TWO ASPECTS OF EMPATHic AWARENESS EN YOUNG CHILDREN: AFFECTWE AND coenmvz ROLE - TAKING Dissertation for the Degree 0f Ph. D. MICHIGAN STATE UNIVERSUY LISA BERNADEFTE PARTYKA 1974 ‘__ -‘__, -. M... Ill/ll 5, 129313004 .f‘ .. fl 8' .3 .. 4- : #41 f". [fig-{vb _ .4 A F 13' 2 '. 114'" ("6 . ‘ ‘00 21?»; fan- I? ~ . ‘~. MIJIJ; f;',-_‘, I .' ." D . . ‘ o- NYE‘ . 01‘47.W‘ This is to certify that the thesis entitled TL-J‘S ASPECTS OF El-‘zPATHIC A'~:.’ARET‘JESS IN Yl"Ui~l[3 CHILDREN: AFFECTIVE AND CCGNITIVE RFLE-TAKING presented by Lisa Bernadette Partyka has been accepted towards fulfillment of the requirements for Phil], degree in Clinical Psychology l / WOW I? M1 CIVLLQ/tvk Major professor Date July 29.4974 ABSTRACT TWO ASPECTS OF EMPATHIC AWARENESS IN YOUNG CHILDREN: AFFECTIVE AND COGNITIVE ROLE-TAKING By Lisa Bernadette Partyka The present study was concerned with the phenomenon of empathic awareness as it is exhibited by young children. Affective and cognitive role-taking were proposed as two components of empathic awareness and were defined respectively as the ability to identify the affective reaction of another and the ability to explain the reasons for the affective reaction. The Affective-Cognitive Role-Taking Task and the Cognitive Inference Task were designed by the author to measure affective role-taking and two levels of cognitive role-taking (simple and complex). Hypotheses concerned the relationship of empathic awareness to age and intelligence, the possibility of sex differences, intercorrelations among the measures, and the effects of type and appropriateness (to the situation) of affects on affective role-taking. Sixty-seven children between the ages of 3 and 6 who were American born and who had English as the only language spoken in their home were administered the role-taking tasks and a picture Lisa Bernadette Partyka vocabulary test. Statistical analyses demonstrated the expected positive correlation between age and all three measures of empathic awareness. When the total sample was considered, intelligence (defined as mental age level) was positively related to all three measures, thus supporting the hypothesis. However, when age groups were examined separately, the correlations decreased and presented a variable pattern: affective role-taking was not significantly corre- lated with intelligence for any of the age groups, simple cognitive role-taking was positively related to intelligence for the 3 year olds, and complex cognitive role-taking correlated positively with intelligence for the 4 and 5 year olds. While no sex differences were expected for any of the measures nor for intercorrelations among them, females performed better than males on the simple cognitive role— taking measure and males showed a stronger correlation between age and the complex cognitive role-taking measure. Intercorrelations among the measures revealed significant cor- relations between affective and simple cognitive role-taking for the 3 and 4 year olds, a significant relationship between simple and complex cognitive role-taking for the 4 and 5 year olds, and non- significant correlations between affective and complex cognitive role- taking for all age groups. These results provided partial support for the proposed hypotheses. Tests on the type and appropriateness of affect dimensions of the affective role-taking measure demonstrated easier identification of positive affect (when compared to negative affect) and no difference in the identification of affects that were Lisa Bernadette Partyka common to the situation and those that were uncommon; these results supported the relevant hypotheses. The findings were interpreted as demonstrating non-egocentric qualities in the young child's ability for empathic awareness. That is, the young child was found to exhibit an awareness of the different affective reactions of others; difficulty in identifying and explain- ing feelings seemed to come from unfamiliarity with the type of affective reaction rather than from the egocentric stance that others react as the self does. The role that language ability has played in measuring egocentrism in past studies of interpersonal role-taking was discussed and the need for a more precise definition of egocentrism was emphasized. Overall, the results were thought to uphold the notion of decreasing egocentrism in the 3 to 6 year old period since empathic awareness increased steadily over this age range. The instruments used in the study were evaluated and ideas for future research were pr0posed. TNO ASPECTS OF EMPATHIC AWARENESS IN YOUNG CHILDREN: AFFECTIVE AND COGNITIVE ROLE-TAKING By Lisa Bernadette Partyka A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1974 To Mom and Dad For your encouragement throughout the years ii ACKNOWLEDGMENTS Many people deserve thanks for their help in the preparation of this thesis: To Dr. Lucy Ferguson, my thesis chairperson and academic advisor, for her guidance, encouragement, and support throughout the five years of my graduate program. She has chaired both my master's thesis and doctoral dissertation, has shared her knowledge, has super- vised my clinical work, and has lent personal support during the fre- quent critical moments in the life of a graduate student. I give her my deepest thanks for her professional help and for her friendship. To Drs. Bertram Karon, Ellen Strommen, Claire Siegel, and Mrs. Elaine Williams, the other members of my thesis committee, for their suggestions, ideas, enthusiasm, and support. To Dr. Robert Docking of the East Lansing Board of Education and to the Preschool Committee at the Institute for Family and Child Study, for their assistance in obtaining subjects. To the teachers at Spartan Elementary School, the MSU Laboratory Preschool, and Spartan Cooperative Nursery, who cooperated so willingly in the schedule for testing. To the parents, who showed an interest in child development research by giving permission for their children to participate. To the children, who shared their thoughts about feelings; they made this study possible. To Debbie Shulman and Mine Lange, for their enthusiasm and skill as experimenters. To Will Roberts, for his artistry in preparing the task pictures. To Drs. Jerry Wilson and Charles Bassos,for their time and patience spent in scoring responses. To Steve Olejnik at the Research Consultation Office, for his talent with the computer. To the Psychology Department, for their financial assistance and to Mr. Roger Halley, for his skill in handling the accounts. To Barb, my roommate and consultant, for many long hours spent in discussing ideas, solving problems, ironing out details, and enjoying successes. Only a friend could give so much. Lastly, to Bill, whose love, support, and understanding throughout the year gave me the strength and motivation to finish. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES INTRODUCTION Research on Empathic Awareness in Children . Piaget's Concept of Egocentrism Borke' 5 Challenge . Rationale for the Present Study The Concept of Decreasing Egocentrism The Shift of Perspective The Role of Language Expression in Cognitive Egocentrism Statement of Hypotheses . . . . . . . METHOD Subjects Experimenters Procedure . Instruments RESULTS Empathic Awareness and Age . . Empathic Awareness and Intelligence Sex Differences in Empathic Awareness Positive vs. Negative Affect . . . . . Appropriate vs. Inappropriate (to the Situation) Affects Relationships Among the Empathic Awareness Measures . DISCUSSION . Age Changes in Empathic Awareness .. Empathic Awareness and Intelligence Sex Differences in Empathic Awareness Page vii viii ix Page Relationship Among the Empathic Awareness Measures . . . 70 Dimensions of the Affective Role-Taking Measure . . . . 75 Evaluation of the Instruments . . . . 77 Implications of the Findings for Child Development Research . . . . . . . . . 80 Additional Ideas for Future Research . . . . . . . . 85 SUMMARY . . . . . . . . . . . . . . . . . . . 88 APPENDICES . . . . . . . . . . . . . . . . . . 9l BIBLIOGRAPHY . . . . . . . . . . . . . . . . . 145 vi Table l. 2. l0. ll. 12. l3. 14. LIST OF TABLES Age and Sex Characteristics of the Sample Correlations Between Age and the Three Empathic Awareness Measures . Ranges, Means, and Standard Deviations of Scores on the Three Empathic Awareness Measures for the Three Age Groups . Correlations Between Intelligence and the Three Empathic Awareness Measures for the Total Sample Correlations Between Intelligence and the Three Empathic Awareness Measures for the Separate Age Groups . . . . . Analysis of Variance for Affective Role-Taking as a Function of Sex and Age Analysis of Variance for Cognitive Role-Taking as a Function of Sex and Age . . . . . . Analysis of Variance for the Cognitive Inference Task as a Function of Sex and Age . Means for Males and Females on the Three Empathic Awareness Measures . . . . Correlations Between the Three Empathic Awareness Measures for Males and for Females . Analysis of Variance for Affective Role-Taking as a Function of Sex, Age, Appropriateness of Affect, and Type of Affect . . . . Correlations Between Affective and Cognitive Role- Taking for the Three Age Groups . Correlations Between Affective Role-Taking and the Cognitive Inference Task for the Three Age Groups . Correlations Between Cognitive Role-Taking and the Cognitive Inference Task for the Three Age Groups . vii Page 30 47 48 49 50 5] El 52 52 53 55 58 59 60 LIST OF FIGURES Figure I. Mean scores for males and females on the Affective Role-Taking Measure plottes at six month intervals . 2. Mean scores for males and females on the Cognitive Role-Taking Measure plotted at six month intervals . 3. Mean scores for males and females on the Cognitive Inference Task plotted at six month intervals viii Page 62 62 62 INTRODUCTION The present study focused on the phenomenon of empathic aware- ness as it is exhibited by young children. Empathic awareness can be defined as a sensitivity to the feelings and thoughts of another. This sensitivity includes recognition and understanding of the thoughts and feelings of another so that the empathically aware child can be said to "know" and sometimes "feel" the emotional states of others. This study was aimed at investigating two aspects of empathic awareness in the young child. These two aspects were defined as two types of role-taking ability and were labeled (by the author) affective and cognitive role-taking respectively. Affective role-taking was operationally defined as the child's ability to identify the affective reaction of characters in stories, while cognitive role-taking was defined as the child's ability to explain the reasons for the char— acter's affective reaction. The study used the term "role-taking ability" because the capacity for empathic awareness is thought to depend upon the child's ability to take the role of the other both affectively and cognitively within an emotion producing situation. This study explored the developmental changes and possible sex differences in these role-taking abilities across the 3 to 6 year age range. Also, intercorrelations between these two types of role- taking abilities and a possible relationship with intelligence were part of the investigation. In short, the quality and degree of l empathic awareness in the child as related to age, sex, and intelli- gence were the main objects of inquiry. Research on Empathic Awareness in Children A child's sensitivity to the feelings of another has been investigated under a variety of labels: empathy (Borke, 1971, 1973; Burns and Cavey, 1957; Cohen, 1973; Dymond, et_el,, 1952; Feshbach and Roe, 1968; Walton, l936); apperceptive reaction (Amen, 1941); awareness of affect (Brandt, 1972; Gilbert, 1969); emotional sensi- tivity (Cheyne and Jahoda, 1971); social sensitivity (Rothenberg, 1970);social perception (Gates, 1923); and interpersonal perception (Borke, 1971). This diversity of labels reflects the differences in focus and methodology which characterize the research. However, all of the studies have investigated some aspect of the child's interpersonal sensitivity. The majority of studies have limited themselves to measuring the child's ability to label emotional states produced either in himself or in a stimulus character within stories, pictures, or dialogue (Borke, 1971, 1973; Brandt, 1972; Burns and Cavey, 1957; Cheyne and Jahoda, 1971; Cohen, 1973; Dimitrovsky, 1964; Feshbach and Roe, 1968; Gates, 1923, 1927; Rothenberg, 1970; Ruckmick, 1922). Several studies have gone further in assessing the child's understanding of the reasons for the affective state (Dupont, 1959; Dymond, et;e1,, 1952; Flapan, 1968). The present study incorporated both aspects of past research in that the child was asked to identify egg explain the reasons for the affective state produced in story characters. The kinds of stimuli which have been used in past research are illustrated by the following variations. Some of the earliest studies (Gates, 1923, 1927; Ruckmick, 1922; Walton, 1936) used the identifica— tion of emotion in faces as an index of affect awareness. Later studies presented pictures of situations in which affect was either clearly or ambiguously portrayed by the character and implied by the elements of the situation (Amen, 1941; Burns and Cavey, l957; Dymond, et_gl,, 1952; Feshbach and Roe, 1968). Still another method was the verbal descrip- tion of situations through prose or dialogue, either alone or accom- panied by a picture (Borke, 1971, 1973; Feshbach and Roe, l968; Rothenberg, 1970). The detection of emotion in voices was likewise researched (Cheyne and Jahoda, l97l; Dimitrovsky, 1964; Gates, 1927). Two additional procedures were the use of affect rating scales and affect tests (Gilbert, 1969) and the pairing of emotion adjectives to lines (Walton, 1936). An overall analysis of the methods used to assess affect identification shows that studies have generally differed in their emphasis on visual and auditory perception or on conceptualization. Subjects have been exposed to such perceptual cues as facial expression, body posture, tone of voice, and action in a situation. In other studies, subjects have been asked to comprehend a verbal description of a situation. The differences in methods point to two facts: first, that different cognitive processes are being used (perceptual or con- ceptual), and second, that the results of the studies are influenced by the ease with which different cues can portray specific affective states. In the present study, both perception and conceptualization were used; subjects listened to stories and were aided somewhat by the perceptual cues provided by the pictorial scene, by the tone of the dialogue, and by the body posture of the characters in the scene. How- ever, performance relied most heavily on comprehension of a verbally described situation since direct perceptual cues were minimal. Another variation in past research is shown by the several dif- ferent ideas about the nature or essence of empathy. Some researchers have used the words empathy and affect awareness interchangeably, while others have held clearly to a distinction. There have been primarily three different ways in which empathy has been conceptualized. First, empathy has been argued to exist ggly_when there is a vicarious emotional response. Feshbach and Roe (1968) hold to this idea and argue that social comprehension may be a prerequisite for empathy but it is not a sufficient condition for empathy. The second view regards empathy as an awareness of the affective state of another without sharing the same emotion. Most of the research falls in this second classification despite the semantic label used. The present study also falls in this second category since the subjects were asked to identify the feelings of story characters. A third view considers empathy to be primarily a cognitive ability to predict someone else's responses on a test instrument. Dymond (1948, 1949, 1950) defined empathy in this cognitive sense and related its presence in an indi-’ vidual to certain personality traits. Past research on empathy has been predominantly concerned with age changes in this ability and with correlations with other personality characteristics. Consistently, a positive relationship between age and empathic ability has been demonstrated for children in the 3 to 14 year old age range. Gates (1923, 1927), Amen (1941), Burns and Cavey (1957), Gilbert (1969), Borke (1971, 1973), and Cohen (1973) included children as young as 3 or 4 in their studies and observed an increase in empathic ability with increasing age. Dymond, etggl, (1952), Dupont (1959), Rothenberg (1970), and Cheyne and Jahoda (1971) used children somewhat older (range of 5 to 14) and also found an increase in empathy with age. In the present study using young children, degree of empathic awareness (both affective and cognitive role-taking) was expected to increase with age. The findings for sex differences in empathic ability have been contradictory. Some studies have shown girls to be more empathic than boys (Borke, 1973; Cohen, 1973; Dimitrovsky, 1964), while other studies have found no sex differences (Borke, l97l; Cheyne and Jahoda, 1971; Rothenberg, 1970). An interaction of sex of subject with sex of stimulus was noted by Feshbach and Roe (1968) and an interaction of sex and age of subject was demonstrated by Gates (1923) and Gilbert (1969). The subjects for the present study came from a nursery school and kindergarten population; since the socialization and educational experiences provided by these institutions may tend to eradicate sex differences, no sex differences in empathic awareness were expected. Empathic ability has shown a positive correlation with intelli- gence in most studies. The recognition of emotion in a voice was positively related to both verbal and non-verbal intelligence in Cheyne and Jahoda's (1971) study, but the correlations decreased with age. At the 6 year old level, the investigators found nonverbal intelligence to be more highly correlated to success on the empathy task than was verbal intelligence. However, Dimitrovsky (1964), originating the method that was later used by Cheyne and Jahoda, found that empathic ability was positively related to yergel_intelligence in children of this age (6 years). Rothenberg (1970) detected an age and 1.0. inter- action in her sample such that for 3rd graders nonverbal intelligence correlated with social sensitivity and for 5th graders verbal intelli- gence was positively related to this measure. Gilbert (1969) found no correlations between affect awareness and either an estimated 1.0. or verbal concept score from the WISC Vocabulary subtest. In the present study, both affective and cognitive role-taking were expected to corre- late positively with intelligence. This correlation was expected because the intelligence measure that was used is also an assessment of receptive language and is thought to give some indication of the child's ability to understand words that are frequently used in social communi- cation-~a skill that may be closely related to the ability to under- stand the thoughts and feelings of another. Empathic ability has also been found to correlate positively with interpersonal adjustment (Rothenberg, 1970), active consideration for others in boys (Cohen, 1973), social insight (Dymond, et_gl,, 1952), and such personality traits as maturity, expressiveness, and imagination (Gilbert, 1969). Class (confounded with ethnic) status was related to affect awareness in Gilbert's study (1969); middle class (Jewish) children were more aware of affect than lower class (Gentile) children. Gilbert also demonstrated an interactive effect of class with ordinal position in family; in her Jewish sample first—barns were more aware of affect than later-barns. Stotland, et_gl, (1971) have proposed that the relationship of the observed person to the subject who is empathiz- ing has an effect on the subject's ability to empathize. For example, they noted that later-borns (when compared to first-borns) tend to empathize better with persons who are similar to themselves or who have interacted with them. Piaget's Concept of Egocentrism Most studies on the development of empathy in children have merely concluded that social awareness increases with age. However, several have been explicit in linking the relatively poor interpersonal skill of the young child (when he is compared with the older child) to the egocentric quality of the young child's thinking during this age period (Burns and Cavey, 1957; Gollin, 1958; Flapan, 1968). The young child has been described as unable to put himself in the shoes of the other and as consequently deficient in understanding and communicating the perspective (thoughts, feelings, ideas, etc.) of the other. Egocentrism is a quality of thought that is said to exist throughout childhood (Flavell,1963), and perhaps during the entire life span (Looft, 1972). Simply stated, egocentrism is an "embeddedness in one's own point of view." While the nature of the embeddedness is different with each new intellectual stage, egocentrism manifests itself as lack of differentiation in some area of subject-object interaction (Elkind, 1970). Although its existence is perennial during childhood, it is also proposed to ebb and flow with