UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The behavioural expression of fear in young children Gilbert-MacLeod, Cheryl A. 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-565505.pdf [ 5MB ]
Metadata
JSON: 831-1.0089750.json
JSON-LD: 831-1.0089750-ld.json
RDF/XML (Pretty): 831-1.0089750-rdf.xml
RDF/JSON: 831-1.0089750-rdf.json
Turtle: 831-1.0089750-turtle.txt
N-Triples: 831-1.0089750-rdf-ntriples.txt
Original Record: 831-1.0089750-source.json
Full Text
831-1.0089750-fulltext.txt
Citation
831-1.0089750.ris

Full Text

THE B E H A V I O U R A L EXPRESSION OF F E A R IN Y O U N G C H I L D R E N by CHERYL A. GILBERT-MACLEOD B . S c , Dalhousie University, 1993 M . A . , The University of British Columbia, 1996  A THESIS SUBMITTED IN P A R T I A L F U L F I L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY In T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Psychology) We accept tfcris-thesis as conforming to the recjuired standard  THE U2&VERSITY OF BRITISH C O L U M B I A September 2000 © Cheryl A . Gilbert-MacLeod, 2000.  In  presenting  degree at the  this  thesis  in  University of  partial  fulfilment  of  of  department  this thesis for or  by  his  or  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  representatives.  an advanced  Library shall make  It  is  granted  by the  understood  that  publication of this thesis for financial gain shall not be allowed without permission.  The University of British Columbia Vancouver, Canada  DE-6 (2/88)  it  agree that permission for extensive  scholarly purposes may be her  for  head of copying  my or  my written  11 Abstract Children, over the course of development, experience numerous situations capable of eliciting fear; however, the behaviours which children exhibit in these situations remain unclear. The investigation presented here pursued the question "how do young children express fear in a non-painful medical situation where they perceive threat from physical harm?". It is important to note that this study differentiated between fear and anxiety, however it did not examine differences between these two emotions. 116 children, between the ages of 12 and 87 months, and their parents participated in the study. Children's fine-grained behavioural responses, (i.e., facial activity) and broader behavioural displays (e.g., crying, protective behaviours) were examined during a fearful situation. The specific threat used to provoke fear was orthopedic cast removal with an oscillating saw. Few people, including adults, who have had a cast removed would challenge the notion that the oscillating saw can effectively elicit fear. Facial activity was measured with the Baby-FACS coding system and global behaviours were assessed with the Observational Scale of Behavioral Distress. Results demonstrated the existence of a constellation of facial actions and a group of more global behaviours indicative of fear in young children. The facial actions and global behaviours identified in the total sample were examined on a subset of the children who were rated as displaying clinically significant fear. The same 13 facial actions were found to cluster together in the sample of children displaying clinically significant fear. Further, global behaviours occurred with a higher frequency in this sub-sample. Age and cast location were found to predict children's fear for both classes of behaviour in that younger children and children with  iii casts on their legs had higher facial action factor scores and OSBD scale scores than older children or those with arm casts. Finally, facial activity and global behaviours appeared to be valid measures of fear as they were both correlated to an independent observer's and the cast technician's ratings of fear. Results are discussed in relation to current theories of emotional development and implications for clinical applications are reviewed.  iv Table of Contents Abstract  ii  List of Tables  vi  List of Figures  vii  Acknowledgements  viii  Introduction  1  Literature Review  4  Psychology of Emotional Expression  4  Characterizing Emotional Expressions  15  Facial Expressions of Emotion  16  Broadband Behavioural Expressions of Emotion  27  Important Factors in the Expression of Emotions  29  Methodological Considerations  33  Overview of the Study  39  Hypotheses  40  Method  41  Results  55  Discussion  81  References  100  Appendix A Observational Scale of Behavioural Distress  116  Appendix B Demographic Questionnaire  119  Appendix C Past Medical History Questionnaire  122  Appendix D Chart Review  124  Appendix E Baby Facial Action Coding System Appendix F Regression Analyses on the Clinically Significant Fear Sample  vi List of Tables Table 1. Child Demographic Information.  42  Table 2. Parent Demographic Information.  43  Table 3. Means and Standard Deviations for Facial Action Fear Segments  ....  Table 4. Eigenvalues and Percent of Variance Accounted for by the Factors . . .  57 59  Extracted in the Principle Component AnalysisTable 5. Facial Action Factor Loadings  61  Table 6. Eigenvalues and Percent of Variance Accounted for by the Factors . . . .  62  Extracted in the Principle Component Analysis for Children Displaying Clinically Significant Fear. Table 7. Facial Action Factor Loadings for Children Displaying Clinically  ...  64  Significant Fear. Table 8. Descriptive Statistics for the Prediction of Children's Fear.  66  Table 9. Bivariate Correlations in the Prediction of Children's Fear.  67  Table 10. Standard Multiple Regression Analysis: The Prediction of  69  Children's Fear. Table 11. Means and Standard Deviations for OSBD Fear Segments.  ....  Table 12. Frequency of OSBD Behaviours and Correlations of Individual  72 74  Behaviours with Total OSBD Score. Table 13. Frequency of OSBD Behaviours and Correlations of Individual  76  Behaviours with Total OSBD Score for the Significant Fear Sample. Table 14. Descriptive Statistics for the Prediction of Children's Fear.  77  Table 15. Bivariate Correlations in the Prediction of Children's Fear.  79  Table 16. Standard Multiple Regression Analysis: The Prediction of  80  Children's Fear.  vii List of Figures Figure 1. Drawing of a Child's Neutral Face.  83  Figure 2. Drawing of a Child's Fear Face.  84  vm Acknowledgements Cheryl Gilbert-MacLeod was supported by a Social Sciences and Humanities Research Council (SSHRC) doctoral fellowship for the duration of this study. Many thanks to the research assistants and fellow graduate students who helped collect and code this data. This includes, Amy Beauchamp, Christine Chambers, Catherine Duffy, Kelly Giesbrecht, Jenny Lang, Christine Lilley, and Bruce McMurtry. Further, I would like to thank Dr. Kenneth Craig, my supervisor, Dr. Charlotte Johnston, Dr. Janet Werker, Dr. Mark Schaller, Dr. Peter McLean and Dr. Eric Mash whose questions and comments directed the final version of this dissertation.  Behavioural expression of fear. 1  The Behavioural Expression of Fear in Young Children Introduction Fear has been described as an emotion evoked by a specific, threatening stimulus which declines when the threat is removed (Rachman, 1998). According to Webster's Dictionary (1957) " fear is a feeling of anxiety and agitation caused by the presence or nearness of danger, evil, pain, etc." Children, over the course of development, encounter numerous situations capable of eliciting fear (e.g., parental separation, being sick or hurt), and, not surprisingly, many researchers have catalogued these sources of fear in children (e.g., Croake, 1969; Scherer & Nakamura, 1968). Most recently, Campbell and Rapee (1994) determined that children's perceptions of threatening stimuli fall into two separate, interpretable factors fear of physical harm and fear of social consequences. These two factors were found to be consistent across gender, age and clinical status. That is, children, regardless of age, gender, or whether they did or did not have an anxiety disorder, reported social consequences and physical harm to be the most prominent sources of fear. Researchers have devoted considerable time detennining the sources of threat in young children. They have also evaluated how children perceive different, threatening experiences (e.g., Broome, 1986; Broome, Hellier, Wilson, et al., 1988; Melamed, Meyer, Gee & Soule, 1976), and looked at ways to help children cope with these situations (e.g., Elkins & Roberts, 1985; Roberts, Wurtele, Boone, et al., 1981; Zachary, Friedlander, Huang, et al., 1985). However, somewhat surprisingly, few researchers have examined  Behavioural expression of fear. 2  how young children communicate fear, behaviourally, during a threatening situation and, consequently, health care practitioners are left with no objective means to quantify fear in young children for whom verbal report is often unavailable or even unreliable. Although systematic evaluation of the behavioural expression of fear in young children has received limited attention, researchers have reacted to the young child's inability to use self-report to communicate feelings of pain, and energy has been spent characterizing the behavioural expression of pain. For instance, studies of infants' reactions to invasive procedures such as intramuscular injections and heel pricks have disclosed a fairly specific pattern of cry, facial activity, and bodily movement (Craig & Grunau, 1993). Similarly, young children respond to short, sharp pain with vigorous body movements and facial activity; preschool aged children often cry (Gilbert, Craig, Rocha, & Mathias, 2000; McGrath, Johnson, Goodman et al., 1985). The benefits of documenting the behavioural elements of the pain expression have been significant. Perhaps most notable has been a rise in the education of nurses and other healthcare practitioners in recognizing pain facial configurations in infants (Grunau, Oberlander, Holsti, & Whitfield, 1998; Guinsberg, Berenguel, Xavier et a l , 1997). Similar advantages are likely if a behavioural expression of fear can be identified. For example, children who respond to situations, in particular medical procedures, with fear could be recognized and taught appropriate strategies for coping with fear (e.g., deep breathing, imagery). Further, determining the behaviours which make up a fear response will help health care professionals and parents establish what emotion a child is  Behavioural expression of fear. 3  experiencing and lead them to the most suitable intervention. This is particularly important in situations where other means of communication are hindered, e.g., postoperative surgery, as health care practitioners are often left guessing what a child is experiencing and consequently, appropriate relief is not administered in a timely fashion. Although the present study will likely provide similar clinical benefits, investigations into, and evaluations of, how young children behaviourally communicate fear must take note of the methodological warnings issued by those who have investigated the behavioural expressions of different emotions. In particular, studies addressing this issue should use comprehensive coding systems for a range of behaviours, define the emotion of interest, and use a stimulus capable of eliciting the desired emotion. This study was designed to delineate the facial actions and the broadband behaviours which constitute a fear reaction in young children. It also addressed methodological issues apparent in earlier research. Using the definition of fear presented earlier, the behaviours constituting a fear response in young children were examined. Cast removal was the stimulus used to elicit fear and the behaviours displayed by the children in response to cast removal were coded with the Baby-Facial Action Coding System (Oster & Rosenstein, 1995) and the Observational Scale of Behavioral Distress (Jay, Ozolins, Elliott, & Caldwell, 1983). The influences of age, gender, and cast location on the fear response were also examined.  Behavioural expression of fear.  4  Literature Review Psychology of Emotional Expression Investigations of the behavioural expression of fear need to be considered within the context of whether behaviours (e.g., facial activity, crying, drawing in of limbs, tensing muscles) convey specific emotional information. Researchers invariably agree that behavioural expressions play a vital role in communication, but there remains much debate as to whether facial actions and other broader behavioural actions can be characterized as discrete and unique expressions of emotion. Most of the controversy has centered around whether there are a limited number of basic, hard-wired emotions which automatically manifest themselves in specific, discrete expressions; whether emotions are more multidimensional, overlapping and complex; or finally, whether expressions are more dependent on the social gains to be had from communicating them. To understand the proposed relationship between the experience and expression of fear, it is necessary to provide a brief overview of some of the major theories of emotion. The theories reviewed here have made assertions as to whether unique expressive features of emotions exist. Differential Emotions Theory. Izard and colleagues (e.g., Blumberg & Izard, 1991; Izard, 1992; Izard, 1997) developed and researched a model of emotional expression labeled Differential Emotions Theory (DET). D E T is a broad theory of emotion, personality and social behaviour. Facial expressions, verbal communications, posture and gesture are conceived "as evolved, genetically influenced but highly modifiable and dissociable component(s) of emotion" (Izard, 1997). They are not however  Behavioural expression of fear. 5  viewed as necessary manifestations, i.e., one could experience an emotion without visibly expressing it. DET postulates a limited number of basic emotions (happiness, surprise, fear, anger, contempt, disgust, and sadness) which are genetically determined (i.e., innate), universal, and discrete. Each of these emotions is purported to have a highly coherent pattern of facial actions and behaviours which maintains the experiential and motivational properties that make the emotion unique, but which are also dynamically related to other emotions (Blumberg & Izard, 1986). DET does not suggest that other emotions do not exist, only that those not listed as a basic emotion are more complex (Ekman, 1992a). Support for the theory comes from several sources. First, Wagner (1990) demonstrated that even in the absence of social cues (which we often rely on to help decipher social situations) individuals correctly determined which emotion was being displayed by examining facial expressions. This suggests that facial actions may indeed be automatic readouts of the underlying emotions. Second, studies demonstrating crosscultural invariance in facial expressions and recognition lend credence to the universality hypothesis (e.g., Ekman & Friesen, 1971; Ekman, 1992b). Third, emotional expressions of infants appear to be both morphologically and functionally similar to those of adults (Izard & Malatesta, 1987). Fourth, untrained college students reliably identified infant expressions of joy, interest, surprise, sadness, anger, disgust, contempt, and fear (Izard, Huebner, Risser, McGinnes, & Dougherty,  Behavioural expression of fear. 6  1980). Together, these findings support the contention that emotional expressions and the accompanying feeling states are pre-wired, one of the basic premises of DET. Fifth, research upholds the assumption that each emotion, both negative and positive, is linked to a different motivational response in observers. In one such study, mothers were shown to respond differently to distinct facial displays (Huebner & Izard, 1988). That is, mothers reported giving more "love, affection, and cuddling" to their infants when they exhibited a response indicative of physical distress than when the infant expressed sadness or anger. Finally, consistent with DET, other investigators have noted changes in physiology which appear specific to different emotions. When subjects were asked to move specific facial muscles, while remaining blind to the emotion the actions represented or reliving a past emotional experience, different patterns of psychophysiological activity were observed (Ekman, 1992a; Ekman, 1992b; Ekman, Levenson, & Friesen, 1983). For example, heart rate increased more during prototypical displays of anger and fear than happiness. Finger temperatures also increased more during anger than happiness displays. A more recent study demonstrated a clear physiological difference between anger and fear (Sinha & Parsons, 1996). Although both heart rate and systolic blood pressure increased significantly during anger and fear, only fear was accompanied by significant decreases in finger temperature. This study also examined facial muscle activity and determined that a) more facial activity was evident during anger imagery, and b) fear and anger expressions differed in the amount of activity at the jaw and cheek muscle sites, with  Behavioural expression of fear. 7  anger eliciting more activity in these sites. Together, these findings support the proposition that qualitatively differentiable, inherent emotional states exist. Although there is support for DET, evidence to the contrary does exist. For example, Oster and colleagues (Oster, Camras, Campos et al., 1996) have recently shown young infant expressions of fear and anger to be different from the prototypical fear face described in adults. This finding is problematic if emotional expressions are indeed innate and "hard-wired". However, instead of discrediting the notion of innate emotional expressions Oster's findings may reflect developmental shifts that actually follow a predetermined pattern. Further, Russell contends that cross-cultural judgment studies have not demonstrated that the face universally expresses emotion, but more that different cultures attribute the same emotion to the same pictures (Russell, 1995). Russell also proposes that the methodology used in cross-cultural research (e.g., forced-choice format, withinsubject designs, pre-selected photographs of posed facial expressions) does not have ecological, convergent, or internal validity and therefore, a direct link between facial expressions and emotion labels has yet to ascertained (1994). It is important to note that rebuttals by both Ekman (1994) and Izard (1994) exist. In summary, Izard contends that Russell's criticisms of the universal nature of facial expressions focused on language-dependent data (i.e., universal semantic attributions) and ignored the actual facial actions displayed by the various cultures in the studies. Similarly,  Behavioural expression of fear. 8  Ekman purports that many of the disagreements put forth by Russell stem from a fundamental disagreement in what is meant by universality. Behavioural Ecology View. Fridlund (1991) presents a Behaviour Ecology View of behavioural expressions which differs from D E T in its belief of how expressions evolved, what they signify, and how they function in our everyday lives. According to the Behavioural Ecology View, the evolution of display patterns is more complex and occurs "in response to specific selection pressures [that] necessarily co-evolve with others' responsivity to them" (Fridlund, 1997, p. 123). Fridlund further contends that one can find meaning in emotional displays only in the context of occurrence and that displays are "issued to serve one's social motives in that context" (Fridlund, 1997, p. 123). In other words, facial expressions are not thought to be related to emotion, but instead function as social regulators (i.e., represent direct efforts by the individual to control the environment) or messages which influence another's behaviour (Fridlund, 1991; 1997). Finally, the Behavioural Ecology View of facial expressions postulates that facial displays are mediated by several factors, the intent of the displayer, the behaviour of the recipient, and the context of the interaction (Fridlund, 1997; Fridlund, Kenworthy & Jaffey, 1992). Support for the position that facial displays are moderated by social factors exists. Fridlund, Kenworthy, and Jaffey (1992) established that facial behaviour was influenced by sociality. In that study, participants were asked to imagine themselves in a number of different emotional situations (e.g., fear, anger, happiness, sadness) at both high and low levels of sociality (e.g., listening to music alone versus attending a funeral of someone you  Behavioural expression of fear. 9  loved). Differences in E M G activity in four areas of the face (forehead, brow, cheek and lip) were evident between the high and low sociality situations, even when emotion was controlled. That is, participants displayed more facial activity when imagining themselves in high sociality images (e.g., professor accuses you of cheating in the middle of class), regardless of the emotion elicited. Hess, Banse, and Kappas (1995) also demonstrated that displays of emotion are related to the sociality of the situation, the intensity of the emotional stimulus, and the relationship between the expressor and audience (i.e., friends versus strangers). Participants in that study were more likely to display facial expressions when someone else was present, especially if that person were a friend, and in response to more intense stimuli. The importance of the interaction between the three variables was stressed. Hess et al's results (1995) are consistent with both Fernandez-Do Is and RuizBelda's later findings (1997). In that study, Olympic gold medallists smiled more during social interactive stages (standing on the podium interacting with authorities and the public) than in nonsocial stages (waiting time behind the podium, listening to the national anthem). Similarly, Chovil (1991) has demonstrated that facial activity displayed by individuals listening to another person's description of a personal close-call experience increased during more social situations (e.g., face-to-face versus listening to a tape recording). In summary, one of the main tenets of the Behavioural Ecology View is that displays do not exclusively serve underlying emotions, but instead are reflective of social  Behavioural expression of fear.  10  motives. Further, this theory contends that observers, without knowing the context of a situation, are unable to accurately interpret what another displays. However, it is difficult to disregard the findings that, without context, .people can reliably determine what emotion someone is experiencing by examining facial activity (e.g., Izard, 1994; Wagner, 1990). Perhaps Hess and