1968 Empathy In Six And Seven Year Old.pdf

EMPATHY IN SIX- ANO SEVEN-YEAR-OLDS NORMA D. FESHBACH and KIKI ROE

University of California, Los Angeles

This study tests the hypothesis that similarity between the sex of the subfect and the sex of the stimulus faciUtates empathic behaoior in children, where empathy Ls defined as a vicarious affective response. The effects of sex differences and of variations in affect obserood were also assessed. 6and 7-year-old boys and girls reported their feelings after obseroing slidesequence stories depicting boys or girls in different affective states. Tite results strongly support the experimental hypothesis and further distingui.sh between the recognition of an affective statc in others and the empathic response to that perception. This investigation is concemed with conditions facilitating empathic behavior in cbildren. The phenomenon of empathy has received very little attention in this age group despite its probable relevance to social comprehension, moral development, and the acquisition of positive social values. Apart from methodological difficulties, the study of empathy has been complicated by the variety of meanings given to the term. lnvestigations of empathy in adults have been concemed with rather diverse processes, depending upon the particular delinition employed in the specific study. One common usage of "empathy is similar to social insight. The measure of empathy is the subject's ability to predict another person's responses on a personality test or in sorne social situation (Dymond, 1949, 1950; Milgrim, 1960; Strayer, 1960). This approach conceives of empathy as primarily a cognitive phenomenon and fails to consider the affective connotations of empathy as a vicarious cmotional response of a perceiver to the emotional experience of a perceived object. Althougb social insight and understanding may be inBuenced by empathy, the two are by no means equivalent. Empathy as a vicarious affective response may be contingent upon the comprehension of a social event, while social understanding may be independent of an affective response. N. D. Feshbach's address: Department of Education, Unfversity of California,

Los Angeles, California 90024.

CHILD DEVELOPMENT The studies of Berger ( 1962), Stotland and Dunn ( 1963), and Stotland aod Walsh ( 1963) ha ve d emonstrated the theoretical aod empirical utility of restricting the use of empathy to the experiencing of an emotion similar to that of another person as a consequence of perceiving feeling in the other person. Using self-reports and pabnar sweating as indexes of affective states in an adult population, Stotland and bis co-workers (Stotland & Dunn, 1963; Stotland & Walsh, 1963) have found that empathy is positively related to self-esteem, birth order, and degree of similarity between empathizer and the stimulus person. Tbe few investigations of empathy in children have not focused on its vicarious affective aspects. Thus, Lois Murphy's ( 1937) classic research on sympathy in young children was only incidcntally concemed with empathy as such, while Dymond's (1949) comparative study of fust and sixth graders used as the criterion of empathy, measures of sensitivity to the emotional states of others rather than the emotiooal experience of the "empathizer." The purpose of the present study is to investigate stimulus conditions inBuencing empathic responses in first-grade children, where empathy is defined as a vicarious response. One of these conditions is the sex of the stimulus person and its similarity to the sex of the subject. There is empirical evidence from Stotland and Dtmn's (1963) research to support the significance of similarity as a facilitator of empathic behavior. The basis of similarity in Stotland's study was competence in a sk:ill, and one would expect that similarity on a fundamental dimension such as se."t would be no less signi1icant. Studies of related processes, such as modeling (Kagan, Pearson, & \Velch, 1966) and projection (Feshbach & Feshbach, 1963). provide further evidencc for the relevance of the similarity variable. From a theoretical standpoint, empathy appears to be one aspect of imitative behavior or identification. Similarity on one dimension of personal characteristics shouJd enbance generalization of other attitudes and behavior, including emotional experiences, from the stimulus person to the perceiver. In addition to similarity effects, the study will also yield information regarding sex di.Herences in empathic behavior. Also, the type of affect will be manipulated in order to appraise the extent to which empathic response will vary with the specific emotion observed. METHOD

Subjects The subjects were 46 fi.rst-grade children (23 boys and 2~ girls ),t attending the University Elementary School at the University of California, There were initially 48 children in the study, but one boy and one girl were eliminated due to equipment fnilure. 1

13.(

NORMA D. FESHBACH ANO KIKI ROE

Los Angeles.2 The children carne from middle-class baclcgrounds, were above average in intelligence (mean IQ of 121) and ranged in age from 6 years and 2 months to 1 years and 7 months.