passage from one developmental stage to another. That is, there is a surge of egocentrism at the beginning of each new intellectual stage and a decline as the stage is being traversed and another stage is being approached. At the follow- ing transition point, there is again a burst of egocentrism, but of the kind that is to characterize the new developmental stage. Since egocentrism changes character with each stage of intel- lectual development, it will be discussed as it is manifested in each stage. In the sensory-motor stage of intellectual development (0-2), the infant does not differentiate between the object and his experience of it. That is, he does not think the object exists when it is out of his view--when he is not experiencing it. The major cognitive task of this stage (as described by Elkind, 1970) is the conquest of the object. By use of the symbol (or mental representation), the infant realizes that the object has an existence of its own. The object is seen as existing externally to and apart from the infant when it can be kept in his mind by a mental representation. The major cognitive task of the pre-operational period (2-6) is the conquest of the symbol. While the child entering the pre-operational stage is capable of using the symbol, he fails to differentiate between the symbol and its referent. He thinks that the name (symbol) inheres in the object (referent) and that the symbol does not have an existence apart from that referent. At this stage the child thinks that an object cannot have more than one name, i.e., an object cannot belong to more than one class of things. Also, the child assumes that words carry much more information than they actually do, so his explanations of situa- tions, objects, etc. are often vague and imprecise. It is for this reason that the child is said to be unable to take the view of another; he does not seem to take into account the perspective of the other and to give adequate information to the other. An additional characteristic of this period is the dominance of perceptual processes over conceptual ones. The child remains fixed on single (salient) perceptual Characteristics of an object or situation and does not notice, or at least does not consider, other less salient character- istics which would provide a more accurate understanding of the object or situation. Toward the end of this intellectual stage, the child becomes capable of holding two perceptual properties in his mind at the same time and of attempting to reason about the nature of what he is seeing or experiencing, e.g., conservation tasks. It is the ego- centrism of the pre-operational period that was of interest in the present study. During the concrete operational stage (7-11), egocentrism mani- fests itself as a lack of differentiation between assumption and fact. While the concrete operational child can formulate hypotheses and explanations, he fails to distinguish between what he thinks and what the empirical facts are. While the major cognitive task of this period is to master classes, relations, and quantities (according to Elkind, 1970), it would also seem that the concrete operational child still needs to differentiate between the factual evidence and hypotheses. The major cognitive task oftfiuaformal operational period (11-15) is the conquest of thought. The early adolescent is capable of con- ceptualizing his own thought and of creating contrary-to-fact proposi- tions. Flavell (1963) uses the term “naive idealism" to describe the early adolescent's tendency to reshape and rethink reality without lO proper regard for the impracticability of his proposals and for the powerlessness of his thought to effect such changes. The egocentrism of the formal Operational period is manifested most blatantly by a failure to differentiate between what others are thinking and what the early adolescent is concerned with. For example, it is typical for the early adolescent to think that others are as preoccupied with his appearance and behavior as he is himself. Borke's Challenge The concept of egocentrism with regard to the pre-operational child was challenged by Helene Borke (1971, 1972) when she found that children as young as 3 or 3% could correctly identify the affective reactions of children in stories at better than chance expectancy. Borke presented 3 to 8 year old subjects, divided into 6 month group- ings, with stories in which child characters were made to feel happy, sad, angry, or afraid. The subjects were asked to predict the affective reaction of the characters by pointing to faces depicting the four emotions. Borke found that 3 to 3% year old subjects could accurately predict the emotions of "happy," "sad,” and "angry" at better than chance expectancy, while it was not until 3% to 4 that subjects met this criteria for the afraid feeling. Borke questioned the notion of the total egocentrism of the child at this age since even her youngest subjects showed an awareness of the affective reactions of others. Borke's challenge to the notion of egocentrism in the young child did not go unanswered; Chandler and Greenspan (1972) published a reply in the same journal a year later. These authors criticized Borke's study on the grounds that she did not make a clear distinction 11 between non-egocentric and egocentric thought. They defined non- egocentric thought as the "ability to anticipate what someone else might think or feel precisely when those thoughts and feelings are different from one's own" (p. 105). They suggested that Borke's study might have been measuring egocentric thought in that the subjects may have been giving their own affective reactions to the stories and that projection or sterotyping may have produced the correct results. The controversy between Borke and Chandler and Greenspan raises some interesting questions about the nature of egocentrism for the pre- operational child. Also, it highlights the question of how to measure egocentrism for the child of this age. While egocentrism is primarily a quality of thought, it is said to pervade all areas of the child's functioning, so that the child thinks, speaks, perceives, behaves and perhaps feels egocentrically. The child is said to be tied to his own point of view; thus he operates purely from his own perspective. Flavell (1963) has described this state and stage of egocentrism in the following words: The young child is . . . the unwitting center of his universe. Only his own point of view-~his schemas, his perceptions, etc.-- can really figure in his various activities, since he is unaware that others see things differently, i.e., that there are points of view of which his is only one (p. 274). Further, Flavell (1963) adds that it is social interaction which pro- vides the impetus for overcoming egocentrism: One of Piaget's firmest beliefs, repeated over and over in scores of publications, is that thought becomes aware of itself, able to justify itself, and in general able to adhere to logical-social norms of noncontradiction, coherence, etc., and that all these things and more can emerge only from repeated interpersonal interactions (and especially those involving arguments and disagreements) in which the child is 12 actually forced again and again to take cognizance of the role of the other. It is social interaction which gives the ultimate gggp_de_g:ege_to childish egocentrism (p. 156). From Flavell's words it would seem that the pre-operational child is limited both by his immature cognitive capacities and by his lack of social experience. Thus, he is egocentric. As mentioned pre- vidusly, this egocentrism manifests itself in all areas of the child's functioning (since they are based on the child's thinking)--percethal, conceptual, and affective. Questions raised by Borke's and Chandler and Greenspan's argument with regard to the egocentrism of this age child are: How is egocentrism manifested in each of these areas separately, i.e., perceptual, conceptual, and affective? Further, is the child equally egocentric in all of these areas, or is the child more capable of taking the role of the other in one or another of these spheres? Finally, is it reasonable to make the blanket statement that a child j§_or j§_ggt_egocentric without qualifying the degree of ego- centrism in each of the areas? Before these questions can be answered, it is necessary to decide upon the criteria for non-egocentrism. The query becomes: What thoughts, behaviors, etc. must be manifested by a child before he is said to have successfully overcome egocentrism and to show non-egocentric functioning? In short, when is a child non-egocentric? This question is relevant to deciding upon the best possible measure for differentiat- ing egocentrism from non-egocentrism. Also, it is essential to answering the controversy posed by Borke and Chandler and Greenspan. When Chandler and Greenspan responded to Borke's challenge, they included a study of their own which purported to uphold the concept of 1 ul ‘ll ill I I'll, .il l3 egocentrism in the pre-operational child. Chandler and Greenspan asked their subjects (6 to 13 years of age) to view three cartoon sequences involving a character in an affect-producing situation. The subjects were then requested to identify the resultant affect in the character. Sequels to the initial cartoon sequences were next presented. In these sequels, the character, having been joined by a second child, was depicted as behaving in a way that was congruent with his previously induced affective state but which was incomprehensible to the late- arriving child who had not witnessed the earlier preceding_situatidn. The subjects were then asked to relate the story first from their own viewpoint and then from the viewpoint of the second child. As expected, almost all children correctly identified the affective state of the first character but they varied according to age in their ability to differentiate their own point of view from the viewpoint of the late- arriving child. The authors cited an 85% error level in 6 year olds and a 4% error level in 13 year olds. In order to perform successfully on their cognitive egocentrism task, Chandler and Greenspan's subjects had to put aside information from their own viewpoint and to shift to the perspective of the late arriving child who had less knowledge than they. The task truly seemed to create a situation in which the salience of the child's own viewpoint was put to the test. If egocentrism is defined primarily as an inability to make this kind of shift, and if the criterion for non-egocentrism would be success in making this type of perspective change, then the pre- operational child would definitely be called egocentric. Accordingly, the answer to the earlier questions posed would be that while the child 14 may show increasing and varying ability in the perceptual, conceptual, and affective spheres during the early preschool years, his functioning is still predominantly egocentric. However, contending that the child remains egocentric until well into the middle childhood years does not rule out the value of investigating the process of decreasing ego- centrism (or growth toward non-egocentrism). Also, it does not rule out the possibility that the child can make perspective shifts (or efforts toward them) in the perceptual, conceptual, and affective spheres. The complexity of the shift would seem to determine success in making it. Chandler and Greenspan's task was more complex than Borke's. Thus, taking the role of the other was more difficult in Chandler and Greenspan's experiment than in Borke's. From this fact alone, it may be concluded that children were quite capable of the skill measured in Borke's study while they were still developing the ability that was of interest to Chandler and Greenspan. Looking at the two studies more closely, it becomes evident that Chandler and Greenspan included a shift in perspective and measured the child's ability to state this perspective difference verbally. In contrast,Borke did not build in a necessary perspective change and she measured the child's understanding by a nonverbal method of pointing to the correct answer. Even though children in Borke's study may have used the words happy, sad, angry, or afraid to state their answer, this process of labeling is not as difficult as Chandler and Greenspan's requirement of giving verbal descriptions of another's viewpoint. 15 Rationale for the Present Study The difference in Borke's and Chandler and Greenspan's methodologies, and in their resultant controversy, points to the fact that these investigators were assessing skills that differed in their complexity and which therefore would be expected to have different developmental timetables. Also, the studies reveal that the young child may be capable of early non—egocentric interpersonal skills that precede the complete passage of egocentrism. Traditionally, studies on egocentrism have used subjects in the middle childhood range, since the younger child was thought to be per- vasively egocentric, and it was during middle childhood that he showed a marked change in this quality. However, more recently, several' researchers (Selman, 1971; Shantz and Watson, 1971) have turned their attention to the preschool era and have attempted to assess the degree of egocentrism and the process of decreasing egocentrism during this time. Borke's study could also fall in this category since she investi- gated an interpersonal skill that indicated some degree of perspective awareness. The issue of the young child's capacity for role-taking (as a measure of egocentrism), which was highlighted by Borke's and Chandler and Greenspan's controversy and by the more recent studies using pre- school children,was at the heart of the present study. The research design to be presented was aimed at investigating the young child's capacity for role-taking in the affective sphere (labeled empathic awareness). The study contained several unique features which are proposed to correct criticisms of past research and to present a l6 methodology which is appropriate for studying role-taking in the 3 to 6 year old child. (These features will be described briefly at this point and will be explained in more detail in the next two sections of the introduction when related hypotheses are presented.) The present study proposed affective and cognitive role-taking as two aspects of empathic awareness. Affective role-taking referred to the child's ability to understand the affective reaction of another, while cognitive role-takingreflected his ability to state the reasons for the affective reaction. The affective role-taking measurewas identi- cal to Borke's task. That is, the child was presented with a story of a character in an affect-producing situation (accompanied by a picture) and was asked to indicate the affective reaction of the character by pointing to a happy, sad, angry, or afraid face. The cognitive role-taking measurewas somewhat similar to Chandler and Greenspan's task in that the child was asked to state the reasons for the affective reaction (the viewpoint) of the character. Both of these measures were included because theywere thought to represent two slightly different levels of complexity, viz. understanding and expression, in cognitive role-taking skill. Also, they incorporated both of the skills measured separately by Borke and Chandler and Greenspan. The present study also included another feature which was aimed at correcting for Chandler and Greenspan's criticism of Borke's study. That is, a shift in perspective was built into the present tasks so that a subject was asked to identify and explain two different affective reactions (of two characters) within the same scene. Initially, the. child was requested to give his own affective reaction to the situation 'Il Ill l I ll ll 1" ‘Il'. A all! I ‘1' ll [{[1’1 [ff «I‘ll! 11 1 '12 \ {{I. { (l\ IIIII‘lrlll p I 17 (so that his own reaction cou1d become salient) and then he was pre- sented with the different affective reactions of two characters. This shift in perspective was incorporated to measure the child's ability to identify and explain a reaction different from his own. Chandler and Greenspan criticized Borke's study on the grounds that the child might have been giving his own reaction. Further, they stated that egocentrism can only be measured when the perspective of another is different from one's own. Overall, the inclusion of this shift intended to present a task that more accurately measured role-taking skill (and level of egocentrism), since the ability to make a perspective shift is the ' essence of role—taking. The Concept of Decreasing Egocentrism As mentioned above, several studies, focusing on the role-taking ability of the preschool age child, have attempted to delineate the process by which the yound child progresses on the road toward non- egocentrism. These studies have described the steps of decreasing or waning egocentrism (each step itself being a measure of egocentrism). From studying perceptual and conceptual role-taking in 4, 5, and 6 year old children, Selman (1971) preposed a 4 level progression in the manifestation of this skill over this age range. He maintained that initially the child is not aware of perspective differences. The child then realizes that differences exist but does not think that he can know another's perspective. The next step is also a realization of a differ- ence, but the child attributes his own perspective to another by reason- ing that the other thinks as he does. Finally, the child makes an attempt to predict the different perspective of the other while 18 simultaneously realizing that he may be in error. Selman added that his subjects were more successful at perceptual role-taking than at con- ceptual role-taking, suggesting that skill in these two areas may not develop commensurately. Shantz and Watson (1970, 1971) also proposed a step-wise process in the development of accurate perceptual role-taking. Their steps are quite similar to Selman's. Initially, the child has no aware- ness of visual perspective differences. Then the child recognizes that things look different from different positions, but he has no specific expectancies as to how they should look; they are merely same or dif- ferent. Thirdly, the child can predict one-to-one spatial relations but he does not organize objects into an integrated spatial framework until he has reached the next stage of development. Just as Selman's and Shantz and Watson's studies attempted to delineate the perceptual and conceptual role-taking capacities of the young child, so too the present study focused on an assessment of the capabilities of the young child with regard to affective and cognitive role-taking (as components of empathic awareness). Rather than ask a blanket question as to whether the young child j§_or i§_fl9t_empathically aware, a more reasonable inquiry would be to ask in what ways is he empathically aware or how empathically aware is he. Specific to the present studywere such questions as: Is the young child merely aware in the sense of being able to identify (label) the affective reaction of another? Or is he aware also in the sense of being able to understand and in addition explain the reasons for the affective reaction? 19 Both of these questions point to the two types of role-taking ability posited for investigation in this study. The first question refers to affective role-taking and reflects the child's capacity to put himself in the shoes of the other affectively and to know how the other is reacting (along this dimension). This measure was purposely limited to merely identifying the affective reaction, for some young children may be capable of comprehending feelings without being able to explain (in words) the reasons for the feelings. The analogy of an aphasic individual who understands but who cannot express himself verbally can be cited to illustrate this point. The use of a nonverbal method of pointing to the proper affective reaction (depicted by different faces) was aimed at minimizing the verbal requirements of the affective role-taking task. The second question refers to the cognitive role—taking measure which assessed the child's ability to express verbally his understanding of the thoughts and feelings of another. Empathic awareness was divided into these two types of role- taking abilities for several reasons. First, these abilities are thought to be somewhat distinct although related. Second, it is quite likely that these abilities emerge at different times developmentally. That is, affective role-taking (as defined in this study) may be an earlier and more rudimentary ability than cognitive role-taking (also as defined in this study). Third, studies of interpersonal role-taking have tradi- tionally relied heavily on language ability and it is thought that this emphasis has confounded interpersonal awareness or understanding with verbal expression and has not allowed for an assessment of the young child's more primitive role—taking capabilities in the interpersonal 20 sphere. The division into the two types of role-taking ability in the present study was thought to provide a method and conceptualization which are more suitable for assessing the capabilities of the ygygg_child than has been possible in many previous studies on interpersonal role-taking. While affective and cognitive role-taking, as defined in this study, were distinct abilities, the were related. Both were based on an understanding of the thoughts and feelings of another. They required different language functions in that cognitive role-taking demanded facility with language expression for successful performance. In con- trast, affective role-taking relied only on receptive language since the subject was required to merely point to his answer (the affective reaction). This factor of language skill which differentiated the two tasks was expected to produce a modification of the relationship between affec- tive and cognitive role-taking for the three age groups. That is, affective and cognitive role-taking were expected to be positively related for the 4 and 5 year olds, but not for the 3 year olds. It was thought that more 4 and 5 year olds would have sufficient capacity for language expression to perform equally well (or poorly) on the two tasks, while 3 year olds would show more variability in their performance. The ~3 year olds with good language expression skills would probably perform similarly on both tasks, while those with poorer expressive language would probably do relatively better on the affective role-taking task. In short, when the child has good expressive language skills, his per- formance on the two tasks is thought to be based mostly on his eggegy standing of the thoughts and feelings of the story character. In 11.! 