Empathy, Emotional Expressiveness, and Prosocial …faculty.tru.ca/wlroberts/prosocial.pdfEmpathy, Emotional Expressiveness, and Prosocial Behavior William Roberts Cariboo College

Empathy, Emotional Expressiveness, andProsocial Behavior

William RobertsCariboo College

Janet StrayerSimon Eraser University

ROBERTS, WILLIAM, and STHAYER, JANET. Empathy, Emotional Expressiveness, and ProsocialBehavior. CHILD DEVELOPMENT, 1996,67,449-470. Relations between emotional expressiveness,empathy, and prosocial behaviors are important for theoretical and practical reasons. In thisstudy, all 3 areas were assessed across methods and sources. Emotional expressiveness andempathy were evaluated in 73 children in 3 age groups (5-, 9-, and 13-year-alds) by measuringfacial and verbal responses to emotionally evocative videotapes and hy ratings from best friends,parents, and teachers. Measures of emotional insight and role taking were also obtained, Proso-cial behaviors were assessed by 3 laboratory tasks and by ratings from best friends, parents, andteachers. Confirming expectations, latent variable path analyses (LohmSller, 1984) indicated thatemotional expressiveness, emotional insight, and role taking were strong predictors of latentempathy (multiple R^ = .60). Boys' empathy, in turn, was a strong predictor of prosocial behavior,R^ = ,55. In contrast, girls' empathy was related to prosocial behaviors with friends, R^ ~ .13,but not to cooperation with peers. Thus present findings provide important support and clarifica-tion for certain theoretical expectations, and also raise issues that need clarification.

Relations between emotional expres-siveness, empathy, and prosocial behaviorsare important for both theoretical and practi-cal reasons. Because of their complexity, wehave found it useful to diagram these rela-tions. The model presented below will beused to organize this section of our paper,and it also summarizes the expectations weexamine in later sections. As will be seen,the three constructs of primary interest inthis study were assessed across methods(laboratory observations, interviews, andquestionnaires) and sources (children,friends, parents, and teachers).

A Theoretical ModelIn describing the model presented in

Figure 1, it will be convenient to postponediscussing age-related changes until wehave defined and discussed the other majorconceptual domains it incorporates. We nowturn to the first of these, emotional expres-siveness, the intensity of experienced anddisplayed emotion.

Emotional ExpressivenessEmotional expressiveness and its inter-

actions with empathy and behavior consti-

tute a complex set of phenomena, whosestudy is characterized by diverse theoriesand ambiguous data. One basic distinctionin this domain involves the type of emotionbeing expressed. Following others (e.g,,Bowlby, 1982; Lazarus, 1991), we take theview that positive affect is generally associ-ated with competent, situationally appro-priate behavior and refiects an evaluation bythe individual that things are going well. Incontrast, negative emotions (fear, sadness,and anger) are often associated behavioralresponses that are less than optimal and re-fiect evaluations that things are not goingwell. Roberts (1984, 1995) and Roberts andStrayer (1987) have discussed some of theparticular processes by which high levels ofnegative emotions can disrupt behavior, in-cluding empathic responses and prosocialbehaviors, a view consistent with other re-search (e.g., Batson, Fultz, & Schoenrade,1987; Eisenberg et al., 1989).

From a functional point of view, nega-tive emotions themselves need to be differ-entiated. Sadness, fear, and anger are subjectto different display rules regarding their ex-pression, depending on the social context as

This research was supported by a grant from the Social Sciences and Humanities ResearchCouncil of Canada. We would like to express our thanks to the parents, teachers, and the childrenand their best friends, whose participation made this project possible.

[Child Development, 1996,67,449-470. © 1996 by the Society for Research in Child Development, Inc.All rights reserved. 0009-3920/96/6702-0015$01.00]

450 ChUd Development

FIG, 1.—Empathy and related variables as causes and predictors of prosocial behavior; a theo-retical model.

well as the gender and age of the child ex-pressing the emotion (Brody, 1985; Lewis &Michalson, 1983; Saarni, 1979). For girls, torexample, expressing sadness may be "appro-priate" in the eyes of others and the self,whereas expressing anger may not. Becausechild, teacher, and parent ratings of emo-tional expressiveness in life contexts are pre-sumably influenced by implicit displayrules, high scores for expressiveness on suchmeasures may reflect "inappropriate" ex-pressiveness, rather than simply a greaterwillingness or ability ofthe child to engageon an emotional level.

Another reason for differentiating nega-tive emotions is that their outcomes arethought to differ (Campos, Campos, & Bar-rett, 1989). In contrast to sadness or fear, forexample, moderately intense or prolongedanger tends to disrupt or even sever socialrelationships (Bowlby, 1982). Also, it is clearthat anger is usually incompatible with em-pathic responses and prosocial behaviors(Hoffman, 1975).