Procedure Each child was taken individually from thc classroom to the experimental room by a female experirnenter. Prior to entering the experimental room, the children were told that they were going to observe pictures and hear stories about children their own age. The instructions were repeated whiJe the child was seated in front of a table on which was placed a Hughes videosonic projector. A series of slide sequences, paired with narrativa material, was projected on the screen. Immediately following each sequence, the child was simply asked to state how he felt. The specific instructions used were "How do you feel?" and "Tell me bow you feel." Each child's direct verbal report was recorded verbatim and constituted the primary index of empathy. In addition, for 27 of the subjects, the sequences were presented again with instructions to indicate the feelings of the central child in each slide-story sequence. Tbe speci6c question used was ..How does this chiJd feel?" The responses to this question were used as a checlc on the children's comprehension of the affective situation and also on the presence of the appropriate labels in their verbal repertoire.

Affective Situations A series of slide sequences depicting 7-year-olds in one of four different affective situations, was developed. Each sequence consisted of three sudes. There were two sequences for each of the aHects of happiness, sadness, fear, and anger. In addjtion, two altemate sets of these eight situations were prepared, each set identical in content but different in terms of tbe sex of the stimulus figure. The following themes were used for each of the four affects: ( 1) Happiness: birthday party, winning a television contest; (2) Sadness: a lost dog, social rejection; (3) Fear: child lost, frigbtening dog; ( 4) Anger: the toy snatcher, false accusation. Accompanying each slide sequence was a short narration, matched for number of words over a1l affects, describing the events reflected in the slides. The narrations were so constructed that the use of speciBc or general affective labels were completely avoided. The following narration, accompanying the male, sadness slide sequence, typifies the construction of the series.

1: Hcre is a boy and hJs dog. This boy goes everywhere with bis dog, but sometimes the dog tries to run away. Slfde D: Here the dog is running away again. SUde 111: This time the boy cannot flnd him, and he may be gone and Jost forever. SlJdo

1

The authors wish to express their apprcclotfon to John Coodlad, director of the school; to Madelfne Hunter, principal of the school; and to the school staff for their assistance in facilJtat:ing this investigation. 135

CHILD DEVELOPMENT

Siml/Qrlty Variatitm Tbe maJe or female stimulus series was randomly assigned to subjects of both sexes. The series featuring the fernale stimulus child was presented to 12 girls and 11 boys, while the series portraying the rnale stimulus child was presented to 12 boys and 11 girls. Tbe variation of sex of the stimulus figure and the sex of the subject resulted in four experimental groups: Boys Observing Boy Stimuli {BB ): Boys Observing Cirl Stimuli {BC); Cirls Observing Boy Stimuli {CB ) : and Cirls Observing Cirl Stimuli {CC). In order to heighten the impact of the affective conten t, the two sequences depicting a particular affective situation were always presented consecutively. However, ord er of presentation of the four types of affect was controlled by initially determining the 24 perrnutations of the four affective categories and randomly assigning each permutation to one hoy and to one girl. To reduce the residual carryover from one affect to another, a brief sorting task was given between each of the three changes in affective categories. The tasks entailed such activities as putting jacks into various color categories and assembling jaclcs into piles of two, three. and four according to directions. Approximately 2 minutes were spent on eacb ínter· polated task, whiJe the time devoted to the administration of an affective category ( consisling of two slide-sequences) was of similar duration.