colleagues (Hess et al., 1995) said it best, "facial displays should be considered messages, with a signal as well as a symptom value" (p. 286). Dynamical Systems Theory. This model of emotional development focuses on facial actions. However, it is likely the hypotheses put forth in the Dynamical Systems Theory (Camras, 1992) extend to other behavioural forms of expression (e.g., crying, pulling in of limbs). Expressions in this model are viewed as elements of a larger collection of motoric and physiological variables where expressions are believed to arise from the dynamics of movement itself, and not necessarily because of a direct expression-emotional state link (Camras, Lambrecht, & Michel, 1996; Michel, Camras, & Sullivan, 1992). That is, actions in one part of the face, or body, are often accompanied by other related actions because of a motor or physiological connection and not necessarily because both actions are related to the emotional state. Although a direct link is not postulated, facial actions are thought to provide information about ongoing changes in the system that they help to constitute (Messinger, Fogel & Dickson, 1997). Further assumptions of the dynamical systems theory include: a) in childhood, affect only gradually becomes differentiated into specific emotions, and b) facial  Behavioural expression of fear.  11  expressions are not "automatic read-outs of emotion" but instead are displayed when they are deemed appropriate and/or adaptive by the system (Malatesta-Magai, Izard, & Camras, 1991; Michel, Camras, & Sullivan, 1992). The latter hypothesis seems to allow for the observation that in some situations, an expected facial expression is not visible and would seem to accommodate Fridlund's position that displays are adapted to social contexts. The majority of support for Dynamical Systems Theory comes from research in the area of infant development. For example, Camras and colleagues have shown that facial activity can be viewed as part of larger coordinative structures (Camras et al., 1996). They found that when infants opened their mouths, they then proceeded to raise their brows, suggesting that mouth opening engages brow action. Further, knit brows were rarely observed without nose wrinkling. In regard to the postulation that the expression of distinct emotions does not occur at birth, a contextual analysis of an infant, Camras' daughter, during the first two years of life offers some validation (Camras, 1992). Specifically, these observations demonstrated that differential patterns of emotions do exist, but that they develop later in childhood, i.e., two years of age. It is important to note that Camras did observe some differences in the patterns displayed, but that these differences were better accounted for by the descriptors "negative" and "positive" and not the typical emotion terms we use (e.g., anger, fear, sadness, happiness). Oster et al.'s findings (1996) that young infants did not display the prototypical adult expression of either fear or anger offers further evidence. However, it  Behavioural expression of fear.  12  is possible that young infants have their own displays which develop, over time, to become the more sophisticated expressions observed in adults. Dimensional-Contextual Perspective. This account supposes that behavioural expressions do not signal emotional state directly, but instead convey two types of information which observers use to infer emotions (e.g., Bullock & Russell, 1986; Carroll & Russell, 1997; Russell, 1997). First, observers obtain "nonemotional, quasi-physical information from the face" (p. 298); e.g., tone of voice, degree of concentration, what he/she is doing). Second, observers conceptualize the facial expression according to overall levels of pleasure (pleased vs. displeased) and arousal (agitated vs. sleepy). This second type of information (i.e., judgments about the degree of pleasure and arousal the individual is experiencing) is best understood multi-dimensionally on a circumplex model where the pleasure and arousal dimensions are bipolar (e.g., Russell, 1997; Russell & r  Bullock, 1986). For example, the emotion fear would be situated in the high arousal low pleasure quadrant, as would anger and frustration. According to Russell (1997), these two types of information help observers to determine which, if any, emotion is being expressed. Russell also places considerable weight on the role of context in interpersonal judgements. He contends that expressions are not constant in meaning and that the same configuration of facial actions will be viewed differently depending on the expressor's and the observer's contexts (Russell, 1997). Indeed, he states "if the facial expression does not signal any specific emotion but rather provides, quasi-physical information, pleasure,  Behavioural expression of fear.  13  and arousal, then situation information would determine the specific emotion inferred by the observer" (Russell, 1997). Research by Carroll and Russell (1996) supported this view. In that study, information from a detailed story altered the way participants labeled photographs. For example, a picture of someone displaying a fear expression was chosen as such by observers for whom information about the situation was withheld, however, if the participant heard a story which suggested anger but did not exclude fear, the same facial expression was labeled as anger significantly more often. It is important to note that whether the observer's judgement is isomorphic with the subjective state of the expresser remains moot but, the emphasis on the meaning of behavioural displays to observers leaves open the question as to whether there are discriminable subjective states. Componential View. This theory stresses that "emotional expressions have a systematic, coherent, and meaningful structure" and that "consideration of the individual components which contribute to that structure can directly reveal much about the underlying properties of the emotion being expressed" (Smith & Scott, 1997, p. 231). In this model, behavioural expressions are: a) believed to convey more information than what is represented in the sum of its parts; and b) thought to be a highly organized system which goes beyond merely indicating one's emotional state. Support for this theory comes from studies which have examined different parts of the face (i.e., different facial actions) to see whether or not they convey information over and above that of emotional state. For example, facial activity in the cheek region, as  Behavioural expression of fear. 14  measured by E M G , has been found to be correlated to participants' ratings of subjective pleasantness and motivational congruence. On the contrary, brow region activity was positively correlated with evaluations of motivational incongruence and perceived goal obstacles/discrepancies and negatively correlated with pleasantness (Pope & Smith, 1994). These findings suggest that smiles and frowns convey more information than merely an emotional state. Minimal Universality. Although there are many different philosophies about behavioural expressions and what meaning they may or may not convey, an integrative and parsimonious account, or what Russell describes as "a baseline position", of the current evidence does exist (Russell & Fernandez-Dols, 1997). Russell has named this account Minimal Universality (Russell, 1995). Minimal Universality does not suppose the existence of an innate emotional signaling system as DET does, but predicts that there is a degree of similarity in how individuals express emotion and interpret facial expressions. It assumes: a) facial movements are coordinated with psychological states; b) certain patterns of facial muscle movement occur in all humans; and c) that most people can infer something of another's psychological state from facial movement, much like they can from anything else another person does or says (Russell, 1995; Russell & Fernandez-Dols, 1997).' Minimal Universality cautions researchers that actions are not necessarily signals caused by, or aspects of, emotion, nor are they necessary or sufficient for an emotion (Russell, 1995). Finally, this account predicts that: a) photographs of facial movements  Behavioural expression of fear. 15  will be associated with psychological state with agreement that is greater than chance; b) people are sometimes accurate in the inferences that they make on the basis of facial movements; and c) there will be similarities across cultures in what is inferred from movements. Although these hypotheses are directed to facial activity alone, it is possible to presume they can be extended to include other behavioural expressions. Therefore, Miriimal Universality, with the broader definition of behavioural expression, is the theoretical model upon which this study was based. In other words, we too assumed that children in an emotional situation would be likely to exhibit a configuration of actions which caregivers and health care practitioners could reliably use to help determine which emotion the child was experiencing. We also believed that a lack of behaviour does not necessarily mean that no emotional/psychological state was present. That is, a relationship between emotion and behavioural expressions exists, but it is not a perfect relationship. It is important to note that this study is not attempting to evaluate any of these theories. Instead, its aim is to gain a more complete understanding of the behavioural expression of fear, and, in order to accomplish this goal, it was felt the reader needed an understanding of how different theories view emotional expressions. Characterizing Emotional Expressions There are two classes of measurement with demonstrated value in characterizing emotional expressions: fine-grained (e.g., discrete facial actions) and more broadband (e.g., cry, tensing, pulling in of limbs) oriented analyses of behaviour. The study presented  Behavioural expression of fear.  16  here used both classes of measurement but, consistent with the literature, the two types of behavioural analyses are discussed separately. The value of combining the information to obtain a multidimensional perspective is then argued. Facial Expressions of Emotion. According to a relatively recent review of current theory and research on emotional development, facial expressions are "fundamental building blocks for the development and elaboration of human social relationships" (Malatesta-Magai, Izard, & Camras, 1991), and, with little hesitation, one can consider facial activity to be one of the most consistent forms of non-verbal communication. Indeed, it has long been accepted by most people, including health practitioners and scientists, that the face represents a relatively coherent and open communication system (McGraw, 1943; Sherman, 1927; Tassinary & Cacioppo, 1992). Therefore, it is not surprising that parents and other adults devote considerable time and attention to studying the faces of infants and pre-lingual children for clues as to their psychological status and well-being (Craig, Lilley, & Gilbert, 1996). Nowhere is this dedication to the study of facial activity more readily observed than in the pediatric pain literature (Craig, 1998). Researchers have intensively examined the pre-lingual child's response to such events as heel lancing (Grunau & Craig, 1987), intramuscular injections (Grunau, Johnston, & Craig, 1990; Cassidy, McGrath, Reid, Chambers, Gilbert et al., 1996), more invasive and persistent pain situations including circumcision (Taddio, Stevens, Craig et a l , 1998) and postoperative pain (Gilbert et a l , in press).  Behavioural expression of fear.  17  Results of these studies and others have demonstrated a relatively consistent configuration of facial actions in reaction to painful stimuli. Johnston and Strada (1986) found that infants undergoing immunization by needle injection displayed a pain configuration typical for their age, i.e., lowering of the brow, broadening of the nasal root, an angular and squarish mouth, and tightly closed eyes. Indeed, facial activity was found to be the most consistent response to pain in comparison to cry patterns, heart rate, and body movements. Similarly, in a study of neonates experiencing heel lance, Grunau and Craig (1987) observed, in addition to the above facial actions, parted lips, taut tongue, and chin quiver. More recent studies (Craig, Hadjistavropoulos, Grunau, & Whitfield, 1994; Lilley, Craig, & Grunau, 1997) have identified the following action units as differentiating invasive (heel lance/immunization) from non-invasive procedures in infants: brow lower, cheek raise, lip corner pull, and jaw drop. Research investigating toddler and older children's responses to acute painful stimuli (e.g., vaccinations, blood withdrawals) found a number of similarities to that of infants. Children, like infants, lower the brow, deepen the nasiolabial furrow, open their mouth and stretch it horizontally and vertically. However, children were more likely than infants to squint their eyes instead of squeezing them shut, flare their nostrils, wrinkle their nose, and raise their cheeks and upper lip in response to acute pain events (Cassidy et al., 1996). Longer episodes of pain behaviour have also been examined (Gilbert et a l , in press). In that study, children were observed during the hour following surgery. In  Behavioural expression of fear. 18  addition to displaying lowering of the brow, deepening of the nasolabial furrow, cheek raising, opening of the lips, vertically and horizontally stretching the mouth, children also displayed pulling upward of the lip corners. These findings suggest that a number of commonalties exist across pain situations, providing further evidence for the clinical worth of analyzing facial actions in other contexts. Facial action research in other situations does exist. Rosenstein and Oster (1988) examined infants' facial activity in response to tasting sweet, sour, salty, and bitter substances. The results revealed that facial actions reliably discriminated between sour and bitter stimuli and sweet versus non-sweet substances - lip and mouth actions differentiated the sour and bitter tastes from each other. As well, infants typically raised their cheeks, narrowed or tightly closed their eyes, wrinkled the nose, and either raised, lowered, or both raised and pulled the brows together when tasting non-sweet liquids. Using the same infant expressions, adult observers were able to detennine the type of liquid (sweet versus non-sweet) the infant was consuming. In general, facial actions that appeared more positive were attributed to sweet substances, whereas more negative expressions were thought to be indicative of non-sweet tastes. Observers had more difficulty discriminating between the non-sweet tastes, in particular sour and salty solutions. Several researchers have used the strange situation paradigm as a stimulus for emotional expression. Gaensbauer, Mrazek, and Emde (1979), using a 9-point rating scale for six different emotions (fear, anger, sadness, distress, interest, and pleasure),  Behavioural expression of fear. 19  found that 12-month-olds typically displayed fear or distress during the separation procedure. However, observers had a difficult time differentiating between distress and the other negative emotions. In another study, using both the System for Affect Identification by Holistic Judgments (AFFEX; Izard & Dougherty, 1980), which assigns emotion labels to global facial expressions, and the Maximally Discriminative Facial Movement Coding System ( M A X ; Izard et al., 1980), which codes discrete facial movement in terms of preconceived categories of emotion like anger, fear, surprise, joy and pain, patterns of emotional expression during separation in the strange-situation procedure were identified (Schiller, Izard, & Hembree, 1986). In particular, three monthold infants displayed negative emotions like anger and sadness following separation from their mothers. A study examining 13 and 18 month old infants' reaction to the strange situation (Hyson & Izard, 1985) found A F F E X identified anger and sadness, but also interest during the separation part of the procedure. Again, these studies are limited by their inability to specify the discrete cues instigating judgements. There is only limited work on the facial expression of fear or anxiety, but, at least with adults, it has the advantage of detailed coding and description of the facial activity. The adult facial expression of fear consists of raised and drawn together eyebrows and/or lips stretched horizontally so that they form a rectangular shape (Ekman & Friesen, 1978). More intense fear stimuli have also been found to evoke raised, tensed, upper eyelids which widen the eyes. Recent work on facial displays of anxiety in adults (Harrigan & O'Connell, 1996) has determined that fewer non-enjoyment smiles, more fear movements  Behavioural expression of fear. 20  (a more precise delineation of the movements was not provided), and more eye blinks occur in high state versus low state anxiety situations. A description of the facial actions which constitute a fear response in infants or children is limited to the M A X and A F F E X coding systems and a few studies which looked at both infants' and children's responses to fear evoking stimuli. Both the M A X and the A F F E X coding systems consider the following facial actions to be representative of a fear expression: 1) inner and outer brow raise, 2) brow knot and lower, 3) upper lid raise, 4) cheek raise, 5) tightening of the eyelids, 6) lip stretch, and 7) lips apart (Izard & Dougherty, 1980; Izard et al., 1980). It is noteworthy that these accounts were not based on studies systematically describing children's reactions using objective, anatomically based coding systems, but instead represent the cumulative impression of these experienced investigators. Hiatt, Campos and Emde (1979) compared the facial patterning of fear in infants between the ages of 10 and 12 months with the emotional expressions of happiness and surprise. They used two different stimuli for each emotion. The events used to elicit happiness were a collapsing toy operated by the infant's mother and playing peek-a-boo with the mother. Having a toy switched unexpectedly by the experimenter and seeing an object vanish instantaneously were used for surprise. A visual-cliff apparatus and a stranger approaching were used as means to elicit fear. Facial responses were scored by trained coders who were taught to identify discrete facial configurations in the eyebrow/ forehead, eyes/lids, and nose/mouth regions.  Behavioural expression of fear. 21  Results showed that infants a) reliably smiled and wrinkled the skin under their eyelids during the happiness tasks, b) typically raised their eyebrows and opened their eyes when surprised, and c) opened their mouth and stretched back their lips when scared. Further analyses demonstrated that infants displayed more happiness responses during the happiness elicitor events than the other situations. Infants also displayed more fear responses during the visual cliff situation than in any of the other emotion eliciting conditions. However, infants' surprise reactions were not specific to the surprise eliciting conditions nor were the reactions to the stranger approach. Finally, untrained adult observers had no trouble labeling the happiness reaction as such, but had some difficulty with both the fear and surprise reactions. A later study looked at preschoolers' arid second graders' reactions to films designed to produce either anxiety or apprehension, empathic sadness, or cognitively induced sympathy (Eisenberg, Fabes, Bustamante et a l , 1988). In that study, coders judged whether certain emotions were present based on the Ekman and Friesen Unmasking the Face criteria (1975). Results show that children displayed the most fear, gaze aversion, and anxiety in response to the anxiety film and these actions were significantly correlated with self - reports of fear. As well, children's global facial reactions to the three separate films were rated as significantly different by adult observers (Eisenberg, et al., 1988), suggesting that children's responses looked different to observers. Unfortunately, specific facial actions for fear and the other emotions were not delineated.  