1.11 llllulll'lltll'l'll l‘l.||lllll.ll[{ll [flu 21 contrast, when the child does not have the expressive language skills necessary for the cognitive role-taking task, his performance on the two tasks probably would be more variable andnfightrtfiflect his language gkill_more than his understanding. This latter child may understand the thoughts and feelings-just as well as the first child (one who has the expressive language skills) but he may not be able to convey this under- standing as well. The Shift of Perspective The shift of perspective in the Affective-Cognitive Role-Taking Task was created by including two different affective reactions (of two different characters) in the same background situation. The subjects were asked to identify the two affective reactions and to explain the reasons for both. Two dimensions were included in the task to insure a shift in perspective. These two dimensions were type of emotion (positive vs. negative) and appropriateness to the situation (appro- priate vs. inappropriate). The shift in perspective was accomplished by having one character with a positive affective reaction (happy) and one with a negative affective reaction (sad, angry, or afraid). Also, one of these reactions (either positive or negative depending on the particular story) was considered more common or appropriate to the situation and the other was considered less common or inappropriate to the situation. Two hypothesesrelated U3 these dimensional qualities of the task. First, it was expected that there would be more correct identifi- cations of positive as compared to negative affect across all ages. Past research upholds the notion of easier identification of positive 22 feelings (Alexander, e§_el., 1971; Borke, 1971; Feshbach and Roe, l968; Rothenberg, 1970). Second. no difference was expected in the identi- fication of affects that were considered appropriate to the situation and those whichwere considered inappropriate. Errors in identification were thought to result from unfamiliarity with type of affect (positive vs..negative, or specific negative affect) rather than from the appro- priateness dimension. Even though Borke's study did not include a shift in perspective, it was thought by this author that Borke's find- ings illustrate the young child's awareness of others' affective reactions whether they are similar or different from the child's own reaction. One study in the literature on empathic awareness did utilize a type of perspective shift for the measurement of affect identification. Burns and Cavey (1957) presented 3 to 6% year old children with a series of pictures, two of which were considered "crucial” situations. The crucial situations, a birthday party and a doctor's office with a doctor holding a needle in his hand, were presented to the child and he was asked to state how he would feel in the situation. The child was shown the scene again, but the scene this second time included the figure of a same sex child whose facial expression was depicting an emotion dis- crepant from that which the situation implied, i.e., sad at the birthday party and happy in the doctor's office. The subject was then asked how the child in the picture was feeling. Burns and Cavey found more accurate descriptions of the affective state of the figure in the older age group of his subjects. 23 The present study is similar to Burns and Cavey's in that one of the affective reactions was different from that which is commonly shown, but this affective reaction was comprehensible in light of the dialogue and description of extenuating circumstances given in the story. Also, Burns and Cavey's task relied wholly on perceptual cues for affect identification while the present task relied most heavily on comprehen- sion of a story. The Role of Language Expression in Cognitive Egocentrism The role of language ability in measuring egocentrism has been discussed briefly at two previous points in the introduction. First, it was mentioned that Chandler and Greenspan's assessment of their subjects' verbal descriptions of another's point of view was quite different in complexity from Borke's pointing method. Second, when hypothesizing that affective and cognitive role-taking would not be correlated for 3 year olds, it was mentioned that receptive and expressive language may vary in the same child so that some children may understand more than they can express verbally. While the trend in the literature is to say that the young child's language is markedly egocentric, several more recent studies indicate that the child of this age is capable of some understanding of another's viewpoint and of adapting his speech accordingly. Maratsos (1973) studied referential communication in 3, 4, and 5 year old children. His design required subjects to describe one object from a series of arrays. The main focus in his study was on the differences in quality of communication between a normal-vision and blocked-vision condition. 11 I] A l 1' l lull III: Ylll '11! ff {'1 {l - [([III‘ II“ I ‘( A [i {ll (1 24 The subjects were asked to describe one toy from among 3 or 4 to either a sighted experimenter or to one who put his hands over his eyes. For each age level, adequacy of description was significantly better in the blocked-vision condition, suggesting that the young child was able to take listener attributes into account When formulating his communication. Also, in the normal-vision condition, pointing to the toy rather than describing it was quite common, although pointing decreased with age. In the blocked-vision condition, there was almost a complete absence of pointing for all age levels. This aspect of the subjects' behavior suggests that the young child may have the ability for accurate verbal encoding, but that he may not use this ability unless the situation clearly necessitates his doing so, for instance when pointing is inade- quate for communicating his idea. Garvey and Hogan (1973) videotaped 15 minute play sessions of dyads of children, ages 3% to 5. With a focus on the behavior and speech of the children within the interaction, they found a high level of mutual responsiveness between children in this age range. The children were capable of adapting their speech to the verbal and non- verbal behavior of their partner. While the results are interpreted as not contradicting the presence of egocentric speech in the child of this age, they do emphasize the contention that genuine social behavior occurs in this age group. This study seems to illustrate the concept of decreasing egocentrism in the child of preschool age. In the two studies cited, children of preschool age did manifest some ability to take the role of the other. Success in these studies seems to reflect the level of task complexity. In contrast to these 25 relatively simple designs are more complex cognitive tasks like those used by Chandler and Greenspan (1971) and Feffer and Gourevitch (1960). Chandler and Greenspan's design has already been described so a brief account of Feffer and Gourevitch's will be given. Feffer (1959) developed the role-taking task which has been used in most studies measuring interpersonal role-taking. The task was first demonstrated with children by Feffer and Gourevitch (1960). The Role-Taking Task (RTT) consists of two background scenes from Schneidman's Make a Picture Story (MAPS). The child is asked to choose at least three figures for each background from a wide assortment of men, women, children, and animals and to tell a story using the three figures. After the two stories are related, the first is read back to the child and the child is asked again to tell the story, but this time from the per- spectives of each of the three characters (creating three related stories). This procedure is repeated with the second story. The four protocols from each story are scored in terms of the degree of "balanced decentering" that is evident. This decentering includes refocusing, elaboration, and change of perspective among the stories for the different characters in one background scene. Role-taking ability is reflected by the balanced decentering score since this ability is reasoned to measure the degree to which the subject can "put himself in the shoes of the other"--to think his thoughts, feel his feelings, etc. It is noteworthy that neither Chandler and Greenspan nor Feffer and Gourevitch used children below 6 years of age in their experiments. The choice of age seems to reflect task complexity and the importance of language ability in the skills being measured. Both sets of experimenters found that interpersonal role-taking ability increased with age. 26 The roles of language ability and task complexity were intro- duced at another point in the present experimental design by the inclu- sion of a Cognitive Inference Task. This task was separate from the Affective-Cognitive Role-Taking Task and consisted of four stories in which a main character felt sad inla happy situation or happy in a sad situation. This contrast of feelings to the context was included to insure a shift of perspective and to maximize the role-taking require- ments of the task. The Cognitive Inference Task was similar to the cognitive part of the Affective-Cognitive Role-Taking Task in that it included a per- spective shift and it required the subject to explain the reasons for the character's affective reaction. However, it differed from the Affective-Cognitive Role-Taking Task in several important ways. The most important difference lay in the greater degree of complexity of the Cognitive Inference Task. The reasons for the character's affective reaction had to be inferred by the circumstances of the stories in the Cognitive Inference Task. In contrast, the reasons for the character's affective reaction in the Affective-Cognitive Tole-Taking Task were contained more blatantly in the character's own dialogue within the story. It was thought that success on the Cognitive Inference Task would require a greater degree of reasoning and role-taking ability than would success on the Affective-Cognitive Role-Taking Task. The Cognitive Inference Task was also different from the Affective-Cognitive Role- Taking Task in that only two affects were included (happy and sad) and the subject did not have to identify the affective state of the character since this was already described and labeled in the stories. 27 These two cognitive role-taking tasks were used in the study to illustrate the point that young children may be capable of simple cognitive role-taking but not of more complex cognitive role-taking. The child's ability for two different levels of cognitive role-taking‘ was thus investigated. The cognitive role-taking measured by the Affective-Cognitive Role-Taking Task was expected to be easier and within the capabilities of 3 year olds. The cognitive role-taking reflected by the Cognitive Inference Task was expected to be difficult for some 5 year olds. Further hypotheses concerned the relationships between per- formance on the Cognitive Inference Task and cognitive and affective role-taking respectively. It was expected that the cognitive role- taking measure (of the Affective-Cognitive Role-Taking Task) and the Cognitive Inference Task would show a positive correlation for the 5 year olds but not for the 3 and 4 year olds. Similarly, a positive relationship between affective role-taking and the Cognitive Inference Task was expected for the 5 year olds, but not for the 3 and 4 year olds. Correlations for the 5 year olds were expected since it is likely that children at this age level would be capable of doing well on the Cognitive Inference Task, while performance at the lower ages would be poorer and more variable. The earlier hypotheses proposed for the affective and cognitive role-taking measures were pr0posed for the Cognitive Inference Task. That is, performance on the Cognitive Inference Task was expected to be positively related to age and intelligence of the child and was not expected to be affected by sex of the subject. 28 Statement of Hypotheses In summary, the following hypotheses were proposed for the present research: First, affective and cognitive role-taking ability and performance on the Cognitive Inference Task would increase over the 3 to 6 year age range. Second, intelligence would be positively related to the affective and cognitive role-taking measures and to the Cognitive Inference Task. Third, no sex differences would result for performance on any of the measures nor for intercorrelations between the measures. Fourth, positive affect would be correctly identified signifi- cantly better than negative affect. Fifth, no difference would result in the identification of affects considered appropriate to the situation and those considered inappropriate to the situation. Sixth, affective and cognitive role—taking would be positively correlated for the 4 and 5 year olds, but not for the 3 year olds. Seventh, affective role-taking would be positively related to performance on the Cognitive Inference Task for the 5 year olds, but not for the 3 and 4 year olds. Eighth, cognitive role-taking would be positively related to the Cognitive Inference Task for the 5 year olds, but not for the 3 and 4 year olds. METHOD Subjects Subjects were obtained from two preschools and an elementary school affiliated with Michigan State University. Families sending their children to these schools were thought to be fairly homogeneous in class status since all lived in or near the East Lansing residential community and most were directly associated with the university. Subjects between and including the ages of 3 years 0 months and 5 years 11 months were considered eligible for participation in the study if they were American born and had English as the only language spoken in their home. In calculating ages, the number of days was dropped so that ages were recorded in monthly intervals, e.g., a subject 4 years 5 months 25 days old was recorded as 4 years 5 months. Older subjects were obtained from the Spartan Village Elementary School, while the majority of subjects came from the Laboratory Preschool, and some from Spartan C00perative Nursery. Letters describing the research and requesting permission for participation were sent home to the parents of 34 children in the kindergarten grades at Spartan Elementary School, of 79 children attending the Laboratory Preschool, and of 13 children at Spartan Cooperative Nursery. All of these children met the criteria specified above for inclusion in the study. Copies of the letters sent to the parents of children at the preschools and at the elementary school can be found in Appendix A. 29 WWIIIYIIl‘llllllllull(([illlllllillffl'1. 1.1. . 30 0f the 126 permission slips distributed, 90 were returned to indicate agreement for participation. One mother among the 90 asked for additional information about the study before she agreed to her son's participation; all other parents sent the slip back without questions. The final sample consisted of 67 subjects. (The age and sex characteristics of the subjects are presented in Table 1.) Fourteen children were dropped from the study after testing was attempted: 5 children (3 and 4 year olds) could not identify the faces used in the Affective-Cognitive Role-Taking Task even after repeated trials; 7 (mostly 3 and 4 year olds, but one 5 year old) did not complete the testing because of attention span difficulties or because of time demands on the experimenters, and 2 did not understand the instructions. Testing was not attempted with 8 children for whom permission slips had been returned because of time limitations at the end of the school year. The remaining child was tested but the data sheet was lost after its completion. TABLE l.--Age and Sex Characteristics of the Sample. 3 year olds 4 year olds 5 year olds Total Males ll 7 ll 29 Females 11_ l__ 13_ 38_ Total 22 21 24 67 Mean age 3.63 4.56 5.50 Median age 3 yr. 8 mo. 4 yr. 7 mo. 5 yr. 6 mo. 31 Experimenters Three female experimenters administered the tasks in the present study. Two of the women were undergraduate psychology majors who had no formal training in working with preschool children but who showed a natural skill in relating to children of this age. The third experimenter was the author who holds a masters degree in psychology and who has considerable clinical experience with young children. An attempt was made to distribute subjects evenly by age for the three experimenters. This attempt was successful for the 4 and 5 year olds, but not for the 3 year olds. The author tested twice as many 3 year olds as either of the other two experimenters. This imbalance resulted when a significant number of 3 year olds did not successfully complete the testing and more subjects in this age range were needed for the sample. Because of class schedules, the other two experimenters did not have the time to administer the tasks to these 3 year olds so the author tested these additional subjects. To test for a possible experimenter effect, a one way analysis of variance was performed on a subset of 51 subjects matched for age. Seventeen subjects for each experimenter were inlcuded since 17 was the total number tested by one of the experimenters. The results indicated no significant experimenter effect for any of the three empathic aware- ness measures . Procedure Parental permission slips were distributed to all eligible children at the Laboratory Preschool and in two kindergarten grades at Spartan Village Elementary School. The subjects from Spartan 32 Cooperative Nursery consisted only of 3 year olds and were recruited when more younger children were needed to round out the sample size. After the permission slips were distributed and returned, arrangements were made to begin testing at the school facilities. In the two preschools, experimenters spent time in the class- rooms of the subjects they were to test before beginning the actual testing. This procedure was aimed at giving the subjects an opportunity to familiarize themselves with the experimenter. This initial time period in theclassrooms varied, but the experimenters averaged about an hour in classrooms containing more than three subjects and about 20 minutes per subject when only one or two subjects were part of a class- room. Experimenters made an effort to engage in play or conversation with the subjects they were to test. This preliminary acquaintance period was not carried out at Spartan Village Elementary School since it would not fit well into the classroom structure and it was thought that older children would not need this preliminary period to reduce fearfulness. In all of the classrooms, an effort was made to take children for testing at times that were not disruptive to the classroom program. The schedule of testing was requested by the teachers and followed by the experimenters in order to minimize the child's chance of missing the most important classroom activities and to maximize his motivation for completing the tasks. The tasks were administered at several different locations in each of the schools involved. The testing sites were set apart physically from the classroom, e.g., a consulting room, empty classroom, 33 etc., and were generally quiet enough to avoid distractions, but occasionally the voices of children could be heard from nearby class- rooms. In initially approaching a subject for testing, an experimenter introduced herself again and said that she had some stories for boys and girls to listen to. Depending on the subject's age and demeanor, the experimenter either asked the subject if he/she would like to hear the stories or she simply stated that it was their time to listen to the stories. The experimenter added that the stories would be on a tape recorder and that she was interested in what children thought about these stories. The