These considerations lead us to expectdifferent pattems of correlation for differentmeasures of emotional expressiveness in thepresent study. On one hand, because ourteacher, parent, and self-report ratings pri-marily assess the expression of anger, andbecause high ratings on these measures mayreflect episodes of very intense anger, weexpect them to show negative correlationswith empathy. On the other hand, the rela-tively moderate ievels of emotional engage-ment that we can ethically induce in our lab-

oratory measures, and the fact that ourstimulus materials elicit sadness, fear, andhappiness more than anger, lead us to expectthat these measures of expressiveness willshow positive correlations with empathy.Low levels of expressiveness will indicateeither lack of engagement or emotional over-control, and in either case should be associ-ated with lower levels of empathy; greaterexpressiveness (still moderate by compari-son with the more intense emotional experi-ences that can be encountered in life outsidethe laboratory) should result in greater em-pathy.

These contrasting expectations can beunified by viewing the relation betweenemotional expressiveness and behavior as aninverted-U function. As expressiveness in-creases from low levels, it indicates an in-creased engagement with the situation be-ing experienced, and thus an increasedprobability of empatliic and prosocial re-sponsiveness. On the other hand, high levelsof expressiveness are disruptive of empathicand prosocial responses, because they implyeither a focus on the self (i.e., personal dis-tress) or movement against the other (in Aecase of anger). It is at moderate levels of ex-pressiveness, then, that empaîc respon-siveness, which involves a balance betweenone's own affective exjwiience and tfiat ofthe other, will be maximal (Strayer, 1987).Thus moderate levels of emotional engê-ment motivate interventions white not beingintense enough to disrupt behavioral plansand their execution (Roberts, 1984,Roberts & Strayer, 1987).

Roberts and Strayer 451

In assessing emotional expressiveness,we take the view that it comprises both emo-tional and cognitive processes of which theindividual may be either aware or relativelyunaware (Bowlby, 1982; Lewis & Michal-son, 1983). For these reasons we have as-sessed empathy and emotional expressive-ness by facial as well as verbal measures inthe present study, and we have includedmeasures of emotional insight (i.e., chil-dren's reported awareness of their emotionalstates as reflected in their facial expres-sions).

Emotional Insight and Role TakingIn the model presented in Figure 1, two

additional factors contribute to empathy:children's recognition of their own emotions(their insight) and their role-taking abilities.Emotional expressiveness and emotional in-sight may share some causal factors. Chil-dren experience socialization pressures con-cerning the control and display of emotions(Brody, 1985; Roberts, 1995; Roberts &Strayer, 1987; Saami, 1979), and these pres-sures may afFect their understanding of theirown emotional experiences (Lewis & Mi-chaison, 1983). It is, for example, commonfor researchers studying facial and verbal in-dices of empathy to find relatively low levelsof correspondence between children's fa-cially expressed emotions and the emotionsthat they attribute to themselves (Chisholm& Strayer, 1995; Eisenberg et al., 1989), withsome verbal responses conforming to genderrole stereotypes (i.e., girls reporting moresadness and fear than boys, and boys moreanger than girls; Strayer & Roberts, in press).Consistent with the view advocated byLewis and Michalson (1983) and Schachterand Singer (1962), we consider this to be anexample of a socially biased interpretationof the nonverbal components of one's emo-tional response.

Denial or dissimulation is another com-mon distortion of one's emotional experi-ence. Like gender-linked discrepancies infacial and verbal measures of emotions, de-nial can also be attributed to socializationpressures (Blanck, Rosenthal, Snodgrass,DePaulo, & Zuckerman, 1981; Lewis & Mi-chalson, 1983).

In the context of our laboratory proce-dures assessing emotional expressivenessand empathy, we were able to observe twoofthe three aspects of emotional insight justmentioned. The accurate recognition ofone's own emotional experience is reflectedin our measures as the congruence of facially

expressed and self-attributed emotions. De-nial or dissimulation is reflected in facial ex-pressions of sadness, fear, or anger coupledwith verbal assertions of feeling happy orfeeling nothing at all. (Dissimulation im-plies more awareness of emotional experi-ence than denial, but this distinction is notimportant here: Regardless of level of aware-ness, the negative emotion is judged to beunacceptable—either socially or to the self.)We did not attempt to assess gender-relateddistortions of emotional experience. Al-though such patterns are evident in thegroup data (see Strayer & Roberts, in press),identifying individual instances seemedmore problematic than for accurate recogni-tion and denial because such judgmentswould be based in part on children's gender,not just their observed behavior.

As indicated in Figure 1, we expectedemotional insight to affect role taking andempathy. Accuracy in identifying one's ownemotions and lack of denial should berelated to role taking because greater un-derstanding of one's own emotionalexperiences should contribute to greaterunderstanding of others' experiences. Emo-tional insight is also expected to be relateddirectly to empathy, in that the occurrenceof denial precludes an empathic response.High levels of accuracy also suggest a his-tory in which others have responded posi-tively to the child's displayed emotion, a his-tory that leads to empathic responding(Zahn-Waxler, Radke-Yarrow, & King, 1979).

Because better understanding of othersshould enhance empathy, we expect roletaking to contribute directly to empathy. Al-though understanding the plight of others(without being affectively involved) mightbe expected to lead to some sort of concretehelpful or cooperative response, we followthe lead of other researchers (Feshbach,1975; Krebs & Russell, 1981) in holding itunlikely that role taking directly motivatesprosocial behaviors. Rather, as shown in Fig-ure 1, we expect role taking to lead to greaterprosocial behavior because it enhances em-pathy (Davis, Hull, Young, & Warren, 1987;Hoffman, 1982).