Behavioural expression of fear. 22  Izard and colleagues (1980) have also looked at fear, although not specifically. In a series of studies they investigated a) whether infants could produce emotional expressions, and b) if observers could identify a range of emotions, including fear. Infants displayed facial expressions consistent with emotion-specific criteria for the following emotions: fear, anger, disgust, interest, joy, surprise, sadness, and contempt. Further, observers (e.g., college students and health care professionals), when shown slides of infants displaying these emotion-specific facial expressions, accurately selected the emotion which pertained to the facial displays. However, once again, no specific facial criteria were listed. The most recent study to look at fear was completed by Oster et al. (1996). Using a comprehensive coding system designed for infants, 11 -month-old infants, in response to a toy gorilla head that Opened and closed its mouth and emitted raspy, growling noises (fear), displayed eye widening and brow raising (with or without brow knitting), more often than when gentle arm restraint by the experimenter (anger) was employed. Further, infants tended to fixate on the fear elicitor (i.e., the gorilla) significantly longer than on the experimenter who was restraining their arm (i.e., the anger elicitor). Finally, infants displayed more facial stilling (a cessation of facial movement, typically with a slack, slight open jaw and sometimes widely open eyes and slightly raised brows) when the toy gorilla was present than during arm restraint. The results described above suggest that infants and children are capable of producing a discriminable facial expression to fear stimuli, yet, the actions constituting the  Behavioural expression of fear. 23  fear face in young children remain unclear. This is likely due to a variety of methodological concerns delineated below. First, there has been confusion about the definitions and differences between the emotions of fear and anxiety. According to Rachman (1998) there are only limited commonalties between fear and anxiety (e.g., anticipation of danger or discomfort, elevated arousal and uneasiness), yet many researchers fail to make a distinction or blur the boundaries between these two concepts. Anxiety should be seen as something which a) results from an elusive source of threat; b) tends to be prolonged and persistent; c) has an uncertain onset and offset; and d) is often accompanied by heightened vigilance. Fear on the other hand can be denned as: a) having a specific source of threat; b) usually episodic in nature; c) having a detectable offset; d) declining with the removal of threat; and e) an understandable connection between threat and fear exists (Rachman, 1998). Studies by Harrigan and O'Connell (1996) and Harrigan and Taing (1997) examining the facial displays indicative of state anxiety failed to make an appropriate distinction between fear and anxiety. Indeed, they seemed to have blurred the boundaries and appeared to be treating the two emotional states as if they were one. Both studies defined anxiety as a "compound emotion involving fear, distress, frustration, guilt, apprehension, nervousness, and interest" and reported "elements of fear responses" in high state anxiety situations. Similarly, the Eisenberg et al. study (1988) used both fear and anxiety labels interchangeably to describe children's reactions to the anxiety/apprehension film. In addition, although they claimed to be measuring anxiety, they specifically asked  Behavioural expression of fear. 24  the children how much "fear" they felt while watching the films. Having neglected to make the discrimination between fear and anxiety, one cannot be certain which emotional reaction was induced or displayed by the children in that study. It is important to note that, to date, there is no available data to support a physiological or behavioural difference between fear and anxiety. This is because research designed to answer this question has yet to be conducted; however, given that a theoretical difference does exist, the two should be treated as if they were separate and distinct emotions. If, at a later date, the theoretical distinction is found to be untrue, then findings from fear research can be applied to anxiety and vice versa. Yet, if fear and anxiety are treated as one and a difference does exist, it becomes difficult to tease apart which behaviours are representative of fear and which are more illustrative of anxiety. Secondly, none of the studies which looked at the facial expression of fear, with the exception of the most recent Oster et al. study (1996), used a comprehensive, finegrained coding system to determine which facial actions consistently appear during fearful situations. To date, when researchers found evidence of a "fear" expression (e.g., Izard et a l , 1980), they used a coding system which based its emotion-specific criteria on subjective/theoretical judgments about what the emotion should look like and not on objective criteria. For example, the M A X system (Izard et al., 1980) codes only those configurations of facial actions that Izard believed to correspond to the universally recognized facial expressions of emotion (e.g., fear, anger, surprise, disgust).  Behavioural expression of fear. 25  One major problem inherent in facial coding systems based on a priori assumptions is that they fail to detect everything the face can do. Instead, these systems only detect what the system theorized the face could do. This is particularly problematic for negative findings. For example, in Hiatt et al.'s study (1979), only two of the hypothesized "fear" facial actions were found, and these actions were only specific to fear when the visual-cliff was the fear-eliciting condition. Perhaps had a more comprehensive coding system been used, a greater degree of facial action specificity would have been found. That is, if all possible facial actions had been coded, facial actions uniquely characteristic of fear may have been identified. Similarly, had researchers used pictures of facial expressions which reflected anatomically based, objective criteria, instead of choosing expressions based on subjective criteria, observers may have been more successful at distinguishing photographs of fear expressions from other negative emotional expressions. Other methodological problems intrinsic to theoretically based coding systems place reasonable doubt on earlier research findings. Oster, Hegley, and Nagel (1992) have suggested that the M A X coding system does not accurately represent complex negative emotions like pain, fear, and anger, nor are the facial expression templates significantly different from one another. Research findings support their claim. For example, when adults were presented with pictures of infants displaying selected M A X facial configurations and asked to choose any emotional label to identify the child's state, they chose non-predicted labels more often than chance. In the same study, the MAX-specified infant expressions of discrete negative emotions were perceived by adults as showing  Behavioural expression of fear. 26  either global, nonspecific distress or blends of several different negative emotions. Similarly, a study by Matias and Conn (1993) demonstrated that trained coders were only able to judge infants' M A X positive emotions as being discrete (surprise, happiness) and had more difficulty differentiating between the negative emotions (sad, angry, fear). These findings suggest that M A X formulas do not adequately capture infants' expressions of negative emotions. More specifically, the M A X fear face (a negative emotion) has been criticized. Specifically, Oster (1999, personal communication) believes there is no evidence that the predicted facial expression of fear in the M A X system actually reflects a fear face, i.e., it has poor validity. Camras and her colleagues have also demonstrated that the global judgments implicit in the A F F E X coding system (Izard & Dougherty, 1980) do not capture discrete negative emotions (Camras, 1991,1992; Camras, Malatesta, & Izard, 1991; Camras, Sullivan, & Michel, 1993). In particular, their research has shown that some of the negative expressions (e.g., anger, fear) are identified in a wide range of negative affect situations, thereby questioning the specificity of the facial templates for discrete emotions. It appears that no researchers have looked at the expression of fear in young children without, inadvertently, using an inappropriate measurement system or confusing anxiety and fear (e.g., Eisenberg et al., 1988). Despite these methodological concerns, the findings from the studies reviewed can, and should, be used as a place from which to develop hypotheses.  Behavioural expression of fear. 27  Broadband Behavioural Expressions of Emotion. Although facial actions are the most fine-grained behavioral measures available to researchers, a variety of other sources of information are available to the observer. For instance, information about an individual's present affect can be found in verbal report, paralinguistic vocalization (e.g., noncontent parameters of speech, crying, screaming, moaning, or sighing), autonomic reflexes (e.g., blanching, flushing, panting, or vomiting), voluntary and involuntary motor activity (e.g., withdrawal, gesticulations), and postural adjustments (Craig & Prkachin, 1983; Prkachin, Currie, & Craig, 1983). The pain literature contains many examples of studies which have investigated the utility of these other sources of information. For example, research has shown that a matrix of pain behaviours in infants and toddlers, in addition to a specific facial expression, exists. These behaviours include motor movement (e.g., movements of the extremities), general body responses, and vocal communication (McGrath et a l , 1985; Mills, 1989; Fuller, Thomson, Conner, & Scanlan, 1996). Studies examixiing the pattern of these global behaviours across a variety of painful medical situations have found similar behaviours. McGrath et al. (1985) deteirnined that large movements or behaviors such as crying or grimacing and verbal reports were indicative of postoperative pain in young children. Children undergoing bone marrow aspirations typically cry and make verbal comments of both fear and pain. They also display verbal resistance, muscular rigidity, nervous behaviour and seek information (Jay, Ozolins, Elliott, & Caldwell, 1983).  Behavioural expression of fear. 28  Only a few researchers have looked at the behaviours which occur in response to fear provoking situations (e.g., anesthesia induction, separation from parent). In these situations behaviours like crying, turning away, and tensing were evident (Lumley, Melamed, & Abeles, 1993). Although there is limited research detailing the broadband behaviours occurring during fear situations, an investigation into these behaviours would contribute to our understanding of the behavioural expression of fear. However, these behaviours cannot be looked at in isolation. Consistent with this formulation, Legerstee, Corter, and Kienapple (1990) demonstrated that an infant's body actions and facial expressions were often used in tandem to communicate various affective states. In fact, infants' responses appeared to be organized in such a way as to form unique combinations of hand actions and facial activity to indicate which of four social or nonsocial contexts were engaged. For example, when infants were interacting with a passive mother, they typically had a closed hand, distressed face, and averted gaze. In contrast, when infants were involved in an interaction with an active mother, a behavioral organization involving pointing, gazing, and vocalizations was more common. This supports the position that other behavioural forms of communication work in parallel with facial actions. Further, the combination of fine-grained (i.e., facial actions) and broadband behaviours is of critical importance as children's more global behaviours do not always provide an accurate assessment of fear in a stressful situation  Behavioural expression of fear. 29  (Strachan, 1993). Therefore, to fully investigate the behavioural expression of fear, one must use both forms of measurement systems. Important Factors in the Expression of Emotions Age. The process of ontogenesis leads to remarkable biological and behavioural changes over the course of childhood. Research examining changes in the facial expression of emotion has shown fairly specific changes early in life, i.e., within the first year and a half. For example, Hyson and Izard (1985) determined that 18-month-old infants use more interest expressions, anger, emotion blends, and change their expressions more frequently in response to the strange situation than 13-month-olds. Similarly, Camras, Oster, Campos, Miyake and Bradshaw (1997) demonstrated that 12-month-old infants showed more evidence of distress (i.e., more facial activity, body movements, and vocalizations), more quickly than 5-month-olds when their arms were restrained. Further, the 12-month-olds displayed more negative facial expressions than 5-month-olds. These studies suggest that older infants display more facial expressions than younger infants, providing evidence for the notion that facial expressions vary with chronological age, at least early on. It is important to note that the majority of these studies focused on whether the intensity or emotional category of facial expressions changed across different ages and not on whether individual actions differed. Developmental changes in the more specific facial expression of pain also have been documented. Research has shown that the pain reaction appears to become less intense as children age, i.e., infants and younger children exhibit different and more intense  Behavioural expression of fear. 30  reactions to the same stimuli than older children (e.g., Fradet, McGrath, Kay et al., 1990; Lilley, Craig, & Grunau, 1997; Goodenough, Kampel, Champion et a l , 1997). For example, infants and young children squeeze their eyes shut, and while this often is observed in older children too, they are more likely to squint (Craig & Grunau, 1993). In contrast, adults typically leave their eyes open, although the eyes do appear somewhat narrowed. Developmental changes in the facial expression of fear have not been studied well. Eisenberg et al. (1988) completed one of the few studies to examine age-related changes in children's anxiety/fear responses. Results from this study did not suggest a developmental shift. Preschoolers and second graders displayed the same generalized facial reactions to a film designed to elicit anxiety and/or apprehension. However, as global ratings of emotion and not particular facial actions were used, the possibility of specific developmental changes in facial actions still exists. Further, as very young children and/or infants were not included, the likelihood remains that a developmental shift may occur earlier. Therefore, to date, we remain uncertain as to whether older children display qualitatively or quantitatively different facial actions than young children. Research on more overt behavioural responses has demonstrated that, in general, older children display fewer behaviours indicative of emotion. For example, distress responses have been found to diminish as children age (Katz, Kellerman, & Siegel, 1980), and by 6 or 7 years, observable distress levels in children have dropped dramatically (Jay et al., 1983). Indeed, young children typically display a fairly intense distress response  Behavioural expression of fear. 31  that is more overt and motoric and is of longer duration than older children. This is quite similar to what researchers have found regarding the global pain response. A n explanation for age-related changes found in both the fine-grained and global behavioural responses may lie in the concept of display rules. Display rules can be conceptualized as socialized instructions for regulating emotional responses (Hora-Ratner & Stettner, 1991) and, children appear to develop both a better understanding of these rules and the need to use them as they age. Gnepp and Hess (1986) have documented the age at which children appear to understand the importance of regulating emotions. In their study, children in grades 1 through 5 and adolescents in grade 10 were assigned to one of three conditions. One group heard stories where the main character was alone when experiencing an emotional response. The second group heard stories where the main character was with an audience, and the third group also heard stories where the main character had an audience, but it was made very clear that the main character did not want the others in the story to know how he/she felt. Participants were then asked: what the main character said, why he/she said that, what kind of face the main character made, and why he/she made that kind of face? Results showed that older children believed the main character should, when an audience was present, hide his/her true emotions if it would hurt someone else. Having the prompt to hide feelings generally increased all children's reported use of display rules. Finally, no change in the frequency with which participants cited display rules as explanations as to why you should regulate emotions was found between children in grade  Behavioural expression of fear. 32  5 and adolescents in grade 10, suggesting that by approximately 10 years of age, children have a fairly sophisticated understanding of display rules. In addition to learning the importance of display rules, older children also appear to use these rules more frequently than younger children. Saarni (1984), in his landmark study, demonstrated that third and fifth grade students were better able to hide their displeasure at receiving a disappointing reward for participation than first graders. However, although research suggests that older children have a better conceptual understanding of display rules, Izard (1982) has cautioned researchers who use facial expressions that social learning (i.e., display rules) and cultural factors can and do influence emotion expression, possibly as early as two years of age. Gender. In addition to developmental changes, children's responses also may differ on the basis of gender. In general, little research has examined the effects of gender. on facial displays of emotion. One of the few studies which has studied gender effects suggests that differences between males and females do exist (Zander-Malatesta & Haviland, 1982). In an interactive play session between infants and their mother, infant females showed more interest faces than males. Further, males displayed only two expressions, knit brow and enjoyment, whereas females showed four, knit brow, interest, anger, and enjoyment. In the pain literature, gender is seen as having a significant effect on both global and fine-grained behavioural responses. It appears that girls display a more intense behavioural response and report higher levels of pain than boys (Chambers, Giesbrecht,  Behavioural expression of fear. 33  Craig et al., 1999; Goodenough et a l , 1997). Girls have also been found to report more anticipatory pain than boys (Carr, Lemanek & Armstrong, 1998). Research regarding sex differences in global fear responses has been contradictory (Eisenberg et al., 1988; Jay et al., 1983; Katz et al., 1980). Some researchers have found that girls, on self-report measures of fear situations, reported higher levels of fear than boys (e.g., Broome & Hellier, 1987; Hellier-Wilson & Yorker, 1996); however, others have not. For example, Campbell and Rapee (1994) found that girls reported higher levels of fear than boys, but only when the fear situation was a social consequence and not when it was fear of physical harm. Consistent with the lack of a gender effect for physical harm . situations found by Campbell and Rapee (1994), Carr, Lemanek and Armstrong (1998) found that boys and girls reported the same level of anticipatory fear prior to allergy skin testing, a physical harm threat. Methodological Considerations Model of Fear. Establishing which behaviours comprise a fear response requires a model which is capable of eliciting the emotion.. In the past, several researchers have used emotion-eliciting stimuli which were unsuccessful in evoking fear. For example, Hiatt et al. (1979) predicted that having an approaching stranger and use of the visual cliff apparatus would be strong enough to arouse fear in infants; however, only the visual cliff evoked facial actions hypothesized to be specific to fear. Further, the frequency with which infants displayed these actions was far less than expected, suggesting that, although  Behavioural expression of fear. 34  the visual cliff elicited some emotional response, it was not a particularly effective stimulus. This study used cast removal as the fear elicitor. For those unfamiliar with what occurs during cast removal, it involves a relatively brief period (approximately 1-3 minutes depending on the size of the cast) of cutting the cast with an oscillating saw. The saw cannot cut skin unless the child jerks his or her appendage away, an infrequent occurrence. In general, children are not restrained unless they are moving to such a degree that they may hurt themselves. If that were the case, parents are usually asked to hold their child. Few people, especially children who have undergone cast removal, would disagree that the sights and sounds of the oscillating cast saw are frightening. Indeed, other researchers have also classified cast removal as a stressful procedure (e.g., Zachary et a l , 1985). Furthermore, Johnson, Kirchoflf, and Endress (1975) have demonstrated that young children find cast removal a fearful experience. In their study, one group of children heard procedural information about what was to occur during cast removal (e.g., there would be a saw, scissors), another group received no information, and the third group was provided with sensory details (e.g., what they would hear, smell, and see). A l l children also reported howfrighteningthey thought cast removal would be and rated their level of fear during the procedure. Results showed that approximately 50% of the children reported some fear of cast removal prior to the procedure. These children had higher distress scores, as rated by a treatment - blind observer, and rated the actual  Behavioural expression of fear. 35  experience of cast removal as scarier. Further, children in the sensory information group displayed fewer distress behaviours and had significantly lower heart rates than either of the other two groups. Interestingly, children in the control and procedural information groups also responded to cast removal with large increases in pulse rate. The finding that an intervention decreases distress behaviours highlights the assertion that cast removal is a frightening experience. Cast removal was considered an ideal way to elicit fear in children as it is a naturalistic situation where the child perceives very real physical danger to Wmself, yet there is no associated pain. Peterson and Mori (1988) have called cast removal a "fruitful area" for researchers who wish to study fear in young children as "the demands of orthopedic cast removal, while threatening, involve no actual discomfort" (pp. 476). Together, these findings confirm that young children perceive cast removal as a fearful experience, and, when coupled with the fact that cast removal is not a pairrful event, it becomes an ideal means to characterize the fear expression in young children. It is important to note that the fear investigated here is specific to physical harm. Children in this study were most likely afraid that the saw would cut them or hurt them in some way. They may also have been frightened of the sounds. The expectation of fear reactions specific to a situation is consistent with Endler's Multidimensional Interaction Model of Anxiety (Clewes & Endler, 1994; Endler, 1997) and Campbell and Rapee's two-factor model of fear (1994). In Endler's model, anxiety is considered a result of situation by person interactions and anxiety is only experienced  Behavioural expression of fear. 36  and/or expressed when a threatening situation is congruent with an individual's trait anxiety, i.e., defending a dissertation will result in significant levels of anxiety if the individual has evaluation apprehension. According to Campbell and Rapee (1994), fear occurs in response to two different types of situations, fear of physical harm and fear of social consequences. However, neither research groups have determined if fear or anxiety look different in these various situations. Measurement of Fear. There are many ways to measure fear. For example, researchers have used self-report, fine-grained behavioural analyses (i.e., facial activity), and more broadband behavioural