YOUNG PRESCHOOLERS' ABILITY TO IDENTIFY EMOTIONS IN EQUIVOCAL SITUATIONS

Young childrens' ability to perceive others' feelings has previously been assumed to be limited; such suppositions have, however, been themselves limited by non-contextualized measurement. Thus, it was predicted that preschoolers will show a non-chance ability to use personalized information about others' emotions. This ability will be age-related, and greater where: (a) the other's and their own likely emotions differ along positive/negative valences; and (b) certain pairs of self/other emotions are contrasted. Preschoolers (N=53) were presented puppet vignettes in which they were to identify the puppet's emotion; happy, sad, angry, and afraid emotions were contrasted. Subjects often made Personalized inferences to correctly indicate the puppet's feelings. Each hypothesis was supported.

Young childrens' ability to perceive others' feelings has long been assumed to be limited due to cognitive immaturity and egocentrism (Chandler & Greenspan, 1972). In recent years, however, this assumption has been challenged by use of more contextually valid measurement tools, and more careful sub-analyses of the abilities involved in such affective sensitivity (Denham, 1986; Denham & Couchoud, 1990; Gnepp, McKee, & Domanic, 1987).

Earlier research on young preschoolers' emotion knowledge has suggested that young preschoolers can infer basic emotions (e.g., happy, sad, angry, and afraid), either from expressions or from situations. By the end of the preschool period, if not earlier, they also can quite easily identify the linkage between emotional expressions, situations which unequivocally cause emotion (e.g., being happy to receive an ice cream cone), and the consequences of such emotional expressions (Denham, 1986; Denham & Couchoud, 1990; Gnepp, Klayman, & Trabasso, 1982; Michalson & Lewis, 1985; Stein & Jewett, 1986; Strayer, 1986). Specifically, this research indicates that happy expressions and situations are more readily identified than negative expressions or situations, particularly fear (Denham & Couchoud, 1990).

Current research also portrays younger children as beginning to identify other persons' emotions, even when their own reactions might reasonably differ (Denham, 1986; Gnepp, 1983; Gnepp & Gould, 1985). Gnepp calls this the ability to make inferences of appraisal; in her measures, such inferences require the use of personalized knowledge of the person to conclude that s/he is feeling a non-consensual emotion (i.e., most people would not feel this way; e.g., feeling happy at seeing a spider, or sad to go on a family hike). This ability increases with age, from kindergarten through elementary school (1983; Gnepp & Gould, 1985). Increasing social experience, and the cognitive ability to process information gleaned from such experience, presumably underly this age change.

Discovery of these abilities highlights the need to investigate preschoolers' capabilities to use expressive cues to identify protagonists' emotions in more complex situations. Kindergarteners have some trouble making nonconsensual inferences of appraisal, and younger children may have particular difficulty doing so, perhaps because the bases for such inferences are out of the realm of their experience. Gnepp's procedures can, however, be modified to test the proposition that children younger than six years old are beginning to make correct inferences of appraisal.

Perhaps young preschoolers find it difficult, despite their developing theories of mind (Harris, 1989), to make inferences that are as personalized as Gnepp's. In contrast, perhaps if Gnepp's procedures were modified to focus on situations where protagonist's emotions, rather than quite unusual and personal (e.g., loving spiders), were within the realm of experience (e.g., being happy or afraid to go in a swimming pool), greater ability in younger children would be discovered. Thus, in this study emotionally equivocal situations rather than nonconsensual reactions to typically unequivocal situations were presented to the preschool-aged subjects.

In these equivocal situations, any given child could easily feel either of two possible emotions (e.g., one could reasonably be afraid or quite happy to enter the water at a swimming pool, depending on personal experience and preference). Most preschoolers would have experienced the possibility of differing reactions to certain different emotional situations; they know that people have differing beliefs and desires (Gnepp, 1989a; Gnepp et al., 1987; Harris, 1989). Thus, this level of personalized inference of appraisal might be easier for them to make. It was predicted that, given such modifications, the ability to make personalized inferences in such situations would exceed chance expectations in young preschoolers, although many subjects would evidence some difficulty in making these inferences. It was also predicted that, even within a relatively restricted age range, this ability would increase with age.

Further, the downward extension of procedures used to evaluate children's abilities to make inferences of appraisal must focus on contextualization of measurement. To this end, engaging the preschooler's attention, minimizing verbal response requirements, and integrating the experience into ongoing play were undertaken in this study.

