Journal of Abnormal Child Psychology, 1973, 1, 2, 139-151

Facial Expressions and Interpretation of Emotion-Arousing Situations in Deaf and Hearing Children ~ PENELOPE B. ODOM, 2 RICHARD L. BLANTON AND CLAIRE LAUKHUF 3

Vanderbilt University Deaf and hearing children were given two tasks: (a) sorting faces portraying nine emotions and (b) matching those faces with drawings of appropriate emotion-arousing situations. The deaf children performed as the hearing children did on the first task but did not match the faces to the situations as well as the hearing children. It appeared that the deaf children were unable to analyze and interpret emotion-arousing events adequately. Possible reasons for this finding are presented and discussed in detail.

An emotional immaturity, often described as lack of empathy and deficient impulse control, is one of the most consistent findings resulting from clinical investigations of deaf subjects (Altshuler, 1963, 1964, 1967; Baroff, 1963; Levine, 1956, 1960). These results are thought to be due to an insensitivity to the feelings of others on the part of the deaf; Levine (1956) used the expression "cheerful indifference to the feelings of others [p. 152] ." A related set of findings stem from previous work by Blanton and Nunnally (1964a) in which they found, in a forced-choice word association test, that the deaf chose significantly fewer evaluative associations than did their hearing counterparts. In a subsequent study (Blanton & Nunnally, 1964b), they

1This research was supported, in part, by Social Rehabilitation Services Grant No. RD-2552. 2Requests for reprints should be sent to Dr. Penelope Odom, John F. Kennedy Center, George Peabody College, Box 163, Nashville, Tennessee 37203. 3Now at University of Kentucky Medical Center, Lexington, Kentucky. The authors are indebted to Dr. Lloyd Graunke, Delmas Young, and Warren Flower of the Tennessee School for the Deaf, Elizabeth Stallings of the Monroe Harding Children's Home, Rev. Lucius Hart and Rev. Hudlow of the Baptist Children's Home, Dr. Lloyd Funchess and Jerome Freeman of the Louisiana State School for the Deaf, W. W. Wallace, Milton Lillard, and Wilburn Kelley of the Wiltiamson County Tennessee School System, and Charles Barbara of Tennessee Preparatory School.

139 Copyright 9 1973 by V. H= Winston & Sons, inc.

140

ODOM,BLANTON, AND LAUKHUF

administered Miller's Locus of Evaluation Scale to 137 deaf children and 302 hearing children. The results indicated that deaf children have a more external locus of evaluation relative to the hearing children. This tendency of deaf children to externalize responsibility for the evaluation of environmental conditions was interpreted by the authors to indicate the relatively weak associative strength of evaluational terms in the deaf . . . . Our results support the idea that the deaf not only have a weaker pool of evaluative terms and associations, but feel inadequate to make meaningful evaluative judgments and tend to rely on others for such ' discriminations [Blanton & Nunnally, 1964b,

p. 893]. While it may come as no surprise that the deaf have some noticeable emotional problems, the underlying causes of such problems and their specific connections with hearing loss have not been studied by experimental methods. The purpose of the research reported here was to examine some possible factors contributing to this problem. A definition of empathy that most people would agree with is probably that of English, H. B., and English, A. C. (1958): "apprehension of the state of mind of another person without feeling as the other feels." In the display of empathy, a person should possess several skills that are important: (a) he should be able to recognize the cues in the environment which signal emotion in others, both in the situation causing the emotional reaction and in the reaction, both physical and verbal, of the person experiencing the emotion; (b) he should be able to interpret or evaluate the cues as to their implicaitons, i.e., he must do more than merely acknowledge "another instance of x"; and (c) perhaps he should know that the demonstration of empathy is significant in interpersonal relations, often eliciting positive affect and other reinforcing responses from other persons. It is assumed, although not necessarily so, that the person must have had past experience with a particular feeling state. The research by Blanton and NunnaUy (1964, a,b) described previously would seem to imply that the deaf are deficient in the ability to interpret and evaluate emotion-arousing situations. There is little reason to believe that lack of direct or indirect experience with particular emotions is an explanation for their lack of empathy. It can be argued that they experience emotionally arousing situations more often than the hearing due to the frequent negative reactions on the part of hearing people to the deaf. The expressive behavior of the deaf is quite intense and vivid according to most observers. The studies reported here were planned to investigate the probable source of the lack of empathy attributed to the deal The first study was designed to examine the deaf person's ability to interpret emotion-arousing scenes. The task required deaf and hearing subjects to look at a picture of an emotional situation and select an appropriate facial expression for one of the characters in the situation.

