sheep, goats), pigs, poultry (chickens), fur animals (foxes, mink) and horses. Although horses are primarily used for sport, leisure or therapy they are farmed as draught, food or breeding animals in many countries. Literature on the HAR in other species was relatively scarce so they receive no further mention here.
Content may be subject to copyright.
Assessing the human–animal relationship in
farmed species: A critical review
Susanne Waiblinger
a,
*
, Xavier Boivin
b
, Vivi Pedersen
c
,
Maria-Vittoria Tosi
d
, Andrew M. Janczak
e
,
E. Kathalijne Visser
f
, Robert Bryan Jones
g
a
Institute of Animal Husbandry and Welfare, Department of Veterinary Public Health and Food Science,
University of Veterinary Medicine, Veterina
¨
rplatz 1, A 1210 Wien, Austria
b
INRA de Clermont-Ferrand/Theix, 63122 St.-Gene
´
s Champanelle, France
c
Biological Institute, Animal Behaviour Group, University of Copenhagen, Tagensvej 16,
2200 Copenhagen N, Denmark
d
Istituto di Zootecnica, Facolta
`
di Medicina Veterinaria, Universita
`
Degli Studi, 20133 Milano, Italy
e
Department of Animal and Aquacultural Sciences, Norwegian University of Life Sciences,
P.O. Box 5003, N-1432 A
˚
s, Norway
f
Animal Sciences Group, Wageningen University Research Centre, P.O. Box 65,
NL-8200 AB Lelystad, The Netherlands
g
Consultant in Animal Welfare, Edinburgh EH9 3HH, Scotland
Accepted 1 February 2006
Abstract
The present paper focuses on six main issues. First, we briefly explain why an increased
understanding of the human–animal relationship (HAR) is an essential component of any strategy
intended to improve the welfare of farmed animals and their stockpersons. Second, we list the main
internal and external factors that can influence the nature of the relationship and the interactions
between human beings and farm animals. Third, we argue that the numerous tests that have been used
to assess the HAR fall into three main categories (stationary human, moving human, handling/
restraint), according to the degree of human involvement. Fourth, the requirements that any test of
HAR must fulfil before it can be considered effective, and the ways in which the tests can be validated
are discussed. Fifth, the various types of test procedures that have been used to assess the HAR in a
www.elsevier.com/locate/applanim
Applied Animal Behaviour Science xxx (2006) xxx–xxx
* Corresponding author. Tel.: +43 1 25077 4905; fax: +43 1 25077 4990.
E-mail address: Susanne.Waiblinger@vu-wien.ac.at (S. Waiblinger).
0168-1591/$ see front matter # 2006 Published by Elsevier B.V.
doi:10.1016/j.applanim.2006.02.001
APPLAN-2512; No of Pages 58
range of farmed species are reviewed and critically discussed. Finally, some research perspectives
that merit further attention are shown.
The present review embraces a range of farmed animals. Our primary reasons for including a
particular species were: whether or not general interest has been expressed in its welfare and its
relationship with humans, whether relevant literature was available, and whether it is farmed in at
least some European countries. Therefore, we include large and small ruminants (cattle, sheep,
goats), pigs, poultry (chickens), fur animals (foxes, mink) and horses. Although horses are primarily
used for sport, leisure or therapy they are farmed as draught, food or breeding animals in many
countries. Literature on the HAR in other species was relatively scarce so they receive no further
mention here.
# 2006 Published by Elsevier B.V.
Keywords: Human–animal relationships; Farm animals; Tests; Assessment; Welfare
1. The impact of the human–animal relationship (HAR) on welfare and
productivity
Domestication is a ‘process by which a population of animals becomes adapted to man
and to the captive environment by genetic changes occurring over generations and
environmentally induced developmental events recurring during each generation’ (Price,
1984). Despite countless generations of selective breeding the potentially most frightening
events that many farm animals are likely to experience are exposure to human beings and to
sudden changes in their social or physical environments (Boissy, 1995; Jones, 1996;
Hemsworth and Coleman, 1998). More specifically, unless they have become accustomed
to human contact, of either a neutral or positive nature, the predominant reaction of most
farm animals to people is still one of fear (Duncan, 1990; Jones, 1997). Not unexpectedly,
exposure to rough, aversive and/or unpredictable handling can exacerbate the problem.
Furthermore, it has been proposed that animals often perceive contact with a human being
as a predatory encounter (Suarez and Gallup, 1982; Jones, 1997; Boissy, 1998). Indeed,
many of the occasions on which animals and humans interact in current farm practice are
negatively reinforcing, e.g., veterinary treatment, restraint, depopulation, etc., while, other
than feeding, few are positively reinforcing. It has also been suggested that contact with
humans could become even more distressing if the increasing use of labour-saving
technologies, e.g., automation, result in reduced opportunities for the animals to become
habituated to people (Duncan, 1990; Rushen et al., 1999a). One of the primary reasons for
differences in the HAR found between farms is variation in the number, duration and nature
of daily interactions between stockpeople and the animals (Hemsworth and Coleman,
1998). The stockpersons’ behaviour is a major variable determining animals’ fear of or
confidence in human beings and, hence, the quality of the HAR. The nature/quality of
human–animal interactions can range from frequent, calm and ‘friendly’ to infrequent and
predominantly negative ones (Hemsworth and Coleman, 1998; Waiblinger et al., 2002).
Negative handling and fear of humans have a number of undesirable consequences for
the livestock, farmers and consumers. For instance, the sudden, intense or prolonged
elicitation of fear can seriously damage the welfare, productivity, product quality and
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx2
profitability of farm animals. These undesirable consequences and their likely
underpinning mechanisms are described in greater detail elsewhere (e.g., Jones, 1996;
Hemsworth and Coleman, 1998; Mills and Faure, 1990). For present purposes though, we
present just a small number of illustrative examples.
In poultry, inappropriate fear reactions, like panic or violent escape attempts, not only
waste energy and thereby impose a metabolic cost but they can also result in injury or even
death when the birds run into obstacles or pile on top of and claw each other. This is a major
welfare insult because injuries can lead to infection, chronic pain, debilitation and social
withdrawal (Jones, 1996, 1997). High fear of humans is also associated with reduced egg
production, growth, food conversion efficiency, product quality and sexual activity, with
increased aggression and handling difficulties, and with immunosuppression (Komai and
Guhl, 1960; Gross and Siegel, 1982; Shabalina, 1984; Barnett et al., 1992, 1994; Jones
et al., 1993; Rosales, 1994; Jones, 1996). Fear of humans accounted for 28% and 20% of
the variation in food conversion efficiency in broiler chickens (Jones et al., 1993) and in egg
production by commercial layers (Barnett et al., 1992), respectively, suggesting that fear of
people could cost the broiler and layer industries several million pounds each year (Jones,
1996).
In pigs, negative handling increased adrenal weight (indicative of chronic stress),
impaired growth and reproductive performance and induced high fear of humans, both in
the laboratory (Gonyou et al., 1986; Hemsworth and Coleman, 1998; Seabrook and Bartle,
1992) and on commercial farms (Hemsworth et al., 1989a, 1993a, 1999). The magnitude of
the negative correlation coefficients between the avoidance of people and the pigs’
farrowing rates demonstrated that fear of humans is a major limiting factor on productivity
(Hemsworth and Coleman, 1998).
A similar picture emerges in farm herbivores. Negative interactions or fear of humans
are associated with reduced milk yield or milk let down in dairy cows and goats (Seabrook,
1972; Lyons, 1989; Knierim and Waran, 1993; Rushen et al., 1999b; Breuer et al., 2000;
Waiblinger et al., 2002). Chronic and acute stress responses, traumatic incidents, injuries,
death and poorer meat quality are also more prevalent in cows, heifers and calves that have
been handled negatively and that show elevated fear of humans (Fordyce et al., 1985;
Hemsworth et al., 2000; Breuer et al., 2003; Lensink et al., 2001b). Regular gentle handling
counteracted some of these undesirable effects (Lensink et al., 2000b,c).
High fear of humans also has harmful effects on farmed fur animals, e.g., non-handled
foxes were more frightened of humans and novel stimuli, had enlarged adrenals and lower
reproductive success than handled ones (Pedersen, 1993b, 1994), whereas regular rewards
(tit bits) reduced fear in silver fox vixens and enhanced the cubs’ growth and behavioural
ontogeny (Bakken, 1998). Foxes or mink selected over several generations for tameness
showed lower basal and stress-induced hypothalamo–pituitary–adrenocortical activity,
higher reproductive success or reached sexual maturity earlier and were easier to mate than
more fearful animals (Plyusnina et al., 1996; Jeppesen and Pedersen, 1998; Nikula et al.,
2000; Malmkvist, 2001a).
In horses, fear of humans and/or exposure to negative interactions can cause serious
accidents, e.g., using a whip during steeplechasing increases the risk of falling (Pinchbeck
et al., 2004). Conversely, early handling improves manageability and reactivity to humans
(e.g., Lansade et al., 2004; Søndergaard and Halekoh, 2003). Despite this, most training
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 3
techniques are based on traditional methods with a surplus of punishments, although
innovative methods using more positive reinforcement are slowly being developed (Waran
et al., 2002).
From the stockpersons’ point of view, fearful animals are often more difficult to handle
and manage (Gonyou et al., 1986; Grandin et al., 1987; Pedersen and Jeppesen, 1990;
Boivin et al., 1992b); e.g., defensive reactions make the handler’s work more difficult and
sometimes cause his or her injury or death (Le Neindre et al., 1996). This, in turn,
exacerbates the problems encountered during procedures like routine examination,
artificial insemination and translocation, thereby decreasing job satisfaction, motivation,
commitment and self-esteem (Jones, 1996; Hemsworth and Coleman, 1998). A negative
feedback cycle might then be established whereby the stockpersons’ attitudes and
behaviour towards the animals in their care worsen and thereby increase the livestocks’ fear
of humans.
Conversely, the development of a positive HAR (low levels of fear or high levels of
confidence in people) can be beneficial. For example, the presence of a familiar human,
contingently providing gentle handling, may calm the animals in potentially aversive
situations (e.g., isolation, tethering, rectal palpation, insemination) thereby reducing
distress and the risk of injury to the animal and the human (Korff and Dyckhoff, 1997;
Pedersen et al., 1998; Boivin et al., 2000; Waiblinger et al., 2004) and potentially
enhancing reproductive performance. A high quality HAR clearly requires a certain level
of positive human contact, and this is most likely in husbandry systems that involve regular,
intense and long-term contact with humans; dairy or sow farms provide good examples.