experimental materials consisted of a tape recorder, cassette tapes containing the recorded stories (for the Affective- Cognitive Role-Taking Task and the Cognitive Inference Task), a series of 12 pictures and 4 faces (depicting the emotions of happiness, sadness, anger, and fear) for the Affective-Cognitive Role-Taking Task, a set of 4 larger pictures for the Cognitive Inference Task, the manual and stimulus cards from the Peabody Picture Vocabulary Test (PPVT), and an answer sheet and PPVT form for each subject. These materials had been arranged in a convenient order on a table or desk to the side of the experimenter and subject before the subject was brought to the testing room. For all subjects, testing was begun with the Affective- Cognitive Role-Taking Task. The number of pictures administered (of a total of 12) during the initial session depended upon the attention span of the subject, but no more than 6 pictures were given to a subject in 34 any one session. The PPVT was administered after an appropriate number of pictures for the Affective-Cognitive Role-Taking Task were given. For all but one subject, the Cognitive Inference Task was administered last because it was thought to be more difficult than the other tasks. Testing sessions varied both in length and number. Many sub- jects were tested on only two occasions since they were able to complete half of the Affective—Cognitive Role-Taking Task and the PPVT during the first session and the second half of the Affective-Cognitive Role—Taking Task and the Cognitive Inference Task on the second occasion. These sub- jects usually took about 70 to 80 minutes to complete all of the testing. The maximum number of sessions needed to finish the task was 7 while the average number of sessions was 3 or 4. The interval between start and finish of the testing was as brief as only one day and as long as five weeks (when one child was bedridden with pneumonia). The average interval was about 8 days. Each subject's age was recorded as of the first day of testing. Instruments The instruments used in the present study were the Affective- Cognitive Role—Taking Task, the Cognitive Inference Task and the Peabody Picture Vocabulary Test (PPVT). Each of the instruments will be dis- cussed separately in a section below. Affective-Cognitive Role-Takingglask The Affective-Cognitive Role-Taking Task assessed both affective and cognitive role-taking by means of two different measures. An affect identification score was taken as a measure of affective role-taking and 35 and an assessment of the subject's understanding of the reasons for the affective reaction reflected cognitive role-taking. The task consisted of a set of faces, a series of pictures, and a series of stories composed by the present author to accompany the pictures. As a method of measuring affect identification (taken from Borke, 1971), four faces, depicting the emotions of happiness, sadness, anger, and fear, were drawn in blue ink on a white background and placed separately on 5x5 inch pieces of white cardboard. The subject was required to point to the appropriate face to indicate his response for affect identification. These faces are similar to those used by Brandt (1972) when she studied affect awareness in 3 year old children. Duplications of these faces have been inserted into Appendix B. A series of 12 pictures illustrating familiar background scenes such as a zoo, classroom, living room, outdoors, etc. were also drawn in blue ink on a 9x11 inch white background and placed in an acetate cover with a piece of white background cardboard for support. These pictures are contained in Appendix C. All persons in the pictures had blank faces so as to minimize these perceptual cues for the subjects. Four different stimulus figures were made for each picture; two for these figures were females and two were males. These figures were approximately 2 to 3 inches in height and were drawn in blue ink on a white background and attached to rice paper that would adhere to the acetate when the figure was placed on the background scene. The figures had different body postures but all had blank faces. (The subject was to choose the appropriate face for the character from among the four given.) The figures were kept in an envelope glued to the back of the 36 picture and were placed at specific but different locations on the background picture when they were used. Female figures were used with female subjects and male figures were used with male subjects. This arrangement was created in order to eliminate a possible interactive effect of sex of subject and sex of stimulus and to heighten the chances that the subject would identify with the stimulus figure. Samples of the stimulus figures are placed within the background scenes contained in Appendix C. Each background picture was accompanied by a brief description of the scene in the picture and by two brief stories about the two characters (stimulus figures) involved in the scene. These stories depicted characters in affect-producing situations and contained dialogue and a description of amplifying circumstances for that particu- lar character. When the description and stories were composed, an effort was made to use situations and langauge that are familiar and comprehensible even to 3 year olds. The idea of using stories in which characters are made to feel a certain emotion was taken from Borke (1971) but the inclusion of two different characters within the same situation was the author's innovation. Since pre—testing revealed that young children are sensitive to the affective cues in a person's voice, the descriptions and stories were tape recorded by the author so as to provide a uniform stimulus for all subjects. The text of the stories is contained in Appendix 0. Subjects were initially pretrained to identify the feeling (happy, sad, angry, or afraid) depicted by each of the four faces. Pre- training was accomplished by first asking the subject to name each of 37 the feelings that the faces represented. If the subject did not use the correct label, he was told what the face was feeling. (It was found that the happy face was more easily identified than any of the negative emotion faces.) After the four faces were identified, the subject was asked again to name the feeling for each face. This procedure was con- tinued until the subject could correctly and spontaneously identify the feeling of each of the faces on a single trial. Number of trials needed to name the faces was recorded, with the first trial beginning after the experimenter had named the faces. The mean number of trials was 1.2 and the maximum number for any one subject was 3. Since administration of the tasks was completed over several sessions,pretraining preceded administration for each session, but number of trials was recorded only for the initial session. Pretraining was used as a criterion for decid- ing if the subject was capable of understanding and completing the Affective-Cognitive Role-Taking Task. Five subjects did not succeed in correctly naming the faces even after repeated pretraining trials and were consequently eliminated from the sample pool. After a subject successfully named the faces to be used in the Affective-Cognitive Role-Taking Task, he was considered ready to begin the formal testing procedure. (A detailed description of the testing procedure, including verbatim instructions, can be found in Appendix E.) The subject was first shown a background picture from the task and he was requested to listen to the brief description of the scene in the picture. At the end of the description, the subject was asked how he thought he would feel in the situation. After indicating his own feeling by pointing to the'appropriate face, the subject was asked "How come?" 38 in order to elicit his reasons for his own affective reaction. These questions about the subject's own affect were aimed at involving the subject in the task and at making his own affective reaction salient for him. Also, his responses were thought to be potentially useful in determining which of the affective reactions (represented by the two later characters in the background situation) was discrepant for him. The subject'sresponses to questions about his own affect were also used to determine the number of “egocentric" errors (errors for the two characters that matched the subject's response about his own feeling in the situation) that the subject made. The story for the first character was presented after these first questions were completed. The story included a description of amplifying circumstances for that particular character and contained dialogue (by the character) which was revealing of how the character felt. The subject was asked to indicate how he thought the character was feeling by again pointing to an appropriate face. If he hesitated, the experimenter coaxed him along by repeating the question. Once a response was given, the subject was asked'Wknv come" the character was feeling that way. Again, if the subject had difficulty responding, the experimenter waited patiently and repeated the question while point- ing to the face that the subject had indicated. The story for the second character followed in like manner. After both characters and stories had been presented for the background picture, the experimenter asked the subject to rearrange the faces so that their spatial positions would be changed. This step was included in order to control for a possible position effect in responses. 39 The procedure described above was repeated for all 12 back- ground pictures for the Affective-Cognitive Role-Taking Task. The pictures were presented over several sessions since administration of 12 pictures at any one time was thought to be too fatiguing for children of this age. The child's identification of affect for the character was used as the measure of affective role-taking, while his response to the "how come" question was taken as the measure of cognitive role-taking.‘ In order to insure a shift in perspective (as discussed in the introductory section of the study), both a positive (happy) and negative (sad, angry, or afraid) affect were presented for each background scene. One character had a positive affective reaction in the scene and the other had a negative reaction. This positive-negative dichotomy was used in order to maximize the conditions for a shift. The combinations of affects studied were: happy-sad, happy-angry, happy-afraid, sad- happy, angry-happy, and afraid-happy. Two background scenes were used to depict each combination of affects. While the combinations of affects, e.g., happy-sad and sad-happy, seem repetitive, both combina- tions were included because the first affect listed was intended to be the most appropriate affect suggested by the situation itself. The study tested the hypothesis that young children do not differ in their identification of affects that were appr0priate or inappropriate (to the situation) affects. In order to control for an order effect, the 12 background pictures were arranged into three different random orders of presenta- tion for the subjects. However, it was decided to test whether an order 4O effect might be produced when the affect that was considered appropriate to the situation was presented before the affect that was considered inappropriate, and vice versa. Thus, the sample was divided into those who were first presented with the character exhibiting the appropriate affective reaction and those who were first shown the character with the affective reaction considered inappropriate (or less common) to the situation. Originally it was intended to divide the sample in half; with the odd numbered sample size, a nearly even split was produced: 34 children received the appropriate (to the situation) affect first, while 33 received the inappropriate (to the situation) affect first. T tests between the means for the two groups on the affective and cognitive role- taking measures showed no significant difference in performance. Thus, the order of presentation did not produce an effect on correctness of response. The randomization of order (3 different sequences) and the division of the sample into those who would receive the appropriate or inappropriate affect first produced 6 different arrangements of stimulus stories for the sample. In the study, the six sets of stimulus stories were evenly distributed by age across subjects. In addition, stories were recorded separately for boys and girls since gender of pronoun and type of toys needed to be changed. As a result, 12 different sets of stories accompanied the Affective-Cognitive Role-Taking Task. Scoring for Affective Role-Taking_ Two sets of scores were computed to assess affective role- taking. Both consisted of a simple dichotomous system where the response was scored as either ”right" or "wrong.” A score of l was given 41 for "right" answers and a score of 0 represented "wrong" answers. A subject could thus achieve a total score from 0 to 24 since two affects were presented in each of 12 background pictures. The first set of scores reflected the subject's accuracy in identifying the affect precisely. That is, the subject had to point to the happy face for the positive emotion and to the appropriate negative face (sad, angry, or afraid) for the negative emotion. The subject had to differentiate correctly among the negative emotions in order to achieve a ”right" score. The second set of scores was more lenient than the first. That is, for the negative emotion in the story, the subject had to point to one of the negative faces, but not to the exact one. For example, if the appropriate response was sad and the subject pointed to angry, he was given a l for this response. However, for the positive emotion he always had to point to the happy face in order to receive credit. Although only the first set of scores was intended to be used in the formal analysis, the second set was used to test for age differ- ences in categorization of affects along the positive vs. negative dimension. Scoring for Cognitive Role-Taking The scoring for cognitive role-taking was based on the adequacy of the subject's response in conveying his understanding of the reasons for the character's affective state. The answers were recorded verbatim by the experimenters. A 4 level scoring system was used such that a higher score denoted better cognitive role-taking. 42 The scores of 3, 2, l, and 0 were used to denote level of the subject's response. Criteria for placing responses in each of the levels are explained below and scoring examples for each of the 24 stories are contained in Appendix F. Level 3: The subject gives a full explanation of the reasons for the character's affective reaction. In most stories, it is an event, i.e., getting to do something or not getting to do it, which is responsible for the character's feeling. In these cases, the subject must state that the event occurred and must state or imply strongly how the character felt about the event, i.e., whether the character had or had not wanted the event to occur. In several of the stories, where the event is not as crucial to determining the affect, but rather "liking" or "not liking" something is more important, a description of the character's "wanting” or "not liking" was placed in this level. These latter cases were discriminated when there was no implication in the story that the character was able to do something special that he otherwise might not have been able to do or that he was kept from doing something he wanted to do. Level 2: The subject implies the reasons for the affective reaction but does not state them directly. For example, in the cases where an event precipitated the feelings, a statement of a character "liking" or "not liking" to do something (without stating that he was or was not able to do it) belong in this level. This level contains responses indicating a character's internal state (which is one element of the level 3 response when two are needed). The internal state or feelings about the event implies the occurrence of the event itself. Also falling in this level are responses which are related to the main reason but not central to it; these responses are determined by the particular story. Level 1: The subject merely gives a statement of the action (event) in the story, e.g., "he's going outside," ”he's going home," without indicating how the character felt about the event. This type of response is considered to be indicative of lower role-taking skill because it is merely describing external events without getting ”inside” the character and telling how the character feels about the event. However, depending upon the story, some descriptions of events are more important and are placed in the 2 level. Responses which are tangentially related to the main reason are also categorized in this level. 43 Level 0: The subject shows no real understanding of the reasons for the character's affective reaction. Inaccurate information, no response, “I don't know," and a restatement of the affective state belong here. Also, responses stated in the first person, e.g., "because I like dolls,” are given a 0 score because they are considered to be egocentric. (Responses stated in the second person are considered non-egocentric and are scored according to the above criteria.) In addition, responses which accompany affect identification that is in the wrong direction, i.e., a negative affect indicated when the correct response is happy and a happy response when the affect is negative, are automatically given a 0 score. It was decided to give these latter cases a 0 score because of the probability that the subject misunderstood the story and was merely repeating action or dialogue from the story. Inter-rater reliability on the items of the cognitive role- taking measure was arrived at by computing the correlation between two scorers on half the total number of items. Since the correlation coefficient was .918, the scores from the first rater were considered valid and were used in the formal analyses. Since item reliability had been obtained, the Spearman-Brown formula (Walker and Lev, 1953) for predicting reliability when increasing the length of a test was followed and a correlation coefficient of .997 for the instrument was recorded. Cognitive Inference Task The Cognitive Inference Task consisted of four pictures and four stories depicting a character as feeling either sad within a happy situation or happy within a sad situation. The pictures were drawn in blue ink on a 10x12 inch white background and included four pictorial frames per picture that depicted the sequence of action in the story. Females were presented with pictures and stories of girls and males were presented with pictures and stories of boys. Faces as were used in the Affective-Cognitive Role-Taking Task were not part of the Cognitive Inference Task since the affect of the character was labeled and 44 depicted as part of the story and picture. However, the subject was asked to state the reasons for the affective reaction of the character and this was taken as the measure of role-taking on the Cognitive Inference Task. The role-taking required on this task was considered to be much harder than the cognitive role-taking from the Affective- Cognitive Role-Taking Task because the reasons for the affective reaction had to be inferred from the circumstances of the story and were - not stated blatantly in the character's dialogue as they had been in the Affective-Cognitive Role-Taking Task. The choice of affects, i.e., happy in a sad situation and sad in a happy situation, was included to produce a type of perspective shift for the subject. That is, the subject had to discriminate the affective reaction of the character and reasonsfronlthe affective reactions of the other children in the story who represented the more common reaction in the situation. For example, sadness occurred in the situations of a birthday party and of a class who was going to the store for ice cream cones, and happiness was set in a situation of a class not being able to go to the zoo and a class not being able to keep a bunny rabbit. The text of the stories for the Cognitive Inference Task is contained in Appendix H and the pictures are included in Appendix G. In order to control for the possibility of an order effect, 3 different orders of presentation of the four stories were used and were evenly distributed across subjects by age. Since stories and pictures were adapted to each of the sexes, 8 pictures and 8 accompanying stories were included in the Cognitive Inference Task--4 for females and 4 for males. The stories were presented on cassette tapes so as to make the 45 stimulus the same for all subjects. As the stories were being presented, the experimenter pointed to each new frame on the picture at the appro- priate moment in the story. Scoring for the Cognitive Inference Task As it was for cognitive role-taking in the Affective-Cognitive Role-Taking Task, scoring for the Cognitive Inference Task was based on the adequacy of the subject's response in conveying his understanding of the reasons for the character's affective reaction. Since the character's direct dialogue was not part of the Cognitive Inference Task, a 3 level scoring system was used to rate the response. The numbers 2, l and 0 were used to score the response. A 2 response reflected the highest level of role-taking and was recorded when a clear and accurate statement of the reasons was given by a subject. A 1 score reflected some understanding of the reasons for the reaction, but not a totally clear and comprehensive one. Responses related to the reasons but not directlyexplanatorylaf the reasons fell in this level. A 0 response reflected a totally inappropriate response, an "I don't know" answer, no response at all, or a restatement of the question or affective state. A more detailed explanation of the scoring system, including examples of the various levels, can be found in Appendix I. Inter—rater reliability for the Cognitive Inference Task was arrived at by following the method outlined for the cognitive role—taking 46 measure. The computed correlation was .915, while the predicted reliability for the instrument was .977. Peabody Picture Vocabulary Test The Peabody Picture Vocabulary Test (PPVT) was used as a measure of intelligence in the present study. The test consists of an instruction manual and a set of stimulus cards which contain four pictures each. The child being administered the test is required to point to the picture (from among the four) of the word that the tester says. The test actually measures the receptive vocabulary of the child and it was specifically used in this study because it is thought that some early efforts toward role-taking may reflect a child's ability to understand another's point of view before being able to express that viewpoint verbally. Thus, a measure of receptive language, also used as the assessment of intelligence, was thought to be appropriate for the sample used. The raw score from the PPVT was used as the measure of intelligence since raw score can be considered a mental age equivalent. RESULTS Empathic Awareness and Age Hypothesis 1 concerned the relationship of empathic awareness to age. It was expected that affective and cognitive role-taking and successful performance on the Cognitive Inference Task would increase over the 3 to 6 year age range. Pearson product moment correlations between continuous age points (age in months) and each of the measures were used to test this hypothesis. The correlations, presented in Table 2, are significant (p < .001) for each of the measures. TABLE 2.