FACIAL EXPRESSIONS AND INTERPRETATION OF EMOTION

141

The second study was undertaken to determine whether differences obtained in the previous study were attributable to the deaf subjects' inability to interpret situations or their inability to categorize facial expressions. The latter skill, categorizing facial expressions, was the focus of this study. Deaf and hearing subjects sorted a stack of photographs of people portraying nine different emotions. Should the deaf be deficient in this task, it could be concluded that their purported lack of empathy is a very general attribute, perhaps having its source in basic perceptual processing. If the deaf subjects perform similar to the hearing subjects, however, it can be concluded that the differences found in the previous task are caused by difficulties in evaluating emotion-arousing situations. EXPERIMENT I This task was one which seemed to measure the degree to which a subject could infer from emotionally loaded situations what affect the person in that situation was experiencing and could match it with a facial expression appropriate to that emotion. For several years, Izard (1971) has been compiling a series of photographs of individuals depicting various emotions. To date, using these standardized materials, he has found evidence for nine affects that are detectable by their facial expressions. These are listed in Table 1. His photographs, along with a set of drawings of situations, were the materials used in the study. Deaf and hearing children were compared in their ability to select a facial expression that was appropriate for a situation. The experiment was nonverbal so that the hearing subjects would have no advantage with respect to understanding instructions or establishing rapport with the experimenter. Method

Subjects Fifteen deaf subjects were selected from a population of about thirty 7- to 8-year-olds at the Tennessee School for the Deaf, a residential school. All subjects had a hearing loss of 75 dB in their best ear and were prelingually deafened. The mean IQ for the group was 97.6 as measured by the WAIS or WISC. There were seven gifts and eight boys. The 30 hearing subjects came from two grade levels, (15 subjects from each): a private kindergarten and the second grade of a rural public elementary school. No IQ data were available on the kindergarten children, but aB were considered average but not superior by their teachers. The mean IQ for the second graders was 98.8 as measured by the LorgeThorndike Test. Eight boys and seven gifts comprised each group of elementary school children; nine boys and six girls made up the kindergarten group.

bJ

Anger:

Distress:

Joy:

Surprise:

Fear:

Lion chasing child Man pointing gun at figure Ghosts scaring a child Coiled snake near child Genie appearing from bottle Surprise birthday party Jack-in-the-box Child's bubble gum popping Child eating ice-cream cone Christmas day scene Child playing with puppies Figure receiving lollipop Child who has fallen from bicycle Truck running over dog Child looking at de-headed snowman Person in hospital with cast on leg Child hitting the other Child hitting another with a bat Child being put to bed and struggling Child grabbing toy car out of another's hands

Emotion and situation

95 93 89 79 90 88 84 83 95 94 92 81 88 78 76 75 90 87 85 77

response agreement

Percent of

88 74 72 71 80

Child holding dirty diaper Person holding an apple with a worm in it Chewing gum stuck on foot Child looking at pastry covered with flies

Contempt: Group of children ignoring a lone child Children looking at figure who missed the ball Children watching child tangled in jump rope Children watching child topple a pile of blocks

35

37

65

85 84

Disgust:

Shame:

Child watching his mother put on makeup Construction being watched by child Child watching mother cook Child looking at monkey Child spilling milk Child caught tracking dirt in house Mother shaking hand at child who has broken piggy bank Person being spanked

100 100 92 77 95 92

Percent of response agreement

Interest:

Emotion and situation

PERCENT AGREEMENT FOR SITUATIONS USED TO MATCH WITH FACIAL EXPRESSIONS

TABLE 1

FACIAL EXPRESSIONS AND INTERPRETATION OF EMOTION

143

Since the deaf are also in an institutional setting, it seemed appropriate to have a control group of institutionalized hearing children. Administrators at two residential institutions, a private and public orphanage, allowed their children to participate. After screening for low IQ scores and eliminating two children for not meeting criteria on the practice items, only 13 of these children qualified-3 girls and 10 boys. The mean IQ of six members of this group was 81. The remainder of the children for whom IQ scores were not available had reading achievement scores at or above grade level.4

Materials Situation drawings were made by giving a list of the affect categories to an artist with suggestions for appropriate situations. She made 67 sketches, approximately seven situations for each affect. The 67 drawings were shown by an opaque projector to 70 college students who were instructed to pick (from the list provided) the word that best described the way the designated figure felt in the situation. The four situation drawings eliciting the most agreement in each affect category became the test stimuli. Two of the drawings, fear and surprise, are included as Figures 1 and 2. It was originally planned to have a criterion of 70% agreement with no more than one-sixth of the responses in any other affect category. This criterion was met for at least four drawings in every category except contempt. The percentages are shown in Table 1. Additional drawings were made following suggestions for appropriate contempt situations solicited from the same subjects. These were embedded in a subset of 10 of the other acceptable drawings and tested again. All the additional drawings had to be rejected because of failure to reach criterion. After several such attempts, it was decided to use the four situation drawings that received the most agreement. They are the ones shown in Table 1. (It was decided to retain the emotion of contempt because some of the matching materials, the faces, were standardized. No conclusions were based on the results of individual emotions.)

Training Before each subject was tested on the actual matching task, 12 training items were administered. Each consisted of a picture of an object or situation drawn on a 24.1 x 17 cm poster card that the subject was to look at before selecting from a choice of 3 colors or objects (all presented on a 35 x 16.3 cm poster card), the one most appropriate for the picture. The early training items paired

4The matching or selection of subjects based on IQ is partially perfunctory because deaf and hearing subjects' scores are of questionable comparability. The problems associated with gathering valid IQ data from children in institutional settings are well known.

144

ODOM,BLANTON, AND LAUKHUF

FIG. 1. Drawing representing fear. a banana with 3 colors, a pumpkin with 3 colors, etc. The later items paired a snowman with 3 different activities (the correct picture was a child in a snowsuit throwing a snowball) and a beach picture with 3 varied scenes. The last item concerned an appropriate face for a Santa Claus. Test

The test materials consisted of the 36 situation drawings-four representing each of the 9 affect categories and the 36 triads of photographs of facial expressions used by Izard (1971). The triads were arranged so that every correct facial expression was seen once with a facial expression from every other affect category. The position (left-middle-right) of the correct face was randomly determined with the restriction that it occurred equally often in each position and that the same position was not the correct response for all four presentations of an affect category. Furthermore, no triad contained all three faces that belonged to the same sexed person. The four correct choices for each affect category were represented by two different faces-two male and two female. Procedure

Each subject was tested individually in one session which lasted about 30-45 minutes. He was seated at the table with the experimenter and asked if he would

FACIAL EXPRESSIONS AND INTERPRETATION OF EMOTION

/

].45

f

FIG. 2. Drawingrepresenting surprise. like to play a game. The hearing subjects were also told that the game had one rule: "We can't talk while we play. You'll have to figure out from the way we point (on the first training trials) how you're supposed to play the game." The 12 training items were shown to subjects. The first two items were demonstrated by the experimenter; the rest were indicated by the subjects. Subjects had to respond correctly on five out of the last six training items to participate in the test. On the 36 test trials, subjects were shown the situation drawing that illustrated a particular affect and a triad of three faces, each portraying a different emotion. Subjects pointed to the face in the triad that illustrated the emotion in the situation. No feedback was given during the test. Each subject had a different order of test stimuli; and the triad accompanying a given situation drawing was systematically switched within an affect category such that four subjects in each group were shown a particular triad of faces with, for example, the drawing of a holdup. Four other subjects saw another triad with that drawing and the first triad with another fear situation. There was no time limit imposed on the subjects. Results The mean number of correct choices (agreements with college-age raters) is shown by groups and emotions in Figure 3. Considering only the 7- and 8-year-olds, the noninstitutionalized hearing subjects performed best, the institutionalized children were in the middle, and the deaf scored lowest on every emotion. An analysis of variance for unequal frequency per cell on number correct by all groups (kindergarten, deaf,