investigations. A discussion of these measurement systems follows. There are several self-report measures assessing the different things children fear. For example, The Medical Fear Questionnaire (Broome, 1986), The Child Medical Fear Scale (Broome et a l , 1988), and The Hospital Fear Questionnaire (Melamed et al., 1976) ask children about specific medical situations. More general fears have also been investigated (Croake, 1969; Scherer & Nakamurer, 1968). As this study investigated the behavioural expression of fear and was not concerned with hospital events other than cast removal, the utility of such measures was questionable. Tools to assess children's self-report of the quantitative and qualitative nature of their feelings of fear also exist. One of the most popular self-report techniques is the "face-type" scale where children are asked to point to the face which most closely represents how they feel (e.g., LeBaron & Zeltzer, 1984; Venham & Gaulin-Kremer,  Behavioural expression of fear. 37  1979). Although these self-report measures are often considered the gold standard, they are best used with school-age children and rarely are used with children less than 6 years of age. Recent studies (Chambers & Craig, 1998; Chambers et a l , 1999) have determined that children younger than 7 years of age have a difficult time understanding what the faces represent and often confuse different emotions when reporting how they feel (e.g., pain and sadness). Therefore, this type of measurement system was inappropriate for use in this study as the children were all less than 7 years of age. It is important to note that the limitations cited above deal exclusively with the self-report of emotions. Younger children can effectively use a face scale when it is applied to non-emotional situations (e.g., Schrans, Werker & Brown, 1990). A more useful tool may be observational measures of behaviour. These tools are often effective in children for whom verbal communication of emotional states may be difficult, i.e., young children. Further, broadband behavioural assessment systems are advantageous as a single, discrete behaviour is not generally sufficient, or indeed necessary, to indicate an emotional state (Lewis & Michalson, 1982). A well-validated and reliable broadband behavioural coding system exists that measures both fear and distress behaviours during medical events. The Observational Scale of Behavioral Distress (OSBD; Jay et a l , 1983) differs from other behavioural coding systems, e.g., Children's Hospital of Eastern Ontario Pain Scale, (McGrath et al., 1985), Toddler-Preschooler Postoperative Pain Scale (Tarbel, Cohen, & Marsh, 1992), in that it can reliably be used in situations other than painful procedures. The OSBD codes  Behavioural expression of fear. 38  behaviours such as cry, verbal fear, verbal pain, and seeking emotional support. Originally, the OSBD was designed to measure children's generalized distress responses to painful medical procedures (e.g., bone marrow aspiration), but, more recently, it has been used successfully in situations where there is perception of fear from physical danger (e.g., medical exam). The OSBD has been found to be a reliable measure of distress (Elliott, Jay & Woody, 1987; Jay et a l , 1983). It also has demonstrated concurrent and construct validity as it correlates significantly with parent and child self-reports of state anxiety (Jay et al., 1983). It is important to note that Elliott, Jay and Woody (1987) have suggested that the OSBD may be a better overall measure of medical fear (although the authors use the term anxiety, it is better conceptualized as fear) than either pain or generalized distress. Given these findings, the OSBD appears to be the most psychometrically sound tool for assessing fear and/or global behavioural distress in a number of situations. Although global measurement systems provide invaluable data, more fine-grained behavioural analyses, i.e., facial action coding, have also proven to be beneficial. To date, researchers often have used facial coding systems which have their roots in subjective or theoretical beliefs about what emotions should look like (e.g., Izard & Dougherty, 1980; Matias & Cohn, 1993). However, a comprehensive, anatomically based system which measures all possible facial movement is available. The Facial Action Coding System (FACS: Ekman & Friesen, 1978) has been used the most extensively with dozens of empirical articles supporting its validity (Rosenberg, 1997).  Behavioural expression of fear. 39  Oster and Rosenstein (1995) adapted the original FACS for infants, creating BabyFACS. A n adaptation was deemed necessary as it was noted that infant faces have different morphological features that make the adult version of the FACS system inappropriate (Oster et al, 1992). This is consistent with the growing realization that children's facial expressions may differ qualitatively from those of adults. Typically, Baby-FACS has been used with infants (e.g., Lilley et al., 1997), but, given that it encompasses all the features of F A C S while sensitizing coders to differences in facial structure in the growing child, it is reasonable to assume that it would be appropriate for use with young children too. In addition, Baby-FACS is the only comprehensive system available for populations other than adults. It has also maintained the most important feature of the FACS - it allows for the discovery of new configurations of movement which may be specific to situations or emotions. Therefore, Baby-FACS was judged to be the ideal system for reducing the fear display to its component structure. Overview of the Study. Hospitals are unfamiliar and threatening places to children in and of themselves, particularly so when children must undergo numerous procedures (Astin, 1977). Admission, blood withdrawal, intrusive procedures, transport to the operating room, anesthesia induction, and returning from the recovery room have repeatedly been demonstrated to be scary situations for young children (e.g., Broome & Hellier, 1987; Visintainer & Wolfer, 1975). Young children experience even greater threats to their  Behavioural expression of fear. 40  emotional well-being than older children as they are often isolated from caregivers, are unable to verbally communicate their level of discomfort to proxy caregivers, and may even be afraid of negative repercussions if they do communicate their distress (Craig, Lilley & Gilbert, 1996). Further, health care practitioners, e.g., nurses and physicians, are placed in a difficult position when they attempt to determine what young children may be feeling. Indeed, lack of vocabulary, comprehension, and production capacities in young children often make it difficult for adults to accurately label children's emotions (Izard, 1994). By deterniining the nonverbal behaviours constituting a fear response, health care providers may be better able to help alleviate fear in young children. The purpose of this study was to a) determine what constitutes an expression of fear in young children using comprehensive facial and behavioral coding systems and b) examine possible gender and age changes in the fear expression. Measurement systems included the Baby-FACS, to examine facial activity, and the Observational Scale of Behavioral Distress (OSBD), to assess other verbal and non-verbal responses to fear. The fear elicitor was removal of a cast with an oscillating saw. Hypotheses 1.  Young children will display a cluster of discrete facial behaviours in response to cast removal that can be interpreted as fear.  2.  Young children will display a set of global behaviours in response to cast removal that can be interpreted as fear.  Behavioural expression of fear. 41  3.  Older children will display a less intense reaction, marked by a lower facial action factor score and OSBD scale score.  4.  Girls will have higher facial action factor scores and OSBD scale scores. Method  Participants A total of 116 children between the ages of 12 months and 87 months (M  =  50.63), all of whom had a cast removed, participated in this study. Tables 1 and 2 provide complete demographic information on the children and the participating parent. The majority of children and parents, as reported by a parent, were either Caucasian or Asian. Nine children were said to have a chronic illness (osteogenesis imperfection, n = 4; asthma, n = 2; celiac disease, n = 1; diagnosis not reported, n = 2) and four children were taking medication (cortisone cream, n = 1; Pulmicort, n = 1; Tylenol, n = 1; medication not reported, n = 1). According to the parent's report of education level and occupational status, using Hollingshead's two factor index of social position (1977), the mean social class of the families in this study was 3.23 (SD = .92; Range 1 - 5), suggesting most families were middle class.  Tables 1 & 2  All children were recruited through the British Columbia Children's Hospital with parents providing consent and the children assenting to participate. Children were  Behavioural expression of fear. 42  Table 1 Child Demographic Information. Demographic Variable Child's Age in Months  M = 50.63  Child's Gender  Males = 53.4%  SD = 20.54  Females = 46.6% Child's Ethnic Identity  Caucasian = 61.0% Asian - 16.2% Other = 12.4% Indo-Canadian = 6.7% Native American = 2.9% African American = 1.0%  Diagnosis of a Chronic Illness  8.8% (n = 9)  Children Currently Taking Medication  3.4% (n = 4)  Behavioural expression of fear.  Table 2 Parent Demographic Information.  Demographic Variable Parental Marital Status  Married = 79.4% Divorced/separated = 8.8% Never Married = 7.8% Remarried = 1.0% Other = 2.9%  Relationship to Child  Mother = 68.9% Father = 29.1%  Parent's Ethnic Identity  Caucasian = 64.4% Asian - 17.8% Indo-Canadian = 6.9% Native American = 5.0% Other = 5.0% African American =. 1.0%  Parent's Age  M = 35.09 SD = 5.63  Socio-economic Status  M = 3.23 SD = .92 Range = 1-5  Behavioural expression of fear. 44  accompanied by at least one parent. Both parents and children were required to be fluent in the English language. In total, 15 families declined to participate. Children with major developmental difficulties, e.g., Down's Syndrome, and/or congenital facial abnormalities were excluded. Measures Coding Systems. Facial actions were coded using Baby-FACS (Oster & Rosenstein, 1995). This is an adaptation of the Facial Action Coding System (FACS; Ekman & Friesen, 1978), which assesses a maximum of 46 specific action movements without reference to global patterns indicative of emotions. Oster and Rosenstein (1995) adapted the original, comprehensive and anatomically based FACS coding system by adding actions and providing alternative descriptions of what coders may see in infants. A n example of an adaptation would be cautioning coders that infants and young children often have more full and firm cheeks than adults due to extensive deposits of subcutaneous fat and connective tissue. In this study, one principle Baby-FACS coder, who had fulfilled the rigorous training requirements and was considered to be a reliable Baby-FACS coder, completed all the facial action coding. In addition, reliability in this study, as per Ekman and Friesen's suggestion (1978), was calculated using Ekman and Friesen's reliability calculation, a conservative means of calculating reliability because it does not include actions that were not coded by either coder. Reliability between the principal coder and another proficient coder, on 20% of the data was considered to be .77. It is important to note that  Behavioural expression of fear. 45  (2 x Actions coded by both Coder 1 and 2 as present) Reliability = Actions coded by Coder 1 + Actions coded by Coder 2 agreement was defined as both coders reporting that the same action occurred in the same specific time interval.  This is considered an acceptable level of agreement by Ekman and  Friesen (1978). Further, the principal coder was not aware of which children would be coded for reliability purposes. Using the same calculation, the reliability for each of the 21 facial actions entered into the Principle Components Analysis was determined. Reliability ranged from .67 1.00. More general behaviours were coded using the Observational Scale of Behavioral Distress (OSBD; Jay & Elliott, 1986; Jay et al., 1983). The OSBD codes 10 complex behaviors: verbal fear, cry, scream, restraint, verbal resistance, information seeking, verbal pain, flail, muscular rigidity, and emotional support (see Appendix A for a complete description of each behaviour). It has been found to be both a reliable and valid measure of fear and distress (Jay et al., 1983). In this study presented here, reliability between the principle OSBD coder and the reliability coder, using Ekman and Friesen's reliability calculation (1978), was .93 across all categories. As with the facial action data, reliability was calculated on 20% of the data. The reliability for each of the six OSBD behaviours used in the analyses were: cry = .96, scream = .92; physical restraint = .97; verbal resistance = .90; emotional support = .91; and flail = .83. The OSBD coders were not the same as those who coded the Baby-FACS data.  Behavioural expression of fear. 46  Questionnaires. Parents completed a demographic questionnaire and another documenting their child's past medical history (see Appendix B & C). The child medical history questionnaire yielded two separate variables: 1) quantity of medical experience (i.e., number of times the child had undergone common medical procedures, e.g., doctor's appointments, as well as more infrequent ones, e.g., surgery), and 2) quality of medical experience (i.e., an estimate of the child's reaction to each medical procedure he/she had undergone). The research assistant also completed a chart review to provide information concerning the type of fracture, the length of time the child was in the cast, and whether pain medications had been prescribed (see Appendix D). Procedure Families were given an information sheet by the clinic secretary when they came to the casting clinic to have their child's cast removed. This information sheet summarized the study and requested their participation. If parents expressed an interest in hearing more about the study, the clinic secretary notified the research assistant. The research assistant would then describe the study in more detail and seek written consent from parents and assent from children older than three years of age. If consent and assent, when applicable, were provided, parents were asked to complete questionnaires. The chart review was completed by the research assistant following the completion of the participant's involvement. Videotaping was completed in the casting room. Two video cameras examining the child's full body and facial reaction, one capturing the former, the other the latter,  Behavioural expression of fear. 47  were used. The full body camera was in a fixed location where the best view of the child on the cast removal bed was possible; the camera for recording the facial reactions was hand - held. Children were instructed to act normally and pretend, as much as possible, the video cameras were not present. A baseline phase for each of the coding systems was captured with the one hand held camera operated by the research assistant on a set of chairs in a different part of the cast removal room. The baseline phase was recorded prior to cast removal. The research assistant video-recorded 20 seconds of each child's face and then 20 seconds of each child's full body prior to the start of cast removal after asking the children to remain as still as possible. The baseline phase served as an estimate of each child's neutral face and it represented, as much as possible, an unemotional reaction. Given this, an independent observer rated the child's reaction during the baseline phase on a ten-point scale (1 = negative reaction, 5 = neutral reaction and 10 = positive reaction). Up to a maximum of 60 seconds following the moment when the saw first touched the cast constituted the fear situation. At the conclusion of the cast removal, the technician was asked to rate how scared he felt the child had been on a ten point Visual Analogue Scale (VAS; 1 = not at all scared and 10 = very scared). Parental presence and the use of comforting or distraction (e.g., earplugs, walkman) were documented. Selfreport data was not collected from the children as it has not been shown to be a valid measure in children less than 7 years of age (Chambers & Craig, 1998).  Behavioural expression of fear. 48  For coding purposes, the fear segment start and end time were identified on the two videotapes (one for the face, one for the entire body) by a research assistant who manually viewed the tape and noted when the oscillating saw first touched the cast. The research assistant then calculated the start and end times for a maximum of three, 20second fear segments and each baseline segment for the Baby-FACS and OSBD tapes. The baseline segments for the Baby-FACS and OSBD coding were identified vocally on the facial activity tape when the child was videotaped to help the research assistant identify the appropriate 20-second baseline segments. Time codes for each segment were then randomly ordered on a coding sheet and each coder coded the segments in the random order presented. When using Baby-FACS, the coder was blind to the event the child was experiencing; however, when coding the OSBD behaviours, environmental cues made it impossible for the coder to remain blind. For the tape capturing facial actions, the coder examined each 20-second segment in two-second time periods to establish the presence or absence of each action unit for the facial coding system. For the tape capturing the whole body, coders examined each 20-second segment in its entirety for the occurrence of any of the actions described in the OSBD. Finally, videotapes were viewed by one individual who was blind to the hypotheses of the study in order to rate the amount of fear each child expressed during cast removal. This person rated the amount of fear expressed during each of the fear segments with a 10 point V A S scale where 1 = no fear and 10 = most amount of fear possible.  Behavioural expression of fear. 49  Manipulation Checks Both prior to and after data collection, manipulation checks using Visual Analogue Scales (VAS) were completed to ensure that cast removal elicited fear responses in children. Research on pain has determined that a rating of 3 on a 10 point V A S , where 1 = no pain and 10 = worst pain possible, is clinically significant (Collins, Moore, & McQuay, 1997; Gauthier, Finley, McGrath, 1998). Indeed, three is considered an appropriate rating at which to medicate a child or adult for pain (Gauthier et al., 1998). Although similar research has not been conducted with fear, applying the same guidelines to fear seemed appropriate as, in the treatment of child anxiety disorders, children are often taught to rate the situations they are afraid of and often avoid (e.g., touching aunt with cancer, eating food that is considered contaminated) on a "fear thermometer" (March & Mulle, 1998). On this tool, a three is considered the start of fearful behaviours and is labeled "I am a little uneasy" (p. 211). Further, a rating of three is considered a reasonable level at which to encourage children to use coping strategies (e.g., deep breathing, imagery). Therefore, a fear rating of three in this study was considered clinically significant. The first manipulation check occurred prior to the start of the main study. Twenty children between the ages of 7 and 15 years were asked, immediately following removal of their cast, how scared they had been during cast removal. These children, on average, gave self-report ratings of 3.2 on a 10 point scale (1 = not at all scared; 10 = very scared), with 12 children rating their fear as greater than or equal to three.  Behavioural expression of fear. 50  Although over half of these children reported a clinically significant amount of fear, these ratings are likely an underestimate of the amount of fear the children actually experienced. Recent research has shown that children, when asked to rate their pain, often underreport the actual amount of pain they are experiencing (Chambers et al., 1999). Children often underreport their level of pain because of fear of possible consequences (e.g., if I tell the nurse it hurts, I will have to get a needle; Craig, Lilley, & Gilbert, 1996). Although the consequences are different in this situation (e.g., implicit expectations of being labeled a "scaredy cat"), these children also likely underreported the level of fear experienced. Second, after data was collected, a research assistant rated, on a 10 point V A S , how fearful the videotaped children appeared to be over all three-fear segments. The mean rating was 3.72 (SD = 3.47) and 47% of these ratings were greater than or equal to 3.0. Again, although roughly half of the children in the study were rated as experiencing clinically significant fear, the research assistant may have underestimated the actual amount of fear experienced by the children. Indeed, given that a) parents, when asked to rate their child's level of pain, undervalue the amount of pain their child is actually experiencing (Chambers et a l , 1999), and b) parents are less accurate at reporting child internalizing symptoms, e.g., depression, anxiety, (Achenbach, McConaughy, Howell, 1987), it is probable the observer, who would be even unfamiliar with the child and his or her behavioural signals, may have underrated the children's level of fear.  Behavioural expression of fear. 51  Statistical Analyses Two main sets of analyses were performed, one with the Baby-FACS data, i.e., facial activity, and one with the OSBD data, i.e., broadband behaviours. Analysis One: Facial Action Data. The analysis followed several steps: 1) The fear reaction was captured during the first 60 seconds after the saw first touched the cast. It was concluded that if children displayed a similar level of activity during a fear segment and the baseline phase* the children were no longer reacting to the stimulus and data from those fear segments should be excluded from further analyses. Further, a decrease in the amount of facial activity over the three fear segments was expected, but, as long as the fear segments were significantly different from baseline, the facial activity data was considered acceptable. Therefore, to establish whether more facial activity occurred during the fear segments than during baseline, a one-way repeated measures analysis of variance (ANOVA) with follow-up tests was performed. This test compared the total number of facial actions coded during each of the four segments, baseline, fear segment 1, fear segment 2, and fear segment 3. Only data from those fear segments which significantly differed from baseline were used in subsequent analyses. 