Empathy and Prosocial BehaviorsIn contrast to sympathy and persona!

distress, empathy is marked by shared affectand a balanced focus between the self andother (Strayer, 1987). In general, we expectthe relations shown in Figure 1 to holdwhether empathy is measured as affectmatching alone (Feshbach & Roe, 1968) or

452 Child Development

whether its assessment includes cognitivemediators of affect (Hoffman, 1982; Strayer,1993). This extends also to children's self-appraised emotional responsiveness andempathy (Bryant, 1982). We are less certainhow well others (best friends, teachers, andparents) can assess children's empathy (incontrast to their prosocial behaviors), but weanticipate that similar relations will obtain.

We expect links between empathy andprosocial behavior to be clearest for verbalmeasures of empathy (as opposed to physio-logical or facial measures) because verbal as-sessments more closely reflect the cognitiveappraisals that are involved in planning andcarrying out behavior (Bowlby, 1982; Con-nolly & Bruner, 1974; Miller, Galanter, &Pribram, 1960). That is, to the extent thatprosocial responses involve evaluations(e.g., of the demands of the situation, of re-sources available from others, of the emo-tional states of others and the self, and possi-ble alternative responses) and decisions(i.e., the selection and execution of a re-sponse), cognitive processes are involved ofwhich we have some degree of awareness.Thus we agree with Lewis and Michalson(1983) that our understanding of our emo-tional experience (tapped by verbal mea-sures) may be more important for behaviorthan nonverbal components of emotional ex-perience (tapped by facial or physiologicalmeasures). It is for this reason diat we fo-cused on verbal responses in assessing em-pathy across laboratory measures and reportsfrom children, best friends, and teachers.However, we also assessed empathy on aless cognitively mediated level by includinga measure of facial expressive empathy simi-lar to those employed by other researchers(e.g., Eisenberg et al., 1988, 1994).

Although we have not differentiatedprosocial behaviors in Figure 1 for the sakeof simplicity, most work indicates that thisconstruct is multidimensional. Some re-searchers have found that context and re-spondent are important features. For exam-ple, Strayer (1981) reported that youngchildren's naturally occurring prosocial be-haviors fell into two groups, one directed topeers, the other to adults. Other researchershave distinguished prosocial behavioral do-mains, most frequently identifying positiveresponses to others' emotional distress (car-ing, comforting), helping or aiding others,sharing or donating, and cooperation(Radke-Yarrow, Zahn-Waxler, fie Chapman,1983). One would expect empathy to be

most closely linked to the first domain (givenits affective nature) and least to the last,given the reciprocal and social norms thatsurround cooperation (Craves & Graves,1983; Trivers, 1983). In this study, we at-tempted to assess these domains acrossmethods (laboratory observations and rat-ings of real-world behavior) and sources(parents, teachers, children, and bestfriends), with a focus on prosocial behaviorswith peers.

Although it is plausible to suppose thatempathy enhances prosocial responding(e.g., Hoffman, 1982), empirical evidencehas been inconsistent, with most ofthe sup-porting evidence coming from adult samples(Barnett, 1982, 1987; Batson et al., 1987;Eisenberg & Miller, 1987a, 1987b; Un-derwood & Moore, 1982). It has been sug-gested that inconsistent findings may be duein part to shortcomings in assessing basicconstructs. In contrast to much of this earlierwork, the present study has multiple mea-sures of empathy as well as multiple mea-sures of prosocial behavior. It is, therefore,better able to assess the underlying con-structs and their relations independently ofartifacts of method and biases of source.

Another possible expianation for incon-sistent findings is that important moderatorvariables have not been identified. For ex-ample, it has been suggested that processesrelating empathy and prosocial behaviormay differ for girls and boys (e.g., Radke-Yarrow et al., 1983), although specific detailshave not been proposed. Data indicate thatthere are dependable gender differences (fa-voring girls) in mean amount of empathy onmeasures that involve cognitive appraisals(as compared to facial or physiological mea-sures; Lennon & Eisenberg, 1987); that girlsand boys experience different socializationpressures for expressing and controllingemotions (Brody, 1985); and that such social-ization pressures are related to children'sprosocial behaviors (Roberts, 1995; Roberts& Strayer, 1987). Moreover, gender differ-ences in empathy are not paralleled by simi-lar differences in prosocial behaviors(Radke-Yarrow et al., 1983), and the absenceof a dose-response relation (in which in-creasing levels of a putative cause produceincreasingly strong effects) suggests that em-pathy may not be strongly linked to girls'prosocial behaviors. For these reasons, wecarried out separate analyses for each gen-der, and they will be reported below whenthey differed significantly.


Age-related ChangesHaving defined our central constructs

and discussed some ofthe issues associatedwith them, we return to the question of age-related changes. Our expectations concern-ing developmental changes in emotional ex-pressiveness and insight reflect thecomplexity of the phenomena and the rela-tive lack of research in the area.