Neutral relationships, where fear of humans is low but the animals still avoid physical
contact, can also be found and probably develop via frequent neutral or mildly positive
human contact, a lack of negative contact, and none or few intense positive interactions
(e.g., in dairy cows: Waiblinger et al., 2003b).
The HAR is also sensitive to stockpersons’ decisions on management and housing.
Stockperson behaviour and attitudes were related to the emphasis placed on taking the
animals’ needs into account when making management or resource design decisions, and
to subsequent injuries or disease prevalence (Waiblinger et al., 2001; Lensink et al., 2001a).
There are at least two explanations. Firstly, attitudes towards animals influence ways of
interacting with them and decisions made about them. Secondly, increased contact
improves the stockpersons’ knowledge of the animals and facilitates the early recognition
and solution of any problems (Seabrook, 1984; Waiblinger et al., 2001).
A poor HAR, high levels of fear and inappropriate management and housing represent
bad news for the animals, the farmers and the concerned public. Clearly, we need to
develop effective, practical strategies for alleviating the animals’ fear of humans and for
promoting a more positive HAR. These strategies might include genetic selection for
increased adaptability and tractability, increased human–animal contact and the
modification of the stockpersons’ attitudes and behaviour through educational initiatives.
Not only would such efforts likely improve the quality of life for livestock and farmers but
by engendering the perception of farmers as benevolent guardians rather than unfeeling
jailers, it would help to address societal concerns about farm animal welfare. Increased
public esteem may also serve to attract more caring people into the industry (English et al.,
1992).
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx4
2. The human–animal relationship—concept and influential factors
The HAR can be defined as the degree of relatedness or distance between the animal and
the human, i.e., the mutual perception, which develops and expresses itself in their mutual
behaviour (Estep and Hetts, 1992). It is a dynamic process with the catalogue of previous
interactions between the animal and humans forming the foundation for an established
relationship that then exerts a feedback effect on the nature and perception of future
interactions.
In principle, a relationship develops between two individuals that know each other
(Estep and Hetts, 1992), in particular the caretaker and an animal in his/her care. Such
relationships require mutual individual recognition and are therefore limited to systems (or
experiments) enabling sufficient contact. However, animals might also generalise their
experiences with one human to other humans (Jones, 1994; Tanida et al., 1995; Hemsworth
et al., 1996a), although pigs (Tanida and Nagano, 1998; Koba and Tanida, 1999), poultry
(Davies and Taylor, 2001), cattle (Taylor and Davis, 1998; Rybarczyk et al., 2001) and
sheep (Boivin et al., 1997) can discriminate between different people. Stockpeople may
also show generalised attitudes and behaviour towards their animals (Hemsworth and
Coleman, 1998). Thus, if individual recognition is precluded or generalisation occurs a
general HAR may develop. Of course, both individual recognition and generalisation of
response can operate within common test situations. For example, lambs that were bottle-
fed and received gentle handling showed less isolation distress when a known or unknown
shepherd was present, though the effect was greater with a familiar shepherd (Boivin et al.,
1997). Handled piglets also interacted more with familiar and unfamiliar humans than non-
handled ones, but made contact sooner and more often with the familiar handler and were
less agitated when caught by him than by an unknown person (Tanida et al., 1995).
2.1. The animals’ perspective—the animal–human relationship
Human–animal interactions can involve visual, tactile, olfactory and auditory
perception, and human contact on farm can be subdivided into five main types: (a)
(stationary) visual presence, (b) moving between the animals without tactile contact (but
maybe using vocal interactions), (c) physical contact, (d) feeding (rewarding), and (e)
invasive, obviously aversive handling.
An animal may perceive an interaction as negative, neutral or positive; this is influenced
by its existing relationship with humans which is, in turn, based on previous interactions
(De Passille
´
et al., 1996; Munksgaard et al., 1997). However, even if the HAR is very
positive, some interactions are aversive because they are painful or otherwise distressing
(dehorning, beak-trimming, etc.). In contrast, a high quality HAR might reduce the
perceived aversiveness of traumatic events like isolation and restraint (e.g., Hinrichsen,
1979; Grandin, 1984; Boivin et al., 2000).
The period of the animal’s life during which human contact occurs can be important,
although conflicting results have caused debate (Jones, 1995b; Burrow, 1997; Boivin et al.,
2003). For example, no durable effects of early handling were found in dairy calves (Boissy
and Bouissou, 1988), foxes (Pedersen, 1992) or horses (Williams et al., 2002), whereas
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 5
goats (Lyons, 1989; Boivin and Braastad, 1996), beef cattle (Boivin et al., 1992b, 1994),
sheep (Markowitz et al., 1998) and foxes (Pedersen, 1994) showed long-term effects.
Previous experience of a specific interaction and the controllability or predictability
associated with it may also be influential. Firstly, for example, previous aversive
experiences and unfamiliarity with a squeeze can hamper attempts to lead a cow to the
apparatus and to confine it in the head gate (Lewis and Hurnik, 1998). Secondly, silver
foxes associated pleasant or unpleasant interactions with the colour of the humans’
clothing (Bakken et al., 1993): foxes captured with neck tongs by someone wearing white
clothes showed greater hyperthermia to the mere sight of white rather than blue clothing.
The animals’ perception of humans and their responses to certain interactions are also
strongly influenced by their underlying personality traits, e.g., fearfulness/emotionality
(Jones et al., 1994; Jones, 1996; Visser et al., 2001). Indeed, the substantial variation
between and within breeds of several species in the animal’s responses to humans or
handling illustrates the powerful effect of the background genome (e.g., Murphey et al.,
1981; Hemsworth et al., 1990; Le Neindre et al., 1993; Grandin and Deesing, 1998; Jones
and Hocking, 1999).
Hediger (1965) described the five most common roles or ‘meanings’ that animals may
ascribe to humans: predator, prey, part of the environment without social significance,
symbiont, and conspecific. Estep and Hetts (1992) suggested that some of these roles may
not be mutually exclusive, and that an animal probably perceives a human in terms of a
combination of the above roles and according to the current situational factors. However,
some of these terms may more realistically describe observed behaviour than actual
perception; this applies especially to symbiont but it is also questionable if animals actually
see humans as conspecifics (Boivin et al., 2003), except for hand-reared animals directing
courtship behaviours to humans (e.g., Sambraus and Sambraus, 1975). An emotion-based
classification of animals’ perception of humans results in three main categories:
frightening (indicated by fear, avoidance and stress responses in the presence of a human;
Hemsworth and Barnett, 1991; Hemsworth and Coleman, 1998), neutral (no signs of fear
or positive emotions; Waiblinger et al., 2003b), or a source of pleasant emotions (e.g.,
reassurance in aversive situations; Boivin et al., 1997; Visser et al., 2002). These categories
can also overlap or vary according to the person or location (Rushen et al., 1998, 1999b;
Jago et al., 1999).
In short, different emotions and motivations are involved in the perception of and
reaction to humans. They belong to two dimensions: positive/pleasant and negative/
unpleasant (Fig. 1). Their relative strengths determine an animals’ relationship to humans,
from negative through neutral to positive.
The nature of any ‘communication’ between an animal and a human can profoundly
influence the way in which the HAR develops. Humans may unconsciously emit calming
signals or ones of danger, often overlooking resultant signs of fear, aggression or calmness
in the animal, and subtle differences in human behaviour may be crucial (Hennessy et al.,
1997, 1998). Species-independent body signals may be important, e.g., threatening or
submissive behaviours are often associated with making the body appear larger or smaller,
respectively (Eibl-Eibesfeldt, 1999). Conversely, imitating species-specific animal signals
has been recommended for effective control of farm animals (Grandin et al., 1983;
Grandin, 1987). Indeed, it is widely used in training and behavioural therapy of dogs and
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx6
may provide the basis for the success of Fulani herdsman in the control of cattle
(Hinrichsen, 1979; Lott and Hart, 1979). Clearly, in-depth investigation of human–animal
communication is required. In the meantime, the human’s posture, facial expression or
vocal communication must be considered as likely influential variables (see Section 3.2).
2.2. The humans’ perspective—the human–animal relationship and underlying
determinants
Starting with Seabrook (1972), substantial literature now reveals the impact that the
caretaker’s behaviour, personality and attitude can have on farm animals’ relationships to
humans and on their welfare and performance. This is not unexpected because the human
mostly determines the number and nature of the interactions and, hence, the relationship;
the animals more often react to humans’ actions rather than initiate them. Further,
stockpeople differ considerably in the type and amount of their interactions with the
animals under their care (Hemsworth and Coleman, 1998; Lensink et al., 2000a;
Waiblinger et al., 2002). The housing or production system can be constraining, but in the
dairy, pig or veal industries the most important factors determining the behaviour of
stockpeople were personality and attitude (Seabrook, 1984; Hemsworth et al., 1989a;
Coleman et al., 1998; Breuer et al., 2000; Lensink et al., 2000a; Waiblinger et al., 2002).
Personality is the individual’s unique system of traits that affect how he/she interacts
with the environment. Farmers’ personality characteristics (aggressiveness, agreeableness,
self-confidence, etc.) were correlated with their management, interactions with the
animals, and animal productivity (Seabrook, 1972, 1995; Seabrook and Darroch, 1990;
Waiblinger, 1996; Waiblinger and Menke, 1999; Waiblinger et al., 2002). Unlike attitudes,
personality characteristics are relatively stable over time (Costa and McCrae, 1986).
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 7
Fig. 1. The two dimensions (unpleasant; pleasant) contributing to the human–animal relationship and examples of
influential variables. Increased levels of pleasant emotions improve the relationship and vice versa.