2) A facial action score was created for each action coded during each of the significant fear segments by counting the number of times the action was coded as present. Each facial action score from the first fear segment was then summed with the corresponding facial action scores from the other significant fear segments to create total facial action scores. 3) The frequency with which each facial action was displayed was then investigated. Actions  Behavioural expression of fear. 52  which occurred less than 5% of the time were dropped from further analyses as it was decided that these actions would have limited clinical implications. 4) The remaining total facial action scores were then subjected to a Principal Components Analysis (PCA). A factor score was calculated by summing the total number of actions coded as present for each action that loaded on the factor greater than or equal to 0.35. This created a facial action factor score that served as the dependent variable in farther analyses. 5) Steps 3 and 4 were repeated using only those children who were rated as experiencing clinically significant fear (i.e., V A S ^3). 6) Pearson product-moment correlation coefficients were calculated to describe the relationship between each predictor (age, gender, and cast area) and the facial action factor score from the total sample PCA. A standard multiple regression was used to delineate the relationships among the predictor variables and the facial action factor score. All predictor variables were entered together. The contribution of predictors was presented as standardized regression coefficients (standardized P). 7) The convergent validity for the facial action factor scores was evaluated by examining the bivariate correlation between mean V A S ratings for the fear segments and the facial action factor score. Validity was further assessed by examining the relationship between the cast technician's ratings and the facial action factor score. Analysis Two: OSBD Data. As with the facial action analyses, the OSBD analyses followed several steps: 1) Again, it did not make conceptual sense to include OSBD data from a fear segment which was not significantly different from baseline. Therefore, to establish whether more global behaviours occurred during the fear segments than during  Behavioural expression of fear. 53  baseline, a one-way repeated measures analysis of variance (ANOVA) with follow-up tests was performed comparing the total number of OSBD actions coded during each of the four segments, baseline, fear segment 1, fear segment 2, and fear segment 3. Only the data from those fear segments which significantly differed from baseline were used in subsequent analyses. 2) A n OSBD behaviour score was created for each behaviour, for each significant fear segment, by counting the number of times each behaviour was coded. Each behaviour score from the first fear segment was then summed with the corresponding behaviour scores from the other significant fear segments to create total behaviour scores. 3) Using the procedure outlined in the OSBD manual (Jay & Elliott, 1986), the frequency of each of the 10 total behaviour scores was examined and bivariate correlations between the individual behaviours and the total OSBD score were computed. Those behaviours which occurred more than 5% of the time and had a correlation with the OSBD total score greater than .30 were included in further analyses. A n overall OSBD scale score was then computed by summing the data from the total behaviour scores which met the inclusion criteria. 4) The same procedure was then completed using only the data from those children who were rated as experiencing clinically significant fear (i.e., V A S >3). 5) Pearson product-moment correlation coefficients were calculated to describe the relationship between each predictor variable (age, gender, and cast area) and the OSBD scale score (dependent variable). A standard multiple regression was used to delineate the relationships among the predictor variables and the OSBD scale score. All predictor variables were entered together. The contribution of predictors was presented as  Behavioural expression of fear. 54  standardized regression coefficients (standardized P). 6) Validity of the OSBD scale score was evaluated by examining the correlations between the mean V A S ratings for the fear segments and the OSBD scale score, and the cast technician's ratings and the OSBD scale score. Sample Size Justification Factor analyses are improved when a wide variety of people are present in the sample (Gorsuch, 1997; Tabachnick & Fidel, 1996), and, in the past, the recommendation for sample size was given as a function of the number of variables (e.g., 10 cases for every variable). However, information now exists which suggests that the "10-case-per-item" rule sets the sample size above the niinimum needed and that a ratio of 3 to 1 is sufficient for up to 40 - 50 variables (Gorsuch, 1997; Guadagnli & Velicer, 1988). Gorsuch (1997) has authoritatively stated the current consensus as follows, "the sample size needed is a function of the stability of a correlation coefficient without any correction needed for the number of variables, at least for situations of up to 40 or 50 variables" (p. 541). However, Gorsuch recommends that researchers check whether the sample size is too small for the factors extracted by using Bartlett's significance test (Bartlett's Test of Sphericity) where any factor of interest should be highly significant (1997). In this study only 21 variables were entered into the overall P C A and Bartlett's Test of Sphericity determined that the sample size of 113 was appropriate. The sample size and the number of variables entered in the second P C A (60 children with a VAS'2:3, 16 actions) was also determined to be appropriate by Bartlett's Test of Sphericity.  Behavioural expression of fear. 55  Results The children in the study had fractures to either their arm (n = 65), wrist (n = 9), leg (n = 34), or ankle (n = 5). Ninety percent of the fractures were simple. Seventeen children required surgery to set the fracture. All participants were outpatients. The mean number of days children spent in the cast was 24.06 (SD = 11.39). In general, the child's parent was present during cast removal (96.4%); however, only 38.1 % and 24.8% of these parents were observed to engage in soothing or distraction behaviours, respectively, as rated by the research assistant. The casting technician, on average, rated the fear level as a 3.22 on a ten-point V A S with 1 = no fear and 10 = most amount of fear possible. Forty-five percent of the cast technician's ratings were greater than or equal to three. Facial Action Analysis Preliminary Facial Action Analyses. As discussed earlier, it was important to determine whether the amount of facial activity occurring during the fear segments was significantly different from baseline. First, an independent observer rated the baseline segments on a ten-point V A S with 1 = negative, 5 = neutral, and 10 = positive. 71.9% of the children were rated as having a neutral expression (VAS = 4-6). Second, the total number of facial actions coded during each of the four different segments (i.e., baseline and the three fear segments) was calculated. These totals were then compared using a one-way, repeated measure A N O V A . The analyses indicated there was a significant difference between these totals (F (3) = 21.60; g <.0001; see Table 3). Tukey's follow-up tests showed that fear segments 1 and 2 were significantly different from baseline  Behavioural expression of fear. 56  Table 3  (HSD = 10.62; p <.001; HSD = 5.26; p <.01). Fear segment 3 was not (HSD = 1.06; p > .10). Fear segment 1 was also significantly different from fear segment 2 and 3 (HSD = 5.36; p < .01; HSD = 8.77; p < .001). Fear segment 2 did not differ from fear segment 3 (HSD = 3.42; p>. 10). As indicated earlier, a decrease in facial activity was expected across the fear segments. However, it was decided that as long as fear segments were significantly different from baseline, summing the data from those fear segments was acceptable. Therefore, data from fear segments 1 and 2 were used in all subsequent analyses. Factor Structure of Facial Actions. Only 32 of the Baby-FACS actions were coded as present during the fear segments. The number of times each action was coded as present during fear segment 1 and 2 was calculated (i.e., facial action scores). The facial action scores from each segment were then summed together to create total facial action scores. A l l subsequent analyses used these total facial action scores. Frequency counts for each of the total facial action scores were then examined. Eleven facial actions occurred less than 5% of the time and were dropped from further analyses as per the earlier rationale. The Bartlett Test of Sphericity was significant, indicating that there were significant correlations among the variables so as to warrant a PCA. Therefore, to determine whether any of the remaining facial actions clustered  Behavioural expression of fear.  Table 3 Means and Standard Deviations for the Facial Action Baseline and Fear Segments. Mean  SD  Baseline  17.80  15.15  Fear Segment 1  31.19  21.70  Fear Segment 2  24.44  16.03  Fear Segment 3  20.13  7.17  Segment  n= 116  Behavioural expression of fear. 5 8  together and which actions best represented a facial expression of fear, a Principle Components Analysis of the total facial action scores for the remaining 21 facial actions was undertaken. See Appendix E for a description of these facial actions. The most appropriate solution to the P C A was a single factor with an eigenvalue of 5.87. A single factor was deemed most suitable as both the scree plot and the percent of variance accounted for by subsequent factors suggested one main factor. Table 4 presents the eigenvalues and the percent variance for all factors extracted. The factor was not rotated as this was a one factor solution and the first factor provides the most internally consistent and thus most reliable combination of a set of variables (Tabachnick & Fidell, 1996).  Table 4  The factor accounted for 28% of the variance in Baby-FACS actions. Thirteen actions loaded positively on the factor and had factor loading values greater than .35 (see Table 5). The actions were, in rank order of importance, brow knot and lower, nose wrinkle, lip stretch, squint, brow knit, cheek raise, chin raise, tongue shapes, mouth stretch, nasolabial furrow, tongue protrusion (without touching lips), eyes closed, and lower lip depress. The single factor solution indicated that a unitary construct, interpretable as fear, underlies the various facial actions coded. Follow-up analyses were conducted using the sum of the number of times each significant action was coded as  Behavioural expression of fear. 59  Table 4 Eigenvalues and Percent of Variance Accounted for by the Factors Extracted in the Principle Components Analysis. Component  Eigenvalues  Percent of Variance  1  5.9  27.93  2  2.1  10.04  3  1.6  7.78  4  1.4  6.46  5  1.3  6.17  6  1.2  5.89  7  1.1  5.15  Behavioural expression of fear. 60  present. This index represented the amount of fear expressed.  Table 5  Factor Structure of the Facial Actions Using Children who Displayed Clinically Significant Fear TV AS > 3). Frequency counts for the 21 facial actions entered into the total sample P C A were examined using the data from only those children who were rated as displaying clinically significant fear was used (n = 60). Five of these facial actions occurred less than 5% of the time and were dropped from further analyses. The most appropriate solution to the P C A was a single factor with an eigenvalue of 5.9. A single factor was deemed most suitable as both the scree plot and the percent of variance accounted for by subsequent factors suggested one main factor. Table 6 presents the eigenvalues and the percent variance for all factors extracted. This factor, as with the  Table 6  earlier PCA, was not rotated as this was a one factor solution and the first factor provides the most internally consistent and thus most reliable combination of a set of variables (Tabachnick & Fidell, 1996). The factor accounted for 37% of the variance in Baby-FACS actions. The same 13 actions from the total sample P C A loaded positively on the factor and had factor  Behavioural expression of fear. 61  Table 5 Facial Action Factor Loadings. Facial Action  Loading  Brow Knot and Lower (FA #4)  .83  Nose Wrinkle (FA # 9)  .82  Lip Stretched Laterally (FA # 20)  .81  Squint (FA # 44)  .79  Brow Knit (FA #3)  .77  Cheek Raise (FA #6)  .73  Chin Raise (FA # 17)  .64  Tongue Shapes (FA #75)  .64  Lip and Jaw Opening (FA # 26)  .57  Nasolabial Furrow Deepener (FA #11)  .55  Tongue Protrusions (FA #19)  .44  Eyes Closed (FA # 43)  .42  Lower Lip Pulled Down (FA # 16)  .41  Behavioural expression of fear. 62  Table 6 Eigenvalues and Percent of Variance Accounted for by the Factors Extracted in the Principle Components Analysis for Children Displaying Clinically Significant Fear. Component  Eigenvalues  Percent of Variance  1  5.9  36.86  2  1.5  9.38  3  1.5  9.26  4  1.3  7.97  5  1.1  7.16  Behavioural expression of fear. 63  loading values greater than .35 (see Table 7). The actions were, in rank order of importance, brow knot and lower, lip stretch, nose wrinkle, cheek raise, squint, brow  Table 7  knit, mouth stretch, tongue shapes, chin raise, nasolabial furrow, tongue protrusion (without touching hps), lower lip depress and eyes closed. The single factor solution indicated that a unitary construct, interpretable as fear, underlies the various facial actions coded. It is important to note that the children in the clinically significant fear sample did not differ in terms of age and, as with the total sample, a larger percentage of children had casts on their arms or wrists and there were more males than females. Relationship between the Facial Action Factor Score and Background Variables. Relationships between the total sample facial action factor score and background variables were examined using bivariate correlations, t-tests, or one-way A N O V A s . No significant relationships were found between the facial action factor score and quantity of medical experience, quality of medical experience, ethnicity, length of time the child was in his/her cast, parental presence in room, mother versus father present in the room, or the number of people in the casting room (a > .05).  Behavioural expression of fear.  Table 7 Facial Action Factor Loadings for Children Displaying Clinically Significant Fear. Facial Action  Loading  Brow Knot and Lower (FA #4)  .87  Lip Stretched Laterally (FA # 20)  .84  Nose Wrinkle (FA # 9)  .83  Cheek Raise (FA # 6)  .82  Squint (FA #44)  .78  Brow Knit (FA #3)  .75  Lip and Jaw Opening (FA # 26)  .66  Tongue Shapes (FA #75)  .6.3  Chin Raise (FA # 17)  .55  Nasolabial Furrow Deepener (FA #11)  .48  Tongue Protrusions (FA #19)  .47  Lower Lip Pulled Down (FA #16)  .42  Eyes Closed (FA # 43)  .35  64  Behavioural expression of fear. 65  The Prediction of Children's Fear. The independent variables entered into the multiple regression analysis were as follows: (1) the child's age in months, (2) the child's gender (female = 1, male = 0), and (3) the location of the cast (arm = 1, not arm, i.e., leg, = 0). Means and standard deviations for each of these variables are listed in Table 8. The dependent variable was the facial action factor score derived from the P C A on the total sample (n = 113).  Table 8  The regression analysis was conducted using standard multiple regression strategies. The three predictors were entered in one block; the order of entry was not specified. Data were screened with respect to the assumptions of multiple regression. As the facial action factor score was found to be moderately and positively skewed, a square root transformation was applied in order to improve normality, linearity and homoscedasticity (Tabachnick & Fidell, 1996). Pearson product-moment correlations were calculated between all pairs of variables in the analysis (see Table 9).  Table 9  Behavioural expression of fear. 66  Table 8 Descriptive Statistics for the Prediction of Children's Fear. Mean  SD  Facial Action Factor Score  5.62  2.79  Child's Age in Months  50.6  20.54  Child Gender  Female:46.6%  Male: 53.4%  Cast Location  Arm: 65.5%  Leg: 34.5%  Continuous Variables  Categorical Variables  n= 113  Behavioural expression of fear. 67  Table 9 Bivariate Correlations in the Prediction of Children's Fear.  Fear  Fear  Age  Gender  Cast Location  1.00  -34**  -.12  -.27*  .13  .26*  1.00  .11  Age Gender Cast Location n= 113  Note: * = E<.01; ** =E<.001.  1.00  1.00  Behavioural expression of fear. 68  The results of the regression analysis are presented in Table 10. Age, gender, and cast location had significant predictive power, accounting for 15% of the variance (F = 6.57; p_<.001) for the total facial action factor score. According to the standardized regression coefficients (standardized P), significant independent contributions to the prediction of fear were made by child's age and cast location . 1  Table 10  It is important to note that age and cast location were also significantly and negatively correlated with each other suggesting that older children were more likely to have arm casts than younger children. However, as regression analyses evaluate the unique and independent contribution each predictor makes and cast location had a significant standardized P, cast location, on its own, should be considered an important predictor of the fear expression. Validity of the Facial Action Factor Score. A n independent observer made global V A S ratings of fear for each of the three fear segments; however, as only the first two fear segments were significantly different from baseline, only ratings from these two segments  1  Regression analysis conducted only on those children displaying clinically significant fear  produced the same results (see Appendix F).  Behavioural expression of fear.  Table 10 Standard Multiple Regression Analysis: The Prediction of Children's Fear. Variable  Block One: Age  R_  F  .15  6.57*** •  Standardized [3  -.28**  Gender  -.07  Cast Location  -.19*  _n= 113 Note: Standardized betas reflect the unique contribution of the variable controlling for other variables in the regression model. * -p<.05. ** =p<.01. *** = p<.001.  69  Behavioural expression of fear. 70  were used. The mean fear rating was 4.39 (SD = 3.05; range = 1-10), with 58% of the ratings falling within the clinically significant range. Overall, these global ratings of fear were significantly and positively correlated with the total facial action factor score (r = .60; p <,001). Similarly, the cast technician's ratings of fear were significantly and positively correlated with total facial action factor score (r = .47; p < .001). Global ratings of fear were significantly and positively related to each of the 13 facial actions (Range: r = .24 -.58; p_<.05 - p<.001) with the exception of lower lip pull down (r = .16; p_= 10) and tongue protrusion (r = .15; p =.13). Similarly, cast technician's ratings of fear were significantly and positively related to each of the 13 facial actions (Range: r = .21 -.55; p_<.05 -p<.001) with the exception of lower lip pull down (r = .18; g=.06) and tongue protrusion (r = 17; p=.08). Further, global ratings of fear were significantly and negatively related to child's age in months (r =-.47; p<.01) and children with casts on their legs had higher V A S ratings than those children with cast on their arms (t (105) = -2.52; g<.01; M  =  5.2, sd =  3.3; M = 3.8, sd = 2.7). Cast technician's ratings were also related to child's age in months (r = -.24; g<.05) and, as with global ratings, children with leg casts were rated as displaying more fear than children with arm casts (t (108)= -2.56; p_<.01; M = 4.1, sd = 3.0; M = 2.8, sd = 2.1). Global Behaviour Analysis Preliminary Global Behaviour Analyses. As discussed earlier, it was important to establish whether global behaviours occurred more frequently during the fear segments  Behavioural expression of fear. 71  than during baseline. To establish this, the number of times an OSBD behaviour was coded was calculated for each of the four different segments (i.e., baseline and the three fear segments). These scores were then compared using a one-way, repeated measures A N O V A . The analyses indicated there was a significant difference between these scores (F (3) = 25.73; p <.0001; see Table 11). Tukey's follow-up tests showed that  Table 11  fear segments 1, 2 and 3 were all significantly different from baseline (HSD = 10.19; p <.001; HSD = 8.95; p <.01; HSD = 10.89; p <.001) but that none of the three fear segments differed from each other. Therefore, data from all three fear segments was used in subsequent analyses. Scale Structure of the Broadband Behaviours. The number of times each of the ten OSBD behaviours were coded as present in each of the three fear segments was calculated. The scores for each behaviour, for each segment, were then summed together to create ten total behaviour scores. To ensure consistency with previous studies which have used the OSBD coding system, the same procedure as outlined by Jay and Elliott (1986) in the construction/ revision of the OSBD was used. The manual lists the following as elimination criteria: (1) category scores occurring in less than 5% of the subjects, and (2) category scores with an  Behavioural expression of fear.  