First, age should be related to greateremotional insight and greater intensity ofnegative emotional experience in our labora-tory measures. Because children's copingand cognitive skills increase with age, mak-ing negative emotions less disruptive, weexpect older children to permit themselvesto experience and report a greater intensityof negative emotions. In addition, older chil-dren should experience more intense emo-tions in our laboratory procedures becausethey understand situations more fully, andthey have a greater store of their own emo-tional experiences which may be reactivated(Flapan, 1968; Flavell, 1985; Gilbert, 1969;Roberts, 1995; Roberts & Strayer, 1987).However, we do not expect age-relatedchanges to be due to increases in basic un-derstanding of our stimulus materials(Strayer, 1993). There is a growing body ofevidence that by the end of the preschoolperiod children can identify basic emotionalexpressions and situations, and are able totalk meaningfrilly about their own and oth-ers' emotions (Denham, Zoller, & Cou-choud, 1994; Saarni, 1990). To take an exam-ple from the current stimulus set, althougholder children may understand more fullythe implications of having an abusive parent,even 5-year-oId children can understand theemotional response of the video-child whois slapped across the face.

In contrast to negative affect evoked inthe laboratory, we expect ratings of chil-dren's expression of anger in everyday lifeto decline with age, as socialization pres-sures for emotional control are internalized(Cole, 1986; Dunn & Brown, 1991; Roberts,1995; Saarni, 1989) and more effective waysof coping with frustration and peer confiictare acquired (Piaget & Inhelder, 1969).

These same considerations lead us toexpect that age will not be related to inten-sity of positive emotions over the range sam-pled in the current study (5 to 13 years). Be-cause understanding of happy events isacquired early, probably before the age of 5(Dunn & Brown, 1991; Lewis & Michalson,

1983), understanding of simple stimulus ma-terials portraying happiness (and hence in-tensity of positive expression) would not beexpected to change in important ways dur-ing middle childhood. In addition, socializa-tion pressures to control displays of happi-ness are probably linked to contexts (suchas sermons and funerals, for example) ratherthan to age, so that changes in emotional reg-ulation should not produce any obvious age-related changes in laboratory measures ofpositive affect.

In contrast to these tentative expecta-tions, both theory and past research indicatea strong relation between age and role-taking skills, especially for the measure usedhere, developed by Selman (Piaget & In-helder, 1969; Selman & Jaquette, 1977).Thus, as shown in Figure 1, we expect age-related changes in empathy (Lennon &Eisenberg, 1987) to be mediated by age-related changes in role taking (as well aschanges in expressiveness and emotional in-sight). Similarly, we expect age-relatedchanges in prosocial behaviors (Radke-Yarrow et al., 1983) to be mediated by age-related increases in intervening variables,notably role taking and empathy.

In summary, following developmental-ists such as Hoffman (1982), it is our con-tention that empathy is an important way inwhich emotional engagement motivates pro-social behaviors. In our model, we considermany of the paths antecedent to empathy tobe predictive: In the current state of theoryand research, we can only specify lines ofinfiuence, not causal processes. Yet we dobelieve that some links diagrammed in Fig-ure 1 are causal in nature. Both theory andempirical evidence support the contentionthat role-taking ability enhances empathy,and that empathy, in turn, is a cause of proso-cial behavior (Barnett, 1982, 1987; Batson etal., 1987; Davis et al., 1987; Iannotti, 1978).Thus, although some parts of the proposedmodel are strongly supported by theory andresearch, other parts are more tentative orexploratory.

MethodSubjects

A total of 73 children in three age groupsparticipated in this study. Croup 1 consistedof 15 boys and 18 girls (M = 5.1 years, SD= .3). Croups 2 and 3 each consisted of 10boys and 10 girls. For Group 2, M = 8.8years, SD = .4; for Croup 3, M = 13.1 years,SD - .4. Children came from predominantly


white, middle-class backgrounds. Mean agefor mothers was 37 years (SD = 5.3), and forfathers, 39 (SD = 6.2).

Measures and ProceduresWe begin by describing the laboratory

procedure used to derive four of our sevenmeasures of emotional expressiveness andtwo of our five measures of empathy.

Children individually viewed six emo-tionally evocative videotaped vignettes. Thevignettes (a brief description is provided inTable 1) portray primarily dysphoric affect,as assessed by adult and child judges(Strayer, 1993). Positive emotions occurbriefly across vignettes, and are prevalent in"Circus." Details of selection and pretestingare given in Strayer (1993).

Vignette interviews.—Ghildren wereindividually interviewed after first watchingall vignettes. Each story was cued by a pic-ture, and children described the vignette'scontent in their own words, as a check onmemory and comprehension. They werethen asked to report each character's emo-tion and its intensity, and whether theythemselves had felt neutral ("OK," "just reg-ular") or an emotion (and its intensity) inresponse to the vignette. The memory checkand interview were carried out for each vi-gnette in turn. Eight emotion categorieswere used across all ages: neutral, happy (in-cluding positive surprise), sad, angry, afraid,startle (including negative surprise), con-cerned-worried, and disgusted.^ Intensity ofemotion was rated on a 3-point scale: 0 =none, 1 = a little, 2 = a lot. Additional de-tails are available in Strayer (1993).