Attitudes towards farm animals and their development have been extensively reviewed
(Hemsworth and Coleman, 1998). Herein, we simply identify the most important aspects
that should be considered for a better understanding of HAR. Attitudes express a positive
or negative evaluation of ‘an entity’ (species or particular animal), a tendency for or
against, a like or dislike, etc. (Hemsworth and Coleman, 1998). Beliefs, emotions and
behavioural intentions with regard to animals are different aspects of human attitude that
are generally consistent with each other and with human behaviour (Fishbein and Ajzen,
1975; Hemsworth and Coleman, 1998). For example, if a stockperson has an underlying
general positive attitude about cows (beliefs) and thinks they are intelligent, learn easily,
and like to be stroked, that person is likely to enjoy contact with the cows (emotion), to
favour handling animals patiently (behavioural intention), to believe that regular positive
contact is important, and to show positive behaviours towards the cows (Waiblinger et al.,
2002). Behavioural attitudes are generally considered to be better predictors of the
expression of a particular behaviour than are general attitudes, which mainly act on
behaviour indirectly by affecting the formation of behavioural attitudes. However, studies
on dairy and pig farms found correlations between general attitudes and behaviour,
especially that involving close contact with the animals (Coleman et al., 1998; Waiblinger
et al., 2002). Attitudes are learned, through experience with or information about the
animals, and they can change with new experiences or information (Ajzen,1988;Pauland
Serpell, 1993; Hemsworth and Coleman, 1998). Thus, the daily interactions may affect
attitude: if a caretaker believes a pig is difficult to move he tends to use more aversive
handling thereby initiating a vicious circle where the pigs’ fear of humans and its
difficulty of handling are likely to increase (Hemsworth and Coleman, 1998). Attempts to
change attitudes can improve the HAR. Indeed, cognitive–behavioural intervention
methods have improved stockpersons’ attitudes and behaviour towards their animals in
the Australian pig and dairy industries (Hemsworth et al., 1994a, 2002; Coleman et al.,
2000). However, attitudes can also worsen, e.g., the positive attitudes of new staff towards
pigs can deteriorate if they work in a system where the pigs are treated as machines
(Seabrook, 2001).
Other factors that can impact strongly on human behaviour, either directly or via
changing attitudes, include knowledge of the job, experience of particular animals and the
system, job satisfaction, the possibility of performing a particular behaviour or adopting an
alternative one, the behaviour of colleagues, the perceived consequences of their
behaviour, time constraints, and psychological strain in the work environment or home life
(Hemsworth and Coleman, 1998; Lensink et al., 2000a; Seabrook, 2001; Coleman et al.,
2003; Waiblinger et al., 2003a). All these factors could therefore influence the HAR. They
merit continued investigation.
3. Methods of assessing the animal–human relationship
Measuring the attitudes and behaviour of stockpeople gives insights into their
relationships with the animals. Attitudes cannot be measured directly but can be inferred
from responses to a series of statements in a questionnaire (Hemsworth and Coleman,
1998). The farmers’ behaviour can be observed directly during routine day-to-day
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx8
interactions like milking, moving animals or provision of food. Careful instruction is
necessary to achieve valid responses or observations.
Measuring animals’ reactions to humans enables us to reach conclusions about how they
perceive specific human beings or people in general. The animal’s reactions reflect a
mixture of different emotions (see Fig. 1). Fear is likely to be of primary importance,
depending on the type of animal and husbandry system, but inferences can also be drawn
about its social attachment to humans, the nature (positive, neutral or negative) of its past
experience with people, and the quality of stockmanship (including overall management
and environmental design decisions). In the present paper, where animal welfare is a key
issue, we concentrate on tests aimed at evaluating the HAR from the animal’s perspective.
Many researchers have measured animals’ behavioural and physiological reactions to
human beings to illuminate selected aspects of the HAR. These include: fear and avoidance
of humans (Hemsworth and Barnett, 1989; Hemsworth et al., 1989a, 2000; Jones and
Waddington, 1993; Pedersen et al., 2002), confidence in or attachment to humans
(Pedersen and Jeppesen, 1990; Boivin et al., 2000; Lensink et al., 2000b), ease of handling
(Boivin et al., 1992b; Lensink et al., 2000c) and/or the potential for positive relationships to
reduce the animals’ distress during aversive events (Rushen et al., 2001; Waiblinger et al.,
2004). Many experiments, particularly in the laboratory, focussed on the effects of different
types of handling treatments (e.g., rough, gentle, mixed) on the animals’ reactions to people
(e.g., Hemsworth et al., 1989b; Pedersen, 1993a; Boivin et al., 2000; Hemsworth and
Barnett, 1991; Jones and Waddington, 1993; Jones, 1995a). Studies carried out at
commercial farms largely examined the relationships between measures of approach/
avoidance and potentially influential variables such as the stockpersons’ behaviour and
attitude, the type of management and housing, or animal characteristics such as breed or
age (Hemsworth et al., 1989a, 2000; Waiblinger et al., 2003b). In addition, individual
differences in selected personality traits, such as general reactivity, fearfulness, coping
style, temperament or docility (e.g., Tilbrook et al., 1989; Jones et al., 1992a, 1994; Erhard
et al., 1999; Visser et al., 2001, 2002) have been evaluated. Therein, animals with the same
history of human–animal interactions are compared in their reactions to a human or to
handling.
Tests measuring the animals’ reactions to human beings fall into three main categories:
(1) reactions to a stationary human, (2) reactions to a moving human and (3) responses to
actual handling. In the latter category, in addition to specially designed tests, observations
taken during routine handling can yield valuable information. As outlined below, the
relative importance of possible confounding motivations may differ between the test
categories. For example, when testing the animals’ approach reactions towards a stationary,
unknown human, its motivation might be strongly influenced by its level of curiosity or
interest, i.e., the motivation to explore, whereas such motivations seem subordinate to the
avoidance reaction when the animal is approached by a human being (Murphey et al., 1981;
Marchant et al., 1997; Waiblinger et al., 2003b).
Within each of the categories described above, the precise tests employed may also
differ according to the test location, e.g., whether it is familiar or not. Indeed the physical
and social environment can strongly influence the test outcome. For instance, the animals’
reactions to the test human might be confounded or swamped for a number of reasons
including: (a) either fear-induced flight or behavioural inhibition elicited by enforced
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 9
exposure to novel, and hence potentially frightening environmental stimuli; (b) distraction
of attention by such stimuli; (c) memory of handling associated with the test location or a
similar one; (d) human contact incurred in moving the animal from its home cage to a test
arena (De Passille
´
et al., 1996; Rushen et al., 1998; Jago et al., 1999). All these variables
must be considered when choosing the most appropriate test for assessing the HAR.
Before discussing these influential variables and the tests in greater detail, we describe
the concept of validity and some ways of assessing the validity of HAR tests.
3.1. Validity and reliability
Measures used to study human–animal relationships should ideally be established as
reliable and valid prior to their use (see Table 1). Validity refers to the relation between a
measured variable and what it is supposed to predict, in this case, the animals perception of
humans. Validity is determined by accuracy, specificity and scientific validity (Martin and
Bateson, 1993; Table 1).
Accuracy refers to the degree of freedom from systematic errors that might otherwise
cause over- or underestimation of animal characteristics. Assessment of the accuracy of
measures of the HAR may involve registering whether different stockpersons or
professional observers score the behaviour of the same animals in the same way. If there is
a systematic disagreement then the indicators or recording methods may have low
accuracy.
Specificity is the extent to which a variable reflects what it is supposed to and nothing
else. It is useful here to draw on the principles of construct validation (Cronbach and
Meehl, 1955; John and Benet-Martinez, 2000) based on convergent and discriminant
validity. Convergent validation involves a search for convergence across independent
measures of the same conceptually related construct. In practice, this can be done by testing
for predicted correlations between alternative measures of either fear/avoidance or
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx10
Table 1
Definitions of the different components of reliability and validity
Subject Question
Validity Is the measure accurate, specific and scientifically valid?
Accuracy Is the measure free from systematic errors?
Specificity Does the measure reflect what it is supposed to and nothing else?
Convergent validity Are conceptually related measures empirically associated with one another?
Discriminant validity Are independent conceptually unrelated measures empirically independent?
Scientific validity Does the method give scientifically relevant information and answer the
research question?
Internal validity Does the method answer the research question?
External validity Does the method have relevance in other situations and does it have
practical relevance?
Reliability Does the measure have consistency and high resolution, and is it precise
and sensitive?
Consistency Do repeated measures of the same construct produce the same results?
Sensitivity Does the measure change with small changes in the true value?
Resolution What is the smallest detectable change in the true value?
Precision How free is the measure from random errors?
attraction to humans that have been recorded in different tests thought to measure the same
thing. Discriminant validation, on the other hand, searches for divergence across
independent measures of different (conceptually unrelated) constructs. This could, for
example, involve studies showing that a measure of personality such as general fearfulness
or sociality is not correlated with a measure thought specifically to reflect aversive and/or
pleasant experiences with a human.
Scientific validity in the present context refers to whether the method and response
variable actually tells us anything of scientific importance about some component of how
an animal perceives humans. It can be useful, as suggested by Lehner (1996), to subdivide
scientific validity into internal validity (which characterises how well the research
methodology answers the question in a given study) and external validity (which reflects
how applicable the results of a given study are to other situations (times, places) and their
practical relevance). Relevance to situations outside of the experiment is especially
important for methods developed for on-farm studies.
Measuring animal’s reactions to humans involves measuring a number of different
emotions, including fear (see Fig. 1), which vary in intensity due to the existing
relationship. Because most researchers focussed on fear of humans, it will be used in the
following as an example of how methods can be validated. Methods of evaluating the
internal validity of putative measures of fear of humans involve testing for predicted effects
of a treatment thought to affect it. Aversive treatment would be predicted to increase
avoidance/reduce approach, indicating increased fear of humans, whereas pleasant
treatment would be expected to reduce avoidance/increase approach indicating reduced
fear. If treatment effects are not in the predicted direction the sensitivity (see Table 1)orthe
internal validity of the putative measures may be considered low. Confirmation of predicted
effects on the response variable would entail a partial internal validation. Several
experiments have compared the effects of positive and negative handling treatments on
avoidance or withdrawal distance (see Hemsworth and Barnett, 1991; Breuer et al., 2003),
but if no control treatment (neutral contact) is included, such experiments cannot by
themselves show whether the measures have internal validity and sensitivity for measuring
effects of only positive or only negative handling.
External validity can be assessed by determining if recorded measures of fear of humans
predict zootechnical performance (milk production, egg production, growth, immune
function) or other aspects of animal behaviour or physiology thought to be sensitive to
variability in fear at on-farm locations. For example, human–animal interactions can
markedly affect the productivity of farm animals (Hemsworth and Barnett, 1987;
Hemsworth et al., 1993a; Hemsworth and Coleman, 1998; Janczak et al., 2003; Jones,
1996), and a negative relationship between fear, as indicated by avoidance of humans, and
productivity was found in pigs, cattle and foxes (Hemsworth et al., 1981a,b; Jeppesen and
Pedersen, 1998; Breuer et al., 2000; Nikula et al., 2000).