Dependent Measures Two diHerent scoring procedures were utilized in deternúning the subjects' verbal responses for empathy. In order for empathy to be scored in the first procedure, the feeling re8ected in the response had to be a speciflc match with the aHective situation observed. Thus, in response to the anger sequences, replies such as "not so good" to the question "How do you feel?" were not scored, while '1 feel mad• constituted an empathy response. The subjects received a score of 1 for each specific match; and, since there were two sequences for each affect, the empathy score for a speci.Sc affect could range from O to 2 while the total empathy score, over four affects, could range from O to 8. The second procedwe for scoring empathy was broader in that empathy was scored if the affective category and verbal response were consistent in teJms of their negative or positive connotations. For example, if in response to a fear sequence the child indicated that he felt oad,.. empathy was scored, while it would not have been under the criteria for the 6.rst scoring procedure, wbere onJy responses such as ..alraid," "'frightened," or "scared.. would have been acceptable. These two scoring procedures were similarly applied to the responses of tbose 27 subjects who were readministered the story-slide sequences and who were asked to report the feelings of the central figure in the situations. These latter scores constituted the social comprehensio n index. 136

NORMA D. FESHBACH ANO KIKI ROE

Because the range of verbal responses to each affectivo situation W3S quite limited, it was relatively simple to establish scoring prooedures which could be uniformly applied by diHerent raters. The two authors independ ently rated the protocols for 10 of the children and obtained at least 96 per cent agreemen t for each of the scoring procedures for the empathy and social comprehension indexes. Any discrepancy between the two raters in scoring a response was considered a disngreement. The reliability index, then, consisted of the number of agreements divided by the total number of responses scored. Separate reliability estimates were also determin ed for each of the four affective categories, and a similar high degree of consistency was obtained for each of the individual affects. RESULTS

The mean empathy scorcs for boys and girls in response to maJe and female stimuli, are presented separatel y by affect in Table l. Inspection of Table 1, in which the data are based on tbe more speci6c scoring system, reBects the predicted interaction benveen the sex of tbe subject and sex of the stimulus for each affect except fear. With that one exception, the empathy seores of boys observing boys ( BB ) are greater than those of boys observing girls (BG); empathy scores of girls observing girls (GG) are greater than the scores of girls observing boys ( GB). In the case of fear, both boys and girls report experiencing more anxiety when observing girls in a fearful situation. Because the range of scores within any affect category is restricted from O to 2, the assumptions permittin g nn analysis of variance of these distributions cannot be properly met. Severa1 alternative procedures for demonstrating tbe interaction behveen sex of the subject and sex of the stimulus were utilized. For each a.fJect, the BBB and GGG groups wero combined and compared witb the combined BG and CB groups. permittin g a contrast of the effects of similarity and dissimilarity between the subject and the stimulus. The differences between the similarity and dissimilarity groupings were evaluated by menos of a nonparametric procedure, the Mano Whitney U test (two-tailed ), and, as Table 1 iodicates, tbe diHerence s are signiBcan t at the .05 level for hnppiness, the .10 Jevel for sadness, and tbe .01 level for anger. In order to permit an overall comparison of the empathic behavior of the two sexes, a total empathic score, based on the sum of empathic responses to all of the olfcctivo situalions, was detemún ed for each child. The menos of these total empnthy scores are presented in Table 2. Since the range of scores is from O to 8, it was possible to use an analysis of variance. A marked interaction efiect, signi6cant at the .01 level (F 8.16), was obtafned between sex of the stimulus and sex of the subject. The t tests used to evaluate cell diHerences indicated that, while the eHect was most striking for the girls (CC vs. CB: p < .01). it also held for tbe boys (BB vs. BC: p < .05). Of

=

137

TABLE 1 MEAN EwPATBY ScoRES, BAsEo ON SPECif'IC MATCDING OF AFFECT, AS A FuNCTION ov SEx OP STtwULus, SEX OF SUBJECT, ANO 'I'vPE OF AFFECT (ScORING PROCEDURE 1) HAPPlNESS

SADNESS

FEAR

ANGER

SUDJECT

B

G

B

G

B

G

D

G

B .. . ....... . .

1. 92 (N- 12) 1.64 (N- 11)

1.36 (N - 11) 2. 0 (N- 12)

1. 58 (N - 12) 1.36 (N - 11)

1.36 (N.,. 11) 1.90 (N - 12)

0 .92 (N - 12) . 18 (N - 11)

0 . 27 ( N • 11) 1.0 (N- 12)

0 . 17 (N- 12) 0 . 27 (N- 11)

0 . 64 (N .. 11 ) 0 .58 (N- 12)

G ...... ..... .