It also is important to examine children's abilities to identify others' emotions in equivocal situations across several different emotions, since earlier research with unequivocal situations supports a discrete emotions view (i.e., that some emotions would be more difficult to understand than others, due to the differential communicative value of each; Izard, 1971). This research, thus, extends that of both Denham (1986) and Gove and Keating (1979) in its examination of personalized inferences involving different emotions.

The following predictions also expand on this earlier work: Given young preschoolers' advanced understanding of happiness relative to negative emotions, it was expected that the other's emotional reaction would be easier for young preschoolers to discern where happiness or negative emotions are possible, as opposed to where two negative emotions are possible. Further, sadness is often used by preschoolers as an omnibus negative emotion; it may be inferred when other negative emotions are actually expressed. Also, of the four basic emotions, fear is the most difficult for preschoolers to correctly identify. Therefore, it would be reasonable to assume that when the two possible emotions for any situation includes the emotion of fear, the preschoolers' nascent ability to make personalized inferences of appraisal would break down. This would be true whenever fear is a possibility, whether in equivocal situations where happiness and fear, or fear and another negative emotion, are involved.

In this study, three hypotheses were therefore addressed. First, preschoolers will show a greater than chance ability to make personalized inferences of appraisal, and this ability will be age-related. Next, personalized inferences of appraisal will be more readily made by preschoolers for certain contrasted emotions. For example, children might most easily be able to make inferences of appraisal if the emotions inherent in the equivocal situation vary along positive/negative valences. Making inferences of appraisal would be more difficult where the two possible emotions are negative. In either case, situations where fear is possible should be the most difficult.

Method

Subjects

The sample consisted of 53 preschoolers with a mean age 44.88 months (range: 33 to 56 months). There were 27 boys and 26 girls. They attended five classrooms in a university laboratory preschool in a suburb of a major metropolitan area.

Measures and Procedures

Children were escorted to a quiet area by a female tester whom they knew well. They were presented with twelve vignettes via puppetry; the situations in each were planned to be equivocal (i.e., two emotional responses to each event were possible, such as feeling happy or sad to come to preschool), so that the child had to attend to the puppet/experimenters' emotional cues to make the correct inference. To this end, the puppeteer's systematically controlled, standard facial and vocal expressions of emotions were included during each vignette. These peak facial expressions followed prototypes of discrete emotions theory (Izard, 1971; see Table 1 for examples of expressions and vocalizations, for each specific emotion). The 12 items represented two contrasts each of possible emotional responses to specific situations: happy or sad, happy or angry, happy or afraid, sad or angry, sad or afraid, and angry or afraid (see Table 2).

Situations portrayed in each item were chosen based on questionnaire responses from 41 adults. These adults were given a pool of 30 situations and asked to identify what emotion a child would feel in each situation, from a choice of the four emotions previously mentioned. The 12 vignettes chosen for inclusion in the measure were those for which the adults' responses were approximately evenly divided between two emotions as paired previously, suggesting that the situations were indeed equivocal.

Order of the situations was randomized for each subject. Children were asked to respond to the question of "How does the puppet feel?" by affixing a felt face with the appropriate emotional expression on the puppet (from a choice of happy, sad, angry, and afraid). They received I point for a response which depicted the emotion actually shown in the situation, and 0 points for any other response. The pair of items for each emotion contrast were summed, yielding six emotion contrast scores per subject.

The emotion shown by the puppet/experimenter in each vignette was determined by in formation given by subjects' mothers, who had indicated the emotional reactions that their child would show in each equivocal situation. The same-sex puppet depicted the contrasting emotion in each situation. Information from mothers was used to maximize the necessity for the child to make personalized inferences of appraisal. Even if the mothers' responses were at times incorrect, the situations enacted by the puppets were equivocal in nature, so that the child would have to attend to the puppet/experimenters' emotional cues to make correct personalized inference of appraisal. In Table 3 the number of children shown each emotion for each vignette, and the number of errors made for each emotion portrayal are shown.