--Correlations Between Age (Computed in Months) and the Three Empathic Awareness Measures (df=65). Affective ' Cognitive Cognitive Role-Taking Role-Taking Inference Task .657*** .743*** .642*** ***p < .001 In order to detect the magnitude of the differences between means for the separate age groups, 2x3 analyses of variance for unequal cell A frequencies (least squares solution) and Scheffe's method for post-hoc comparisons were performed on the means. Tables 6, 7, and 8 show that age as a main effect is significant (p < .0001) for all three measures. Cell means for the analysis of variance tables are contained in 47 48 Appendix J. Ranges, means, and standard deviations of scores on the three measures for the three age groups are presented in Table 3. TABLE 3.--Ranges, Means, and Standard Deviations of Scores on the Three Empathic Awareness Measures for the Three Age Groups. Measure 3 year 4 year 5 year (Range) Scores olds olds olds minimum 5 14 14 Aff RT maximum 18 23 24 (0-24) mean 13.05 18.76 19.21 S.D. 3.79 2.55 2.75 minimum 0 35 49 Cog RT maximum 57 62 66 (0-72) mean 29.73 50.71 57.17 5.0. 15.85 7.18 4.48 minimum 0 O 3 Cog Inf maximum 6 7 8 (0-8) mean 2.18 4.10 5.38 5.0. 1.56 2.07 1.53 When the means for just 4 and 5 year olds are compared by Scheffe's method for post hoc comparisons, a significant difference results only‘ for the Cognitive Inference Task (p < .05). However, when the means for the 3 year olds are compared to the means for the 4 and 5 year olds respectively, significant differences show up for’; all three measures (p < .01). Thus, the analyses show that the 3 year olds performed quite differently from the 4 and 5 year olds on all three measures, while the 4 and 5 year olds generally performed the same except for the Cognitive Inference Task. The comparisons reveal that for the affective and cognitive role-taking measures the variance for the main effect of age can be accounted for largely by 49 the difference between the 3 year olds and each of the two older age groups. The variance for the Cognitive Inference Task is more evenly distributed. Hypothesis 1 is thus strongly supported by the data. Affective and cognitive role-taking and successful performance on the Cognitive Inference Task increased over the 3 to 6 year age range when both con- tinuous age points and yearly intervals were considered. Empathic Awareness and Intelligence Hypothesis 2 concerned the relationship of empathic awareness to intelligence. It was expected that affective and cognitive role- taking ability and performance on the Cognitive Inference Task would be positively related to intelligence. Since subjects at different chronological ages were included in the study, intelligence was defined as mental age level and was measured by the PPVT raw score. Pearson product moment correlations between PPVT score and each of the measures were used to test this hypothesis. When all subjectswereiconsidered, correlations for all three measures are significant (p < .001). Table 4 presents the correlations for the total sample. TABLE 4.--Correlations Between Intelligence (PPVT Score) and the Three Empathic Awareness Measures for the Total Sample (df=65). Affective Cognitive Cognitive Role-Taking Role-Taking Inference Task .583*** .676*** .622*** ***p < .001 50 When age is held constant, the correlations between intelligence and the three measures are generally lower and show a variable pattern. The correlations are presented in Table 5. As the table indicates, affective role-taking is not significantly related to mental age level for any of the age groups. Cognitive role-taking shows a significant correlation for the 3 year olds, and the Cognitive Inference Task is positively related to mental age level for the 4 and 5 year olds. TABLE 5.--Correlations Between Intelligence (PPVT Score) and the Three Empathic Awareness Measures for the Separate Age Groups. -_m___.¢‘. _____ 3 year olds 4 year olds 5 year olds df=20 df=l9 df=22 Affective Role-Taking .275 .164 .339 Cognitive Role-Taking .439* .300 .136 Cognitive Inference Task .174 .643** .540** *p < .05 **p < .01 Thus, Hypothesis 2 is only partially supported when age groups are considered separately, but is strongly supported when all subjects are included. Sex Differences in Empathic Awareness Hypothesis 3 concerned the possibility of sex differences in empathic awareness. No differences between males and females were expected for performance on any of the three measures nor for correla- tions between the measures. 51 The first part of the hypothesis was tested by 2x3 analyses of variance for unequal cell frequencies. Tables 6, 7, and 8 present the results of these analyses for the three measures. Cell means for the analysis of variance tables can be found in Appendix J. TABLE 6.--Ana1ysis of Variance for Affective Role-Taking as a Function of Sex and Age. H H H Source SS df MS F Sex (S) 6.45 l 6.45 .65 Age (A) 519.98 2 259.99 26.26**** S X A .50 2 .25 .03 Respondents: S X A 603.90 61 9.90 ****p < .0001 TABLE 7.--Analysis of Variance for Cognitive Role-Taking as a Function of Sex and Age. Source SS df MS F Sex (S) 536.02 1 536.02 5.11* Age (A) 9024.80 2 4512.40 43.04**** S X A 91.48 2 45.74 .44 Respondents: S X A 6395.24 61 104.84 *p < .05 ****p < .0001 As Tables 6 and 8 indicate, the main effect of sex is non- significant for the affective role-taking and Cognitive Inference measures. However, sex as a main effect is significant for the cognitive role-taking measure (p < .05, Table 7), with females showing 52 greater cognitive role-taking ability than males. On all three measures, females scored higher than males, but the difference is significant only for the cognitive role-taking measure. Means for the males and females for the three measures are presented in Table 9. TABLE 8.--Analysis of Variance for the Cognitive Inference Task as a Function of Sex and Age. Source SS df MS F Sex (5) 10.02 1 10.02 3.66 Age (A) 115.12 2 57.56 21.02**** S X A 16.46 2 8.23 3.01 Respondents: S X A 167.14 61 2.74 ****p < .0001 TABLE 9.--Means for Males and Females on the Three Empathic Awareness Measures. Males Females Affective Role-Taking 16.69 17.32 Cognitive Role-Taking* 42.90 48.61 Cognitive Inference Task 3.48 {4.26 *significant difference The second part of the hypothesis was tested by comparing the correlations between the measures for males and females. The correla- tions are presented in Table 10. 53 TABLE 10.--Correlations Between theThree Empathic Awareness Measures for Males (lower half of matrix; df=27) and for Females (upper half of matrix; df=36). Affective Cognitive Cognitive RT RT Inference Affective RT .790*** .420** Cognitive RT .703*** .510** Cognitive Inference .602*** .781*** **p < .01 ***p < .001 As the table indicates, all of the correlations are significant (p ranges from < .01 to < .001). To test the hypothesis, Fisher Z transformations were performed on the correlation values and t tests were used to detect any significant differences. The results indicate no significant difference between males and females for any of the correlations between measures. However, since the correlations of the Cognitive Inference Task with the other two measures were lower for females, other correlations for the Cognitive Inference Task were inspected and a significant difference between males and females was detected for the correlation of the Cognitive Inference Task with age (t = 2.77; df = 65; p < .01). The respective correlations were .839 for males and .464 for females. In general, intercorrelations for males and females tended to be the same on the affective and cognitive role-taking measures, while differences were noted for the Cognitive Inference measure. 54 Hypothesis 3 is thus partially supported by the data. No sex differences were found for affective role-taking (affect identification), for the Cognitive Inference Task, and for intercorrelations between the empathic awareness measures. However, for the cognitive role-taking measure and for the correlation between the Cognitive Inference Task and age, sex differences became apparent. Females demonstrated greater cognitive role-taking ability than males (on the cognitive role-taking measure). For the Cognitive Inference Task, males showed a stronger correlation with age than did females. Since it is possible that the sex difference for the cognitive role-taking measure was reflecting greater vocabulary skill in the females, mean PPVT scores for the males and females were computed and were tested for a significant difference by using a t test. No dif- ference was found in the mean vocabulary scores for the two groups, suggesting that this sex difference is not reflecting the type of language ability that is measured by the PPVT. Positive vs. Negative Affect Hypothesis 4 concerned the type of affect portrayed in the affective role-taking measure. It was expected that positive affect would be identified significantly better than negative affect. This hypothesis was tested by an analysis of variance for repeated measures and unequal cell frequencies with a 2x3 design over subjects and a 2x2 design over measures. The results are presented in Table 11. Means for each of the cells are contained in Appendix J. 55 TABLE ll.--Analysis of Variance for Affective Role-Taking as a Function of Sex, Age, Appropriateness of Affect, and Type of Affect. -—_.—~.- df Source SS MS F Between Subjects Sex (S) 1.61 1 1.61 .65 Age (A) 130.00 2 65.00 26.26**** S X A .14 2 .O7 .03 Respondents (R): S X A 151.28 61 2.48 Within Subjects Appropriateness of Affect (AP) .13 1 .13 .22 Type of Affect (T) 326.93 1 326.93 267.04**** AP X T 2.15 1 2.15 3.52 S X AP .83 1 .83 1.38 S X T .57 1 .57 .47 S X AP X T .09 1 .09 .14 A X AP 1.76 2 .88 1.45 A X T 6.24 2 3.12 2.55 A X AP X T 1.06 2 .53 .86 S X A X AP 1.52 2 .76 1.27 S X A X T .58 2 .29 .24 S X A X AP X T .44 2 .22 .35 R X AP: S X A 36.60 61 .60 R X T: S X A 74.42 61 1.22 R X AP X T: S X A 37.21 61 .61 ****p < . OOOl 56 As the table shows, the main effectfbr'type of affect is significant (p < .OOlL thus supporting Hypothesis 4. Positive affect (happy) was identified significantly better than negative affect (sad, angry, and afraid). The mean of correct identifications for positive affect (on a 0 to 6 scale) was 5.35 and the mean for negative affect (including sad, angry, and afraid) was 3.15. No interactive effects between type of affect and the other variables resulted. To test for age differences in subjects' ability to differen- tiate affects along the positive vs. negative dimension, the affective role-taking measure was scored according to the second scoring system explained in the method section. That is, subjects were given credit for choosing happy when the correct response was happy and choosing any of the negative feelings when the correct response was sad, angry, or afraid. A 2x3 analysis of variance demonstrates a significant main effect for age (F = 26.25; df - 2,61; p < .0001). Scheffe's test for post hoc comparisons reveals that the main effect resulted from the difference between the 3 year olds and each of the two older age groups. Three year olds made significantly more errors than the 4 and 5 year olds in categorizing affects along a global positive vs. negative dimension. Appropriate vs. Inappropriate (to the SituationT'Affects Hypothesis 5 concerned the appropriateness dimension of the affective role-taking measure. No difference in the identification of appropriate and inapprOpriate (to the situation) affects was expected. 57 This hypothesis was also tested by means of the analysis of variance presented in Table 11. The results for the appropriateness dimension can be found in Table 11 and the cell means in Appendix J. As the table indicates, the main effect for appropriateness of affect is non-significant, thus supporting Hypothesis 5. Further sup- port is given by the absence of interactive effects for appr0priateness of affect and the other variables. The results indicate that the appropriate and inappropriate (to the situation) dimension did not produce a significant effect on the correctness of the response; appropriate and inappropriate affects were identified equally well. The mean of correct identifications for appr0priate affects (on a 0 to 6 scale) was 4.22, while the mean for inappr0priate affects was 4.28. An assessment of the "egocentricity" of the subjects' errors on the affective role—taking measure was made by checking the content of the errors against the response that the subject had given about how he would feel in the particular situation. An ”egocentric" error was recorded when the error matched the feeling that the subject had given for himself, and a "non—egocentric“ error resulted when the error did not match the subject's response. These numbers were computed in order to detect the degree to which the subject imputed his own reaction to the character. The percentage of ”egocentric” errors (based on the total number of errors) was .42 for the 3 year olds, .30 for the 4 year olds, and .30 for the 5 year olds. Paired difference tests reveal a signifi- cant difference between the two types of errors for 4 year olds (t = 4.28; df = 20; p < .005) and 5 year olds (t = 3.71; df = 23; 58 p < .005), demonstrating that 4 and 5 year olds made significantly more "non-egocentric" errors than "egocentric" errors, while 3 year olds tended to make both types of errors. Relationships Among the Empathic Awareness Measures Hypotheses 6, 7, and 8 concerned the relationships among the three empathic awareness measures at the different age levels. The ranges, means, and standard deviations of the scores on the three measures for the three age groups have been presented in Table 3. Hypothesis 6 concerned the relationship between affective and cognitive role-taking. It was expected that affective and cognitive role-taking would be positively correlated for the 4 and 5 year olds, but not for the 3 year olds. Pearson product moment correlations between the affective and cognitive role-taking measures for the separate age groups were used to test this hypothesis. The correlations are presented in Table 12. TABLE 12.--Correlations Between Affective and Cognitive Role-Taking for the Three Age Groups. 3 year olds 4 year olds 5 year olds df=20 df=l9 df=22 .619** .442* .205 *p < .05 **p < .01 59 As the table indicates, the correlations for the 3 and 4 year olds are significant (p < .01 and < .05 respectively), while the correlation for the 5 year olds is non-significant. The hypothesis is only partially supported by the significant correlation for the 4 year olds, but actually the correlations show a trend in the direction opposite to that predicted by the hypothesis. The hypothesis suggested that correlations would increase with age when in fact they decrease. Hypothesis 7 concerned the relationship between the affective role-taking measure and the Cognitive Inference Task. It was expected that affective role-taking would be positively related to the Cognitive Inference Task for the 5 year olds, but not for the 3 and 4 year olds. Pearson product moment correlations between the affective role- taking measure and the Cognitive Inference Task were used to test this hypothesis. The correlations for each age group are contained in Table 13. TABLE 13.--Correlations Between Affective Role-Taking and the Cognitive Inference Task for the Three Age Groups. 3 year olds 4 year olds 5 year olds df=20 df=l9 df=22 .256 .203 .157 As the table indicates, none of the correlations is significant. The hypothesis is partially supported by the non-significant correla- tion for the 3 year olds, but again the correlations show a trend in the direction Opposite to that predicted by the hypothesis. The 6O hypothesis suggested that correlations would increase with age when they actually decrease. } Hypothesis 8 concerned the relationship between the cognitive role-taking measure and the Cognitive Inference Task. It was expected that cognitive role-taking would be positively related to the Cognitive Inference Task for the 5 year olds, but not for the 3 and 4 year olds. Pearson product moment correlations between the cognitive role- taking measure and the Cognitive Inference Task for each of the age groups were used to test this hypothesis. The correlations are pre- sented in Table 14. TABLE l4.--Correlations Between Cognitive Role-Taking and the Cognitive Inference Task for the Three Age Groups. 3 year olds 4 year olds 5 year olds df=20 df=19 df=22 .379 .481* .531** *p < .05 **p < .01 As the table indicates, the correlations for the 4 and 5 year olds reach significance (p < .05 and < .01 respectively), while the correlation for the 3 year olds is non-significant. The hypothesis is partially supported by the significant correlation for the 5 year olds. Also, the correlations show a trend in the direction predicted by the hypothesis. That is, correlations increase with increasing age. DISCUSSION Age Changes in Empathic Awareness The hypothesis that all three measures of empathic awareness would increase over the 3 to 6 year age range was strongly supported by the data. Significant positive correlations between age computed in months and all three measures were obtained. Also, age as a main effect was significant when age groups (3, 4, and 5 year olds) were considered. The support for this hypothesis is highly consistent with past research on empathy in children. Regardless of the method used to measure empathy, age changes in this ability have been regularly demonstrated. While most of the researchers have used older children (6 to 14 years old) as subjects and have compared age groups that differed from each other by l to 3 years, Borke (1971, 1973) focused on 6 month intervals in 3 to 8 year old children. She found a signifi- cant and steady increase in affect identification over this age range. Likewise the present study found that empathic awareness was related to monthly increments in age, showing a continuous growth in this ability from 3 to 6 years of age. Figures 1, 2, and 3 illustrate the increase with age by six month intervals. As the graphs indicate, affective and cognitive role-taking increase most sharply during the 3 to 4% year old period, while more complex cognitive role-taking, measured by the Cognitive Inference Task, shows a more pronounced 61 62 24 . 20 J a“ A # VV’V l6 4 12 4 8 " _.._._._ Males -———— Females A . , 3-34 3I-4 4-42 42-5 5-52 54—6 Figure 1.--Mean scores for males and females on the Affective ' Role-Taking Measure plotted at six month intervals. 60 . 52 . 44 . 36 . Males -l —— Females 28 20, T 333: 32-4 4-45 42-5 5-55 5'-6 Figure 2.--Mean scores for males and females on the Cognitive Role-Taking Measure plotted at six month intervals. 6 . 5 . l. 4 3 2 1 ” —--o “8185 -‘ , l ‘s‘,’ ._.._ Females 3’33 31-9 9‘93 93-5 5'51 5&‘6 Figure 3.--Mean scores for males and females on the Cognitive Inference Task plotted at six month 1ntervals. 