If the effects of several different treatments are tested, preferably in different studies, the
results can be used to evaluate the specificity (discriminant and convergent validity) of
putative measures of fear of humans. For instance, we would predict that prior exposure to a
novel object, not associated with or similar to humans, has no effect on fear of humans.
This could establish that conceptually unrelated constructs, novelty-induced anxiety and
fear of humans, are also empirically unrelated, and thus have some discriminant validity. If
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 11
repeated exposure to novelty itself affects measured fear of humans, the measure is likely to
reflect general fearfulness in addition to or instead of fear of humans. The inverse could
also be the case if pleasant or unpleasant handling affects fear of humans in the predicted
direction, but also affects fear of novel objects. In these cases, specificity in the form of
discriminant validity might not be high. Early handling of cattle reduced the distance at
which animals avoided an approaching human, but did not affect reactions to non-human
stimuli (Boissy and Bouissou, 1988), suggesting that tests of avoidance of humans or of
non-human stimuli may have discriminant validity for cattle. Correlations between
different measures can also test for discriminant validity. Here one would predict a lack of
correlation between measures of fear of humans and those of novelty-induced anxiety,
hunger, aggressiveness or other unrelated constructs. Studies specifically testing for
discriminant validity of measures of fear of humans are scarce in the farm animal literature;
this approach merits pursuit.
Convergent validity can be evaluated by similar methods, but in this case one would
predict that different forms of aversive or pleasant treatments would affect different
indicators of fear of humans in the same direction. Cattle handled regularly allowed closer
approach by humans, were easier to lead, and fed more in a novel environment in the
presence of a human (Boissy and Bouissou, 1988), suggesting that these indicators have
convergent validity as measures of fear of humans. A number of studies also showed
convergence between increased cortisol concentrations after exposure to humans,
increases in basal corticosteroid concentration, and changes in adrenal gland weight and
morphology (Hemsworth and Barnett, 2000). Correlations between different putative
indicators of fear may also be used to assess convergent validity. Here one would predict a
positive association between different indicators that are all thought to reflect fear of
humans, as reported in domestic chicks exposed to different handling treatments (Jones,
1993).
Reliability, which is related to the degree to which measures are free from random
errors (Martin and Bateson, 1993), is another important requirement of scientific
measurement, and will be mentioned only briefly here. Reliability is determined by
precision, sensitivity, resolution and consistency (see Table 1). Consistency, e.g., inter-
and intra-observer correlations, can be readily assessed in behavioural studies, but this
may be somewhat complicated by real changes in animal perception and associated
changes in behavioural expression over time. Knowledge about the sensitivity of
measures can also be important when evaluating internal validity. A measure may, for
example, have low sensitivity to small changes in a treatment variable but be strongly
affected by larger changes; thus having little or high internal validity for evaluating small
or large changes, respectively.
We recommend that validation should be given more attention in future studies. As a
general basis for validating measures it is also important to have insight into the general
biology and behaviour of the species in question; a detailed ethogram may be a valuable
starting point. This should ideally include detailed species-specific behavioural
expressions such as posture, head and tail position, ear position and eye movements.
The registration of such detailed behavioural expressions and more comprehensive
validation of test methodology may provide considerable information about an animal’s
emotional state and its perception of humans.
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx12
3.2. Technical problems and solutions: confounding motivations and other factors
Defining a test procedure is never simple. The above validation section identifies the
steps that should be taken when developing a valid test paradigm. As inferred, from initial
design to realisation, many confounding factors could come into play. This is particularly
true of tests designed to measure animal’s reactions to human beings. By definition, many
such tests use a specific person as a ‘standardised test stimulus’, but others may also be
involved, e.g., in bringing the animal to the test situation. The potential impact of ‘general’
human presence on the animal’s behaviour is sometimes minimised through previous
habituation to people, but it is necessary to balance the ‘standardising’ effects of
habituation and the risk of dampening responsiveness to humans to such an extent that it
compromises assessment of treatment effects. In any case, habituation procedures can be
difficult to impose when working with farm animals, particularly larger ones.
The present section identifies those methodological aspects that, in our opinion, merit
particular attention. These include the effects of pre- and post-test conditions, variations in
test duration, repeated testing and exposure to a number of different tests.
3.2.1. Pre-test conditions
First, the animal is often brought to the test environment; this may involve sorting and
isolating it from its social group, catching it, and carrying or leading it to the test arena.
Similarly, physiological tests often require fitment of radiotelemetry devices or the
withdrawal of blood. Such procedures themselves elicit reactions. However, very rarely are
such procedures precisely described or any observations performed during their execution,
despite the potential knock-on effects of variables such as the handler’s familiarity,
personality, attitude, haste and calmness (Boivin et al., 1997, 1998b; Hemsworth, 2003;
Tanida and Nagano, 1998; Seabrook, 2001). Some animals may also react to visible
observers (Boivin and Braastad, 1996), though there was little effect on the open-field or
tonic immobility responses of chickens unless the observer stared directly at the bird or
wore unfamiliar clothing (Jones, 1987a, 1990, 1996).
Researchers must also consider the animals’ expectations during a test. For example,
choice tests measuring animals preferences for or aversion to different handling
procedures (Rushen, 1986; Pajor et al., 2003) indicated that they could predict which
procedure was likely (feed, hit/shout, isolation, etc.) from environmental or human cues.
Experience-dependent variations in the animals’ perception of the test procedure could
conceivably reduce its general value.
3.2.2. Test conditions
3.2.2.1. The physical and social environment. Statistical constraints, such as the need for
a sufficiently large sample, often demand that animals are taken from a group and tested
individually in an environment that differs substantially from their home area. Even if
tested in the home pen or in a group in the novel environment, the animal’s neighbours or
pen mates could influence its behaviour. Here we identify some potentially confounding
variables.
Firstly, the familiarity/novelty of the test environment can vary markedly across studies.
Sometimes novelty and isolation are central features if one wishes to test the reassuring
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 13
properties of human presence (e.g., Boivin and Braastad, 1996; Boivin et al., 2001), but
their description is often neglected. Novelty is hugely important. For example, calves
reactions to a human previously associated with positive or negative reward varied with the
familiarity of the test pen (De Passille
´
et al., 1996). In an attempt to minimise this potential
confound several researchers use prior habituation (varying from minutes to hours or to
repeated exposure with and without peers over several days) to the test pen (Hemsworth
and Coleman, 1998; Jago et al., 1999; Krohn et al., 2001; Lyons et al., 1988a; Visser et al.,
2001, 2002), but the optimum duration of habituation is unknown and it may even depend
on the animal model, its background genome and the husbandry conditions. Repeated
habituation that includes contact with a handler also bears the risk of confounding the
animal’s test response (see above).
Secondly, farm animals are social species and their reactions at test can be strongly
influenced by social factors such as isolation, disruption and/or the identity of the audience.
Social separation is widely known to be highly distressing per se (Jones, 1996; Boissy and
Le Neindre, 1997). Furthermore, the expression of social reinstatement behaviour during
isolation can compromise the interpretation of chickens’ and other animals’ responses to a
wide range of test stimuli (Jones, 1996; Jones and Mills, 1999). Moreover, the nearby
presence of calm or distressed conspecifics can affect the animals’ responses to humans
(Lyons et al., 1988b; Boissy et al., 1998; Munksgaard et al., 2001).
Thirdly, spatial constraints can vary substantially; animals may be tethered or otherwise
restrained in the home cage or test arena while others may be loose or even on pasture. Such
variability normally reflects the species, husbandry system or the precise objective of
specific tests, e.g., if they are used to assess reactivity to motionless or moving humans or to
actual handling. However, many test environments and procedures are commonly used
without a clear understanding of their effects on behaviour. The nature and magnitude of
the animals’ behavioural and physiological reactions may differ substantially if they are
tested in a situation that either enables or precludes flight from the human, and the presence
or absence of shelter may determine whether flight or immobility behaviours are shown
(Jones, 1996). These issues merit further investigation.
3.2.2.2. The characteristics of the human stimulus: discrimination/generalisation. Re-
searchers have asked if animals generalise from their experience with a known human to
other people. Several studies demonstrated that the response to an unknown human is
influenced by previous treatments based on different types of human contact (Jones, 1996;
Hemsworth and Coleman, 1998; Rushen et al., 1999a; Boivin et al., 2003 for reviews), but
some only exposed the animals to one experimenter and thereby only to his or her specific
characteristics (size, weight, sex, odour, etc.). Although our understanding of farm
animals’ perception of humans and of the cues they use to discriminate between people
(colour of clothing, facial differences, height, etc.) is progressing (Rushen et al., 1999a;
Rybarczyk et al., 2003), the precise nature and influence of the mechanisms underpinning
their ability to generalise from familiar caretakers to an unknown person need further
exploration. Moreover, the behaviour of the human stimulus (passive, active, seated,
standing, looking at the animal) often varies despite reports (Gonyou et al., 1986; Kendrick,
1998; Pajor et al., 2003; Erhard, 2003) that a standing person looking at the animals
induced less approach than a seated one who merely glanced at the animals or sat with his
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx14
or her back to them. Similarly, chickens showed shorter tonic immobility fear reactions if
the experimenter averted his gaze (Gallup et al., 1972; Jones, 1990). We therefore strongly
recommend that the physical appearance and behaviour of the human stimuli should be
reported. We may need to standardise such variables, although differences might be useful
when assessing generalisation of response. Logically, Boivin et al. (1998b) suggested that
discrimination/generalisation of response in beef calves could depend on the collective
impact of all incoming sensations at test, e.g., the quality of the situation (perceived as
positive, neutral or negative) and the physical and behavioural characteristics of the human
stimulus. Discrimination (Y-axis) could be plotted on an inverted U-curve with situational
quality as the X-axis. Perception of the test situation as positive or negative might confound
our assessment of discrimination, but the animal would be expected to show measurable
discrimination when overall sensation fell between these two extremes.