BB + GG vs. BG + GB : 1> < .OS

BB BG

+

+

GG vs. GB: p < .10

BB + GG ve. BG + GD : p < .01

BB + GG vs. BG + GB : 1> > .10

NORMA D. FESHBACH ANO KIKI ROE TABLE

2

MBANS OF TOTAL Elll'ATHY ScoRE5 BASED ON SPECIFIC MATCHING OF APFECT STIIlULUS SUDJECT

Boy ..... . Girl. .... .

Boy

Girl

4.58 ( N = 12)

3.64 (N- 11)

( N- 11)

(N., 12)

3.45

5.50

interest, too, is the finding that the mean empathic score for girls observing girls is significantly greater (p < .05) than the corresponding score for boys observing boys. In order to pennit a comparison of empathy responses to the four affective situations, the responses to the four sequences depicting a particular affect were combined. Although the maJe.female stimulus variable is thereby eliminated, the resulting range of scores and distributions met the assum~ tions for the analysis of variance. The mean empathy score for each affect is presented in Table 3. A highly significant main effect (F 84.5, p < .005) for the affect variable was obtained, and the differences among the four affects were tested by the Neuman Keuls procedure (Winer, 1962). This analysis shows that the mean empathy score for bappiness is signi.6· cantly greater than that for sadness, whicb, in tum, is significantly greater tban that for aggression. The difference between aggression and fear is not significant The mean empathy responses based on the second scoring procedure, which required consistency rather than specificity in affective matching between subject and stimulus, are presented in Table 4. It can be seen that tbe previous findings for the happiness and sadness stimuli remain essentially the same, wbile the interaction for the anger stimuli becomes attenuated. However, the sex of subject·stimulus interaction for the empathic response

=

TABLE 3 VAJUATIONS IN EMPATHY RESPONSES BASBD ON SPECIFIC MATCHING OF APFECT, AS A FUNCTION OF TvPE OF AFFBcT

Affect

Mean

Anger

Sadness

Happineas

Fear . . . .. . ... . ....

0.83

0.36

2.28•

2.63• 2.27• 0. 84•

An~er . . .

. . ..•.... •

Sa ness . ... .... .. .

H~piness . . .. .• ...

ritical values

1.19

3. 11 3.46

••• ••• •••

.54

t. lO• • •• •••

.61

• ••

.65

NoTL-cet1 entrles are dilfertnc:es amon• the mean• ol the variou1 a.tfecta. Tllae cWfereDCU mult be ~l~ lhan the critica! values i.a orde~ t o be IIID16cant at the .01 leveL Indica tes lianlf!cant cWferenc:es at t < .01.

139

TABLE 4 MEAN E MPAmv ScoRES, BASED ON CoNsrs'rENT MATCDI NG OP AFPECT, .AS A FuNCTION

ov

Ssx STIMULUS,

SEX OF SUBJECT, AND TYl'E OF AFFECT HAPPINESS

SADNESS

F EAR

ANGER

SUBJECT

B

G

B

G

B

G

B

G

B ... . .........

1. 92 (N- 12 ) 1.64 (N- 11 )

1. 36 (N- 11) 2.0 (N- 12)

l. 75 (N- 12) 1.45 (N - 11)

1. 36 (N- 11) 1.90 (N- 12)

2 .0 (N- 12) 1.45 (N - 11)

1.54 (N- 11) 1.50 (N- 12)

1. 75 (N- 12) 1. 27 (N- 11 )

1.27 (N- 11 ) 1.83 (N- 12)

G .......•....

BB + GG vs. BG + GB: p < .05

BB + GG vs. BG + GB : p < .10

BB + GG vs. BG + GB : p > .10

+

BB GG vs. BG G B : p < .OS

+

NORMA D. FESHBACH ANO KIKI ROE

to the situation depicting fear is now significant. In addition, the mean empathy responses for the different affects are now more similar, in comparison to the previous analysis which reBected a number of significant differenoes arnong the four aHects. Total empathy responses, based on this second scoring procedure, were also calculated, and the means are presented in Table 5. A signi6cant subject by stimulus interaction eEfect is again obtained (F 4.61, p < .05) although these differences are more moderate as compared to the results based on the more specific scoring procedure. lt is of interest tbat, under this scoring procedure, the mean empathy responses for girls observing girls is no longer significantly clifferent from the mean empathy responses for boys observing boys and is, in fact, slightly smaller.