The expressions on the detachable faces have been validated with a sample of 41 adults (i.e., happy, sad, and angry faces were correctly identified by 85 to 95% of the adults, whereas the fearful face was correctly identified by 65% of the sample. Wagner, McDonald, and Manstead [1986] corroborate that fear expressions are the hardest to identify, even for adults). Further, children showed their ability to identify these expressions prior to the 12 vignettes; the 53 subjects in this study obtained an average score of 6.4 (with one point each given for correct verbal and nonverbal identification of each expression). In fact, 41 of the 53 subjects obtained either seven or eight of the eight possible points on this measure. Nevertheless, ability to identify the emotion expressions involved was used as a covariate in the ANOVAs, in order to assess the effect of positive/negative valence, unconfounded by the effect of sheer ability to discriminate the expressions to affix to the puppet.

Results

Subjects quite often spontaneously verbalized about the process of making personalized inferences of appraisal. At times, the children evidenced simple wonderment at the puppets' "odd" perspective (e.g., looking very surprised, or saying, "But why????" "But I love French fries!"; But I love to swim in the water!") They also questioned the nature of the situation, even while indicating the correct emotion (e.g., "Why did the mom say he couldn't touch that?"). They seemed to be grappling with the fact that emotional situations could be equivocal.

At other times, their reactions were more complex. For example, one boy, after indicating that the puppet was indeed happy to eat oatmeal (i.e., making the inference of appraisal), lectured the puppet on its more objectionable qualities, notably its lumps! Some children changed the stories, as in putting the angry face on the dog puppet when the child puppet was afraid (i.e., creating a consensual emotional situation), having the girl puppet suck her thumb when she was sad, and saying "They'll bring her some ice cream" (cf. Gove & Keating, 1979). It is clear that the children struggled with what they felt were the somewhat unusual emotional responses of the puppets. Making personalized inferences of appraisal was not always easy for them.

Hypothesis One: Overall Ability to Discern Equivocal Situations, and Effect of Age

To demonstrate the subjects' overall ability to make personalized inferences for emotionally equivocal situations, the sum of each subject's scores of one or zero for all 12 vignettes was divided by 12 to form a proportion-correct score. The average of these proportion-correct scores for the 53 subjects, .66, was significantly greater than that expected by chance, t (52) = 12.98, p < .001. Next, an Age (3) x Sex (2) mixed model ANCOVA was set up. For the repeated factor, two levels of a Valence factor were created: (a) the mean of positive-negative items, such as Happy-Sad; and (b) the mean of negative-negative items, such as Sad-Angry. Means and standard deviations for these scores are listed in Table 4. Total score for labeling and recognizing the emotion expressions served as the covariate. Gender was not used as a factor in this analysis, because preliminary between-subjects Age (3) x Sex (2) ANCOVAs (with expression labeling/recognition total scores as the covariate) showed that gender never exerted a significant effect on the emotion contrast scores; moreover, there were no significant Age x Sex interactions.

The Age factor consisted of 3-year-old, 3-1/2-year-old, and 4-year-old levels. Older children made more accurate personal appraisals, F (2,49) = 3.10, p < .05. There was also a significant main effect for the covariate of ability to identify emotional expressions, F (1,49) = 4.74, p < .05; children who were less adept at identifying the expressions on the felt faces understandably made less accurate personal appraisals. There was no significant Age x Valence interaction, F (2,50) = 0.89.

Hypothesis Two: Equivocal Situations Involving Positive and Negative Emotions Versus Only Negative Emotions

To determine whether valence contrasted affected scores on the measure, the repeated measure factor in the ANCOVA was evaluated. Results indicated that preschoolers were indeed more able to utilize the personal emotional information regarding the puppet, and make a personalized inference of appraisal, in vignettes with positive-negative emotion contrasts, rather than negative-negative contrasts, even with ability to label the emotional expressions on the faces partialled out, F (1,50) = 28.81, p < .001.

Hypothesis Three: Differences Among Specific Emotions Within Positive/ Negative Emotion Contrasts

The significant F aforementioned confirms that young children were more able to use personalized information for certain emotion contrasts. Planned comparisons using the six emotion contrast scores as dependent variables pinpointed the nature of this effect: Scores were lower for vignettes involving fear as a possibility (see Table 5). Scores for the happy versus sad items were also higher than those for the happy versus angry items.

Moreover, nonparametric analyses of distributions of errors shown in Table 3 corroborate these findings. Subjects made more errors than expected by chance whenever the puppet enacted fear; chi²s (1) for contrasts of afraid versus happy, angry, and sad equalled 9.78, 13.09, and 54.87, respectively, p < .001. Similar analyses also suggested that correctly identifying the puppet's angry emotion was more difficult when the subject's own likely emotion was sadness; chi² (l) equalled 10.12, p < .001.