3.2.3. Consequences of variation in test duration or repetition and the application of
multiple tests
Variations in test duration (commonly from 2 to >10 min, see Tables 2–7), the use of
repeated testing and/or the imposition of several tests may all affect an animal’s reactivity
to humans. Furthermore, responsiveness to humans has often been included as just one of a
battery of ‘personality’ tests, such as social motivation or neophobia, without always
balancing the test order (e.g., Romeyer and Bouissou, 1992; Vierin and Bouissou, 2002;
Visser et al., 2001, 2002). In many cases, the use of cross-over or Latin square designs has
often enabled a number of experimental procedures to be carried out on the same animals,
while in others the tests have been repeated at different ages, sometimes before or after a
handling intervention (Boivin et al., 2001; De Passille
´
et al., 1996; Markowitz et al., 1998).
This is not a problem in well-designed experiments. However, unless they are deliberately
built into the experimental question(s) we must consider the possible effects of habituation
(decreased responsiveness to humans or the test situation), sensitisation (increased
reactivity), frustration or reinforcement that may accompany repeated testing. Of course,
exposing the animal to repeated or multiple tests may also change its perception of humans
simply through increased contact with people. Encouragingly, although the level of
responding decreased with repeated testing and age, the consistency of test variables was
high in horses (Visser et al., 2001, 2002).
3.3. Tests used for assessing the HAR
Farm animals frequently encounter familiar and/or unfamiliar humans during their
everyday life; these may be stockpersons, veterinarians, inspectors, catching crews, etc.
The human–animal interactions that take place at these times may be voluntary or
involuntary and can involve visual, auditory, tactile and olfactory stimulation. Imposition
of a painful surgical procedure by a veterinarian represents one (negative) end of a response
scale while a stockperson feeding the animal represents the other extreme, while a
stationary stockperson probably occupies an intermediate position. On this basis, tests
involving various human actions have been developed to measure the animal–human
relationship in numerous species, including farm animals. Tables 2–7 list the main tests
used to assess the HAR in cattle, sheep and goats, pigs, poultry, fur animals and horses,
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 15
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx16
Table 2
Tests of responsiveness to humans and handling in cattle
References Test-type
a
sizes
Time Context
b
Species/
type
c
Procedures and other factors
d
Variables Validity
e
Main confounding
factors/motivations (mot)
f
1 Munksgaard et al. (1997,
1999, 2001), Rushen
et al. (1998, 1999b)
RSH-H 60 s G; U/F; K Dairy
cows
P stands still for 60 s, hands in
pockets; 0.8 m or 0.5 m in front
of bar. Scores at 5 s intervals
Cow’s position scored from
1 (contact with P) to 6
(muzzle behind tie bar and
head turned away from P)
NEG,
POS
Interference by
neighbouring cows.
Exploratory mot
2 Lensink et al. (2000a,b,
2001b,c)
RMH-H,
App MoveHd
I; F/U; K Veal
calves
P approaches from the side 10 s
after A starts to drink or eat,
stands still 5 s, 0.5 m behind bucket,
then touches calf’s forehead
Calfs reactions to appearance/
to touch: none or withdrawal
(5-point score); latencies to
resume drinking or feeding
POS Social mot; feeding mot
3 Lensink et al. (2001b) RMH-H I; U; K Veal
calves
P passes behind the crates and
touches the calf’s hip
Reaction score from 1 (no
movement) to 5 (escape attempt)
REL Startle
4 Waiblinger and
Menke (1999),
Waiblinger et al. (2003b)
RMH-H; App G; U; K Dairy
cows
P approaches A in the feeding rack
from front, 1 step/s, hand held at 458,
until A withdraws
Distance of withdraw (DW),
i.e., between hand and head/nose,
when cow withdraws percentage
of animals with DW of 0
CONV,
REL
Interference by
neighbours;
feeding mot
5 Jago et al. (1999),
Krohn et al. (2003),
Lensink et al. (2000b,
2001c), De Passille
´
et al. (1996)
RSH-H;
0.9 2.2;
2.1 1.85
2–5 min I; F/U; K Calves P stands in front of pen 10 s, P enters
pen and stands still for 2.5 or 10 min.
A are separated in their box if not
housed singly
Latency to approach and
contact human. Frequency
and duration of bouts of
contact. Orientation to human.
Position in pen
POS,
CONV
Exploratory mot
6 Murphey et al. (1981) RSH-H G; U; K Cows P approaches largest concentration of
cows on pasture and lies on ground
Approach/avoidance
responses and behaviour
directed to observer
Exploratory mot
7 Waiblinger and
Menke (1999),
Waiblinger et al.
(2003b), Rousing and
Waiblinger (2004)
RSH-H 15 min G; U; K Dairy
cows
P enters cowshed and stands still
at a central place
Latency to approach;
proportions of standing
animals that approach to
1 m, make contact
ACC,
CONV,
REL
Exploratory mot;
feeding mot
8 Breuer et al. (2003) RSH-H 30 s I; F; K Heifers P enters pen and stands still at its
centre for 30 s. Blood samples taken
via fixed catheters with extension
12 times for 40 to +90 min
Plasma cortisol
concentration
9 Sambraus (1974),
Waiblinger and
Menke (1999),
Waiblinger et al.
(2002, 2003b),
Murphey et al. (1980)
RMH-H; App G; U; K Cows 30 min habituation to P, P slowly,
1 or 2 step/s, approaches standing
animals from front, flank or within
visual field, hand held overhand 458
or hang
Distance of withdraw (DW), i.e.,
between human’s body or hand
and cow’s head/nose herd value:
percentage of animals with DW
of 0; median of DW
CONV,
REL
Available space;
feeding mot
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 17
10 Rousing and
Waiblinger (2004)
RMH-H; App G; U/F; K Dairy
cows
P approaches standing A from front,
1 step/s, arms by side, stops at 1 m,
after 10 s reaches to touch cow
Categories of withdrawal (>2m,
1.5–2, 1–1.5, accepts arm stretched,
accepts touch)
Available space;
feeding mot
11 Boissy and
Bouissou (1988)
RMH-H; App G; ?; K Heifers;
loose
P approaches lying A from front
(6 per animal)
Score from 6 (flight at >2m)
to 1 (remains lying, tolerating
touching) to 0 (approaches human)
POS,
CONV
Lying mot
12 Breuer et al. (2003),
Hemsworth et al.
(2000, 2002)
RMH-T; App I, U/F; N/K Dairy
cows
Follows RSH-T and carried out in
same arena. P walks furthest from A
and then approaches at 1 m/s
Distance of withdraw Carry-over from
RSH-test; isolation
(novelty)
13 Jago et al. (1999),
Krohn et al. (2001, 2003)
RMH-T; App;
2.4 7.4;
2.2 5.5
I; U; K Calves P enters pen, waits till A looks at
him/her, approaches A until it
withdraws (preceded by RSH-T)
Distance of withdraw;
approach/avoidance/baulking
POS Isolation (novelty)
(carry-over from RSH-T)
14 Lensink et al. (2000b) RMH-T; App;
3.7 4.5
3 min I; F/U; K Calves Preceded by 5 min RSH-T. P approaches
A from behind and tries to touch and
stroke its back
Latencies to, durations and
frequencies of touching and stroking
POS Carry-over from
RSH-test; isolation
15 Boivin et al. (1992a) RMH-T; App 2.5 min A; U; N/K Calves Preceded by sorting test. A spends
30 s alone, 30 s with stationary P,
2 min with P following it
Time spent looking at human,
ambulation (squares crossed)
POS Carry-over from sorting
test; social mot
16 Boivin et al.
(1992a, 1998a)
RMH-T; MoveHd;
PRH-T;
10 2
15, 1.5 min I; F; K/N Calves A isolated in a (small) pen. P tries to
stroke the animal (offers concentrate),
in 1998a combined with RSH
directly before
Times spent accepting stroking,
standing still, vocalising,
sniffing pen, lying down,
playing with human, within
1, 2, 4, 6, 8 m of and touching
handler. Escape attempts
POS Isolation (carry-over
from RSH-test)
17 Boivin et al. (1998b) RMH-T; MoveHd 3 min A; U/F; K P enters pen and stands still 10 cm from
bucket, A released in pen. If A feeds for
10 s P tries successively to touch
shoulder, head, nostril, offers food
for 10 s each
Latency to feed from bucket and
to accept touching on different
body parts
POS Social mot;
feeding mot
18 Boivin et al. (1998a) RSH-T; PRH-T;
10 2
1.5 min I; F; K Calves A left alone in arena for 1 min, then
P enters the pen, stands motionless
for 1.5 min
Times spent standing still, vocalising,
sniffing pen, within 1, 2, 4, 6, 8 m
of and touching handler. No. of
escape attempts
POS Carry-over from
RMH-test, exploratory
mot, isolation
19 Boissy and
Bouissou (1988)
RSH-T;
10 m
2
1–5 min I/A; U/F; K Heifers,
calves
A placed alone in a holding pen for
30 s. P then stands near the feeder
Latency to feed. Times spent
feeding, >1 m from bucket, orienting
to and interacting with human
POS, CONV Feeding, social,
exploratory mot,
isolation
20 Jago et al. (1999),
Krohn et al. (2001, 2003)
RSH-T;
2.4 7.4
A; U; K Calves A left alone for 90 s, P enters pen
and stands still opposite to audience
Latencies to approach and
to touch human. Time spent
<1 m from human
Social mot, exploratory
mot
21 Jago et al. (1999),
Krohn et al. (2001, 2003),
Lensink et al. (2000b)
RSH-T;
2.4 7.4; 2.2
5.5; 3.7 4.5
3.5 min I; F/U; N/K Calves After 24 h familiarisation with test
arena P enters and stands still or (b)
calf released into arena where P is
standing (combined with RMH of 3 min)
Latency to contact, duration
and frequency of contact with
human. Time spent < 1 m from
human. No. of escape attempts,
squares crossed, defecations
In K: POS,
CONV,
DISC
Isolation, exploratory
mot (novelty)
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx18
Table 2 (Continued )
References Test-type
a
sizes
Time Context
b
Species/
type
c
Procedures and other factors
d
Variables Validity
e
Main confounding
factors/motivations (mot)
f
22 Breuer et al. (2000, 2003),
Hemsworth et al. (1987b,
1989b, 1996a, 2000, 2002),
Tilbrook et al. (1989)
RSH-T 2 + 3.5 min I; U/F; N Dairy
cows
A left alone in arena for 2 min,
P enters the pen and sits on a stool
Latencies to approach
and to touch human. Time
spent within 1–3 m of human.
Frequency of physical contact.