Table 11 Means and Standard Deviations for the OSBD Fear Segments. Mean  SD  Baseline  .21  .61  Fear Segment 1  .81  1.15  Fear Segment 2  .74  .95  Fear Segment 3  .85  .52  Segment  n= 113  Behavioural expression of fear. 73  item-total correlation of less than +.30 for the total sample. Following the Jay and Elliott procedure (1986), the ten total behaviour scores were scored for frequency of occurrence and correlated with the total OSBD score (the sum of the ten total behaviour scores). Using the criteria described above, the following behaviours were dropped from further analyses: information seeking, muscle rigidity, verbal pain, and verbal fear (see Table 12). A reliability analysis was then conducted on the six remaining behavioural categories (a =  Table 12  .70). This is consistent with the alpha internal consistency coefficient of .72 reported in the OSBD manual (Jay & Elliott, 1986). To obtain a single index representing the amount of fear expressed, these six total behaviour scores were summed together to create an OSBD scale score. Scale Structure of the Broadband Behaviours with Children who Displayed Clinically Significant Fear (VAS > 3). The same procedure as described above was undertaken with this sub-sample of the data (n = 55). Information regarding the frequency of occurrence and the correlation between each individual OSBD behaviour and the total OSBD score for this sample can be found in Table 13. After inspection of this data, the same OSBD behaviours were dropped from further analyses and an alpha internal consistency was calculated (a = .62). To obtain a single index representing the amount of  Behavioural expression of fear. 74  Table 12 Frequency of OSBD Behaviours and Correlations of Individual Behaviours with Total OSBD Score. OSBD Behaviour  Frequency (%)  Correlation with Total OSBD Score  Cry  33.6  .77*  Scream  14.5  .76*  Restraint  10.0  .70*  Verbal Resistance  5.50  .57*  Emotional Support  12.7  .53*  Flail  5.00  .31*  Verbal Fear  3.60  .28  Verbal Pain  7.30  .26  Muscle Rigidity  1.80  .08  Information Seeking  1.80  -.003  Note. * = E<.01.  Behavioural expression of fear. 75  fear expressed, these six total behaviour scores were summed together to create an OSBD scale score for this sample.  Table 13  Relationship between the OSBD Scale Score and Background Variables. The relationships between the total sample OSBD scale score and background variables were assessed, when appropriate, with t-tests, one-way A N O V A s and bivariate correlations. No significant relationships were found between the OSBD scale score and ethnicity, quantity of medical experience, quality of medical experience, length of time the child was in his/her cast, parental presence in room, mother versus father in the room, or the number of people in the casting room (a = .05). The Prediction of Children's Fear. Independent variables entered into the regression analysis were as follows: (1) the child's age in months, (2) the child's gender, and (3) the location of the cast (leg or arm). Means and standard deviations for each of these variables are listed in Table 14. The dependent variable was the total sample OSBD scale score.  Table 14  Behavioural expression of fear. 76  Table 13 Frequency of OSBD Behaviours and Correlations of Individual Behaviours with Total OSBD Score for the Clinically Significant Fear Sample. OSBD Behaviour  Frequency (%)  Correlation with Total OSBD Score  Cry  63.0  .69**  Scream  25.9  73**  Restraint  20.4  .69**  Verbal Resistance  9.30  .60**  Emotional Support  22.2  44**  Flail  7.50  .31*  Verbal Fear  7.40  .22  Verbal Pain  11.1  .28*  Muscle Rigidity  0.00  .11  Information Seeking  3.70  -.10  Note. * = p < .05; ** = p <.001.  Behavioural expression of fear. 77  Table 14 Descriptive Statistics for the Prediction of Children's Fear. Mean  SD  OSBD Scale Score  0.72  0.94  Child's Age in Months  50.6  20.45  Child Gender  Female:45.6%  Male: 54.4%  Cast Location  Arm: 65.5%  Leg: 34.5%  Continuous Variables  Categorical Variables  Behavioural expression of fear. 78  The regression analysis was conducted using standard multiple regression strategies. The three predictors were entered in one block; the order of entry was not specified. Data were screened with respect to the assumptions of multiple regression. As the OSBD scale score was found to be moderately and positively skewed, a square root transformation was applied in order to improve normality, linearity and homoscedasticity (Tabachnick & Fidell, 1996). Pearson product-moment correlations were calculated between all pairs of variables in the analysis (see Table 15).  Table 15  The results of the regression analyses are presented in Table 16. Age, gender, and cast location had significant predictive power, accounting for 22% of the variance (F = 10.33; g<.001) for the OSBD scale score. According to the standardized regression coefficients (standardized P), significant independent contributions to the prediction of fear were made by child's age and cast location.  Table 16  Validity of the OSBD Scale Score. As all three fear segments for OSBD activity were significantly different from baseline, mean V A S ratings from all fear segments were  Behavioural expression of fear. 79  Table 15 Bivariate Correlations in the Prediction of Children's Fear.  Fear  Fear  Age  Gender  Cast Location  1.00  -.43**  -.05  -.29*  .13  .26*  1.00  .11  Age Gender Cast Location  n=113 Note: * = p<.01. ** =p<.001.  1.00  1.00  .  Behavioural expression of fear. 80  Table 16 Standard Multiple Regression Analysis: The Prediction of Children's Fear. Variable  Block One:  "  R_  F  .22  10.32**  Standardized p  Age  -.38**  Gender  .03  Cast Location  -.20*  n= 1.13 Note: Standardized betas reflect the unique contribution of the variable controlling for other variables in the regression model. * =p<.05. ** =p<.001.  Behavioural expression of fear. 81  used. As with the facial activity factor score, the global ratings of fear were significantly and positively correlated with the total sample OSBD scale score (r = .89; p <.001). Similarly, the cast technician's ratings of fear were significantly and positively correlated with the OSBD scale score (r = .72; p <.001). Global ratings of fear were also positively and significantly correlated with each of the six OSBD behaviours (Range: r = .28 - .83; p<.05 - p_<.0T). Similarly, cast technician's ratings of fear were positively and significantly correlated with each OSBD behaviours (Range: r = .33 - .65; p_<.05 - p<.01), with the exception of flail (r = .14; E  =.14). Global ratings of fear and cast technician's ratings were also significantly and  negatively correlated with child's age in months (r = -.47, p<.01; r = -.24; p<.05, respectively). Cast location was also related. Both independent observer's and the cast technician rated children with leg casts as displaying more fear (t = -2.55, p <.01; M =5.23, sd = 2.69; M = 3.72, sd =2.69; t = -2.56, p<.01; M =4.05, sd =3.03; M = 2.79, sd = 2.12, respectively). The final statistical analysis performed was a bivariate correlation between the facial action factor score and the OSBD scale score. Results showed that these two behavioural scores were positively and significantly correlated (r = .49; g < .0001). Discussion Behavioural Description of Fear  Behavioural expression of fear. 82  This study demonstrated that children, between the ages of 1 and 6 years, display a constellation of facial actions and a group of global behaviours interpretable as fear. These facial actions, brow knit, brow knot and lower, cheek raise, nose wrinkle, nasolabial furrow, lower lip depress, lips stretched laterally, chin raise, mouth stretch, closed eyes, squint, tongue protrusion, and tongue shapes (see figure 1 for a characterization of a neutral face and figure 2 for a depiction of a fear face), and OSBD behaviours, physical restraint, verbal resistance, flail, emotional support, scream, and cry, occurred in response to cast removal.  Figures 1 & 2  The same facial actions and global behaviours were displayed by the children in the clinically significant fear sample. This suggests that these behaviours are indeed measuring fear as: (a) the P C A on this sub-sample of children produced the same results and (b) the broadband behaviours occurred almost twice often. The most methodologically sound prior study on the facial actions occurring during fear was conducted by Oster et al. (1996). Infants in that study typically displayed raised brows, with or without brow knit, and widened eyes when confronted with a fear evoking stimulus (a toy gorilla which emitted raspy, growly sounds). The only other study to examine youngsters' facial actions during a fear event found that infants open their  Behavioural expression of fear. 83  Figure 1.  Behavioural expression of fear.  Behavioural expression of fear. 85  mouths, stretch their lips, and widen their eyes when afraid (Hiatt et al., 1979). A n a priori list of suspected fear facial actions in children also exists. Both the M A X and A F F E X coding systems posit that infants and children who are afraid will exhibit inner and outer brow raise, brow knot and lower, upper eyelid raise, cheek raise, tightening of the eyelids (i.e., squint), lip stretch, and lips apart (Izard & Dougherty, 1980; Izard et al., 1980). The children in this study, in both the total sample and in those children rated as displaying clinically significant fear, exhibited some of the same actions either found or predicted to occur in the literature (e.g., brow knit, cheek raise, squint, lip stretch, mouth stretch), but, although these consistencies were apparent, there were notable differences. First, the children in this study did not raise their eyebrows or widen their eyes, yet infants in the Oster et al. study (1996) displayed both of these actions. Second, these children did not raise their inner and outer brows, predicted actions in both the A F F E X and M A X coding systems. Third, the children in this study closed their eyes and/or squinted whereas the infants in the Hiatt et al. study (1979) widened their eyes. Reasons for these differences may be found by closely examining the fear eliciting stimuli. Children in this study perceived the source of threat as the oscillating saw, whereas infants in the Oster et al. study (1996) reacted to a different form of threat, they were likely afraid of the toy gorilla. Perhaps fear of being hurt or fear of a medical procedure (i.e., cast removal) produces lowering of the brows and eye closing/squinting whereas a more generalized threat which encompasses many different fear sources, e.g.,  Behavioural expression of fear. 86  fear of animals, safety, physical harm, (i.e., a toy gorilla; visual cliff) leads to brow raising and eye widening. The theoretical position that the fear expression is specific to the type of situation is plausible for several reasons. First, we often see brow lowering, nose winkling, vertical and horizontal stretching of the mouth, and closed eyes or squint in children during painful medical procedures. Therefore, perhaps some of the facial actions attributed to pain may also occur during situations where a child perceives pain as likely. One would then also predict that these "pain" actions would not occur during other fear evoking stimuli where physical pain is not a possible outcome (e.g., separation from a parent). The fact that a few of the facial actions which occurred in response to cast removal, a situation where children often perceive pain as likely, resembled what one sees during pain events supports this hypothesis, at least in part. It is important to note that although there were similarities between the typical pain face and what these children displayed during a fear situation, differences existed. Children in this study pulled their lower lips down and raised their chins, actions not typically observed in pain expressions. Further, they did not display some of the actions usually seen during pain responses: upper lip raise and flared nostril. These differences highlight the fact that pain and fear are not represented by the same facial displays. The second reason this theoretical position is likely can be found in research by Campbell and Rapee (1994). They factor analyzed the Child and Adolescent Worry Scale and determined that two distinct fear factors exist, fear of physical threat (e.g., not being  Behavioural expression of fear. 87  able to breath, being blind, being hurt, breaking arm or leg) and fear of social consequences (e.g., being teased, not looking good, being late for school). Given that children rate fear of physical harm and fear of social consequences as separate constructs, it is credible that they may respond to them with different facial actions. Furthermore, Campbell and Rapee's finding that fear evoking situations fall into two separate constructs is also consistent with Endler's (1997) multidimensional interaction model of anxiety which posits that anxiety occurs in four situations: social evaluation, physical danger, ambiguous situations, and daily routines. Finally, health care professionals and caregivers are often observed to tell young children to look away during potentially painful medical procedures (e.g., needles), whereas it is considered advantageous to closely watch sources of fear like a gorilla emitting growly, raspy sounds or a stranger approaching. These two behaviours, looking away or closely watching, in and of themselves, would call for different facial actions in the eye region. This idea that situations may "call for" different behaviours is consistent with Fridlund's Behavioural Ecology View (1991; 1997) which posits that displays evolve in response to specific selection pressures. That is, if looking at something during a threatening situation is found to be more ecologically advantageous then looking away, than that behaviour should continue to be displayed in those situations. The assertion that there is situational specificity of facial expressions to different fearful poses problems for the theoretical position that facial actions of fear are a unitary phenomenon. According to the Differential Emotions Theory (DET; e.g., Izard, 1992;  Behavioural expression of fear. 88  Izard, 1997), fear is a basic emotion and, therefore, it has a genetically determined pattern of universal and discrete facial actions. The idea that situation or context influences fear facial actions is in direct contrast to DET. Instead, it supports Fridlund's (1997) position that facial displays are mediated by the context of the interaction, the intent of the displayer, and the behaviour of the recipient. Indeed, the previous point that others may in fact be encouraging different behaviours during different situations (e.g., health care practitioners encouraging children to look away during painful medical procedures) strengthens Fridlund's argument that sensitivity to the reactions of others often leads to optimally adaptive behaviour (1997). Further, the occurrence of situation-dependent facial expressions reinforces Russell's (1997) standpoint that facial expressions provide only quasi-physical information, and it is the context of the situation which determines the specific emotion inferred by someone else. If fear expressions change according to the type of fear, i.e., the situation, then at least one of the basic premises of D E T is called into question and, instead, the hypothesis that context helps observers correctly infer an emotional state and may alter facial displays is supported. Indeed, the discrepancies between the findings from this study versus other studies support the contextual theories of emotional expression. In addition to the role of contextual or situational variables as determinants of the expression of fear, developmental processes are also contributing. That is, although there may be differences in facial activity across fear situations, differences across age groups may also exist. Oster et al. (1996) found, by comparing results from her research with  Behavioural expression of fear. 89  that of Ekman and Friesen (1973) that infant fear expressions were significantly different than adult expressions. That study found that infants widen their eyes, raise their brow and display more facial stilling (i.e., a cessation of facial movement, typically with a slack, slight open jaw and sometimes widely open eyes and slightly raised brows). In contrast, although adults also raise their brow and widen their eyes, adult expressions also consisted of brow knot and lower, lip stretch and jaw drop/mouth stretch (Ekman & Friesen, 1978). Not surprisingly, differences between the fear expression observed in the present study and the adult description also exist. The children in this study closed their eyes, squinted, raised their cheeks, wrinkled their noses, bowed their nasolabial furrows, and raised their chins, actions not seen in the adult fear face. Further, as described earlier, the children in this study displayed different facial actions than those actions typically found in infants. Given that developmental shifts in facial actions occur in pain expressions (Fradet et al., 1990; Goodenough et a l , 1997) changes in fear facial actions are also likely sensitive to developmental age. Consistent with this, research has found that younger children endorse more worry items when the focus is on the consequences of physical harm than social consequences (Campbell & Rapee, 1994). Therefore, if age is related to the type of items children endorse as fearful, it is plausible to think that age may also be related to the actions children display. A developmental shift in fear expressions occurring across the life span undermines the DET position that facial displays of emotion, including fear, are universal, discrete, and innate. Instead, it suggests that facial expressions are subject to many of the same  Behavioural expression of fear. 90  developmental processes of aging as other aspects of communication. However, instead of debating whether the differences in the facial expressions observed here are due to situational specificity or developmental changes, a parsimonious or complementary explanation of the two differing points exists. Perhaps developmental changes in the facial display of fear (i.e., changes in the actions themselves and adjustments in the intensity of the display) follow a predetermined or "hard-wired" pattern present from birth, and, although this pattern may drive the process, context may modify the display at any age. That is, perhaps infant and young children's facial expressions of fear have only a few similarities with those of an adult (differences between infants and young children would also be likely), and these differences would occur as a result of predetermined developmental shifts hard-wired into the brain. But, expressions at any age would be sensitive to the context or situation in which a facial expression is exhibited. A genetically determined developmental pattern of changes, which, at times, is sensitive to context is plausible. Indeed, other forms of communication (e.g., talking) are often affected by both processes. For example, the vocalizations of an infant are different, but related to what a young child or adult says, and few people would argue whether situational variants play an important role in when and what we say. Indeed, observers often misjudge overheard comments because they are unaware of the context in which something was said. This theoretical stance is in contrast to many of the current theories of emotional expression as, for the most part, the theories reviewed in the introduction  Behavioural expression of fear. 91  appear to place importance on either innate processes or situational context, not a combination of the two. In addition to a constellation of facial actions, children in this study displayed global behaviours (e.g., cry, seeking emotional support, scream) which are interpretable as fear. Previous research exanrining behaviours in fear provoking situations (e.g., anesthesia induction, separation from parent) have found that children often cry, turn their heads away, and tense their muscles (Lumley et al., 1993). Children in this study cried, but they did not exhibit muscular rigidity. Indeed, the children who were rated as displaying clinically significant fear were never coded as displaying muscular rigidity. Again, differences in the behaviours seen here, as compared to previous studies may be justified. That is, perhaps the behaviours exhibited by the children in this study were representative of fear, but a specific fear, i.e., fear from physical harm. Therefore, as with facial actions, one would expect a few similarities with what is typically observed during painful situations, especially as pain is multidimensional in nature with a substantial emotional component (Melzack & Torgensen, 1971). Consistent with this, there was a resemblance between the behaviours displayed by the children in this study and those exhibited by children undergoing painful medical procedures. For instance, children undergoing bone marrow aspirations often cry, verbally resist the procedure, make both fear and pain statements, engage in nervous behaviour, seek information, and tense their muscles (Jay et al., 1983). In this study, young children, when confronted with a fear of physical harm stimulus (i.e., cast removal), resisted the  Behavioural expression of fear. 92  procedure verbally and cried. However, they also flailed their arms and/or legs, sought emotional support (e.g., "help me"), screamed, and, at times, needed to be restrained. Perhaps verbal resistance and crying are sensitive to both the threat of physical harm and pain, but flailing, seeking emotional support, screaming, and requiring physical restraint are specific to fear of physical harm. Further, perhaps the threat of physical harm is perceived as more frightening than other fear stimuli used in previous research, and this increased severity of fear may account for why children engaged in different and/or more behaviours than what has been shown to occur in response to fear situations. That is, as fear increases, children recruit additional behaviours to communicate their feelings. This hypothesis is supported by the fact that the particular behaviours observed here have not been seen in other fear provoking situations (Lumley et al., 1993). Considering that it is not adaptive to (a) scream or seek emotional support in all fear provoking situations, especially those situations where increased attention is not a positive consequence, or (b) to be restrained in all frightening events (e.g., gorilla, stranger approaching), global behaviours may well be sensitive to context. Although this study did not aim to find supporting evidence for one theory of emotional expression over another, the data can shed some light on the subject. First, the idea that the behaviours displayed by the children in this study were specific to fear of physical harm, as described above, is in direct contrast to the Differential Emotions Theory and instead supports aspects of both the Behavioural Ecology View, the Dynamical  Behavioural expression of fear. 93  System's Theory, and Russell's Dimensional-Contextual Perspective (1997). However, the prospect remains that the differences found between the infant facial expressions of fear described in the literature, the children's expressions in this study, and the adult, empirically