Discussion and Conclusion

These results corroborate those reported by Gnepp and Gould (1985), where kindergarteners responded to equivocal situations correctly over 60% of the time, and abilities to make personalized inferences of appraisal increased with age through sixth grade. This study confirms similar trends with younger subjects and equivocal, rather than nonconsensual, situations. In less taxing procedures than Gnepp's, these subjects accurately identified emotions in approximately 66% of the equivocal situations. Contextualized measurement appeared to make this task accessible to younger preschoolers (cf. Denham, 1986; Denham & Couchoud, 1990). It may be quite important that actual facial and vocal cues were available to the children to aid them in their decisions (Gnepp, 1989a). Subjects were not, however, merely "reading" off the puppet's emotional expressions; their spontaneous verbalizations, and the significant Valence effect, despite partialling of ability to identify emotional expression, argue against this simple an explanation.

Happiness/nonhappiness is the first differentiation made among both emotional expressions and emotional situations (Borke, 1971; Bullock & Russell, 1985; Denham, 1986; Denham & Couchoud, 1990; Michalson & Lewis, 1985; Ridgeway, Waters, & Kuczaj, 1985). Happy expressions and situations may be also accentuated by socializing agents more than negative expressions and situations (Malatesta, 1981). It is not, therefore, surprising that this developmentally early differentiation is also found in a more complicated task.

Stein and Jewett's (1986) model would also support these findings based on dimensions differentiating the causes of happiness, fear, anger, and sadness. In states of happiness, one is focused on attainment of a desired state or avoidance of an undesired state; one anticipates goal enjoyment. In all negative emotional states, one focuses on an undesired state which is eminent. Thus, there is a clear distinction between happiness and the negative emotional states. Sadness is likely to be the first, more or less global, negative emotion differentiated on this basis.

Young children's slow identification of specific negative expressions and situations also would suggest that vignettes where two negative emotions were possible would be more difficult, as was found here. Earlier analyses of kindergarteners' understanding of equivocality have suggested that young children do realize there are situations in which different persons feel different ways. Preschoolers may fail, unless prompted, however, to apply this general knowledge to specific emotional situations (Gnepp, 1989a, 1989b; Gnepp et al., 1987).

In particular, earlier research suggested that young children find stories where the consensual emotional reactions are positive, and personalized reactions are negative, easier than those which contrast two negative emotional reactions (Gnepp and Gould, 1985). This finding is conceptually similar to the main effect of valences found here.

Subjects' confusion over situations which pitted a positive and a negative emotion may be qualitatively different than confusion over two possible negative emotions. The differentiation among negative emotional states is more individual, more dependent upon one's interpretation of the situation. For example, although both fear and sadness focus on the consequences to self of an undesired state, loss is emphasized in sadness, and anticipation of harm in fear. In the equivocal situation of receiving punishment for misbehavior, then, some children may focus on the anticipation of harm from a spanking (and choose fear); others on the loss of "face" or of closeness with the parent (and choose sadness). Further, as corroborated in chi-square analyses here, the same situation can cause anger or sadness depending on the individual's focus (e.g., one is angry when one's blocks are knocked down if one focuses on the violation of one's wish by another, or one is sad if one focuses on the loss of the beautiful tower). Thus, the possibility of individual differences in negative emotional situations may have rendered children's choices more difficult in such situations.

Fear is one of the last negative emotions to be accurately identified (Michaelson & Lewis, 1985); not unexpectedly, children in our study had difficulty making personalized inferences of appraisal when fear was involved. This relative deficiency may stem from different socialization of fear than for other emotions. For example, parents say, "Don't be afraid." and people talk much less about fear than about other emotions. Full-blown fear is also not often observed in others, and the experience of this emotion may become more inward than anger and sadness; thus these children may have had less experience dealing with the fear emotion. This lack of experience, coupled with the relative perceptual difficulty of differentiating the fear expression, make for ample justification of this study's findings.

The hypotheses addressed were thus supported. More research on the substantial affective knowledge of preschool children is needed to address certain issues, such as fine-tuning the measure introduced here, by asking children how they would feel in these situations and later enacting the puppet vignettes with the alternate emotion, and by more intentionally equating emotional intensity for each vignette. Young preschoolers' differentiation of negative emotions should also be more finely delineated.