Percentage of animals within
1.3 m of human
CONV Isolation, novelty,
exploratory mot
23 Becker and Lobato (1997),
De Passille
´
et al. (1996)
RSH-T (testing
discrimination)
90 s, 5 min I; F/U; N/K Calves A left alone for 30 s/5 min, two
P (U/F or positive/aversive handler)
enter pen, sit down in centre or stand
at either side of pen
Time spent moving, looking at
experimenter. No. of escape
attempts, aggressive actions.
Latency and frequency to interact
POS Isolation, exploratory
mot
24 Rybarczyk et al. (2001) RSH-T (testing
discrimination)
A; U/F; K Cows trained at operant conditioning
apparatus (rewarder against empty
chamber). Tested with rewarder against
unfamiliar P
Correct choices
25 Lensink et al. (2000c) RHd-T I; F + U; N/K Calves Calves are loaded individually onto a
cart and transported for 2 min
Time needed for loading.
Defecation; score of struggling
during transport
POS,
CONV,
DISC
Isolation, novelty
26 Lensink et al. (2001b,c) RHd-T I/G; U; N Calves P moves calves individually to truck,
loads, transports and unloads
them. P restricted to special
behaviour (e.g., pushing and vocal
command)
Effort required to load calves. No.
of turns, buck kicking, running
per m. Latencies to time get calf
out of crate, to move it to truck
and to load it. Number of
potentially traumatic incidents.
Plasma cortisol. Heart rate
POS,
CONV
Isolation, novelty
27 Breuer et al. (2003),
Tilbrook et al. (1989)
RHd-T;
48 m
I; U/F; N/K Heifers A moved individually along a route
to a crush or from home pen to test
arena; (after RMH; RSH-tests)
Latency to reach crush. No. and
time of interactions used by
experimenter. No. of animals
baulking. Distance from human
maintained by animal
Carry-over effects from
RMH and RMS tests,
isolation
28 Breuer et al. (2003) RHd-T;
2.8 0.8
I; U; K? Heifers P who had moved the A in the crush
stands 0.5 m besides the head and
scores the A’s restlessness
Score reactions from 0
(quiet, no movement) to 3
(vigorous movement)
Carry-over effects
of multiple tests,
isolation
29 Boissy and
Bouissou (1988)
RHd-T G; ?; K Heifers
15 mth
P catches A within a group and
places a halter on it
Time to capture and put a
halter on the animal
POS,
CONV
Social mot
30 Lewis and Hurnik (1998) RHd-H G; ?; K Dairy
cows
P places a halter on the cow while
in the tie stall
Score from 1 (holds head
still) to 5 (aggressive)
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 19
31 Boissy and Bouissou
(1988), Lewis and
Hurnik (1998)
RHd-T;
20 m/76 m
I/G; ?; K/N Heifers,
dairy
cows
A are taken out of the pen or stall
after the halter was placed and led
through a corridor (past herd members
in the second paper)
Ease of leading; relative time of
walking voluntarily, walking when
coaxed, running and refusing
to walk; or score from 1 (none to
mild hesitation) to 5 (escape or
aggression)
POS,
CONV
Carry-over from test of
placing halter, isolation,
social mot
32 Lewis and Hurnik (1998) RHd-T G; ?; N Dairy
cows
A are scored three times:
(1) being moved into the squeeze,
(2) head gate closed, (3) being backed
out of squeeze
Score from 1 (none to mild
hesitation) to 5 (escape or
aggression) for each of the
three situations
Carry-over from test of
placing halter and leading,
social mot
33 Boivin et al.
(1992a,b, 1994)
RHd-T;
100 m
2
Maximum
3min
G; U; N/K Beef
cattle
A group of around 10 A is placed in
pen. P separates each animal (moving
it out of the pen) in pre-determined
order
Time needed to separate the
animal from the group
ACC,
POS,
DISC
Social mot
34 Boivin et al. (1992a,b,
1994), Le Neindre et al.
(1995), Grignard et al.
(2000, 2001)
RHd-T;
55 m
2
,6 6, 5
5or3.5 5
2.5–3.5 min A; U; N/K Beef
cattle,
calves,
heifers
A successively exposed to 30 s alone in
arena, 30 s with passive P, P tries to
move it to a 2 2 m corner opposite
other A and keep it there for 30 s; then
tries to touch it
Latency to restrain animal in
corner. No. of aggressive
animals and of escape attempts.
Time spent motionless, running,
orienting to human, accepting touch.
Aggregate docility score
ACC, POS;
CONV;
DISC
Social mot; novelty
35 Boivin et al. (1998b) RHd-T I; U/F; N Calves P leads calf to a weighing scale, leaves
it alone for 30 s, then strokes it for 30 s
Time needed to lead onto scale Isolation; novelty
36 Rushen et al. (1999b),
Munksgaard et al. (2001)
RHd-H G; U/F; K Dairy
cows
Cow milked with or without a familiar/
unfamiliar or aversive/non-aversive
P standing nearby
Steps, kicks, tail movement,
defecations, urinations. Heart rate.
Milk yield; milking duration,
residual milk
Partly NEG,
REL
37 Hemsworth et al. (1987b,
1989b, 2002), Knierim
and Waran (1993),
Breuer et al. (2000),
Waiblinger et al. (2002),
Seabrook (1984)
RHd-H G; U/F; K Dairy
cows,
heifers
Observations during regular milkings,
e.g., with relief or regular milkers.
On farm surveys
Flinch, step, kick, tail flick.
Dislodged clusters, assistance
needed by milker. Cortisol in milk,
heart rate; milk yield. Latency
to enter parlour
POS,
CONV,
ACC,
REL
Previous experience of
unfamiliar people
in parlour
38 Rushen et al. (2001) PRH-T I; F; N Dairy
cows
Milking in isolation in a novel room
with/without a P brushing the cow
Steps, kicks, tail movements,
defecations, urinations,
vocalisations. Heart rate; plasma
cortisol and oxytocin. Milk yield;
milking duration, residual milk
CONV Novelty
39 Waiblinger et al. (2004) PRH-H 4 + 5 min G; F/U; K Dairy
cows
Rectal palpation with sham
insemination with/without a P
stroking the cows
Steps, kicks, tail movements,
butts. Licking and leaning at
the person, stretching the neck.
Heart rate
POS,
CONV,
REL
40 Tulloh (1961),
Fordyce et al. (1985)
RHd-T 1 min I?; ?; K? Steers P touches the animal in the crush Temperament score (vigour
of movement, audible respiration,
bellowing, kicking, kneeling,
going down) docile—aggressive
Isolation?
Social motivation
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx20
Table 2 (Continued )
References Test-type
a
sizes
Time Context
b
Species/
type
c
Procedures and other factors
d
Variables Validity
e
Main confounding
factors/motivations (mot)
f
41 Grignard et al. (2001) RHd-T 8 min I; U; N Beef
heifers
A 5 min alone in crush, 30 s
P motionless 1 m in front of animal,
30 s P strokes the animal’s head
Time spent standing still,
moving leg, tail or head. No.
of eliminations, vocalisation,
sniffs and licks at human.
Heart rate
CONV Isolation
42 Grandin (1993),
Grandin et al. (1995)
RHd-T Cattle Observation of A restrained in a
crush for vaccination, ear tagging,
blood sampling, etc.
4- or 5-point score
(from calm, no movement
to violent movement and
vocalisation)
Social motivation
a
Test-type: PRH = test for positive response to human. RMH = test for reactions to moving human. App: human is approaching; Ret: human is retreating; MoveHd: human stands still, but moves the hand. RSH = test for
reactions to stationary human. RHd = reaction to handling (T = in test environment, H = in home environment).
b
Context: social conditions during the test (I = social isolation, G = group, A = audience); experimenter characteristics (F = familiar; U = unfamiliar); familiarity of environment (N = novel, K = known/familiar—at least one
test performed in same test environment before or home).
c
Type = type of production or animal age class, e.g., dairy; beef; cow, heifer.
d
P = person; A = animal; procedures in brackets used only in some references.
e
Codes for validity: ACC = repeatable between observers without systematic error; CONV = convergent validity: relation in expected direction of behaviour with cortisol; heart rate; production, to other tests or correlations
with human behaviour in on-farm studies; DISC = discriminant validity: no relation to other fear reactions, e.g., reaction to a novel object, activity in open field when alone; NEG, POS, NEU = sensitive to ‘negative’ (hitting...),
‘positive’ (stroking...) or ‘neutral’ (habituation) treatment; NEG POS = discriminates between ‘negative’ and ‘positive treatment’; REL = external validity: test is usable on farm.
f
Excluding personality traits.
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 21
Table 3
Tests of responsiveness to humans and handling in sheep and goats
References Test-type
a
sizes
Time Context
b
Species/
age
Procedures and other factors
c
Variables Validity
d
Main confounding
factors/motivations (mot)
e
1 Boivin et al. (1997,
2000, 2001, 2002)
PRH-T;
6 2m
1.2 + 2 +
1.2 min
I; F; N, K Sheep;
lambs
A left alone in arena for 1, 2 min. P enters
and sits for 2 min, calls, stretches the arm
and touches the A if it approaches within
1 m, P leaves and A left alone for 1, 2 min
Latency of contact with the human,
duration in contact (<1 m). Number of
vocalisations of sections entered
POS Isolation; novelty
2 Markowitz
et al. (1998)
PRH-T;
5 1m
5 min I; U; N, K Sheep;
lambs
A put in a box 30 s before the test, A
then released in arena where P sits still.
After the first contact, P presents the hands
allowing the A to reach the fingers
Mean distance from the person, latency
of contact, time spent in proximity (<2m),
in contact with the human (<1 m), number
of sections entered, number of human contact
POS,
DISC,
CONV
Isolation; novelty
3 Le Neindre
et al. (1993)
RSH-T;
4 6m
4 min I; U; N, K Sheep;
lambs;
adult
A released in arena with a P standing
motionless. Included in a battery of tests
in the same arena
Latency, number of sniffs at P.