=

TABLE 5 MEANS OP TOTAL EMPATRY 5coR.ES BASED ON CoNSISTBNT MATCUINC OF AFF ECT STlMULUS SUBJECT

Boy .. .. .. Girl . . ... .

Boy

Girl

7.42

5.82

(N - 12)

(N - 11)

(N - 11)

(N - 12)

5.82

7.25

Relevant to the comparisons of the two scoring procedures for empathy are the judgments made by the 27 subjects who, when readministered the sUde scquences, were asked to state bow the stimulus person felt. As Table 6 indicates, the responses of the boys are quite similar to those of the girls exoept in the case of fear, where their judgments tend to be more accurate. Moreover, all subjects accurately labeled the happiness and sadness stimuli; anger was properly labeled in about half the judgments and fear in about 20 per oent of the judgments. Whcn the broader sooring procedure is followed, 100 per cent accuracy is obtained for all of the judgments. DISCUSSION

The contrast between the empathy data and the findings re]ating to the comprehension and recognition of affective states in others has important methodological and tbeoretical implications. These results indicate that the empathy measure is distinct from the social comprehension measure and tbat tbe variations in empathy cannot be accounted for solely by ability to reoognize the affective experiences of others. Social comprebension may be a neoessary prerequisite for empathy. However, given accurate recogni-

1-'1

TABLE 6

MEAN

A CCURACY OF CoMPRE B ENSJON OF S PECIPIC AFFECT 0 E1'1CTED IN TR E F OUR AFFEC'IlVE S JTUATIO NS HAPPlNESS

SADNESS

FEAR

ANGE R

SUB]ECT

B

G

D

G

B

G

B

G

B . . ... . .. ... .

2.00

2.00

2.00

2 .00

1. 00

(N • 7)

( N • 7)

( N "'" 7)

( N - 7)

( N = 7)

o.71

( N ... 7)

0 . 71 ( N- 7)

( N = 7)

( N = 6)

(N

(N - 6)

1.00

( N • 6)

== 7)

1. 00

0 . 71 ( N = 7) 0 . 17 ( N - 6)

G .. . .. ... .. ..

2.00

2 .00

2. 00

2. 00

( N - 7)

0 .43

( N .. 7)

NORMA D. FESHBACH ANO KIKI ROE

tion of the affective state of the stimulus person, there is still considerable varlation in tbe empathic response to that stimulus. Th1s variability in empathy, moreover, is not random but is systematically related to the similarity between the child and the stimulus person. The interaction between sex of the subject and the sex of the stimulus is, itself, evidence that social comprehension cannot account for the empathic responses that were observed. The empathic responses of both boys and girls to the male and fema)e sequences were signiBcantly diHerent although the sequences were cognitiveJy simüar, except for the düference in sex. A second finding of methodological importance is the diHerent outcomes resulting from the two scoring procedures. The 6rst procedure requiring specific matching of the affect of the perceiver with that of the stimulus person may result in the exclusion of genuine empathy responses if the subject's verbal description lacks precision. The broader scoring procedure will encompass tbese more general empatby responses but may also include other affective responses, such as sympathy. The increase in the social comprehension scores for anger and fear with the second scoring system suggests that, in appraising fear and anger as compared with happiness and sadness, subjects tend to be less discriminating and use more gross verbal labels. The second scoring procedure, then, probably provides a better estímate of empathy responses to the fear stimuli and, perhaps, to the anger situatlons than does tbe more specific scoring method. These dif. ferences in the discrimination and description of aHective states malee it difficult to interpret the ordering of the aHects according to their mean empathy scores. The results of both scoring procedures suggest that the happy sequences eUcit the most empathy while the diHerences obtained between sadness and the other two aHects are largely due to the less discriminate responses to the latter. It seems reasonable that there is more empathy with the happy situation than with the other aHects, since the experience of happiness is probably rewarding while the other aHective responses are unpleasant and are more generally avoided. The use of both scoring procedures, while presenting a more complex outcome for interpretation, also provides a better understanding of the results. Thus, the sex diiierences in empathic responses to same.sex stimuli, observed with the more specific empathy method, are no longer present under the second scoring method. This diHerence was largely due to the fewer speci6c empathy responses of the boys to the fear stimulus. Since the social oomprehension data indicate thnt the boys are no less discriminating than girls in their labeUng of the aHective responses of others, it may be inferred that boys are reluctant to describe themselves as afraid and tend to use more general descriptions, such as "feel bad ... These considerations indicate that boys are no less empathic than girls. Further, any sex düferences are subordinated to the interaction between the sex of the subject and the aex of the stimulus.