63 linear increase with age, especially for males. (One subject was contained in the 3-3% year old male category, so the relatively high scores are reflecting the performance of one subject rather than the mean performance for 3-3% year old males.) In terms of the skills measured by the empathic awareness instruments, the results indicate a continuous growth in the ability to identify feelings produced in a story character, to explain the reasons for the story character's feelings, and to explain the reasons for a different story character's feelings when the reasons are more difficult to understand. While Borke had previously shown regular age increments in ability for affect identification, the present study points out that the ability to explain the reasons for the feeling shows a similar increase over the 3 to 6 year old age range. Also, the present study illustrates that, with increasing age, subjects become more accurate in interpreting affective cues provided by dialogue and situational context of stories. Borke's method included only stories of a character within a situation while the present method used dialogue (tone of voice, content of words) to supplement the situa- tional context. The increase in age for both studies demonstrates that children become progressively more able to interpret affective cues provided by the situational context or manifested in the quality of speech. One explanation for the age increase in empathic awareness can be found in the social experience and cognitive maturity that are con- comitant with increasing age. Usually a child expands his social world as he grows older so that he has more opportunity to learn how and why other people feel as they do. He has the chance to observe a variety of 64 affective reactions and to begin labeling and understanding the reactions he sees. The cognitive maturity provided by age changes in cognitive structures also increases his ability to interpret cues more effectively and, according to Piaget, frees him up to consider points of view that are progressively different from his own. Thus, the child exhibits more awareness as he grows older and it is this awareness which is probably being measured by his steadily increasing performance on tests of empathy. Empathic Awareness and Intelligence The hypothesis that empathic awareness would be positively related to intelligence was supported when the total sample was con- sidered, but received only partial support when age groups were examined separately. That is, when age was held constant, the corre- lations generally dropped and showed a more variable pattern. For the total sample, all three measures were significantly correlated with mental age level, with the correlations for the two cognitive measures slightly higher than the correlation for the affective role-taking measure. This finding is not surprising since the test used to assess mental age level is also a measure of receptive language. Mental age level, as defined in this study, is a function of the subject's language ability, an ability that is also being measured in the cognitive tasks. With age held constant, affective role-taking was not signifi- cantly related to mental age for any of the age groups. This finding suggests that the ability to identify feelings in another person may be primarily dependent on social maturity and experience rather than on 65 cognitive skills. It is interesting that, while non-significant, the correlations for the 3 and 5 year olds are slightly higher than the correlation for the 4 year olds. Two different processes may be Opera- ting at these extreme age groups, with the 3 year olds learning to discriminate positive from negative feelings, and the 5 year olds learning to differentiate more finely among negative feelings. Cogni- tive maturity may help the child discriminate cues more effectively, while social experience may provide the modus operandi for learning about affective cues in the first place. Cognitive role-taking was significantly correlated with mental age only for the 3 year olds. The correlations declined over the 4 and 5 year old groups. Since the 3 year old period is a time for rapid growth of language, it would seem that language skill would be strongly related to cognitive maturity for children in this age group. That is, some 3 year olds may be highly verbal, with others only beginning to manifest variety and fluency in language. The greater cognitive maturity of some 3 year olds would likely be evident in their being more able to use words and sentences correctly. Thus, the correlation between mental age level and cognitive role-taking for the 3 year olds can be understood in light of the relation of cognitive maturity to language skills. To further support this notion, it can be noted that the non-verbal affective role-taking measure did not show a positive correlation with mental age for the 3 year olds, while the verbal cog- nitive role-taking measure did. The decline in correlations between mental age and the cognitive role-taking measure over the 4 and 5 year old groups can be explained in 66 several related ways. First, since the speech of 4 and 5 year olds is more highly developed than for 3 year olds, verbal descriptions of the reasons for another's feelings may not be as pure a reflection of cognitive maturity. Second, 4 and 5 year olds showed similar character- istics of speech in that children at both these age levels tended to leave out information that is necessary in fully explaining the view- point of another. According to the scoring system devised for the cognitive role—taking measure, 72 was the maximum possible score. Five year olds received a mean score of 57.17, while 4 year olds had a mean score of 50.71. The comparable mean for the 3 year olds was 29.73. The comparison of means and ranges shows that the 4 and 5 year olds tended to give verbal descriptions of the reasons for another's feel- ings at about the same level (no significant difference was found between their mean scores). It is noteworthy that the mean for the 5 year olds did not approach the maximum possible score, demonstrating that there are egocentric qualities in the speech of 5 year olds and that 4 and 5 year olds tended to give comparable descriptions. Thus, while the mental age levels (or receptive language skills) of 4 and 5 year olds may be different, their expressive language does not show a noticeable change. It is possible that a more refined scoring system might detect greater differences in the speech of 4 and 5 year olds; such an investigation could provide a focus for future research. The Cognitive Inference Task was positively related to mental age level for the 4 and 5 year olds, but not for the 3 year olds. The difference between the 3 year olds and the two older age groups suggests that, as expected, the Cognitive Inference Task was a very difficult 67 task for most 3 year olds. The Cognitive Inference Task was expected to be more difficult since inference was needed to comprehend and explain the reasons for the story character's feelings. The ability for inference would seem to be somewhat dependent on intellectual capacity. If a certain amount of social and cognitive maturity are required to perform successfully on the task, many 3 year olds may not be able to meet these criteria. The mean and standard deviation of scores on the Cognitive Inference Task were significantly lower for the 3 year olds than for the two older age groups, demonstrating con- sistently poorer performance among 3 year olds. Among the older ' children, who have more social and cognitive maturity, mental age level could then begin to show a positive relationship with performance. The correlations for the 4 and 5 year olds are both strong, indicating that this task was indeed reflecting cognitive maturity for the two older age groups. Overall, the results seem to indicate that subjects who are functioning on similar mental age levels will also show similar per- formance on the three empathic awareness measures, but that the factor of social experience afforded by age may confound the correlations for separate age levels. It is quite likely that there are differential interactions between social and cognitive maturity for certain age levels. The literature presents contradictory findings about the relationship of empathy to intelligence. These inconsistencies have probably resulted from differences in methodology, intelligence measures, and sample size and age range. The possibility of an 68 interactive effect of experience and cognitive skills must also be kept in mind, so that one would expect differential effects of intelligence at different ages. Sex Differences in Empathic Awareness The hypothesis of no sex differences in performance was sup- ported for the affective role-taking measure and for the Cognitive Inference Task, but not for the cognitive role-taking measure. On all three measures, females performed better than males, but the dif- ference was significant only for the cognitive role-taking measure. It is noteworthy that the sex difference appeared on a measure dependent on language ability. As indicated in the Results section, the mean PPVT scores for males and females were compared since this sex difference could just be measuring a difference in verbal capacity. The PPVT scores were not significantly different (in fact, males were slightly higher), so it was concluded that differences in receptive language ability could not account for the discrepancy. However, it is still possible that this sex difference is reflecting varying abilities in expressive language, with females being more verbally fluent than males. This possibility is further upheld by the notice- able trend for females to do better than males on the Cognitive Inference Task (p < .06). Perhaps significant differences did not appear for the Cognitive Inference Task because of its high correla- tion with mental age level (or receptive language). Since the present study did not include a measure for expressive language, such an investigation could be part of future research. 69 As predicted, there was no sex difference for intercorrelations among the measures. Males and females showed similar correlations between tasks, although for females, correlations between the Cognitive Inference Task and the other measures fell from .18 to .26 below those for males. A significant difference between males and females was noted for the correlation between age and performance on the Cognitive Inference Task; males showed a higher correlation than did females. A consideration of the range of scores for the task, as well as an inspection of the means and correlations for the two sexes at the separate age levels, reveals possible reasons for this difference with regard to the Cognitive Inference Task. With the limited range of possible scores for the task (0-8), ceiling effects were observed for the 4 and 5 year old females so that the correlation with age was reduced. The range of scores for the 4 year old females was 1-7, while the range for the 5 year old females was 3-8. The difference in mean scores for the two female groups was .51. In contrast, the difference in mean scores for the 4 and 5 year old males was 2.59, although the range of scores was similar, i.e., 0-7 for the 4 year old males and 4-8 for the 5 year old males. T tests between the 4 and 5 year olds in each of the sex groups revealed a significant difference for the 4 and 5 year old males (t = 2.98; df = 16; p < .005), but not for the 4 and 5 year old females. Thus, the task seemed to discriminate between the 4 and 5 year old males, but not between the 4 and 5 year old females. The lower intercorrelations between the Cognitive Inference Task and the other two empathic awareness measures could perhaps be influenced by the ceiling effects for the 4 and 5 year old females. 70 If the other two measures increased more proportionately from the 4 to 5 year old level, the greater ceiling effects for the Cognitive Inference Task could reduce the correlations. The significance of the correlations indicates some degree of relationship between the measures. Also, the correlations are generally of the same magnitude. However, the relatively small sample size of the present study may not be pick- ing up differences where they might exist. The slight difference in correlations between males and females highlights an area for further research on the empathic awareness measures using separate age and sex groups. The contradictory findings in the literature on sex differences are confusing since no consistent patterns of interaction between age, sex, and methodology emerge. Perhaps the variables interacting with sex to produce differences between males and females have not been discerned. The present study demonstrates that when verbal capacity is a variable in the measurement of empathy, sex differences may appear. Borke (1971) has pointed out that sex differences in affect identifica- tion may have been eliminated over the years by the emergence of social sensitivity as a desirable trait for boys as well as girls. Parents may be emphasizing affect awareness in their children regardless of the sex of their child. Relationships Among the Empathic Awareness Measures The hypothesis that affective and cognitive role-taking would be correlated for the 4 and 5 year olds, but not for the 3 year olds, was only partially supported. The results indicated significant correlations 71 for both the 3 and 4 years old. Actually, the direction of the cor- relations is opposite to that predicted. The correlations decreased over the age groups when they were expected to increase. When postulating the increase over age groups, it was argued that younger children may be able to understand another's viewpoint, but not be able to explain it. Further, it was posited that as language skills increased, the correlation between understanding and explaining would become stronger. The decrease, rather than increase, in correlations over age groups seems to reflect the characteristics of the instruments as well as the variability in skills for the separate age groups. The high positive correlation for the 3 year olds indicates a strong relationship between the ability to identify feelings in another and to explain the reasons for the feelings. This finding suggests that the two skills may develop concomitantly for this age group (since they are both tapping the child's understanding of affective phenomena), but it is also possible that the majority of 3 year olds possessed sufficient language skill to explain the reasons and that the stage of being able to identify feelings but not explain them comes earlier than 3. Casual observations during administration of the tasks revealed that some 3 year olds arkithey "could not tell” or "did not know" the reasons for the feeling when they had been correct in the affect identification of the character. An inspection of the ranges and standard deviations for the affective and cognitive role-taking measures reveals greater variability in these skills for the 3 year olds than for the two older age groups. 72 The possibility remains, however, that understanding and explaining the reasons for feelings follow a dialectic process where they interact and develop simultaneously. The question of the possible priority of understanding feelings as opposed to explaining them thus does not seem to be answered by this study, but is left to future investigation with even younger children. The significant correlation between the affective and cognitive role-taking measures for the 4 year olds can be explained by the argu- ment presented earlier. That is, when language skills are sufficiently developed, children will be able to verbalize their understanding of another's feelings as well as to identify the feeling. The decrease in correlation for the 4 year olds when compared to the 3 year olds probably comes from the greater variability in affective and cognitive role-taking for the 3 year olds. The lower non-significant correlation between the two measures for the 5 year olds, when compared to the 4 year olds, seems to reflect different patterns of variability in performance on the two measures for the two age groups. That is, the 4 year olds showed greater variability on the cognitive role-taking measure (S.D. = 7.18) than did the 5 year olds (5.0. = 4.48). Yet the two groups showed relatively the same variability for the affective role-taking measure (5.0. = 2.55 for the 4 year olds and 5.0. = 2.75 for the 5 year olds). Thus, 4 year olds differed more on the cognitive role-taking measure than did the 5 year olds, while generally they performed with equal variability on the affective role-taking measure. While the means of the two age groups for each measure are not statistically different, inspection of 73 the values and test results shows that the performance of the 4 and 5 year olds was more similar for the affective role-taking measure than for the cognitive role-taking measure. In short, the 4 year olds were still developing their language skills but had already reached the 5 year old level in affect identification. The comparison between the correlations for the 4 and 5 year olds thus suggests that all 5 year olds tend to describe the reasons for feelings in a similar fashion, while they differ in their ability to identify feelings. This pattern may be reflecting con- sistency in level of speech for 5 year olds, but inconsistency in identifying negative feelings in our culture. The hypothesis that affective role-taking and performance on the Cognitive Inference Task would be correlated for 5 year olds, but not for 3 and 4 year olds, was only partially supported by the non- significant correlations for the 3 and 4 year olds. In fact, the correlations for all age groups were non-significant. Again, the results show a trend in the direction opposite to that predicted by the hypothesis. Correlations decreased with age when they were expected to increase. An increase in correlations with increasing age was predicted since the results for the Cognitive Inference Task were expected to be less reliable for the younger subjects and to preclude a consistent correlational pattern. Also, the affective role-taking measure was expected to differentiate more clearly between the 4 and 5 year olds. With these expectations in mind, it was thought that a consistent correlational pattern would emerge for the 5 year olds since they would 74 be more capable of performing successfully on the Cognitive Inference Task. In actuality, the Cognitive Inference Task differentiated among age levels more sharply than was expected. The task was within the capabilities of some 3 year olds and showed an increase with increasing age. The affective role-taking measure essentially reached a ceiling at 4 years of age since 4 and 5 year olds identified feelings with about equal accuracy. Accordingly, the juxtaposition of the diffi- culty of the tasks was changed and the direction of the correlations was reversed. The non-significance of the correlations for all three age groups suggests that the two measures are not tapping the same abilities. The capacity fo identify feelings in another person may not be related to the more difficult task of inferring the reasons for another's feelings, i.e., explaining the viewpoint of another when the explana- tion relies heavily on cognitive skill. It was shown earlier that affective role-taking was not significantly related to mental age level, while the Cognitive Inference Task was strongly correlated with mental age for the 4 and 5 year olds. The discrepancy and lack of significant correlation between the two tasks illustrates the difference in the tasks used by Borke (1971) and Chandler and Greenspan (1972). A child may be able to identify feelings in another person without being able to explain the viewpoint of another when this latter task is made difficult. Even for 3 year olds, identifying feelings and explaining the reasons (viewpoint) of another were significantly related when this latter task was made simple. 75 The hypothesis that the cognitive role-taking measure and the Cognitive Inference Task would be correlated for the 5 year olds, but not for the 3 and 4 year olds, was partially supported by the signifi- cant correlation for the 5 year olds. The correlation for the 4 year olds was also significant, while the correlation for the 3 year olds was non-significant. The differences between correlations are in the direction predicted by the hypothesis. That is, correlations increased with age. The direction and significance of the correlations suggest that as a child's language skills increase, his ability for complex role- taking will more closely approximate his skill for simple role-taking. For the 3 year olds, the two tasks were not significantly correlated, suggesting that 3 year olds were capable of cognitive role-taking on a simple level, but not on a complex level. The increase in correla- tions with age demonstrates the role that language and cognitive skill play in explaining the viewpoint of another. Dimensions of the Affective Role-Taking Measure The hypotheses concerning the type and appropriateness of affect dimensions of the affective role-taking measure were both supported by the data. Positive affect was identified significantly better than negative affect and there was no difference h1identification of appro- priate and inappropriate (to the situation) affects. The finding that positive affect was better identified than negative affect is consistent with the literature (Alexander, e£_gl,, 1971; Borke, 1971, 1973; Brandt, 1972; Dupont, 1959; Rothenberg, 1970). 