Duration in proximity of P. Number
of sections entered, sniffs, vocalisations,
defecations, urinations, rearings
Isolation; carry-over
effects from other tests
4 Lyons et al. (1988a) RSH-T;
1 8m
10 min I; ?; K Goat;
lambs,
adult
2-Day period of familiarisation to the
arena in group before the test. A
restraint in a starting zone for 45 s and
release in the arena with a P standing
Latency of proximity with the human,
duration in proximity (within 2 m),
sections crossed, mean distance from
the humans
POS,
CONV,
ACC,
DISC
Isolation
5 Mateo et al. (1991) RMH-T;
movehd
5 min G; F; N, K Sheep;
lambs
Three lambs (from different
treatments) placed in arena with
P sitting in the middle with hand
outstretched, touching A if they
approached
Latency to sniff human’s hand,
number of contacts
POS Novelty
6
Lyons and
Price (1987)
RSH-T;
1 1m
5 min A; U; N Goat;
lambs
A placed 5 min in the arena, peers
behind a fence, then P enters and stays
still for 5 min. Heart rate recorded
by telemetry
Duration in contact with the human. Number
of vocalisation. Heart rate
POS,
ACC
Isolation; social mot
7 Boivin et al. (2002) RMH-H
movehd
2 min A; F; K Sheep;
lambs
P approaches out of the home pen and
stretches his hand towards the animals
Latency of contact with the human, duration
in proximity close to the human, vocalisation,
number of escapes
POS
8 Romeyer and
Bouissou (1992),
Vandenheede and
Bouissou (1993),
Vierin and
Bouissou (2002)
RSH-T;
4 4,
10 10
4 min I; A; U; K Sheep;
lambs,
juvenile
2–10-Day period of habituation to
the arena with a food trough. A entered
the test arena with the P standing or
sitting still behind the trough
Latency to enter the section in front of the
trough, feeding latency, feeding duration,
number of sections entered. Synthetic
score computed from the variables
recorded during the test
POS,
CONV
Feeding mot.
Test included in
a battery of tests
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx22
Table 3 (Continued )
References Test-type
a
sizes
Time Context
b
Species/
age
Procedures and other factors
c
Variables Validity
d
Main confounding
factors/motivations (mot)
e
9 Lankin (1997),
Lankin and
Bouissou (2001)
RMH-T;
movehd
3–5 min G, I; U; N Sheep A placed in A arena after 12 h food
deprivation. P enters, fills the feeder,
lets A approach and eat for 1 min.
Then P attempts to mark the sheep
on their back at three times
No. of paints placed on sheep’s body (back) REL Feeding mot
10 Le Neindre
et al. (1993)
RSH-T;
4 6
4 min A; U; N, K Sheep;
adult
A placed in arena with P standing
in front of peers behind a fence
Latency and number of sniffs at human.
Duration in proximity with the human
(1 2 m section). Number of section entered,
sniffs, vocalisations, defecations, urinations,
rearing against the walls, looks towards the human
Social mot. Order of
testing; test included
in a battery of tests
11 Fell and
Shutt (1989)
RSH-T;
15 4m
10 min G; F; N, K Sheep;
lambs
Three or four A placed in arena
with P standing in front of peers
behind a fence
Mean, minimum and maximum distance
from the human (distance observed 4/min)
NEG Social mot; novelty
12 Goddard
et al. (2000)
RSH-T;
RMH-T;
4.5 4.5
5 min G; U; N Sheep;
lambs,
adult
Four A left alone in arena for
10 min, P enters and stands still
for 5 min, then P walks around
at constant speed for 5 min. Heart
rate recorded with Polar Sports
Tester
TM
Latency to move from the original position,
duration spent facing the human, number
of sections entered. Heart rate and total
plasma cortisol with plasma samples
taken just after the test
NEU Novelty
13 Boivin and
Braastad (1996)
RMH-T;
Movehd; App
I; F; N, K Goats;
lambs
A left alone in arena for 1 min,
P enters and stands still for
1.5 min, then P approaches and
tries to pet A for 1.5
Duration in proximity (<2 m), in contact
with the human. Vocalisation, sections crossed,
POS Isolation; novelty
14 Lyons et al. (1988b),
Lyons (1989)
RSH-H;
RMH-H App;
12.2 9.8;
1 8
3 3 min;
3 2min
I, G; U; K Goats;
adult
Three successive parts of 2,
3 min: P enters pen and stands
still, P moves back and forth
along the front fence, P tries
to touch A
Latencies to approach the human (<1m)
and to make contact, duration in proximity
(stationary or moving)
POS;
CONV
Isolation
15 Lyons et al. (1988a),
Markowitz
et al. (1998)
RMH-T App
(circular
corridor,
6.1 m radius)
3.5 min I; U; K Goats,
sheep;
lambs
A placed in a circular runway,
P walks (0.5 steps/s) behind it
for 3.5 min; blood sampling
taken 3 days before the test,
immediately after and
3 days after
Mean flight distance, following,
approach, avoidance, vocalisation,
human contact, urination (plasma cortisol)
POS,
CONV,
ACC,
DISC
Isolation
16 Hutson (1982) RMH-T; App I, G; ?; N, K Sheep;
adult
A placed in test corridor. P
approaches the (group of)
A at a constant speed
Flight distance, head orientation
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 23
17 Hargreaves and
Hutson (1990)
RMH-T App;
18 1.5 m
Cage:1.7 0.5
I; ?; N, K Sheep;
adult
wethers
A placed in corridor or
restraint in a cage at the
end of the corridor. P
approaches A at a constant
speed. Heart rate telemetry
Flight distance and heart rate NEU,
POS,
REL
Isolation; novelty; restraint
18 Mateo et al. (1991) RSH-T; App 2 min I; ?; K Sheep;
lambs
A put in a halter, lead to
arena and tethered to a post.
The P sat quietly next to the
Afor2min
Time spent pulling on the halter,
vocalisations
POS Isolation; restraint
19 Lyons (1989) RHd-H 21 days G; F; K Goats;
adult
A milked twice daily for 21
days by two P. Then, the P
scored each goats behaviour
Seven behavioural scales: excitable, tense,
watchful, apprehensive, confident, friendly
to humans, fearful of humans. Milk ejection
POS,
CONV,
REL
20 O’Connor et al.
(1985), Le Neindre
et al. (1998)
RHd-H G; ?; K Sheep Within 24 h after parturition
on pasture, the shepherd
approached the ewes and
tagged the lambs. Responses
of the ewe were scored
Maternal behaviour score:
5-point scale from 1 (flees at the
approach of the shepherd, no return
to the lamb’s) to 5 (stays close to the
shepherd during handling of their lambs)
REL Maternal mot
a
Test-type: PRH = test for positive response to human. RMH = test for reactions to moving human. App: human is approaching; Ret: human is retreating; MoveHd: human stands still, but moves the hand. RSH = test for
reactions to stationary human. RHd = reaction to handling (T = in test environment, H = in home environment).
b
Context: social conditions during the test (I = social isolation, G = group, A = audience); experimenter characteristics (F = familiar; U = unfamiliar); familiarity of environment (N = novel, K = known/familiar—at least one
test performed in same test environment before or home).
c
P = person; A = animal.
d
Codes for validity: ACC = repeatable between observers without systematic error; CONV = convergent validity: relation in expected direction of behaviour with cortisol; heart rate; production, to other tests or correlations
with human behaviour in on-farm studies; DISC = discriminant validity: no relation to other fear reactions, e.g., reaction to a novel object, activity in open field when alone; NEG, POS, NEU = sensitive to ‘negative’ (hitting...),
‘positive’ (stroking...) or ‘neutral’ (habituation) treatment; NEG POS = discriminates between ‘negative’ and ‘positive treatment’; REL = external validity: test is usable on farm.
e
Excluding personality traits.
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx24
Table 4
Tests of responsiveness to humans and handling in pigs
References Test-type
a
sizes
Time Context
b
Species/type
c
Procedures and other factors
d
Variables Validity
e
Main confounding
factors/mot.s (mot)
f
1 Hemsworth et al. (1981a,
1986, 1994b, 1996a,b),
Hemsworth and
Barnett (1992)
RSH-T 2 + 3/5 min I; F/U; N Gilts;
boars; piglets
A left alone in test pen for
2 min, P enters and stands
still for 3 min
Latencies to approach
within 0.5 m and to touch
experimenter. Time spent
near human. No. of
physical interactions
NEG, NEU,
POS, CONV
Isolation; novelty;
exploratory mot
2 Hemsworth et al. (1994b) RHd-T I; F/U; N Gilts P moves pigs individually
along a standard route (100 m)
using a board and positive
interactions. Negative interactions
used if pig baulks and remains
stationary for >5s
Time to move along a
standard route; no. of
baulks, negative
interactions by handler;
score from 0 (very
difficult) to 4 (easy)
for ease of movement
NEG, NEU,
POS, CONV
Isolation; novelty
3 Gonyou et al. (1986),
Hemsworth et al. (1987a),
Paterson and Pearce (1992)
RSH-T 2 + 3 min G/I; U; N Gilts;
young males
A left alone in test pen for
2 min, P enters and stands
still for 3 min. Group testing
preceded individual testing
Latencies to approach
to 0.5 m and to interact
with human. Time spent
near human. Physical
interactions. Plasma cortisol
NEG, NEU,
POS, CONV
Isolation; novelty;
exploratory mot,
social mot
4 Gonyou et al. (1986) RMH-H - G; F; K Breeding sows P enters the pen, walks towards
the pigs, squats and pets an
approaching pig
Measures as for 3.
Ranking of the
different treatments
NEG
POS, CONV
5 Hemsworth et al.
(1989a, 1990)
RSH-T 2 + 3 min I; U; N Breeding sows A left alone in test pen for
2 min, P enters and stands
still for 3 min
Locomotion during the
familiarisation period.
Latencies to approach
within 0.5 m and to touch
human. Time spent near
human. Physical interactions
CONV,
DISC, REL
Isolation;
novelty;
exploratory mot
6 Hemsworth et al. (1999) RMH-H
MoveHd
15 s I; U; K Lactating sows P1 slaps sows in farrowing
crate to make her rise. P2
places a food tray in front
of crate and withdraws. After
pig has fed for 5 s, P2
approaches front of crate and
places hand 5 cm from sow’s
snout
Withdrawal response;
reaction time; time
within 5 cm of food tray
CONV, REL Feeding mot
7 Tanida et al. (1995)
RSH-T 1 min I; U/F; N Weanling pigs A released into arena
where P is sitting
Latency to touch
experimenter
NEU, POS,
CONV
Isolation;
novelty;
exploratory mot
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 25
Table 4 (Continued )
References Test-type
a
sizes
Time Context
b
Species/type
c
Procedures and other factors
d
Variables Validity
e
Main confounding
factors/mot.s (mot)
f
8 Tanida et al. (1995) RHd-T 3 min I; U/F; N Weanling pigs A released into arena
containing P; P lifts A’s
hind legs briefly, then sits
still till the A makes contact,
then P walks after pig
Pig’s response to lifting scored
subjectively (0– 3; low to high
struggling). Latency to approach
NEU, POS, CONV Isolation
9 Tanida et al. (1995) RMH-T Not
specified
I; U; N Weanling pigs P stands still in test room;
A released in waiting box,
enters the test room voluntarily;
when A touches P, he/she starts
to walk along the grids
Average distance from the
experimenter to the pig
NEU, POS, CONV Isolation; novelty
10 Wemelsfelder and
Lawrence (2001),
Wemelsfelder et al.