The abilities discussed here also appear important in the enactment of socially competent behavior. For example, Denham, McKinley, Couchoud, and Holt (1990), and Gnepp (1989b) have found these skills to be related to sociometric peer ratings (even with level of cognitive development partialled out). These abilities also are implicated in self-regulation of emotion, particularly the self-regulation involved in reacting to others' emotions. First, one must safely determine which negative emotion one's peer is emitting (e.g., so that one can either respond prosocially, withdraw, or stand up for one's rights) and second, one must determine whether positive emotion is being emitted (e.g., so that one can share in it). Understanding of equivocal emotional situations may, thus, allow for better cue use in emotional situations, and its development should be further studied to maximize children's social competence.

Footnotes

An earlier version of this paper was presented at the Annual Meetings of the American Psychological Association, August 26, 1987.

Reprint requests can be sent to:

Susanne A. Denham
Department of Psychology
George Mason University
4400 University Drive
Fairfax VA 22030

TABLE 1 Indicators of Emotion Shown by Puppet and Puppeteer

The following chart reads as follows:

Row 1: Emotion; Hand Puppet's "Body Language"

Row 2: Puppeteer's Facial Cues; Puppeteer's Vocal Cues

1 Happy                    "Bounces,"
                            spreads arms

    Broad smile             Relaxed,
                           "pearly" tones

2 Sad                       Wipes eyes,
                            Head downcast

    Eyes & mouth            Whiny,
    down-turned             crying tones

3 Angry                     Clenched

    Eyebrows down,          Gruff, growling,
    lips pursed             clipped, abrupt

4 Fear                      Hands up Rigid

    Eyes wide,              High-pitched
    mouth gaping            unwavering

TABLE 2 Vignette Content

Emotion Contrast   Vignette Content

Happy or Sad       Separating from mother when coming to
                   preschool.

Happy or Sad       Going to airport to see parent off on a trip.

Happy or Angry     Being served favorite or most hated food.

Happy or Angry     Coming in from playing outside when called for
                   dinner.

Happy or Afraid    Meeting a very big, but friendly, dog.

Happy or Afraid    Going into the water at the swimming pool.

Angry or Sad       Some other kids not letting you play.

Angry or Sad       Having to stay home when everyone else goes out
                   to get ice cream.

Sad or Afraid      After doing something naughty, a parent says
if
                   you do it again, you will be punished.

Sad or Afraid      Experiencing the death of a pet.

Angry or Afraid    Being punched by an older sibling, who says
                   (s)he will do it again if you tattle.

Angry or Afraid    Getting a spanking.

TABLE 3 Emotions Enacted for Vignettes

Legend for Chart:

A - Emotion Contrast
B - Correct Emotion
C - N[a]
D - Errors[b]

     A                              B             C            D

Happy/Sad                         Happy           35           5
                                    Sad           70           8

Happy/Angry                       Happy           42           0
                                  Angry           62           6

Happy/Afraid                      Happy           33           4
                                 Afraid           73          19

Sad/Angry                           Sad           37           7
                                  Angry           69          25

Sad/Afraid                          Sad           29           6
                                 Afraid           75          71

Angry/Afraid                      Angry           42          10
                                 Afraid           62          34

a Number of children viewing puppet emitting this emotion in the two vignettes for this emotion contrast. b Number of errors out of N seeing this emotion.

TABLE 4 Descriptive Data for ANCOVA

Legend for Chart:

A - Repeated Factor
B - Youngest[a], M
C - Youngest[a], SD
D - Middle, M
E - Middle, SD
F - Oldest, M
G - Oldest, SD

        A                      B                C            D
                               E                F            G

Positive-Negative             1.24            0.63          1.68
                         (1.31)[b]              --        (1.66)

                              0.29             1.57         0.50
                                --           (1.52)           --

Negative-Negative             0.83             0.57         1.24
                            (0.90)               --       (1.22)

                              0.51             1.33          0.37
                                --           (1.28)            --

a Age groups correspond to 3-year-olds, 3-1/2-year-olds, and 4-year-olds. b Means adjusted for covariate are in parentheses.