based, expressions follow a predetermined pattern of changes. That is, although infants, children, and adults may not show the same fear expression, as predicted by the differential emotion theorists, the changes may follow a prescribed pattern. The idea that emotional expression is, in some way, predetermined would go against the Behavioural Ecology View and the Dimensional-Contextual Perspective, but would lend credence to aspects of the Differential Emotions and Dynamical Systems Theories. Regardless, the fact that a constellation of facial expressions and a set of more broadband behaviours were identified is consistent with the basic theoretical model of Minimal Universality (Russell, 1995; Russell & Fernandez-Dols, 1997). Indeed, the findings that children rated as displaying clinically significant fear exhibited the same behaviours identified in the total sample suggests that the behaviours displayed were representative of fear. Further, the findings are consistent with the belief that most people can infer something of another's psychological state from behavioural displays, much like they can from anything else another person does or says (Russell & Fernandez-Dols, 1997). Prediction of Children's Fear In the prediction of children's fear, as measured by facial activity, child's age in months and cast location made unique and significant contributions. Older children had  Behavioural expression of fear. 94  lower facial action factor scores than younger children, and children with casts on their legs had higher facial action factor scores than children with arm casts. Similarly,' when fear was measured by more broadband behaviours, child's age in months and cast location made unique and significant contributions. Again, older children had lower OSBD scale scores as did children with casts on their arms. Interestingly, gender did not uniquely contribute to the prediction of children's fear with either facial activity or global behaviours. The fact that children's age made a significant contribution was not surprising. Researchers have typically found that older children display fewer facial actions than young children and/or infants in response to painful events (Fradet et al., 1990; Goodenough et al., 1997). Further, younger children, in particular infants, typically have more intense facial displays (Lilley et al., 1997). Research has also found that older children display less intense behavioural responses to distressing situations (Chambers et a l , 1999; Goodenough et a l , 1997; Jay et a l , 1983; Katz et a l , 1980). Interestingly, age was related to the independent observer's and the cast technician's ratings of fear, i.e., older children had lower V A S scores than younger children. It appears that not only are facial actions and more broadband behaviours sensitive to age related changes, but so are observers. The finding that cast location made a significant contribution to the prediction of fear with both facial activity and more broadband behaviours was intriguing. During leg cast removal, children were required to lie down and typically could not see what was  Behavioural expression of fear. 95  occurring while, during arm cast removal, children are sitting up and see everything. Perhaps not seeing what is happening is more frightening for children than seeing. It is also remarkable that both an independent observer's and a cast technician's ratings of fear were higher when children had a leg cast than an arm cast. Furthermore, the two research assistants who collected this data both made the observation that children with leg casts appeared more afraid than children with arm casts. One possible explanation for why gender did not make a unique and significant contribution to the prediction of children's fear may be because children, regardless of gender, have been found to report a similar level of fear in response to situations which threaten physical harm (Campbell & Rapee, 1994). That is, boys and girls appear to be equally concerned when threatened with possible physical harm. Therefore, one would not expect a gender difference with cast removal. Validity of the Behavioural Expression of Fear Evidence from the present study show that certain facial actions and global behaviours were displayed by children when scared. However, one needs to examine the validity of such behaviours if they are to be deemed useful to other researchers and/or health care professionals. First, facial activity and global behaviours were positively correlated with each other suggesting both behaviours offer important and consistent information. Second, the independent observer's and cast technician's fear ratings were related to facial action factor scores. Similarly, OSBD scale scores were related to an independent observer's and cast technician's ratings of fear. Therefore, it appears that  Behavioural expression of fear. 96  both types of measurement systems, i.e., fine-grained and global, are valid means of determining whether a child is afraid. Some may say the degree of correlation between the V A S ratings, the facial action factor score, and the OSBD scale score may make the arduous task of coding redundant. However, the coding systems employed in this study, as compared to a subjective measure like a V A S , are less likely to be affected by cultural values and personal experience. Indeed, the rigorous training one must undertake to effectively utilize both the BabyFACS and OSBD coding systems ensures that a consistent standard of coding, one that is not tainted by values and/or expectations, is used. Further, the correlations were consistent with those typically found in other facial action research (e.g., Gilbert et al., in press). However, instead of pitting one measurement system over another, it makes more conceptual sense to examine the different situations where each method may have greater utility. For example,fine-grainedand broadband behavioural coding systems are perhaps more important when delineating the actions which comprise an emotional expression, whereas visual analogue scales are likely to be more useful during clinical assessments. Overall, it appears that the constellation of facial actions and the pattern of more global behaviours displayed by the children in this study are indicative of fear. Limitations of the Study. The study addressed many of the methodological limitations present in early fear research. For example, it used comprehensive and objective coding systems, defined fear,  Behavioural expression of fear. 97  and used a situation capable of eliciting clinically significant fear. However, there were several questions which could not be answered with the present data. First, research will need to replicate these findings in other fear situations to determine if the assemblage of behaviours in this study is indeed representative of fear in general, or if it is specific to fear of physical harm or cast removal as postulated. Indeed, as this study did not directly test the situation-specific nature of the fear expression it is important to note that many of the arguments put forth suggesting a situational specificity could in fact be accounted for by factors other than context. Second, because the chance of error would have increased if more analyses were conducted, this study could not determine whether the same facial actions and broadband behaviours were displayed by all children across the age range or whether the expressions merely differed in intensity. Investigations exploring whether age effects both facial activity and more global behaviours in a wider range of children are also necessary. Expanding the age range would also allow for self-report measures which would hopefully support the findings that not only do independent observers and cast technicians rate cast removal as a fearful experience, but children do too. Third, research exanririing whether differences and/or similarities between emotions (e.g., pain and fear) exist would be indispensable. In addition, it is conceivable that the behavioural expressions displayed by the children in this sample may have under-represented the true fear expression as children with arm and leg fractures may be more likely to engage in physical risk - taking behaviours and may therefore be less fearful by nature. That is, perhaps using children  Behavioural expression of fear.  98  who had a fractured a bone as a population sample produced a biased sample. However, if this is indeed the case, one would expect the results of this study to be more pronounced in a sample of perhaps less impulsive and cautious children. It is important to note that fear and anxiety, by definition, differ. Most importantly, anxiety tends to be prolonged and persistent, is often accompanied by heightened vigilance, and there is an uncertain connection between anxiety and threat, whereas fear is usually episodic in nature and declines with the removal of threat (Rachman, 1998). However, to date, there is no data available to show an actual distinction between these two emotions (i.e., do they vary in physiological responses, are there different facial reactions or behaviours attributable to one but not the other?). The only difference appears to be the rather important distinctions made by Rachman (1998). Therefore, although this study contends it was measuring fear, it is possible that the behaviours displayed by the children in this study are also indicative of anxiety and perhaps cast removal can be classified as both a fear and anxiety eliciting stimulus. Clinical Implications. Health care professionals and parents are likely to continue to be the young child's most important advocate. By equipping these people with the knowledge that fear does appear to have a unique constellation of behaviours, young children may not have to endure fearful medical procedures with little intervention. Indeed, the presence of the behaviours identified in this study should suggest the need for clinical intervention prior to the start of medical procedures, e.g., relaxation strategies, deep breathing, imagery. However, implementing preventative strategies needs to be  Behavioural expression of fear. 99  considered in a cost-benefit framework. That is, the utility of such interventions to reduce both the behavioural and physiological expressions of fear needs to be examined. Further, as this study identified factors which seem to contribute to increased fear during cast removal (e.g., age, cast area), health care professionals may be able to alter their practices in order to minimize the amount of fear children experience or, at the very least, be aware of how different children may perceive cast removal. For example, children with leg casts were rated by an independent observer and the cast technician as being more afraid. These same children also had higher facial action factor scores and OSBD scale scores. Thus, health care practitioners should either supply children with the necessary skills to cope with the increased fear associated with leg cast removal or strive to find alternate ways to reduce the children's fear (e.g., have them sitting up). Finally, given the percentage of children in this study experiencing clinically significant fear, interventions like the ones studied by Johnson, Kirchoflf, and Endress (1975) would be useful for fearful medical procedures, if not ethically necessary.  Behavioural expression of fear.  100  References Astin, E . (1977). Self-reported fears of hospitalized and non-hospitalized children aged ten to twelve. Maternal-Child Nursing, 6, 17-24. Blumberg, S.H. & Izard, C.E. (1986). Discriminating patterns of emotions in 10and 11-year old children's anxiety and depression. Journal of Personality and Social Psychology. 51. 852-857. Broome, M . E . (1986). The relationship between children's fear and behavior during a painful event. Children's Health Care, 14, 142-145. Broome, M . E . & Hellier, A P. (1987). School-age children's fears of medical experiences. Issues in Comprehensive Pediatric Nursing, 10, 77-86. Broome, M . E . , Hellier, A.P., Wilson, T., Dale, S. & Glanville, C. (1988). Development and testing of an instrument to measure children's fears of medical experiences. In C. Waltz and O. Strickland (Eds.), Measurement of Nursing Outcomes, pp. 201 -214. New York: Springer. Bullock, M . & Russell, J.A. (1986). Concepts of emotion in developmental psychology. In: C.E. Izard & P.B. Read (Eds.), Measuring Emotions in Infants and Children, Volume 12 (pp. 203-237) Cambridge University Press: New York. Campbell, M . A . & Rapee, R . M . (1994). The nature of feared outcome representations in children. Journal of Abnormal Child Psychology, 22 99-111.  Behavioural expression of fear.  101  Camras, L . A . (1991). Conceptualizing early infant affect: View II and Reply. In K. Strongman (Ed.), International Review of Study on Emotions (pp. 16-28, 33-36). New York: John Wiley & Sons. Camras, L . A . (1992). Expressive development and basic emotions. Cognition and Emotion. 6, 269-284. Camras, L.A., Lambrecht, L., & Michel, G.F. (1996). Infant "surprise" expressions as coordinative motor structures. Journal of Nonverbal Behavior. 20, 183195. Camras, L.A., Malatesta, C , & Izard, C E . (1991). The development of facial expressions in infancy. In R. Feldman & B. Rime (Eds.), Fundamentals of Nonverbal Behavior (pp. 73-105). Cambridge: Cambridge University Press. Camras, L.A., Oster, H., Campos, J.J., Miyake, K., & Bradshaw, D. (1997). Japanese and American infants' responses to arm restraint. In: P. Ekman & E.L. Rosenberg (Eds.), What the Face Reveals: Basic and Applied Studies of Spontaneous Expression using the Facial Action Coding System (FACS) (pp. 289-299). New York: Oxford University Press. Camras, L . A . , Sullivan, J., & Michel, G. (1993). Do infants express discrete emotions? Adult judgments of facial, vocal, and body actions. Journal of Nonverbal Behavior. 17. 171-186.  Behavioural expression of fear.  102  Carr, T.D., Lemanek, K . L . , & Armstrong, F.D. (1998). Pain and fear ratings: Clinical implications of age and gender differences. Journal of Pain and Symptom Management. 15. 305-313. Cassidy, K . L . , McGrath, P.J., Reid, G.J., Chambers, C.T., Gilbert, C.A., Finley, G.A., Brown, T.L., Smith, D . Szudek, E., Morley, C , & Morton, B . (1996). Watch needle watch T V : Audiovisual distraction in preschool immunization. Poster Presented at the 8th World Congress on Pain, in Vancouver, August 17th, 1996. Carroll, J.M. & Russell, J.A. (1996). Do facial expression signal specific emotions? Judging emotion from the face in context. Journal of Personality and Social Psychology. 70. 205-218. Carroll, J.M. & Russell, J.A. (1997). Facial expressions in Hollywood's portrayal of emotion. Journal of Personality and Social Psychology.. Chambers, C T & Craig, K . D . (1998). A n intrusive impact of anchors in children's pain faces. Pain. 78, 27-37. Chambers, C.T., Giesbrecht, K . , Craig, K . D . , Bennett, S.M., & Huntsman, E . (1999). A comparison of faces scales for the measurement of pediatric pain: Children's and parents' ratings. Pain. 83, 25-35. Chovil, N . (1991). Social determinants of facial displays. Journal of Nonverbal Behaviours. 15. 141-154. Clewes, J.L. & Endler, N.S (1994). State-trait anxiety and the experience of elective surgery in children. Canadian Journal of Behavioural Science. 26, 183-198.  Behavioural expression of fear.  103  Collins, S.L., Moore, A., & McQuay, H J . (1997). The visual analogue pain intensity scale: what is moderate pain in niillimetres? Pairu 72, 95-97. Craig, K . D . (1998). Oh knowing another's pain. In: P. Ekman & E . Rosenberg (Eds.). What Facial Expression Reveals about Emotions. Development. Psvchopathology *  and Health. Oxford, U K : Oxford University Press. Craig, K . D . & Grunau, R.V.E. (1993). Neonatal pain perception and behavioural measurement. In K.J.S. Anand and P.J. McGrath (Eds.), Neonatal Pain and Distress (pp. 67-105). Amsterdam: Elsevier. Craig, K . D . , Hadjistavropoulos, H.D., Grunau, R.V.E., & Whitfield, M.F. (1994). A comparison of two measures of facial activity during pain in the newborn child. Journal of Pediatric Psychology. 19, 305-318. Craig, K.D., Lilley, C M . , & Gilbert, C A . (1996). Social barriers to optimal pain management in infants and children. Clinical Journal of Pain. 12, 232-242. Craig, K . D . & Prkachin, K . M . (1983). Nonverbal measures of pain. In: R. Melzack (Ed.). Pain Measurement and Assessment (pp. 173-180). Raven Press: New York. Croake, J. (1969). Fears of children. Human Development. 12. 239-247. Eisenberg, N , Fabes, R.A., Bustamante, D., Mathy, R . M . , Miller, P.A., & Lindholm, E. (1988). Differentiation of vicariously induced emotional reactions in children. Developmental Psychology. 24, 237-246.  Behavioural expression of fear.  104  Ekman, P. (1992a). A n argument for basic emotions. Cognition and Emotion. 6, 169-200. Ekman, P. (1992b). Facial expressions of emotion: New findings, new questions. Psychological Science. 3, 34-38. Ekman, P. (1994). Strong evidence for universals in facial expressions: A reply to Russell's mistaken critique. Psychological Bulletin. 115. 268-287. Ekman, P. & Friesen, W.V. (1971). Constants across cultures in the face and emotion. Journal of Personality and Social Psychology. 17. 124-129. Ekman, P. & Friesen, W.V. (1975). Unmasking the Face: A Guide to Recognizing Emotions From Facial Clues. New Jersey: Prentice Hall. Ekman, P. & Friesen, W.V. (1978). Investigator's guide to the facial action coding system Palo Alto, C A : Consulting Psychologists Press. Ekman, P., Levenson, R.W., & Friesen, W.V. (1983). Autonomic nervous system activity distinguishes between emotions. Science. 164. 1208-1210. Elkins, P.D & Roberts, M . C . (1985). Reducing medical fears in a general population of children: A comparison of three audiovisual modeling procedures. Journal of Pediatric Psychology. 10, 65-75. Elliott, C.H., Jay, S.M., & Woody, P. (1987). A n observation scale for measuring children's distress during medical procedures. Journal of Pediatric Psychology. 12, 543551.  7  Behavioural expression of fear.  105  Endler, N.S. (1997). Stress, anxiety, and coping: The multidimensional interaction model. Canadian Psychology, 38,136-153. Femandez-Dols, J.M. & Ruiz-Belda, M . A . (1997). Spontaneous facial behavior during intense emotional episodes: Artistic truth and optical truth. In: J.A. Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expression (pp. 255-274). Paris: Cambridge University Press. Fradet, C , McGrath, P.J., Kay., & Adams, S.B. (1990). Prospective survey of reactions to blood tests by children and adolescents. Pain, 40, 53-60. Fridlund, A.J. (1991). Evolution and facial action in reflex, social motive, and paralanguage. Biological Psychology. 32, 3-100. Fridlund, A.J. (1997). The new ethology of human facial expressions. In: J.A. Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expressions (pp. 103129). Paris: Cambridge University Press. Fridlund, A.J., Kenworthy, K . G . , Jaffey, A . K . (1992). Audience effects in affective imagery: Replication and extension to dysphoric imagery. Journal of Nonverbal Behavior, 16, 191-212. Fuller, B., Thomson, M . , Conner, D.A., Scanlan, J. (1996). Relationship of cues to assessed infant pain level. Clinical Nursing Research. 5, 43-66. Gaensbauer, T.J., Mrazek, D., & Emde, R. (1979). Patterning of emotional response in a playroom laboratory situation. Infant Behavior and Development, 2, 163178.  Behavioural expression of fear.  106  Gauthier, J . C , Finley, G.A., & McGrath, P.J. (1998). Children's and self-report of postoperative pain intensity and treatment threshold: deternrining the adequacy of medication. Clinical Journal of Pain, 14, 116-120, Gilbert, C A . , Craig, K.D., Rocha, E . M . , & Mathias, M . D . (2000). Communication of everyday pain in children with and without developmental delays. Journal of Pediatric Psychology. 25, 301-308. Gilbert, C.A., Lilley, C M . , Craig, K . D . , Bennett, S.M., Montgomery, C , Court, C , & McGrath, P . M . (1999). The Child Facial Coding Scale: A Behavioural Measure of Pain in Children Undergoing Minor Surgery Aged 1 to 5 years. Clinical Journal of Pain, 15, 192-200. Goodenough, B., Kampel, L., Champion, G.D., Laubreaux, L., Nicholas, M . K . , Ziegler, J.B., & Mclnerney, M . (1997). A n investigative of the placebo effect and agerelated factors in the report of needle pain from venepuncture in children. Pain, 72, 383391. Gnepp, J., & Hess, D . (1986). Children's understanding of verbal and facial display rules. Developmental Psychology. 22. 103-108. Gorsuch, R.L. (1997). Exploratory factor analysis: Its role in item analysis. Journal of Personality Assessment. 68, 532-560. Grunau, R.V.E., & Craig, K . D . (1987). Pain expression in neonates: Facial action and cry. Pain. 28. 395-410.  Behavioural expression of fear.  107  Grunau, R.V.E., Johnston, C.C., Craig, K . D . (1990). Neonatal facial and cry responses to invasive and non-invasive procedures. Pain. 42. 295-305. Grunau, R.V.E., Oberlander, T., Holsti, L., & Whitfield, M.F. (1998). Bedside application of the Neonatal Facial Coding System in pain assessment of premature neonates. Pain. 76. 277-286. Guadagnoli, E . & Velicer, W.F. (1988). Relation of sample size to the stability of component patterns. Psychological Bulletin. 103. 265-275. Guinsberg, R., Berenguel, R.V., Xavier, R.C., Branco de Almeida, M.F., & Kopleman, B.I. (1997). Are behavioral scales suitable for preterm and neonatal pain assessment? In T.S. Jensen, J.A. Turner, and A . Wiesenfeld-Hallin (Eds.), Proceedings of the 8th World Congress on Pain (pp. 893-9011 Seattle, WA: IASP Press. Harrigan, J.A. & O'Connell, D . M . (1996). How do you look when feeling anxious? Facial displays of anxiety. Personality and Individual Differences. 21, 205-212. Harrigan, J.S. & Taing, K.T. (1997). Fooled by a smile: Detecting anxiety in others. Journal of Nonverbal Behavior. 21, 203-221. Hellier-Wilson, A . & Yorker, B . (1997). Fears of medical events among schoolage children with emotional disorders, parents, and health care providers. Issues in Mental Health Nursing. 18,57-71. Hess, U . , Banse, R., & Kappas, A . (1995). The intensity of facial expression is determined by underlying affective state and social situation. Journal of Personality and Social Psychology. 69. 280-288.  