76 Young children regularly show better ability to predict happy feelings, probably because of the familiarity and preference for happy feelings in their own lives. The greater difficulty in discriminating among negative feelings has been attributed to individual response styles as well as to a defensive attitude in admitting some of the negative feelings (Borke, l97l; Brandt, 1971; Dupont, 1959; Dymond, e§_gl,, 1952). Some researchers have suggested that feelings are first cate- gorized along a global pleasant vs. unpleasant dimension and that differentiation among particular pleasant and unpleasant affects comes with increasing age (Borke, l97l; Feshbach and Roe, l968; Flapan, 1968; Gates, 1923). In the present study, the two older age groups were more successful in correctly categorizing feelings as either positive or negative, while the 3 year olds tended to confuse these dimensions and to respond with a negative affect when the correct response was happy, and conversely. The dimension of appropriateness to the situation was included in the affective role-taking measure as a test of Chandler and Green- span's criticism of Borke's study. That is, the appropriateness dimen- sion tested whether children could identify feelings that were different from their own. In this study, children as young as 3 years of age could indeed make this perspective shift and could accurately identify two different affective reactions within the same situation when one of those reactions was considered typical (appr0priate) and most likely similar to their own. Thus, the awareness of others' feelings 77 that young children exhibit on tests of empathy can be said to reflect a genuine ability to make perspective shifts in the affective sphere. It was proposed that familiarity with type of affect and not appropriateness to the situation would determine correctness of affect identification. Together, the results for these two dimensions of the task support this notion and suggest that a child's inability to identify the affective reaction of another likely results from unfamiliarity with the type of affective reaction rather than from a failure to realize that others react differently from himself. The predominance of "non-egocentric" errors, as compared to “egocentric" errors for the 4 and 5 year olds, demonstrates that subjects in these two older age groups did not automatically impute their own reaction to the character, but attempted to identify the reaction from the cues provided in the story. Most 3 year olds likewise tended to make "non-egocentric" errors, but the difference between types of errors was not significant for this age group. Since appropriate and inappropriate affects were identified equally well, it can be con- jectured that the "egocentric" responses in the 3 year olds may also be related to ignorance as to type of affect. That is, the 3 year olds may have responded with the same affect because of inability to identify the correct affective reaction. "Egocentric" responses may result when the 3 year old child lacks knowledge about affective reactions in general--either about his own or about others'. Evaluation of the Instruments In general, the instruments used to assess affective and cognitive role-taking provided a viable and valid method for examining these skills in 3 to 6 year old children. For the total sample, the 78 correlations with age and mental age were strong, suggesting that the instruments were measuring phenomena that follow a developmental growth pattern in young children. Individually, the instruments had their own limitations which could be corrected for use in future research. The Affective-Cognitive Role-Taking Task proved to be too lengthy for the subjects. While the majority of potential subjects completed the task in the several ses- sions planned for, some subjects complained about the length and repetitious nature of the stories. For many, the task was not stimu- lating enough after they had listened to half the stories. To correct for this factor, the instrument could be shortened provided a sufficient number of storieS'nskept to insure reliability. While the varying difficulty in the identification of negative feelings is consistent with the literature, it is possible that this discrepancy results partially from a confounding of several feelings in one story. To control for this possibility, the stories can be tested out with older subjects or new stories can be composed from children's own accounts of situations that cause specific feelings. Borke (1973) used this latter method in developing her instrument for 3 to 6 year old children. The ceiling effect observed for the 4 and 5 year olds on the affective role-taking measure could perhaps be eliminated by refining the stories for the negative feelings. However, if this ceiling is a natural stage in the development of empathic awareness, the instrument could be restricted to preschool subjects, with its use extending down to 2% year old children. 79 The similarity in the speech of the 4 and 5 year olds for the cognitive role-taking measure suggests that the instrument could per- haps be tested out with older subjects to obtain criteria for refining the scoring system. If the speech of older children was examined, developmental changes in language skill could be assessed more accu- , rately. The present scoring system is somewhat similar to the method developed by Gollin (1958) and explained by Rothenberg (1970), but possibly a finer comparison of the two systems and an examination of the speech of older children on the cognitive role-taking measure could discern the consistencies and discontinuities in the speech of the 4 and 5 year olds more precisely. While all of the empathic awareness measures differentiated between the 3 year olds and the two older age groups, the Cognitive Inference Task was the only measure which discriminated among the 4 and 5 year olds (when males and females were pooled into one age group). The task was significantly related to mental age for the two older age groups, suggesting that a cognitive skill was being measured. Since ceiling effects for 4 and 5 year old females seem to have been pro- duced partially by the limited range of scores, perhaps the test could be lengthened and a more refined scoring system devised. The instrument may be appropriate for use with older subjects if these modifications are made. However, its appropriateness for 3 year olds must be reconsidered since the task was very difficult for most subjects in this age group. Since all of the instruments have been used for the first time in the present research, the question of construct validity can be 80 tackled more directly by correlating performance on these instruments with other measures of empathy. Reliability studies through a test- retest procedure would also yield information about the validity of the instruments. The sample size in the present study was relatively small so the results should be confirmed with a larger number of subjects. Implications of the Findings for Child Development Research On all three measures of empathic awareness, the 3 year olds showed a different pattern of performance from the two older age groups. The means for the 3 year old group were significantly lower than the means for the 4 and 5 year olds. 0n the more simple tasks, viz. the affective and cognitive role-taking measures, the variability for the 3 year olds was greater than that for the two older age groups, while on the more difficult measure, viz. the Cognitive Inference Task, they were more consistent in their performance. Several 3 year olds tended to make "egocentric" errors on the affective role-taking measure, while most 3 year olds and the 4 and 5 year olds made predominantly "non-egocentric" errors. These results suggest that rapid develop- mental changes are occurring throughout the 3 year old period so that, for the skills investigated in this study, a wide range of abilities was apparent. The rapid changes in skills for the 3 year olds illustrates the value in studying this developmental period more closely. Also, the need for extending this investigation to 2 and 2% year old children is highlighted' because some 3 year olds already exhibited well developed 81 affective and cognitive role-taking skills. An investigation of younger children could reveal norms and steps in the evolution of these skills. The similarity in explanations offered by the 4 and 5 year olds and by the 5 year olds themselves suggests that even the speech of 5 year olds retains egocentric qualities. Five year olds leave out information that is essential for fully explaining the viewpoint of another. While this observation is highly consistent with the litera- ture on egocentrism, it must be noted that while 5 year olds show ego- centric speech, this quality decreased with age. The lack of correlation between mental age level and affective role-taking for all three age groups highlights the predominant role that age and social experience play in the awareness of another's feel- ings. Ability to identify the feelings of others seems to depend primarily on the opportunity to learn the meaning of affective cues and is not related to cognitive skill in the child. In contrast, tasks based on language ability tend to be related to cognitive level, especially when some complexity in skills is involved. Since affective role-taking likely reflects social maturity rather than cognitive maturity, the variable of interaction experiences must be considered and controlled for in future studies of empathy. Most subjects in the present study had been enrolled in a nursery school or kindergarten for at least 8 months. Thus, most had con- siderable opportunities for social experience. The results for affective and cognitive role—taking may be quite different if a non- nursery school population is used. 82 The lack of correlation between the affective role-taking measure and the Cognitive Inference Task points up the need to define the concept of egocentrism more precisely. If egocentrism means an inability to take the viewpoint of another in the affective, cognitive, and perceptual spheres, then the child of preschool age definitely shows non-egocentric functioning in the affective sphere. However, if ego- centrism refers primarily to an inability to describe verbally the view- point of another, then the young child manifests the egocentrism that has been attributed to him. A more precise definition of egocentrism would also clear up confusion about the controversy of egocentrism vs. non-egocentrism that results from the use of different conceptualiza- tions and their related methodologies. (Borke and Chandler and Green- span are a case in point). Perhaps a more rewarding approach to the question of ego— centrism in the young child would be to disregard the label and just investigate the capabilities in the affective, perceptual, and cog- nitive spheres that are typical for the child of this age. The inter- relationship of the different types of role-taking abilities could provide the focus for another study. Flavell, et_el, (1968) and Selman (1971) have both suggested that perceptual role-taking ability may develop earlier than conceptual role-taking ability. It is quite likely that affective role-taking even precedes perceptual role-taking, since affective role-taking seems to be less dependent on cognitive structures. The preschool child's success in identifying the feelings of another can perhaps be understood in light of the etiology of this skill. That is, the child learns to identify feelings because he 83 sees and hears the affective reactions of others. Early in life he observes that others react in a characteristic and unique fashion to events that occur. He learns to interpret the affective cues and to match certain events with specific feelings. In contrast, the child learns about the cognitive and perceptual perspectives of others by listening to differences in viewpoint and by noticing that objects look different when he himself moves about in space. The realization of differences in cognitive and perceptual perspectives may be a longer and more difficult process since the difference must often be inferred and is more dependent on other cognitive skills than is necessary for noticing differences in affective perspectives. It is noteworthy that cognitive role-taking (related to language skill) has been said to follow perceptual role-taking; the difference in dependence on cognitive skills may determine the time of emergence of the particular types of role-taking. In brief, the differences in affective perspectives are more blatant, while differences in perceptual and cognitive perspectives exist on a covert level. Ferguson (personal communication) has suggested that the increasing ability of the preschool child to identify the affective reactions of others may come from a generalization of cues from care- taker to others. The child first learns to interpret the affective cues modeled by his caretaker and then, as his social world expands, he notices similar and dissimilar responses in others. The initial awareness of caretaker's feelings may be related to certain qualities in the caretaker-child relationship. The concept of generalization could account for the frequently heard anecdotes about a young child's extreme sensitivity to the feelings of others close to him. Also, this 84 concept provides an explanation for the increasing accuracy in identification of feelings since the young child probably needs to learn the variations of affective responses in others. Several researchers have highlighted the important role that parents play as models in the development of their child's empathic behavior (Ferguson, 1970; Stover, e£_el., 1971). I The present study fits very well into the framework of decreas- ing egocentrism suggested by Selman (1971) and Shantz and Watson (1970, 1971). Children between the ages of 3 and 6 were increasingly able to identify and explain the affective and cognitive perspectives of another. Role-taking ability in these areas seemed to follow a con— tinuous growth pattern rather than a discontinuous stage model. To relate the present design more closely to the steps of decreasing egocentrism proposed by Selman and Shantz and Watson, a future study could investigate when the child actually becomes aware of a difference in affective perspectives. Such a study could ask questions like: "How would you feel (in this situation)? 00 you know how this boy will feel?" The results may be interesting for it would be the child's verbalization of his awareness that would be investigated. The results could then be used for comparing the steps of decreasing egocentrism for affective role-taking with the steps for the other types of role-taking. Since social interaction for children of preschool age has generally increased through the proliferation of day care centers and nursery schools, the recent trend is to re-assess the interpersonal capacities of the young child. In observing mutually responsive speech 85 in 3-5 year old children, Garvey and Hogan (1973) commented that the young child seems to be egocentric because of a lack of tools for social interaction. This comment suggests that today's preschooler may show less egocentric behavior than his counterpart of a decade ago since the opportunities for social experience have increased. Ferguson (1971) has pointed out that the view of the infant as a social being who shows more reciprocal and social behavior with increasing age and with the proper socialization experiences is one that is quite well accepted in the literature. Additional Ideas for Future Research The methodological differences in past studies on empathy in children illustrate the variety of cues that convey affective reactions. Facial expression, body posture, tone of voice, content of words, and sequence of action are among the stimuli that have been used to convey feelings and to measure the phenomenon of empathy. While it is likely that some types of cues are more easily interpreted than others, no systematic study has looked at the degree to which various types of cues are effective in conveying feelings. Such a study could investi- gate the ease and difficulty of various types of cues, along with setting norms for the detection of such cues by different age and sex groups. Chandler, e§_el, (1973) have pointed out that the medium of presentation of stimuli is important in determining the results obtained in a study. When these investigators presented subjects initially with video- taped sequences of moral dilemma situations, subjects tended to focus on intentions more than when first presented with verbal descriptions of 86 the same situation. Perhaps the subjects could more easily take the role of the other when they became involved in the action of the stories. Likewise, children may be more accurate in perceiving affective cues when these cues are presented in action sequences. A movie sequence is probably the most ideal medium for testing empathic awareness (other than natural observation of the child) since such a stimulus would approximate a social situation more closely than static measures of empathy. However, a movie is costly and time con- suming, but may be worth the investment for future studies of empathy. A movie could provide a range of affective cues and variations could be created by including or omitting dialogue. These variations could help determine the importance of each type of cue in conveying feelings. Following Ferguson's concept of the generalization of cues from caretaker to others, an investigation of the child's awareness of his caretaker's feelings could be carried out by using pictures of different affective states of the caretaker or by questioning the child about his awareness of his caretaker's feelings. The caretaker-child relation- ship could simultaneously be investigated in order to detect possible connections between aspects of the relationship and awareness of feelings. Knowledge of caretaker's feelings could then be compared to awareness of the affective reactions of others. It would be interesting to know the age at which children predict that others will react as their caretaker would and the age at which they begin to realize dif- ferences in reactions. Since the literature contains evidence that role-taking skills can be increased by training (Chandler, 1973), subjects who are low in 87 empathic awareness could perhaps be helped to improve their skills by discussion and practice in identifying the feelings of others. Dif- ferences between individuals could be pointed out and emphasized. SUMMARY This study was concerned with the phenomenon of empathic aware- ness as it is exhibited by young children. Affective and cognitive role- taking were pr0posed as two components of empathic awareness and were defined respectively as the ability to identify the affective reaction of another and the ability to explain the reasons for the affective reaction. The Affective-Cognitive Role-Taking Task and the Cognitive Inference Task were designed by the author to measure affective role- taking and two levels of cognitive role-taking (simple and complex). Hypotheses concerned the relationship of empathic awareness to age and intelligence, the possibility of sex differences, intercorrelations among the measures, and the effects of type and appropriateness (to the situation) of affects on affective role-taking. Sixty-seven children between the ages of 3 and 6 who were American born and who had English as the only language spoken in their home were administered the role-taking tasks and the Peabody Picture Vocabulary Test. Analyses of variance (least squares solution) and Pearson product moment correlations were used to test the hypotheses. Statistical analyses demonstrated the expected positive correlation between age and all three measures of empathic awareness. When the total sample was considered, intelligence (defined as mental age level) was positively related to all three measures, thus supporting the hypothesis. However, when age groups were examined separately, the 88 89 correlations decreased and presented a variable pattern: affective role-taking was not significantly correlated with intelligence for any of the age groups, simple cognitive role-taking was positively related to intelligence for the 3 year olds, and complex cognitive role-taking correlated positively with intelligence for the 4 and 5 year olds. While no sex differences were expected for any of the measures nor for intercorrelations among them, females performed better than males on the simple cognitive role-taking measure and males showed a stronger correlation between age and the complex cogni- tive role-taking measure. Intercorrelations among the measures revealed significant cor- relations between affective and simple cognitive role-taking for the 3 and 4 year olds, a significant relationship between simple and complex cognitive role-taking for the 4 and 