TABLE 5 Planned Comparison Data

Emotion Contrast Scores           M                   SD

Happy/Sad[a]                     1.79                0.50
                            (1.79)[b]                  --

Happy/Angry                      1.53                0.72
                               (1.54)                  --

Happy/Fear                       1.17                 0.80
                               (1.16)                   --

Sad/Angry                        1.51                 0.78
                               (1.51)                   --

Sad/Fear                         0.98                  0.75
                               (O.98)                    --

Angry/Fear                       0.91                  0.84
                               (0.97)                    --

Planned Comparison                            F              P

Happy/Sad vs. Happy/Afraid[c]             29.31[d]         .001
Happy/Angry vs. Happy/Afraid                 11.00         .001
Happy/Sad vs. Happy/Angry                     7.91          .01
Sad/Angry vs. Sad/Afraid                     15.63         .001
Sad/Angry vs. Angry/Afraid                   19.07         .001
Angry/Afraid vs. Sad/Afraid                   0.28           ns

a Scores based on two items, each contrasting the two emotions. Highest possible score = 2. b Means adjusted for ability to label and recognize emotional expressions are in parentheses. c Scores higher for italicized emotion pair. d df = (1,51); all contrasts significant at p < .001 would still be significant if a Bonferroni correction were used.

References

Borke, H. (1971). Interpersonal perception of young children: Egocentrism or empathy? Developmental Psychology, 5, 263-269.

Bullock, M., & Russell, J. A. (1985). Further evidence on preschoolers' interpretation of facial expressions. International Journal of Behavioral Development, 8,15-38.

Chandler, M. J., & Greenspan, S. (1972). Ersatz egocentrism: A reply to Helene Borke. Developmental Psychology, 7, 104-106.

Denham, S. A. (1986). Social cognition, prosocial behavior, and emotion in preschoolers: Contextual validation. Child Development, 56, 197-201.

Denham, S. A., & Couchoud, E. A. (1990). Young preschoolers' understanding of emotion. Child Study Journal, 20, 171-192.

Denham, S. A., McKinley, M., Couchoud, E. A., & Holt, R. W. (1990). Emotional and behavioral predictors of preschool peer ratings. Child Development, 61, 11451152.

Gnepp, J. (1983). Children's social sensitivity: Inferring emotions from conflicting cues. Developmental Psychology, 19, 805-814.

Gnepp, J. (1989a). Children's use of personal information to understand other people's feelings. In C. Saarni & P. L, Harris, Children's understanding of emotion (pp. 151-180). Cambridge: Cambridge University Press.

Gnepp, J. (1989b). Personalized inferences of emotions and appraisals: Component processes and correlates. Developmental Psychology, 25, 277-288.

Gnepp, J., & Gould, M. H (1985). The development of personalized inferences: Understanding other people's emotional reactions in light of their prior experiences. Child Development, 56, 1455-1464.

Gnepp, J., Klayman, J., & Trabasso, T. (1982). A hierarchy of information sources for inferring emotional reactions. Journal of Experimental Child Psychology, 33, 111-123.

Gnepp, J., McKee, E., & Domanic, J. A. (1987). Children's use of situational information to infer emotion: Understanding emotionally equivocal situations. Developmental Psychology, 23, 114-123.

Gove, F. L., & Keating, D. P. (1979). Empathic role-taking precursors. Developmental Psychology, 15, 594-600.

Harris, P. L. (1989). Children and emotion: The development of psychological understanding. Oxford: Blackwell.

Izard, C. E. (1971). The face of emotion. New York: Appleton-Century-Crofts.

Malatesta, C. Z. (1981). Infant emotion and the vocal affect lexicon. Motivation and Emotion, 5, 1-23.

Michalson, L., & Lewis, M. (1985). What do children know about emotions and when do they know it? In M. Lewis & C. Saarni (Eds.), The socialization of emotions (pp. 117-139). New York: Plenum Press.

Ridgeway, D., Waters, H, & Kuczaj, S. A. (1985). Acquisition of emotion-descriptive language: Receptive and productive vocabulary norms for ages eighteen months to six years. Developmental Psychology, 21,901-908.

Stein, N. L., & Jewett, J. L. (1986). A conceptual analysis of the meaning of negative emotions: Implications for a theory of development. In C. Izard & P. Read (Eds.), Measurement of emotion in infants and children (pp. 238-268). Cambridge: Cambridge University Press.

Strayer, J. (1986). Children's attributions regarding the situational determinants of emotion in self and others. Developmental Psychology, 22, 649-654.

Wagner, H. L., MacDonald, C. J., & Manstead, A. S. R. (1986). Communication of individual emotions by spontaneous facial expressions. Journal of personality and Social Psychology, 50, 737-743.

~~~~~~~~

By Susanne A. Denham and Elizabeth Couchoud, George Mason University