1.43

CHILD DEVELOPMENT Tbe major empirical finding of theoretical significance in this study is the consistent evidence supporting the hypothesis that similarity facilitates empathic responses. Boys were more empathic with boys than with girls; tbe converse was found with the girls. These results are in accord with the general theoretical importance that has been ascribed to similarity in predicting social bebavior. Since individuals can and do vary along many dillerent dimensions, the problem· remains of deterrnining which dimensions of similarity are relevant to a particular behavioral domain. With regard to empathy, Stotland and Dunn's ( 1963) data and the present findings r&spectively indicate that similarity of attitude and similarity of sex between an observer and a stimulus person enhance empathic responsiveness. lt can be inferred from these data that exaggeration of differences between individuals tends to reduce empathy. These propositions have sorne bearing on a number of social problems. For example, in integrating a school and in working with childreo from differeot racial and ethnic groups, the focusing oo group differences, including positive group qualities, to foster group ideotity and self-image oí group members, may result in decreased sharing of social feeling and empathy between children of differeot groups and even between teacher and child. Equal emphasis should also be placed on the similarities among children which relate to and promote com.moo affective experiences. The social application of empathy findings must, of course, await further lrnowledge in this area. Little is known concerning the developmental course of empathy, its antecedents and its personal and social consequences . The investigation of these problems would be facilitated considerably by a formulation of a more comprehensiv e measure of empathy, possihly inte grating autonomic, lcinesic, and self-report indexes.

REFERENCES

Berger, S. Conditioning through vicarious instigation. Psyclwlogical Reoiew, 1962, 69, 450-466. Dymond, R. F . A scale for thc measurement of empathic ability. ]oumal of Consulting Psychology, 1949, 13, 127-133. Dymond, R. F. Personality and empathy. Joumal of Consulting Psychology, 1950, 14, 343-350. Feshbach, S., & Feshbach, N. In.Ouences of thc stimu1us object upon tbe complementary and supplementary projection of fear. ]ournal of Abnormal and Social Pst¡chology. 1963, 66 ( 5), 498-502. Kngan, J., Pearson, L., & Welch, Lois. Modifiability of nn impulsive tempo. ]ournal of Educatlonol Psychology. 1966, 57 ( 6), 359-365. Milgram, N. A. Cognitivo and empathic fa ctors in role-taldng by schizophrenic and brain damaged patients. Joumal of AbnormlJl and Social Psychologr¡, 1960,65,219- 224.

NORMA D. FESHBACH ANO KIKI ROE Murphy, L. B. Social behaokn and child personality: an exploratory study of sorne root.s of sympathy. New York: Columbia University Press, 1937. StotJand, E ., & Dunn, R. E. Empathy, self-esteem and birth order. ]ournal of Abnorrru~l and Social Psychology, 1963, 66,532-554. Stotland, E., & Walsh, J. Birth order in an experimental study of empathy. ]oumal of Abnormal and Social Psychology, 1963, 66, 610-614. Strayer, F . K. Empathy and social perception. Dissertation Abstracts, 1960, 21, 244. Winer, B. J. Statistical principies in experiment4l design. New York: McCrawHill, 1962. [ Cisild D~t~llc,mmt, 1968, 39, 133-HS. ® 1968 by the Society for Rcuarch Úl Oúld

Development, Inc. ]

1.45