It is doubtful that any set of stimuli canadequately sample the range of meaningfulor evocative emotional events across age.Our objective was to provide a broader rangeof emotional stimuli (within reasonable ethi-cal restraints) than had previously been usedin such research with children. With thislimitation in mind, the stimulus materials

appear to have been effective elicitors ofemotion for most children: 80% ofthe sam-ple reported emotions for five or more ofthesix episodes. In contrast, nearly half (45%) ofall "neutra!" responses were given by fewerthan 10% of the sample. Thus neutra! re-sponses appear to be a function of child vari-ables rather than primarily an artifact of low-intensity stimulus materials.

Facial expressions.—Using a ceiling-mounted camera, children were unobtru-sively videotaped while they watched thevignettes. A 3-min baseline tajw for eachsubject was viewed initially to familiarizecoders with idiosyncratic facial characteris-tics. Facial expressions were scored by cod-ers trained to recognize expressive changesin upper, middle, and lower facial regions(Izard & Dougherty, 1982).

Two indices were derived from facialscores: emotion valence and intensity, andcategory of predominant emotion (happy,etc.). Valence and intensity were recordedat 10-sec intervals using a scale that rangedfrom +3 (positive valence, high intensity)to - 3 (negative valence, high intensity; seeTable 2 for details). Because our focus is onglobal expressions communicated in realtime, 10-sec durations were selected in pilottesting as a sufficient duration to assess emo-tion communicated in facial expressions.Ninety-one intervals were rated during thefive vignettes with predominantly negativeaffect and 25 intervals during the vignettethat featured positive affect ("Circus").

Exact interrater agreement based on twojudges scoring 25% of the videotapes foreach age group was 93% (84% for the youn-gest group, 99% for the others), kappa = .92.All differences were resolved by discussion.

Coders also judged at the end of eachvignette the child's predominant facial ex-pression during it. The same eight catego-ries were used as for children's reports (neu-tral, happy, etc.).^ Details of coding criteria

' If children's spontaneous attributions of eniotion were unclear (e.g., "feels bad"), theywere asked which of the following terms best applied; happy, sad, angry, afraid, siupvised,disgusted (yucky). We queried "surprised," which was described as most like "ainiid" in allvignettes except "Circus," for which it was described as most like "happy."

^ "Concerned" is not a facial code in Izard's system; rather, it represents a procedural com-promise on our part. Judges were trained using MAX pictures (Izard & Dougherty, 1982), whichportray full-blown extreme expressions of each emotion—extremes which did not occur in oursampie. For example, "startle" in MAX might result from the child's loss of proximal physicalsupport, whereas in the current context it refers to the much milder emotion of "negative sxu-prise." Similarly, "concerned" represents our raters' judgment that "startle" and "fear" were tooextreme for what they saw. It therefore reflects a combination of mo<terate apprehension, nega-tive surprise, and agitated interest on the part ofthe children so coded. In this sense, we believethat it parallels the use of "concerned" by other researchers.


TABLE 1

DESCMPTION OF VIDEOTAPE STIMULUS VIGNETTES

1. Old House: Three children sneak into a yard at night A boy climbs upcreaking stairs to peer through a window into the house. A looming shadowof a man appears above him, and the children run away. (Source: commer-cial film.)PROTAGONIST 1 malePREDOMINANT AFFECT: fear

2. Spilled Milk: A husband and wife have an angry exchange while their daugh-ter watches TV. The man leaves and the woman shouts at the girl to cometo dinner; tiie girl accidentally knocks over a glass of milk and the motherslaps her. (Source: "I2V2 Cents," National Film Board of Canada.)PROTAGONIST; femalePREDOMINANT AFFECT: sadness

3. Jeannie: A young woman talks directly to the viewer about the difficult life sheand her children had with her abusive husband. (Source: "Loved, Hon-oured, and Bruised," National Film Board of Canada.)PROTAGONIST; femalePREDOMINANT AFFECT: sadness, anger

4. Skates: A girl and boy argue over taking turns on her new skates. The boycalls her names and threatens to tattle. She pushes him down; he runs cry-ing to her parents. The boy lies; the father believes his story and gives thegirls' skates to the boy as her punishment (Source; "Our Vines HaveTender Grapes"; obtained from Dorothy Flapan, who used them in a 1968study.)PROTAGONIST: femalePREDOMINANT AFFECT; sadness

5. Canes: A girl talks pleasantly about her life and the fun she has despite herphysical disability. (Source; "I'll Find A Way," National Film Board of Can-ada.)PROTAGONIST: femalePREDOMINANT AFFECT; sadness

6. Circus: A father and daughter go to see the circus train. The elephant is let outto perform some tricks. The girl jumps and laughs excitedly, and is liftedup on the elephant's trunk. (Source: "Our Vines Have Tender Grapes.")PROTAGONIST: femalePREDOMINANT AFFECT: happy

NOTE,—Films are in black and white. Total viewing time is approximately 30 min. Furtherinformation is given in Strayer (1993).

are given in Table 3. Percent agreement ex- Means were 1.2 and 1.4, respectively, forceeded 80% for all categories of facially negative and positive intensity; SDs = .5coded emotion. and .7.