Behavioural expression of fear.  108  Hiatt, S.W., Campos, J.J., & Emde, R . N . (1979). Facial patterning and infant emotional expression: Happiness, surprise, and fear. Child Development, 50,1020-1035. Hollingshead, S.E. (1977). Hollingshead's two factor index of social position. In: Miller (Ed.), Handbook of Research Design and Measurement (pp. 231-239). N Y : David McKay Co. Horn-Ratner, H . & Stettner, L.J. (1991). Thinking and feeling: Putting Humpty Dumpty together again. Merrill-Palmer Quarterly, 37, 1-26. Huebner, R.R. & Izard, C E . (1988). Mothers' responses to infants' facial expressions of sadness, anger, and physical distress. Motivation and Emotion. 12. 185197. Hyson, M . C & Izard, C E . (1985). Continuities and changes in emotion expressions during brief separation at 13 and 18 months. Developmental Psychology. 21, 1165-1170. Izard, C E . (1982). Measuring emotions in human development. In: C E . Izard (Ed.), Measuring Emotions in Infants and Children (pp. 3-18). N Y : Cambridge University Press. Izard, C E . (1992). Basic emotions, relations among emotions, and emotioncognition relations. Psychological Review. 99. 561-565. Izard, C E . (1994). Innate and universal facial expressions: Evidence from developmental and cross-cultural research. Psychological Bulletin, 115, 288-299.  Behavioural expression of fear.  109  Izard, C E . (1997). Emotions and facial expressions: A perspective from differential emotions theory. In: J.A. Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expressions (pp. 57-77). Paris: Cambridge University Press. Izard, C.E., & Dougherty, L . M . (1980). System for Identifying Affect Expressions by Holistic Judgments (AFFEX). Newark, D E : University of Delaware, Instructional Resources Center. Izard, C.E., Huebner, R.R., Risser, D., McGinnes, G . C , & Dougherty, L . M . (1980). The young infant's ability to produce discrete emotional expressions. Developmental Psychology, 16,132-140. Izard, C.E., & Malatesta, C Z . (1987). Perspectives on emotional development 1: Differential emotional theory of early emotional development. In J.D. Osofsky (Ed.), Handbook of Infant Development (pp. 494-554). New York: Wiley. Jay, S.M., & Elliott, C H . (1986). Observation Scale of Behavioral DistressRevised. Available from Children's Hospital of Los Angeles, Division of Hemato logyOncology. Jay, S.M., Ozolins, M . , Elliott, C.H., & Caldwell, S. (1983). Assessment of children's distress during painful medical procedures. Health Psychology. 2, 133-147. Johnson, J.E., Kirchoff, K.T., & Endress, M.P. (1975). Altering children's distress behavior during orthopedic cast removal. Nursing Research. 24, 204-210. Johnston, C.C., & Strada, M . E . (1986). Acute pain response in infants: a multidimensional description. Pain, 24, 373-382.  Behavioural expression of fear.  110  Katz, E.R., Kellerman, J., & Seigel, S.E. (1980). Behavioral distress in children with cancer undergoing medical procedures: Developmental considerations. Journal of Consulting and Clinical Psychology, 48, 356-365. LeBaron, S. & Zeltzer, L . (1984). Assessment of acute pain and anxiety in children and adolescents by self-reports, observer reports, and a behavior checklist. Journal of Consulting and Clinical Psychology, 22, 690-701. Legerstee, M . , Corter, C , & Kienapple, K . (1990). Hand, arm, and facial actions of young infants to a social and nonsocial stimulus. Child Development, 61, 774-784. Lewis, M . & Michalson, L . (1982). The measurement of emotional state. In: C.E. Izard (Ed.), Measuring Emotions in Infants and Children (pp. 178-207). New York: Cambridge University Press. Lilley, C M . , Craig, K.D., & Grunau, R . V . E . (1997). The expression of pain in infants and toddlers: Developmental changes in facial action. Pain, 72, 161-170. Lumley, M . A . , Melamed, B.G., & Abeles, L . A . (1993). Predicting children's presurgical anxiety and subsequent behavior changes. Journal of Pediatric Psychology, 18, 481-497. Malatesta-Magai, Izard, C , Camras, L . (1991). Conceptualizing early infant affect; Emotions as fact, fiction, or artifact? In: K.T. Strongman (Ed.), International Review of Studies on Emotion (pp. 1-36). New York : John Wiley and Sons Ltd. March, J.S & Mulle, K . (1998). OCD in Children and Adolescents. A CognitiveBehavioural Treatment Manual. New York: The Guilford Press.  Behavioural expression of fear. I l l  Matias, R., & Conn, J.F. (1993). Are MAX-specified infant facial expressions during face-to-face interaction consistent with DET? Developmental Psychology, 29, 524531. McGraw, M . B . (1943). The Neuromuscular Maturation of the Human Infant. New York: Columbia University Press. McGrath, P.J., Johnson, G., Goodman, J.T., Schillinger, J., Dunn, D . & Chapman, J.A. (1985). CHEOPS: A behavioural scale for rating postoperative pain in children. In H.L. Fields, R. Dubner, & F. Cerevero (Eds.), Advances in Pain Research and Therapy, Vol. 9. Proceedings of the 4th World Congress on Pain. New York: Raven Press. Melamed, B.C., Meyer, R., Gee, C. & Soule, L . (1976). The influence of time and type of preparation on children's adjustment to hospitalization. Journal of Pediatric Psychology, 4, 31-36. Melzack, R. & Torgensen, W.S. (1971). On the language of pain. Anesthesiology, 34, 50-59. Messinger, D.S., Fogel, A . , & Dickson, K . L . (1997). A dynamic systems approach to infant facial action. In: JA Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expressions, (pp. 205-226). Paris: Cambridge University Press. Michel, G.F., Camras, L.A., & Sullivan, J. (1992). Infant interest expressions as coordinative motor structures. Infant Behavior and Development, 15, 347-358. Mills, N . M . (1989). Pain behaviors in infants and toddlers. Journal of Pain and Symptom Management, 4, 184-190.  Behavioural expression of fear.  112  Oster, H . (1999). Personal Communication. July, 1999. Oster, H . , Camras, L., Campos, J., Campos, R., Ujiee, T., Zhao-Lan, M . , & Lei, W. (1996). The patterning of facial expressions in Chinese. Japanese, and american infants in fear- and anger-eliciting situations. Poster presented at the International Conference on Infant Studies, Providence, RI, April 18-21, 1996. Oster, H . , Hegley, D., & Nagel, L . (1992). Adult judgments and fine-grained analysis of infant facial expressions: Testing the validity of a priori coding formulas. Developmental Psychology, 28. 1115-1131. Oster, H . & Rosenstein, D . (1995). Baby FACS: Analyzing facial movement in infants. Paper presented at the annual meeting of the Society for Research on Child Development. Pope, L . K . & Smith, C A . (1994). On the distinct meanings of smiles and frowns. Cognition and Emotion. 8, 65-72. Peterson, L.J. & Mori, L . (1988). Preparation for hospitalization. In: D K Routh (Ed.), Handbook of Pediatric Psychology, (pp. 460-491). New York: Guilford Press. Prkachin, K . M . , Currie, N . A . , & Craig, K . D . (1983). Judging nonverbal expressions of pain. Canadian Journal of Behavioural Science. 15, 408-420. Rachman, S.J. (1998). Anxiety. United Kingdom: Psychology Press Roberts, M . C . , Wurtele, S.K., Boone, R.R., Ginther, L.J., & Elkins, P.D. (1981). Reduction of medical fears by use of modeling: A preventative application in a general population of children. Journal of Pediatric Psychology. 6, 293-299.  Behavioural expression of fear.  113  Rosenberg, E.L. (1997). Introduction: The study of spontaneous facial expressions in psychology. In P. Ekman & E.L. Rosenberg (Eds.), What the Face Reveals: Basic and Applied Studies of Spontaneous Expression Using the Facial Action Coding System t F A C S l . (pp. 3-18). New York: Oxford University Press. Rosenstein, D., & Oster, H . (1988). Differential facial responses to four basic tastes in newborns. Child Development, 59, 1555-1568. Russell, J.A. (1994). Is there universal recognition of emotion from facial expression? A review of cross-cultural studies. Psychological Bulletin, 115, 102-141. Russell, J.A. (1995). Facial expression of emotion: What lies beyond minimal universality? Psychological Bulletin. 118, 379-391. Russell, J.A. (1997). Reading emotions from and into faces: Resurrecting a dimensional-contextual perspective. In: J.A. Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expression, (pp. 295-320). Paris: Cambridge University Press. Russell, J.A. & Bullock, M . (1986). Fuzzy concepts and the perception of emotion in facial expressions. Social Cognition. 4, 309-341. Russell, J.A. & Fernandez-Dols, J.M. (1997). What does a facial expression mean? In: J.A. Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expression, (pp. 295-320). Paris: Cambridge University Press. Saarni, C. (1984). A n observational study of children's attempts to monitor their expressive behavior. Child Development. 55,1504-1513.  Behavioural expression of fear.  114  Scherer, M . & Nakamura, C. (1968). A fear survey schedule for children (FSSFC): A factor-analytic comparison with manifest anxiety. Behavioural Research and Therapy. 6. 173-182. Schiller, V . M . , Izard, C.E., & Hembree, E.A. (1986). Patterns of emotion expression during separation in the strange-situation procedures. Developmental Psychology. 22. 378-382. Schrans, T., Werker, J.F., & Brown, R.E. (1990). Developmental changes in children's awareness of the causal determinants of behaviours. International Journal of Behavioral Development. 13, 251 -262. Sherman, M . (1927). The differentiation of emotional responses in infants. Journal of Comparative Psychology. 7, 265-284. Sinha, R. & Parsons, O.A. (1996). Multivariate response patterning of fear and anger. Cognition and Emotion. 10,173-198. Smith, C A . & Scott, H.S. (1997). A componential approach to the meaning of facial expressions. In J.A. Russell & J.M. Fernandez-Dols (Eds.), The Psychology of Facial Expression, (pp. 229-254). Paris: Cambridge University Press. Tabachnick, B.G. & Fidell, L.S. (1996). Using Multivariate Statistics. 3 Edition; rd  New York: Harper Collins College Publishers. Taddio, A., Stevens, B., Craig, K . D . , Rastogi, P., BenDavid, S., Sherman, A., Mulligan, P., & Koren," G. (1998). New England Journal of Medicine. 336. 1197-1201.  Behavioural expression of fear.  115  Tarbel, S.E., Cohen, I.T., & Marsh, J.L. (1992). The toddler-preschooler postoperative pain scale: an observational scale for measuring postoperative pain in children aged 1-5. Preliminary report. Pain, 50, 273-280. Tassinary, L . G . , & Cacioppo, J.T. (1992). Unobservable facial actions and emotion. Psychological Science, 3, 28-33. Venham, L . L . & Gaulin-Kremer, E . (1979). A self-report measure of situational anxiety for young children. Pediatric Dentistry. I, 91-96. Visintainer, M . A . , & Wolfer, J.A. (1975). Psychological preparation for surgical pediatric patients: The effect on children's and parents' stress responses and adjustment. Pediatrics. 56. 187-202. Wagner. H.L. (1990). The spontaneous facial expression of differential positive and negative emotions. Motivation and Emotion. 14, 27-43. Webster, S. (1957). Webster's New 20 Century Dictionary. 2 Edition; New th  nd  York: Library Guild Inc. Zachary, R.A., Friedlander, S., Huang, L . N , Silverstein, S., & Leggott, P. (1985). Effects of stress-relevant and -irrelevant filmed modeling on children's responses to dental treatment. Journal of Pediatric Psychology. 10. 383-401. Zander-Malatesta C. & Haviland, J.M..(1982). Learning display rules: The socialization of emotion expression in infancy. Child Development. 53, 991-1003.  Behavioural expression of fear.  Appendix A Observational Scale of Behavioral Distress-Revised  116  Behavioural expression of fear.  117  Information Seeking: Any questions regarding the procedures/event. For example, "when will you stop?", "will this hurt". Cry: Crying sounds and/or onset of tears -usually non-intelligible but can be double coded with verbal categories. For example, sobbing, "boo hoo hoo", crying sounds. Scream: Loud vocal expression at high pitch/intensity, usually non-intelligible, but can be double coded with verbal categories. High pitch distinguishes this category from "cry". For example, sharp, shrill, harsh, high tones, shrieks. Restraint: Child must be physically held down by staff member or parent with noticeable pressure and/or child must be exerting force or resistance in response to restraint attempts by staff. Sometimes it is not clear if the child is exercising pressure back due to tightness of restraint (i.e., child cannot move). In such cases where restraint is obvious and child's resistance in not clear, code restraint. Verbal Resistance: Any verbal expression of delay, termination, or resistance. Must be intelligible. For example, "I want to go...", "no, no", "take me home", "no more", "stop". Emotional Support: Verbal or nonverbal solicitation of hugs, hand holding, physical or verbal comfort by child. Code initiation only for physical behaviours. For example, "hold me", "I love you", "momma please", "help me". Verbal Pain: Any words phrases, or statement which refer to pain, damage, or being hurt or discomfort. Must be intelligible. Can be anticipatory as well as actual. For example, "that hurt", "Owwh", "oh", "you are killing me".  Behavioural expression of fear.  118  Verbal Fear: Any words phrases, or statement which refer to fear, being scared or worried. Must be intelligible. Can be anticipatory as well as actual. For example, "this is scary", "you are scaring me", "that machine is scary". Flail: Random gross movements of arms and legs or whole body. Flail often occurs in response to restraint. Must be random. For example, pounding fists, kicking legs repeatedly and randomly, throwing arms out, moving back and forth repeatedly during procedure. Muscular Rigidity: Child appears tensed or rigid. Can be one body part. Important to distinguish from being still, muscles need to tensed or held tightly. For example, arms are held rigidly, legs straight out in front and muscles are contracted.  Behavioural expression of fear. 119  Appendix B Demographic Questionnaire  Behavioural expression of fear. 120  Demographics. Parent Information Sheet 1. Your Relationship to the Child (circle one number): 1. Mother 2. Father 3. Stepmother 5. Other 2. Your Current Age:  4. Stepfather  (years)  3. Your Ethnicity (circle one): 1. Caucasian 2. African-American 6. Other  3. Asian 4. Native  5. Indo-Canadian  4. Your Current Marital Status (circle one number): 1. Married 2. Divorced/ Separated 3. Remarried  4. Widowed 5. Never married 6. Other  5. Your Education (circle one number): 1. Graduate School/ Professional training 2. University graduate (4 year college) 3. Partial university (at least 1 year) 4. Trade School/Community College  5. High School graduate 6. Some high school (min 10th grade) 7. Junior high school graduate 8. Less than 7th grade  6. Your Occupation (please describe):  7. Your Spouse's Partner's Current Age: 8. Your Spouse's Ethnicity (circle one): 1. Caucasian 2. African-American 6. Other  (years)  3. Asian  4. Native 5. Indo-Canadian  9. Your Spouse's Partner's Education (circle one number): 1. Graduate School/ Professional training 2. College graduate (4 year college) 3. Partial college (at least 1 year) 4. Trade School/Community College  5. High School graduate 6. Some high school (min 10th grade) 7. Junior high school graduate 8. Less than 7th grade  Behavioural expression of fear. 121  10. Your Spouse/Partner's Occupation (please describe):  . Adults (20)  11. Number of Family Members:  Children  For each child in your family please list their age, sex, and whether or not they currently reside in your home. Age: Age: Age: _ _ _ Age: Age:  Sex (circle one): Male Sex (circle one): Male Sex (circle one): Male Sex (circle one): Male Sex (circle one): Male  Female Female Female Female Female  Living Living Living Living Living  at home? (circle one): at home? (circle one): at home? (circle one): at home? (circle one): at home? (circle one):  No No No No No  Yes Yes Yes Yes Yes  Child Information Sheet Please complete these questions in reference to your child. 1. Child's Age:  (years)  2. Child's Sex (circle one): Male  Female  3. Child's Date of Birth:  (month)  4. Child's Ethnicity: 1. Caucasian 2. African-American 6. Other  (day)  3. Asian 4. Native  5. Does your child have any chronic illnesses (circle one): N o  (year)  5. Indo-Canadian  Yes  If yes, please list the illnesses:  6. Is your child currently on any medication (circle one): No  Yes  If yes, please list the medications:  7. Child's Current Grade in School  N/A  Behavioural expression of fear. 122  Appendix C Past Medical History Questionnaire  Behavioural expression of fear.  123  Please indicate how many times your child has experienced each of the following medical procedures: 0 = never 1 = one or two times 2 = three or four times 3 = more than four times Throat Cultures Medical Appointments Dental Appointments Bloodwork (i.e., finger poke and/or venipuncture) Hospitalizations Castings Surgery  0 0 0 0 0 0 0  Please rate your child's reactions to each of these medical experiences: Throat Cultures 1 2 3 4 5 6 7 negative no reaction positive  1 negative  1 negative  2  Medical Appointments 3 4 5 6 no reaction  7 positive  2  Dental Appointments 3 4 5 no reaction  7 positive  6  Bloodwork (le.. finger poke and/or venipuncture) 1 2 3 4 5 6 7 negative no reaction positive Hospitalizations 3 4 5 no reaction  1 negative  2  1 negative  2  3  Casting 4 no reaction  2  Surgery 3 4 no reaction  1 negative  6  7 positive  5  6  7 positive  5  6  7 positive  1 1 1 1 1 1 1  2 2 2 2 2 2 2  3 3 3 3 3 3 3  Behavioural expression of fear.  Appendix D Chart Review  Behavioural expression of fear.  1. Inpatient  Outpatient  2. Trauma Patient  Surgical Patient  3. Date of accident/surgery: 4. Type of Fracture: simple  compound  5. Type of Surgery: 6. Area of cast:  N/A N/A  arm  wrist  leg  ankle other  7. Date casted: Date cast removed: 8. Were pain medications prescribed? If yes, please list type of medications:  No  Yes  125  Behavioural expression of fear. 126  Appendix E Baby-Facial Action Coding System  Behavioural expression of fear.  127  A U 3 Brow Knitting: The muscle involved is the corrugator supercilii which pulls the inner ends of the eyebrows downward and together, "knitting" the brows. A U 4 Brow Knot and Lower: The muscle involved is the procerus which pulls down the inner corners of the eyebrows and "knots" the skin between the brows, just above the root of the nose, and the depressor supercilii which lowers the middle and outer portions of the eyebrows. Note: Brow knitting is a salient facial action in infants and young children. In research where a distinction between the knitting and knotting actions of the brows could be important, we suggest A U 3 for brow knitting, when no knotting is visible, and A U 4 for brow knotting, when no knitting is visible. A U 6 Cheek Raise: A U 6 pulls the skin from the cheeks and temple toward the eyes, raising the cheeks, deepening the infraorbital furrow, and causing bags or pouches in the skin below the eyes. A U 9 Nose Wrinkle: Produced by the levator labii superioris alaequi nasi, which runs along the side of the nose from its origin at the root of the nose to its insertion in the nostril wings and central portion of the upper lip. When it contracts, it pulls the skin along the side of the nose upward and raises the nostril wings and central part of the lip. It also pulls down on the skin above the root of the nose (the glabella), making the procerus muscle bulge. A U 11 Nasolabial Furrow Deepener: Produced by the action of the zygomaticus minor muscle, which originates in the zygomatic arch and attaches in the upper lip midway between the philtrum and the lip corners. The upper lip is pulled slightly upward and laterally just inward form the corners, and the upper lip appears to flatten. The middle portion of the nasolabial furrow is deepened, and the infraorbital triangle is raised slightly. A U 16 Lower Lip Depressor: Produced by the action of the depressor labii. It emerges from either side of the chin and attaches below the lower lip. It pulls the lower Up downward toward the chin, exposing more of the lower teeth; with a strong action, it will expose the roots of the teeth and the gums. It may stretch and pull the lower lip laterally to a slight extent and may flatten the lower Up. A U 16 usuaUy parts the Ups. A U 17 Chin Raise: Produced by the action of the mentaUs muscle. It emerges below the lower Up and attaches to the tissue in the chin boss. PuUs the skin over the chin upward and puUs the lower Up upward. A U 19 Tongue Protrudes: The tongue must protrude from its resting position at least as far as the inner margin of the red part of the Ups.  Behavioural expression of fear.  128  A U 20 Lip Stretcher: Produced by the action of the risorius muscle, which has its origin in cheek fascia and attaches to the skin at the corners of the lower lip. It pulls the lips and the skin beyond the lip corners back laterally. The lips, and particularly the lower lip, become stretched and flattened by the lateral pull. The chin and the area just below and adjacent to the lip corners may appear to be stretched and flattened. The mouth is widened, and the lips are elongated horizontally. The lower portion of the nasolabial furrow is pulled laterally. A U 26 Mouth Stretch: Combination of A U 26 (jaw drop) and 27 (mouth stretched widely open) as they are increasing intensities of the same action, stretching the mouth. A U 43 Eves Closed: The upper eyelid is relaxed completely, the lids touch, and the eyes close. There must be a slight raising of the lower lid, but the lids are not tightly closed, A U 44 Squint: Eye opening is very narrow, and tension is visible in the lower eyelids. A U 75 Tongue Shapes: Any time when the tongue is not in its normal resting shape (e.g., narrow, balled, flat, tip of the tongue is curled, concave and relaxed, funneled). Usually coded when child is crying.  Behavioural expression of fear.  Appendix F Regression Analysis using Data from the Clinically Significant Fear Sample  129  Behavioural expression of fear. 130  Descriptive Statistics for the Prediction of Children's Fear. Mean  SD  Facial Action Factor Score  6.80  3.03  Child's Age in Months  43.70  19.29  Child Gender  Female:36.2%  Male: 63.8%  Cast Location  Arm: 42.0%  Leg: 58.0%  Continuous Variables  Categorical Variables  n=57  Bivariate Correlations in the Prediction of Children's Fear.  Fear Age Gender Cast Location n=57  Fear  Age  Gender  Cast Location  1.00  -24*  -.05  -.33**  1.00  .04  -.01  1.00  -.01 1.00  Behavioural expression of fear.  Standard Multiple Regression Analysis: The Prediction of Children's Fear. Variable  Block One:  R_  F  .17  3.59**  Standardized B  Age  -.24*  Gender  -.03  Cast Location  -.33***  _n-57 Note: Standardized betas reflect the unique contribution of the variable controlling for other variables in the regression model. * =p<.05. ** = p<.05. *** =p<.01.  131  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089750/manifest

Comment

Related Items