5 year olds, and non-significant correlations between affective and complex cognitive role-taking for all age groups. These results provided partial support for the pro- posed hypotheses. Tests on the type and appropriateness of affect dimensions of the affective role—taking measure demonstrated easier identification of positive affect (when compared to negative affect) and no difference in the identification of affects that were common to the situation and those that were uncommon; these results supported the relevant hypotheses. The findings were interpreted as demonstrating non-egocentric qualities in the young child's ability for empathic awareness. That is, the young child was found to exhibit an awareness of the different affective reactions of others; difficulty in identifying and explaining 9O feelings seemed to come from unfamiliarity with the type of affective reaction rather than from the egocentric stance that others react as the self does. The role that language ability has played in measuring egocentrism in past studies in interpersonal role-taking was discussed and the need for a more precise definition of egocentrism was emphasized. Overall, the results were thought to uphold the notion of decreasing egocentrism in the 3 to 6 year old period since empathic awareness increased steadily over this age range. The instruments used in the study were evaluated and ideas for future research were pr0posed. APPENDICES 91 APPENDIX A LETTERS TO PARENTS 92 MICHIGAN STATE UNIVERSITY PSYCHOLOGY DEPARTMENT ' OLDS HALL East Lansing, Michigan A882A April 1, 197A Dear Parents: Empathy has been defined as an understanding of the thoughts and feelings of another. For my doctoral dissertation research, I am interested in studying the development of this type of interpersonal sensitivity with regard to peer interaction. That is, I would like to investigate the young child§ ability to understand the thoughts and feelings of another child as presented in stories. During the months of April and May, I will be carrying out my dissertation research at the Lab Preschool and Spartan Nursery. The Psychology Department and the Preschool Committee of the Institute for Family and Child Study at Michigan State University have approved the design and have consented to letting me use children in the Lab Preschool and Spartan Nursery as subjects in my study. Your child will be included in my study provided you agree to have him/her participate. Each child in the study will listen to a series of stories in which charac- ters will feel happy, sad, angry, or afraid. The child will be asked to identify the feeling and to tell why the character is feeling that way. The stories are not aimed at producing these feeling states in the child himself and so are not constructed to be upsetting or distressing to the child. The committee has approved each of these stories and has considered them appropriate to present to young children. In addition to these stories, the child will be presented with a brief pic- ture vocabulary test. The procedures should run about an hour in total time, though each child will be seen on two or three different occaéonstx: complete the tasks. These procedures will not interfere with your child's program. If you have any questions about the study, please give me a call at either 355-8270 (office) or 351-1954 (home). I will return your call to follow up your request for more information if you are not able to get in touch with me. If you consent to your child's participation, please have your child return the attached permission slip to his/her teacher by Thursday, April 4th. The research will begin as soon as the permission slips are returned. After the study,is completed and written up (during the summer), a summary of the results will be made available at the Preschool Lab and Spartan Nursery for interested parents. Sincere}%, Ms. Lisa Paityka Thank you for your cooperation. I give my permission for to participate in research on empathic awareness conducted by Lisa Partyka. Signature: Date: 94 APPENDIX B FACES USED IN THE IDENTIFICATION OF FEELINGS FOR THE AFFECTIVE-COGNITIVE ROLE-TAKING TASK 95 96 Sad Happy 97 Afrai Angry APPENDIX C BACKGROUND SCENES AND SAMPLE FIGURES FOR THE AFFECTIVE-COGNITIVE ROLE-TAKING TASK 98 ANV aaamc cw pan cowuaooF mama cw mm: muawcaocaa< mesmwmv muawcaocanmcm mczmwm lOl A_v >maz<->aamaz<->aa<= 103 uuaPLaocaam=H mcampu .\ x a I . e ’ .u u“ . $3.15. ”L: .u\ x... I e o u \ I 9:. \\ (in x J / v?» a l\ r 7 .I I a/o \\ ..o’o o 74/7 \f ///// 3J A_v aaaaaa<=->aaz< llO ANV >aa<=-aaaa<= HAPPY IN A SAD SITUATION (2) 133 SAD IN A HAPPY SITUATION (I) T34 SAD IN A HAPPY SITUATION (2) I35 APPENDIX H STORIES FOR THE COGNITIVE INFERENCE TASK T36 APPENDIX H STORIES FOR THE'COGNITIVE INFERENCE TASK Sad in a Happy Situation 1 This is a picture of a birthday party. The children are in the living room playing games. Most of them are having lots of fun; they're enjoying themselves and are feeling real happy. But Johnnie/Mary* who's sitting here in the chair is not having fun. He is thinking about his dog that did not come home last night and he's feeling real sad. He loves his dog and he's afraid that he will not come back at all. Johnnie doesn't want to play games. He just feels too sad to have fun. All of the children gather around to watch the birthday boy open his presents. The birthday boy/girl opens up boxes in which are a football helmet/doll, records for his record player, and ice skates. Then his mother and father tell him that they are going to bring out their present for him. His mother and father come into the living room carrying a puppy. The birthday boy is happy and excited to get a puppy. But when Johnnie sees the puppy, he bursts out crying. How come Johnnie starts to cry when he sees the puppy? 2 This is a picture of Tommy/Patty in his classroom in the afternoon doing some painting. All of the children are busy making different pictures. Tommy is painting a picture of a tree but he staps for a minute and remembers that his mother is going to pick him up from school sooner than she usually does. His mother will come in just a few minutes and Tommy will leave while all the other children will stay for one more hour. Tommy is almost done with his painting; he's sure to be finished in just a few minutes when his mother comes. Tommy continues to paint and then the teacher says that she has a surprise for all of the children. The teacher says that she is going to take the class to an ice cream store so they can get ice cream cones. The children will go to the store at the end of the day, just before they all go home from school. But they will have to keep painting for a long while before it is time to go to the ice cream store. All of the children are happy and smiling when they hear they will be going to the ice cream store later in the afternoon, but Tommy is sad. While the teacher was telling the children about going to the ice cream store. Tommy's mother walked into the classroom. How come Tommy is sad when he hears that the teacher will take the class to the ice cream *Gender of nouns and pronouns was changed for female subjects. Other changes for the two sexes are indicated by slash (I) marks. I37 138 store that afternoon? Happy in a Sad Situation 1 This is a picture of Billy/Susan in the park behind his classroom. It is springtime outside and all the trees and grass are green. It is recess time and while all the other children are playing games, Billy is thinking about what his mother told him that morning. At breakfast, his mother said that their family was planning to go to the lake that weekend and that they would be leaving early Friday afternoon. Billy would have to miss the afternoon at school in order to go with his family. He likes going to the lake to swim and ride in the boat, but he remembered that his class was taking a trip to the zoo on Friday afternoon. The class would go to the zoo only once during the year. Billy also liked going to the zoo and he knew he would have to miss the trip to the zoo to go with his family. At the end of recess, Billy's teacher tells the class that she has some sad news for them. The class will not be able to go to the zoo on Friday afternoon but will have to put the zoo trip off for awhile. All of the children are sad that they will not be going to the zoo on Friday afternoon because they waited a long time to go, but Billy is happy. He is smiling when all the other children are sad. How come Billy is happy when he hears that the class will not be going to the zoo on Friday afternoon? 2 This is a picture of Paul/Sally looking at a book of animals in his classroom. Paul is looking at many different kinds of animals but his favorite animal is a bunny rabbit, just like the real live bunny rabbit that is in his classroom in a cage. As Paul looks at the picture of the bunny rabbit, he remembers that the night before his mother and father told him that he could have a bunny rabbit as a pet, just as soon as they could get him one. They would wait to see if someone wanted to give away a bunny rabbit or they would buy one. Paul walks over to look at the bunny rabbit in the cage in the classroom and he thinks of how much he wants a bunny rabbit. He can play with the bunny rabbit, feed it, and even give it a name. He is glad that his parents are going to let him have a bunny rabbit. The teacher walks over to the bunny rabbit cage too and she asks the children to listen to her. She says that she has some sad news to tell the children. She says that they will not be able to keep the bunny rabbit in the classroom, but they will have to give it away to someone, maybe to someone in the classroom. All of the children are sad to hear that the bunny rabbit will no longer be in the classroom since they like the bunny rabbit so much and they enjoy playing with him. But Paul is happy and smiling when he hears what the teacher says. How come Paul is happy when he hears the teacher say that they can no longer keep the bunny rabbit in the classroom? APPENDIX I SCORING CRITERIA AND EXAMPLES FOR THE COGNITIVE INFERENCE TASK 139 APPENDIX I SCORING CRITERIA AND EXAMPLES FOR THE COGNITIVE INFERENCE TASK Scoring is.baaed on the adequacy of the subject's response in stating the reasons for the story character's affective reaction. It is a 3 level system (0 to 2) with a level 2 response denoting the highest level of role-taking. Story A Question: Answer: Level 2: Level 1: How come Johnny/Mary starts to cry when he/she sees the puppy? Because seeing the puppy reminds him/her of his/her own dog that is lost, or because he/she wants his/her own (or any) puppy cause his/her dog ran away. The subject makes the connection between the prior event of the puppy running away and the character's crying. that is, the thought of the character's own dog or a wish to have that dog back or another dog because his ran away must be mentioned in order to have a 2 response. The character was already feeling sad in the story, so the statement must include reasons why he/she cried when he/she saw the puppy. It is not enough just to state "because her dog ran away" since that is a reason for feeling sad throughout the story. Overall, some connec— tion that the puppy stimulated feelings about the character's own puppy that was lost must be made. Examples: the doggie reminds her of her own puppy she wants the puppy cause hers ran away the puppy isn't his cause his ran away The subject has knowledge (direct or implied) that the puppy ran away but makes no real connection between the crying and the fact that he ran away, or does not indicate that seeing the puppy stimulated thoughts or wishes about wanting his/her own (or any) puppy because the puppy had perviously run away. Examples: he thinks it's his puppy her puppy ran away he wants a puppy because he doesn't have one (without saying that the puppy was previously lost or because he wants one because his puppy was lost) 140 141 Level 0: The subject shows no real understanding of the story Story B Question: Answer: Level 2: Level 1: Level 0: Story C Question: Answer: or of the reasons for the character's feelings. This level includes "I don't know," no response, "just be- cause," "he's sad," and incorrect information. Examples: he's afraid the dog will bite she doesn't like dogs How come Tommy/Fatty is sad when he/she hears that the teacher will take the class to the ice cream store that afternoon? Because he/she will not be able to go to the ice cream store. His/her mother is picking him/her up from school early and he/she will not be able to go with the class. The subject says that the character will not be able to go to the ice cream store. It is not necessary to state that the added reason is because the child will go with the mother, but the subject must show an understanding that the character can't go. Examples: cause he can't go with the class her mother came and is going to take her home so she can't go The subject does not make a clear connection between the mother's coming and the character's not being able to go to the ice cream store, or does not state that the character cannot go. A description of the charac- ter's wish to go to the ice cream store (without saying he can't) and a description that the mother has just come (without saying the character cannot go to the ice cream store) belong here. Examples: his mother just came in the door cause she wants to get an ice cream cone she has to go home The subject shows no real understanding of the story or of the reasons for the character's feelings. This level includes "I don't know," no response, "just be- cause," "he's sad," and incorrect information. Examples: he doesn't want an ice cream cone she wants to go with her mother How come Billy/Susan is happy when he/she hears that the class will not be going to the zoo on Friday afternoon? Because then he/she won't have to miss the trip to the zoo, or because then he/she will be able to go to the zoo when the class goes the next time. Level 2: Level 1: Level 0: 142 The subject makes the connection that the character may be given the bunny rabbit because his parents promised him one (they would let him take one that was given away) and because the class is going to give away their bunny rabbit. Examples: he might get it she can have it now they can give it to her The subject does not make the connection directly but either implies the connection or seems to know that the character will be getting a bunny rabbit of his/ her own. Examples: he's gonna get a bunny rabbit she wants it she wants a bunny rabbit The subject shows no real understanding of the story or of the reasons for the character's feelings. This level includes "1 don't know,".no response, "just because," "he's happy," and incorrect information. Examples: he doesn't like bunny rabbits the class is going to keep the rabbit APPENDIX J CELL MEANS FOR THE ANALYSIS OF VARIANCE TABLES 143 Ce11 Means for Analysis of Variance Tab1e 6. 3 year o1ds 4 year o1ds 5 year o1ds Ma1es 12.91 18.86 19.09 FemaIes 13.18 18.71 19.31 Ce11 Means for Ana1ysis of Variance Tab1e 7. 3 year 01ds 4 year o1ds 5 year o1ds Ma1es 26.00 49.14 55.82 Fema1es 33.46 51.50 58.31 Ce11 Means for Ana1ysis of Variance Tab1e 8. 3 year o1ds 4 year o1ds 5 year o1ds Ma1es 1.46 3.14 5.73 Fema1es 2.91 4.57 5.08 Ce11 means for Ana1ysis of Variance Tab1e 11. Appropriate Inappropriate Positive Negative Positive Negative 3 year 01d 4.46 1.82 4.46 2.18 Ma1es 4 year 01d 5.29 3.86 5.57 4.14 5 year 01d 5.91 3.55 5.82 3.82 3 year 01d 5.09 2.00 4.09 2.00 Fema1es 4 year 01d 5.64 3.57 5.71 3.79 5 year 01d 6.00 3.54 6.00 3.77 144 BIBLIOGRAPHY 145 BIBLIOGRAPHY A1exander, T.; Stoy1e, 3.; Roberge, J.; and Leaverton, P. Deve1op- menta1 characteristics of emotiona] experience. Journa1 of Genetic Psycho1ogy, 1971, 119, 109-117. Amen, E. Individua1 differenceS'hiapperceptive reaction: A study of the response of pre-sch001 chi1dren to pictures. Genetic Psycho1ogy Monographs, 1941, 23, 319-385. Borke, H. Interpersona1 perception of young chi1dren: Egocentrism or empathy? Deve1opmenta1 Psycho1ogy, 1971, 5, 263-269. . Chand1er and Greenspan's “Ersatz Egocentrism": A rejoinder. Deve1opmenta1 Psycho1ogy, 1972, 7, 107-109. The deve1opment of empathy in Chinese and American chi1dren between three and six years of age: A cross-cu1ture study. Deve1opmenta1 Psycho1ogy, 1973, 9, 102-108. Brandt, B. The three-year-o1d chi1d's awareness of affect in se1f and mother. Unpub1ished masters thesis, Michigan State University, 1972. Burns, N. and Cavey, L. Age differences in empathic abi1ity among chi1dren. Canadian Journa1 of PsychoIogy, 1957, 11, 227-230. Chand1er, M. Egocentrism and antisocia1 behavior: The assessment and training of social perspective-taking ski11s. Deve1opmenta1 ngcho1ogy, 1973, 9, 326-332. - , and Greenspan, S. Ersatz egocentrism: A rep1y to H. Borke. Deve1opmenta1 Psycho1ogy, 1972, 7, 104-106. , Greenspan, 5., and Barenboim, C. Judgments of intentiona1ity in response to videotaped and verba11y presented mora1 di1emmas: The medium is the message. Chi1d Deve1opment, 1973, 44, 315-320. ~ Cheyne, N. and Jahoda, G. Emotiona1 sensitivity and inte11igence in chi1dren from orphanages and norma1 homes. Journa1 of Chi1d Psycho1ogy and Psychiatry, 1971, 12, 77-90. 146 147 Cohen, E. Empathy, awareness of interpersonaI responsibi1ity and con- sideration for others in young chi1dren. Unpub1ished doctora1 dissertation, Michigan State University, 1973. Dimitrovsky, L. The abi1ity to identify the emotiona1 meaning of voca1 expressions at successive age 1eve1s. In J. R. Davitz, et a1. (Eds.) The Communication of Emotiona1 Meaning, New York: McGraw-HiTT, 1964, 69-86. Dupont, H. The description and acceptance of fee1ing by chi1dren ages 5-11. Unpub1ished doctora1 dissertation, George Peabody Co11ege, 1959. Dymond, R. A pre1iminary investigation of the re1ation of insight and empathy. Journa1 of Consu1tingngychoIOgy, 1948, 12, 228-233. . A sca1e for the measurement of empathic abi1ity. Journa1 of Consu1ting PsychoIOQy, 1949, 13, 127-133. . Persona1ity and empathy. Journa1 of Consu1ting Psycho1ogy, 1950, 14, 343-350. , Hughes, A. and Raabe, V. Measurab1e changes in empathy with age. Journa1 of Consu1ting PsychoIogy, 1952, 16, 202-206. E1kind, D. Chi1dren and Ado1escents. New York: Oxford University Press, 1970. Feffer, M. The cognitive imp1ications of ro1e-taking behavior. Journa1 of Persona1ity, 1959, 27, 152-168. . , and Gourevitch, V. Cognitive aspects of ro1e-taking in chi1dren. Journa1 of Persona1ity, 1960, 28, 383-396. Ferguson, L. R. Persona1ity Deve1opment. Be1mont: Brooks/Co1e .,Pub1ishing Company, 1970. Origins of socia1 deve1opment in infancy. Merri11-Pa1mer .w, 19719179119'139. Feshbach, N. D. and Roe, K. Empathy in six and seven year o1ds. Chi1d Deve1opment, 1968, 39, 133-145. F1apan, D. Chi1dren's Understanding of Socia1 Interaction. New York: Teacher‘s C011ege Press, 1968. F1ave11, J. H. The Deve1opmenta1 PsychoIogy of Jean Piaget. Princeton: 0. Van Nostrand Company, Inc., 1963. ; Botkin, P. T.; Fry, C. L.; Wright, J. w.; and Jarvis, P. E. The Deve10pment of Ro1e-Taking and Communication Ski11$ in Chi1dren. New York: John NiIey and Sons, Inc., 1968. 148 Garvey, C. and Hogan, R. Social speech and social interaction: Egocentrism revisited. _Child Development, 1973, 44, 562-568. Gates, G. S. An experimental study of the growth of social perception. Journal of Educational Psychology, 1923, 14, 449—461. The role of the auditory element in the interpretation of emotion. Psychological Bulletin, 1927, 24, 175. Gilbert, 0. C. The young child's awareness of affect. Child Development, 1969, 40, 629-640. Gollin, E. Organizational characteristics of social judgment: A developmental investigation. Journal of Personality, 1958, 26, 139-154. Looft, N. R. Egocentrism and social interaction across the life span. Psychological Bulletin, 1972, 78, 73-92. Maratsos, M. P. Non-egocentric communication abilities in preschool children. Child Development, 1973, 44, 697-700. Peabody Picture Vocabulary Test and Manual. Minneapolis: American Guidance Service, Inc., 1965. Rothenberg, B. 8. Children’s socia1 sensitivity and the relationship to interpersonal competence, intrapersonal comfort and intellectual level. Developmental Psychology, 1970, 2, 335- 350. Ruckmick, C. A. A preliminary study of the emotions. Psychological Monographs, 1922, 30, 30-35. Selman, R. L. Taking another's perspective: Role-taking development in early childhood. Child Development, 1971, 42, 1721-1734. Shantz, C. U. and Watson, J. 5. Assessment of spatial egocentrism through expectancy violations. Psychonomic Science, 1970, 93-94. _ . Spatial abilities and spatial egocentrism in the young child. Child Development, 1971, 42, 171-181. Stotland, E.; Sherman, S. E.; and Shaver, K. G. Empathy_and Birth Order: Some Experimental Explorations. Lincoln: University of Nebraska Press, 1971. Stover, L.; Guerney, B.; and O'Connell, M. Measurements of acceptance, allowing self-direction, involvement and empathy in adult- child interaction. Journal of Psychology, 1971, 77, 261-269. 149 Walker, H. M. and Lev, J. Statistical Inference. New York: Holt, Rinehart and Winston, 1953. Walton, W. E. Empathic responses in children. Psychological Mono- graph , 1936, 48, 40-67. HICHIGQN STATE UNIV, LIBRQ‘RIES 11111111111111111111”1111111111111“!M11!" 31293100486608