Congruent with verbal reports of emo- Similarly, facial negative intensity is thetion, coders found that most children were sum of negative facial ratings (Table 2)expressive: 60% of the sample had a facial across the five vignettes with predominantlyemotion coded for five or more of the six negative affect. Facial positive intensity isvignettes. In contrast, more than a third the sum of positive ratings for the vignette(37%) of all "neutral" codes were given to ("Circus") with predominantly positive af-the least expressive 10% ofthe sample. feet. Absolute values were used for facial

Emotional expressiveness.-Tvô sum- "^S^^^^^ intensity, so that for bodi variables,mary emotional intensity scores were de- higher values mdicate greater mtensity. Forrived from the self-report and facial ratings facial negative intensity, mean = 88.4, SDdescribed above. Self-reported negative in- = 37-6; for facial positive intensity, mean =tensity is the mean rating across the five vi- ' 'gnettes with predominantly negative affect Emotional expressiveness in more gen-(see Table 1). Self-reported positive inten- eral contexts was assessed by teacher andsity is the rating for the single vignette with parent ratings and children's self reports,predominantly positive affect ("Circus"). Teachers rated 61 children on a 47-item

TABLE 2

CODES FOB INTENSITY OF FACIAL EXPRESSIONS

+ 3 Maximum euphoric/pleasant expression: broad, open-mouth smile; eyes wide with outside comers deeplycreased or narrowed by raised cheeks; even brow witheyebrows raised and horizontal (no furrow in nasalridge).

+ 2 Moderate euphoric/pleasant expression: broad, moderateclosed-mouth smile; eyes wide and relaxed, outside cor-ners creased.

+ 1 Minimal euphoric/pleasant expression: narrow closed-mouth smile; comers of mouth minimally raised,

0 Neutral expression: typical of baseline observations whenno notable facial changes occurred.

- 1 Minimal dysphoric/unpleasant expression: lips com-pressed; lip biting; minimally knit brow.

- 2 Moderate dysphoric/unpleasant expression: moderatelyknit brow; ridges in upper nasal area; eyes narrowed.

- 3 Maximum dysphoric/unpleasant expression: knit browwith verticaJ ridges or flesh folds between brows, horizon-tal ridges in upper nasal aiea; eyes nanowed; lips com-pressed; mouth comers strained or pulled down; chinbulge; lip biting.

TABLE 3

DESCBIPTION OF FACIAL EMOTION CODES

Emotion Description

Happy Forehead smooth; cheeks raised; corners of mouth back andup.

Sad Inner brows drawn together; vertical furrows or bulge be-tween brows; raised inner comers of upper eyelid; eyessquinted; downward mouth corners; lower lip pushed up-ward by chin muscle (chin puckers).

Anger Brows sharply down and together; vertical furrows or bulgebetween brows; nasal root broadened or bulged; eyes nar-rowed by lowering of brow; rectangular mouth; lips may bepressed together tightly.

Fear Straight brows slightly raised and drawn together; horizontallines or bulge on forehead; nasal root narrowed; eyes nar-rowed or squinted; eyelids raised (white shows more thanusual); mouth corners tight or retracted.

Surprise Brows raised; skin below brow stretched; horizontal furrowsacross forehead; enlarged, roundish eyes; mouth opened(corners drawn back and up when positive; otherwise justrounded moutii).

Disgust Brows together and downward; vertical furrow or bulge be-tween brows; nasaJ root bulged (nostrils enlarged); nasalbridge furrows or bulges; eyes narrowed; cheeks raised;mouth tense (tongue may be forward); lower lip forward.

Concern Brows drawn together and possible slightly downward; verti-cal furrows or bulge between brows; eyes narrowed orsquinted; cheeks may be raised {but no smile); mouth re-laxed; posture may lean forward.

Neutral Expression as at baseline; no emotion (above) expressed.

NOTE.—Descriptions are based on Izard, Dougherty, and Hembree (1983). For "Concern,'see text and note 2.

Roberts and Slrayer 457

Child Rating Questionnaire, which assessedprosocial behaviors, empathy, and peer rela-tionships as well as emotional expressive-ness. Items were taken from the FrosocialBehavior Questionnaire (Weir, Stevenson, &Graham, 1980) and the Affect ExpressionQuestionnaire (Buck, 1977). Teachers ratedeach item on a five-point scale (1 = "not atall characteristic of the child" to 5 = "ex-tremely characteristic"). A score for emo-tional expressiveness (Cronbach's alpha =.78) was derived by averaging four items("Expresses anger or hostilities directly";"Displays anger frequently and sometimesinappropriately"; "Controls his or her emo-tions" [loads negatively]; "Shows a widerange of different kinds and intensities ofemotions"). Scores ranged from 1.0 to 4.5;mean = 2.4, SD = .9.

Mothers also rated their children on theChild Rating Questionnaire, allowing us toconstruct parent scales identical to theteacher scales. For parent-rated child emo-tional expressiveness (Cronbach's alpha =.65), scores ranged from 1.5 to 5.0; mean =3.2, SD = .8.