(2000, 2001)
RSH-T 7 min I; F; K Growing
female pigs
Individual pigs placed in
test room containing a
human squatting in the centre
Observers’ own descriptive
(subjective) terminology
NEG, NEU, POS,
ACC, CONV
Isolation;
exploratory mot
11 Marchant et al. (2001),
Marchant-Forde (2002),
Marchant-Forde
et al. (2003)
RSH-T;
RMH-T
2 + 3 min I; U; N Gilts; 6 months
in groups
A fitted with heart rate
monitor, released to arena
where P stands motionless.
3 min observation after 2 min
familiarisation. P then approaches
gilt to touch her snout
Locomotor behaviour.
Latencies to approach to 0.5 m
and to touch the human. Time
in physical contact. No. of
contacts and of short and
long vocalisations. Heart rate
NEU, CONV, REL Isolation; novelty;
exploratory mot
12 Janczak et al. (2003) RSH-H 1 + 3 min I; F/U; K Gilts and
sows; 8 and 24
weeks in groups
P enters the home box, places
a plywood to separate test
animal from littermates, exits.
A left alone for 1 min.
P re-enters, walks to the
wall opposite the entrance,
stays still
Durations and frequencies of
exploring the P, standing,
walking, exploring the room
CONV Isolation exploratory
mot, social mot
a
Test-type: PRH = test for positive response to human. RMH = test for reactions to moving human. App: human is approaching; Ret: human is retreating; MoveHd: human stands still, but moves the hand. RSH = test for
reactions to stationary human. RHd = reaction to handling (T = in test environment, H = in home environment).
b
Context: social conditions during the test (I = social isolation, G = group, A = audience); experimenter characteristics (F = familiar; U = unfamiliar); familiarity of environment (N = novel, K = known/familiar—at least
one test performed in same test environment before or home).
c
Type = type of production or age class, e.g., gilts, breeding sows.
d
P = person; A = animal.
e
Codes for validity: ACC = repeatable between observers without systematic error; CONV = convergent validity: relation in expected direction of behaviour with cortisol; heart rate; production, to other tests or correlations
with human behaviour in on-farm studies; DISC = discriminant validity: no relation to other fear reactions, e.g., reaction to a novel object, activity in open field when alone; NEG, POS, NEU = sensitive to ‘negative’ (hitting...),
‘positive’ (stroking...) or ‘neutral’ (habituation) treatment; NEG POS = discriminates between ‘negative’ and ‘positive treatment’; REL = external validity: test is usable on farm.
f
Excluding personality traits.
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx26
Table 5
Tests of responsiveness to humans and handling in poultry
References Test-type
a
sizes Time Context
b
Species/
type
c
age
Procedures and other factors
d
Variables Validity
e
Main confounding factors/
motivations (mot)
f
1 Barnett et al. (1993),
Hagedorn et al. (1996),
Jones (1996)
RSH-H 2–3 min I/G; F/U; K Chickens
(layers); quail
P disturbs food in trough at
front of home cage to alert
bird(s), stand in front of cage,
measure reactions at 10 s intervals
Position in cage (front,
mid, rear); orientation
(0–4 for head out of cage,
face front, face side, face
rear, escape); cumulative scores
NEU, POS,
CONV,
DISC, REL
Neophobia to U human;
interference by cagemates
2 Jones (1985, 1987b, 1996),
Keer-Keer et al. (1996)
RSH-T 2–3 min I; F/U; N Chickens; adult A placed in arena containing
P seated or standing in centre.
Measure position every 10 s
Position in one of four areas
of pen at increasing distances
from centrally located human
(1 = near, 4 = far);
cumulative scores
NEU, POS,
CONV, DISC,
Isolation; novelty
3 Jones (1993, 1995a,b),
Jones and Waddington (1993)
RSH-T 5 min I; F/U; N Chickens A placed in rectangular arena
with P seated in front of wire-
mesh end wall. Measure position
every 10 s. Other measures
continuous
Position in one of four
areas at increasing distances
from human seated at front
of cage (1 = near, 4 = far);
cumulative score. Freezing,
vocalisation, ambulation,
pecking
NEU, POS,
CONV, DISC
Isolation; novelty
4 Barnett and Hemsworth (1989),
Hemsworth et al. (1993b),
Jones (1985)
RMH-H;
App
15 s–2 min G; F/U; K Chickens (layers);
quail; juvenile
+ adult
P approaches to 1, 0.5, 0 m
from cage. Measure behaviour
of focal bird at each point or
numbers of birds with head
out of cage
Small groups—behaviour
on scale of 0–5 (head out
of cage, face front, face
side, face rear, escape);
cumulative score. Larger
groups—numbers of birds
with heads out of cage
NEU, CONV,
DISC, REL
Neophobia to U
5 Barnett et al. (1992),
Hemsworth and
Coleman (1998)
RMH-T;
App
1 min I; F/U; N Chickens;
generic;
juvenile + adult
A placed on table at one end
of a corridor. P approaches
from 2.4 m. Measures taken
at 2.4, 1.8, 0.8 and 0 m
Proportions of birds
withdrawing, turning
away at selected distances
CONV,
DISC, REL
Isolation;
novelty;
neophobia to U
6 Jones et al. (1993),
Hemsworth and
Coleman (1998)
RMH-H;
App
Variable G; U; K Chickens
(broilers);
juvenile + adult
P with VCR on shoulder walks
through poultry shed, stops for
3 0 s every 20 paces, videotapes
analysed
Numbers of birds in
75 cm semi-circle in
front of human; cumulative
score of individual scans
CONV, REL Neophobia to U
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx 27
7 Bessei et al. (1983),
Carmichael et al. (1999),
Satterlee and Jones (1997)
RHd-H Variable G; F/U; K Quail; generic
for small/
young birds
Unsighted P catches one
bird at a time from caged
group, transfers it to other
cage, continues till all are
caught. Capture/recapture
procedure repeated up to
20 times
Rank order in which birds are
individually captured/recaptured
from established group
ACC,
CONV, REL
Ambulatory
difficulties;
space restriction
8 Jones et al. (1981) RMH-T; App;
RHd
Variable I; F/U; K Chickens
(layers); adult
Telemetry device implanted.
Bird acclimatised (3–4 days)
to new cage at end of corridor
+ food delivery to ensure
forward orientation. P
approaches slowly from
28 m, ultimately opens
cage and captures bird
Distances at which orient,
withdraw, startle, alarm call,
escape and panic responses are
first shown. Heart rate
throughout P’s approach
CONV, DISC Isolation; neophobia to U
9 Gallup (1979), Jones (1986),
Jones et al. (1991, 1992b)
RHd-T 5–20 min I; F/U; N Chickens;
quail; any
age over 7 days
Birds restrained by hand
laterally on table or
ventrally in cradle for 15 s,
one hand cupping head + one
on sternum. Usually performed
in unfamiliar room
Number of inductions required
to obtain tonic immobility (TI)
lasting 10 s. Duration of TI (i.e.,
till self-righting)
NEG, NEU, POS,
CONV, ACC,
DISC, REL
Isolation; novelty;
neophobia to U
10 Webb and Mashaly (1984),
Jones et al. (1994),
Korte et al. (1997)
RHd-T 5–10 min I; F/U; K Chickens;
quail; any age
Bird removed from home
environment and manually
restrained for 5–10 min
before blood withdrawal
for assay of corticosterone.
Controls bled during
resting conditions
Plasma corticosterone
concentrations
POS, CONV,
DISC, REL
Isolation; novelty;
neophobia to U; different
handling and bleeding skills
a
Test-type: PRH = test for positive response to human. RMH = test for reactions to moving human. App: human is approaching; Ret: human is retreating; MoveHd: human stands still, but moves the hand. RSH = test for
reactions to stationary human. RHd = reaction to handling (T = in test environment, H = in home environment).
b
Context: social conditions during the test (I = social isolation, G = group, A = audience); experimenter characteristics (F = familiar; U = unfamiliar); familiarity of environment (N = novel, K = known/familiar—at least one
test performed in same test environment before or home).
c
Type = type of production, e.g., layers, broilers.
d
P = person; A = animal.
e
Codes for validity: ACC = repeatable between observers without systematic error; CONV = convergent validity: relation in expected direction of behaviour with cortisol; heart rate; production, to other tests or correlations
with human behaviour in on-farm studies; DISC = discriminant validity: no relation to other fear reactions, e.g., reaction to a novel object, activity in open field when alone; NEG, POS, NEU = sensitive to ‘negative’ (hitting...),
‘positive’ (stroking...) or ‘neutral’ (habituation) treatment; NEG POS = discriminates between ‘negative’ and ‘positive treatment’; REL = external validity: test is usable on farm.
f
Excluding personality traits.
S. Waiblinger et al. / Applied Animal Behaviour Science xxx (2006) xxx–xxx28
Table 6
Tests of responsiveness to humans and handling in foxes and mink
References Test-type
a
sizes
Time Context
b
Species
gender age
Procedures and other factors
c
Variables Validity
d
Main confounding
factors/motivations (mot)
e
1 Pedersen and Jeppesen
(1990), Pedersen (1992,
1993a,b, 1994),
Pedersen et al. (2002),
Korhonen and Niemela
(1996)
RSH-H 15 s I; U; K Foxes;
juvenile/adult
P approaches cage, attracts animal’s
attention by waving hand, stays or
retreats 1 m and stand still.
Predominant behavioural
reaction after 15 s is recorded
Body movement and posture, facial
expressions, ear positions, vocalisation
ACC, NEG, NEU,
POS, CONV, REL
Exploratory mot
2 Bakken et al. (1999),
Moe (1996)
RSH-H 5 min I; F; K Silver f