8

MEAN

I I

I

: oJ

o

NUMBER

CORRECT

.to o

RESPONSE5 .~ o

I

.~ o

I

I

u

0

I

~X,\\\\~\\\\\\\\\\\\\-,~ I

~

!

I

,I

0 =

~|

~ D D I ..,,z

z,

~"

~

0 r

"1

N

~\\\\\\\\\\\\\\\\\\\~1 I

3

FACIAL EXPRESSIONS AND INTERPRETATION OF EMOTION

147

second-graders, institutionalized second-graders) and emotions (nine) provided a significant groups effect (F = 7.79, df = 3/55, p < .01); a significant emotions effect (F = 33.33, df = 8/440, p < .01); a small but significant interaction (F = 1.37, df = 24/400, p < .01). A posteriori comparisons indicated the following order of goodness of performance: hearing 8-year-olds > hearing 5-year-olds (F--13.20, df = 1/55, p < .01); hearing 5-year-olds > institutionalized 8-year-olds (F = 4.85, df = 1/55, p < .05) and institutionalized 8-year-olds > deaf 8-year-olds (F = 8.51, df = 1/55, p < .01). Sex differences were not included in the analysis because of its already complicated nature and because Odom and Lemond (1972, p. 360) failed to find differences in performance based on sex. The deaf-hearing difference could be due to a differential understanding of the instructional demands of the task. However, this lack of understanding should result in general decrement in performance. The very high performance of the deaf on certain emotions-i.e., fear and anger-would argue that they could follow instructions on items that they could do well on. It could also be argued that by comparing the performance of deaf subjects with a standard set by hearing adults we are eliminating the possibility that there are common experiences peculiar to the deaf. Perhaps the deaf subjects have learned a different affect associated with a particular emotion-laden situation. To examine the degree of consensus among the deaf subjects, an error matrix was constructed in which erroneous choices of the 7- to 8-year-olds were classified according to which response would have been correct and which was chosen. There were a total of 253 errors committed by the deaf and 203 committed by the hearing children. The most agreement, among the deaf, was attributing a surprise situation with a disgust expression. This accounted for 4% of the deaf's errors and 3% of the errors committed by the hearing subjects. The remainder of the errors were rather evenly distributed over the entire matrix for both groups. EXPERIMENT II

The inferior performance of the deaf on the matching task could be due to two factors: either they could not recognize the facial expressions or they could not interpret the situation as well as the hearing subjects could. To distinguish between these two possible causes experimentally, a second study was conducted. Deaf and hearing subjects were compared on their ability to sort facial expressions into the nine predetermined categories. Method

Subjects There were six groups of subjects in the sorting task. Another sample of second-grade children was selected from the middle IQ range at a public

148

ODOM,BLANTON, AND LAUKHUF

TABLE 2 MEAN NUMBER CORRECT, MEAN NUMBER OF ERRORS, AND MEAN NUMBER OF ERRORS AS A PROPORTION OF TOTAL RESPONSES TO A PHOTOGRAPH BY GROUPS

No.