Three items from the Expression of Af-fect Questionnaire for children (Strayer,1985) assessed expression of anger to familyand friends ("When 1 feel angry, it's hard toshow how I feel to my family"; "When I feelangry, it's hard to show how I feel to myfriends"; "I usually don't show my feelingsto my friends"). These items were chosen toparallel those rated by parents and teachers.Responses were scored as yes (= 1) or no(= 0), reflected so that higher scores indicategreater expressiveness, and summed for a to-tal score (Cronbach's alpha = .62). Scoresranged from 0 to 3, mean = 1.6, SD = 1.1.

Emotional insight.—Because the vi-gnettes procedure yields both facial and selfreports of emotion, we were able to generatetwo indices assessing the degree of corre-spondence between these measures. Accu-racy was the number of exact matches be-tween facially rated and self-attributedemotions across the six vignettes. Denialwas measured by the number of vignettes inwhich facially rated negative emotions (e.g.,sad, angry, afraid) occurred with self-reported emotions of "happy" or "neutral,nothing much." Scores for accuracy rangedfrom 0 to 4, mean = 1.0, SD = 1.0. Scoresfor denial ranged from 0 to 6, mean = 1.9,SD = 1.6.

Role taking.—A general measure of roletaking was provided by Selmau's perspec-

tive-taking measure (Selman & Jaquette,1977). Children's global score was based ontheir responses to a series of story dilemmas,using interview methods and criteria, ascited. Obtained scores ranged from 1 to 7,mean = 4.5, SD = 2.1.

Empathy.—Two measures of empathywere derived from the vignettes procedure.The first of these, the Empathy Continuum,integrates the degree of affective sharing ex-perienced (i.e., degree of match betweenown and stimulus person's emotion) withthe child's cognitive attribution for his or herown emotions (see Strayer, 1993, for a de-scription and scoring procedures). It con-tains seven different levels of cognitive me-diation, derived from models of empathydevelopment (Feshbach, 1975; Hoffman,1975), and levels of interpersonal under-standing (Hughes, Tingle, & Sawin, 198.1;Shantz, 1983). Scores can range between 0and 19. In the current sample, scores (aver-aged across six vignettes) ranged from 0.7 to15, mean = 5.7, SD = 4.7.

The second measure was a nonverbalmeasure of affective empathy (Feshbach &Roe, 1968), based on facial expressions.(Similar nonverbal measures have beenused by Eisenberg et al., 1988, 1989, to in-dex empathy and sympathy.) Facial-expressive empathy was the number of exactmatches between the predominant faciallyexpressed emotions of children and charac'ters. (Predominant vignette emotions, listedin Table 1, were identified in pretests by 30children 5-14 years old and 30 adults. SeeTable 1 and p. 191 in Strayer, 1993.) Facialempathy scores ranged from zero to four,mean = 1.4, SD - 1.1.

Empathy in more general contexts wasassessed by self reports and ratings made byteachers and best friends. The self-reportmeasure was Bryant's (1982) Index of Empa-thy for Children (administered in a sessionseparately from the videotaped vignettes).This questionnaire consists of 22 items forchildren and adolescents and is derivedfrom Mehrabian and Epstein's (1972) ver-sion for adults. A global score for empathyis based on children's agreement ( = 1) ordisagreement (=0) with items tapping attri-butes including emotional expressivity andattitudes (e.g., "sometimes I cry when 1watch TV"), sympathy (e.g., "it makes mesad to see a girl who can't find anyone toplay with"), and empathy (e.g., "seeing a girlcrying makes me feel like crying"). Thequestionnaire measure has satisfactory


short-term test-retest reliability as well as di-vergent and convergent validity (Bryant,1982).

Because the youngest children in thepresent study were 5 years old, and Bryant'smeasure has not previously been applied tochildren younger than grade 1, internal con-sistency was examined in order to detenninewhether the questionnaire was appropriatefor this age group. The obtained Cronbach'salpha (.48) indicated a low level of internalcoherence; thus correlational analyses re-ported below for the Bryant measure wererestricted to the two older age groups. (Anal-yses that included the youngest childrenproduced results similar to those for the old-est children alone but attenuated in magni-tude, a pattem typically found when com-paring measures of higher and lowerreliability.) For the older children (N = 40),scores ranged from 5 to 18, mean = 11.8,SD = 3.5; for all 73 children, mean = 11.6,SD = 3.2.

Teachers rated the empathy of partici-pating children on two items ("Is generallysensitive and responsive to others' emo-tions"; "empathic") from the Child RatingQuestionnaire, described earlier, For thisscale, scores ranged from 1.5 to 5, mean =3.5, SD = .9; Cronbach's alpha = .74.

Best friends rated the empathy of partic-ipating children on six items, such as "Does[child's name] feel bad if s/he sees anotherkid without a friend to play with?" Re-sponses were scored 0 = not at all, 1 = alittle, 2 = a lot. Scores ranged from 1 to 10,mean = 7.5, SD = 2.2; Cronbach's alpha =.78.

Prosocial behaviors.—Like empathyand emotional expressiveness, prosocial be-haviors were assessed across methods andsources. We begin by desciibing our threelaboratory measures.