Mean No.

correct

elrors

No. errors/total No. cards put in category

12.9

14.4

.521

16.2

10.9

.400

16.5

10.5

.388

20.7

6.3

.233

19.5

7.5

.342

17.8

4.1

.277

Mean Group

Eight-year-old institutionalized (N = 14) Eight-year-old hearing ( N = 16) Eight-year-old deaf (N = 12) High school institutionalized (N = 16) High school hearing ( N = 16) High school deaf ( N = 12)

elementary school, but the younger deaf subjects and the younger institutionalized subjects were almost all the same children as were used in the previous experiment. There was a time lag of 2 months between the matching task and the sorting task for the deaf subjects. The institutionalized subjects did both tasks in the same experimental session. Half completed the matching before the sorting; the remainder did the tasks in the reverse order. Some high school age subjects were included in the study to see if there are developmental trends in recognition of facial expressions. They were all of average IQ and were obtained from the same sources as the younger subjects. Since there were unequal numbers of subjects in each group, these numbers are included in Table 2.

Materials The photographs to be sorted were the ones used in Experiment 1. One standard or prototype from each emotion category was laid face up on a large poster board that had black lines dividing it into nine equivalent sections, each 7.5 x 7.5 cm. Subjects were given the rest of the photographs (the order of

FACIAL EXPRESSIONS AND INTERPRETATION OF EMOTION

149

which was determined by shuffling) and were told to put all the people who felt the same way in the same pile. They were instructed to put each photograph under the photograph already displayed. No time limit was imposed; most subjects took about 15 minutes. A different set of standard photographs was displayed for each subject and the position of the emotion on the board was changed for each subject. This means that a subject viewing the board with the nine exemplars exhibited was seeing different faces in different positions from the previous subject. Results There are two ways to measure accuracy in sorting the faces-the number of correctly classified faces and the number of incorrectly classified faces. It is conceivable that a subject could have placed all members of his stack under the same photograph. This would have resulted in a perfect score of three correct on the first measure and an error score of 24 on the second. The means for each measure over all emotions are presented in Table 2. A derived measure, the number of erroneous placements in a pile divided by total number of items placed in that pile, is also presented. It can be seen that all measures seem to exhibit the same fluctuations. An unequal frequency-per-cell analysis of variance (6 groups x 9 emotions) was computed only on the number of photographs correctly sorted, since all the measures appeared to be highly correlated. There was a significant groups effect (F = 7.88, d f = 5/76, p < .01), a significant emotions effect (F = 30.60, d f = 8/592, p < .01), and another small but significant interaction (F = 1.45, d f = 40/592, p < .05). The following a priori comparisons were made: 8-year-old deaf vs. high school deaf, F = < 1 8-year-old deaf vs. 8-year-old hearing, F = < 1 8-year old deaf vs. 8-year-old institutionalized, F = 2.34, df= 5/76, p < .05 8-year-old hearing vs. 8oyear-old institutionalized, F = < 1 High school institutionalized vs. high school deaf, F = < 1. Note that within the two age groups, the deaf and hearing subjects performed similarly, except that the young deaf children sorted more accurately than the young institutionalized children. Although young and old subjects were not compared directly, the means reported in Table 2 show that the older subjects sorted the expressions more accurately. Odom and Lemond (1972) used a slightly different technique but found a similar developmental effect between kindergarten children and fifth-grade students. DIscussioN The major finding of these studies was that deaf subjects showed a relative inability to match an appropriate facial expression to an emotion-arousing

150

ODOM, BLANTON, AND LAUKHUF

situation. This finding, apparently, is not due to some perceptual difficulty in recognizing the facial expression since deaf and hearing subjects performed comparably in the expression sorting task. It can be concluded that their deficiency is probably related to the analysis and interpretation of emotion-laden situations. There is an interesting conceptual asymmetry in facial expressions and situations, although they both provide cues which signal emotion to the subject. Situations which elicit a particular emotion are disjunctive in nature in that one may be quite challenged in trying to find similarities between two instances that produce, say, surprise or interest. Indeed, the only thing these situations may have in common is that they elicit a common emotion. A facial expression, on the surface, however, appears to have much more in common with other instances of the same expression. (Why we choose some facial features as "similars" and ignore others is a related and very important question but will not be discussed here.) It was concluded that the deaf have difficulty analyzing and interpreting emotion-producing situations. The tasks used, however, do not provide measures that distinguish between the ability to recognize cues in the environment which could cause emotion and the ability to evaluate the cues according to their implications. It is interesting to speculate which is a more accurate designation of the source of this problem. Assuming that deaf children experience the same behavioral situations as hearing children and see the subsequent expressions that people display, the question is: "Why can't they match the two as well as the hearing children?" What role, if any, does verbalization and its reception play in developing the ability to evaluate appropriately the nature of an emotionally-arousing situation? Although there may be several explanations, the one that has a great deal of intuitive appeal relies on some form of attentional construct. Verbalization of feelings and attributes of a situation may serve to focus (orient) a child's attention on its salient and relevant aspects. Consider the number of times a mother says: "Don't you e v e r . ("hit your sister again," "interrupt your father," etc.) "You ought to be ashamed of yourself for . ("hitting your sister," "spilling your milk," etc.) "Of course it hurts. Next time you _ _ it'll hurt even more." ("pull her hair," "grab the toy," etc.). This list could be quite lengthy and still not capture the frequency with which a hearing child is told exactly what aspects of a situation are causing emotion in another person. The deaf child, especially one with hearing parents, is especially handicapped because most of this information is not available to him. Although rewards and punishments probably serve to isolate some factors that are important, they may not provide enough feedback to allow him the

FACIAL EXPRESSIONS AND INTERPRETATION OF EMOTION

151

finer discriminations that hearing children are capable of making. It seems that such information could be communicated to the deaf child by the parent or teacher with some effort, given the knowledge that it is important to do so. The data reported here, while certainly not conclusive, suggests that receiving this kind of information is an important factor in the development of mature affective judgment. The lack of emotional empathy in the deaf may well be a model instance of a relationship between language training and an important dimension of personality. In other areas where communication in the domain of affective behavior is said to be disturbed, e.g., the "double-bind" situation, (Bateson, Jackson, Haley & Wealdand, 1956) a method such as the one used in the present research might have value. REFERENCES Altshuler, K. Z. Sexual patterns and family relationships. In J. Rainer, K. Altshuler & F. Kaltmann (Eds.), Family and mental health problems in a deaf population. New York: Columbia University Press, 1963. Altshuler, K. Z. Personality traits and depressive symptoms in the deaf. In J. Wortis (Ed.), Recent advances in biological psychiatry. Vol. 6. New York: Plenum Press, 1964. Altshuler, K. Z. Theoretical considerations in development and psychopathology of the deaf. In J. Rainer & K. Altshuler (Eds.), Psychiatry and the deaf. (U.S. Department of Health, Education, and Welfare) Washington, D.C.: U.S. Government Printing Office, 1967. Baroff, G. S. Rorschach data and clinical observations. In J. Rainer, K. Altshuler, & F. Kallmann (Eds.), Family and mental health problems in a deaf population. New York: Columbia University Press, 1963. Bateson, G., Jackson, D., Haley, J., & Weakland, J. Toward a theory of schizophrenia. Behavioral Science, 1956, 1, 251-264. Blanton, R. L., & Nunnally, J. C. Evaluational language processes in the deaf. Psychological Reports, 1964, 15,891-894. (a) Blanton, R. L., & Nunnally, J. C. Semantic habits and cognitive style processes in the deaf. Journal of Abnormal and Social Psychology, 1964, 68, 397-402. (b) English, H. B., & English, A. C. A comprehensive dictionary of psychological and psychoanalytical terms. New York: Longmans, Green, 1958. Izard, C. F. The face of emotion. New York: Appleton-Century-Crofts, 1971. Levine, E. S. Youth in a soundless world. New York: New York University Press, 1956. Levine, E. S. The psychology o f deafness. New York: Columbia University Press, 1960. Odom, R., & Lemond, C. M. Developmental differences in the perception and production of facial expression. Child Development, 1972, 43, 359-369.

Facial expressions and interpretation of emotion-arousing situations in deaf and hearing children.

Deaf and hearing children were given two tasks: (a) sorting faces portraying nine emotions and (b) matching those faces with drawings of appropriate e...
673KB Sizes 0 Downloads 0 Views