Children's helpfulness toward an adultin need of aid was assessed by a measureadapted from Yarrow and Waxier (1976).While interviewing the child, the experi-menter looked through her sheaf of papersand appeared to be distressed because shecouldn't find something she needed. As sherearranged her papers, she upset a box oicolored paper clips resting on the arm of herchair scattering them on the floor. "Oh, no!she said, and stayed occupied with iier pa-pers Children were scored as helpmg (~ 1)if they began to assist the experimenterwithin 30 sec (64% ofthe sample). Children

who had not helped at the end of 10 sec wereprompted with "I just can't seem to get a!lthis together." Children who failed to helpwere scored as 0.

Sharing was measured by a fishing gameadapted from Yarrow and Waxier (1976) inwhich children have 10 turns to get prizesby bobbing foi rings in a pool, using a fishingpole. Children were tested individually.They were told that two usually play thisgame at once, and they were asked for thename of their best friend. Because the friendwas not present, children were told thattheir friend would be invited to play later.There are only 10 turns in total for both chil-dren, howevei, so subjects must decide ateach turn whether they want to take that turnfor themselves or leave it for their friend.Scores (= the number of turns given up forthe best friend) ranged from 0 to 9, mean =2.7, SD = 2.1.

Cooperation with an unfamiliar, same-sex peer was assessed using a task adaptedfrom Madsen and Connor (1973). A penny isencased in a block divided in the center andattached to strings at eitiier end. Each childholds a string. If both children pull simulta-neously, the block divides and the pennydrops out and goes to neither child. There-fore, in order to accumulate coins, it is nec-essary for the players to use a cooperativestrategy in which one child pulh the cointo her side of the table and the other childpemiits this to happen. Children played for10 trials, and were scored using the systemdevised by Madsen and Connor. Scoresranged from 0 to 30, mean = 8.9, SD = 12.2.

Teachers and parents rated children'shelpfulness, sharing, cooperation, and posi-tive responsiveness to peers' emotional dis-tress using the Child Rating Questionnairedescribed earlier. A scale assessing help-fulness was composed of four items (e.g.,"Offers to help people who are feeling sickor in trouble," "Offers to help other peoplewho are having difficulty with a task or activ-ity"). For teacher ratings, scores ranged from1.0 to 4.8, mean = 3.0, SD = .9; Cronbach'salpha = .84. For parent ratings, scoresranged from 1.2 to 5.0, mean = 3.3, SD =.8; Cronbach's alpha - .80.

The scale for sharing was composed ofonly two items, "Shares play, food, or othermaterials with others" and "Is generous indonating own time or contributing towsurdpurchase of gifts for others, charities, etc.").For teacher ratings, scores ranged from 1.5to 5.0, mean = 3.2, SD = .8; Cronbach's


alpha = .70. The parent version of this scalewas not used due to low internal coherence(Cronbach's alpha = .34).

The scale for cooperative behavior wascomposed of eight items, such as "Tries tobe fair in games or activities," "Is generallycooperative," "Behaves aggressively withother children" (loads negatively). Forteachers, Cronbach's alpha - .85; scoresranged from 2.1 to 5.0, mean = 4.0, SD = .7.For parents, Cronbach's alpha = .79; scoresranged from 2.1 to 4.75, mean = 3.8,SD = .6.

Best friends rated participants' prosocialbehaviors on six items: "Does tchild's name]take turns when kids are playing games? . ..when you are playing games?"; "DoesIchild] share things like food, games, andtoys, with other kids? . . . with you?"; "Does[child] help other kids when they need it,like fixing things, carrying things, withschool work or chores? Does [child] helpyou . . . ?" Responses were scored 0 - notat all, 1 = a little, 2 = a lot. Scores rangedfrom 2 to 12, mean = 8.1, SD = 2.1, Cron-bach's alpha = .80.

Results

In the first part of this section, we testthe model presented in Figure 1 by examin-ing correlations across domains, beginningwith age. We then integrate these findingsusing latent variable path analyses (Loh-moller, 1984). Because functional genderdifferences emerge in these analyses, wewill conclude the Results section by examin-ing mean gender differences in our mea-sures.

Relations across Domains: AgeIn order to control Type I error for these

analyses, binomial tests were used to assessthe omnibus null hypothesis that all correla-tions were zero for a given set of compari-sons. Thus the omnibus p is the exact proba-bility under the null hypothesis of observingn or more significant tests in a set of m com-parisons.

Age and emotional expression.—As ex-pected, emotional expressiveness showedrelations with age that depended on type ofemotion (positive or negative) and methodof assessment. Overall, three of seven com-parisons were significant at .05; binomialtests rejected the omnibus null hypothesisthat all correlations were zero, p < .005.

As expected, Intensity of facial expres-sions of negative emotions when viewing

stimulus videotapes increased with age,r(71) = .27, p < .05, whereas intensity ofpositive expressions did not. (For intensityof facial positive expression, r(71) = - .02;for self-reported positive intensity, r(71) ==-.14.)

In contrast, older children rated them-selves, and were rated by others., as less ex-pressive of negative emotions, refiecting in-creased pressure with age for emotionalcontrol. Self-reported negative intensity dur-ing the stimulus videos declined with age,r(71) = - .25 , p < .05, as did teacher-ratedexpression of anger in school contexts, r(59)= - .22, p < .05. Consistent with this, par-ent-rated expression of anger declined mar-ginally with age, r(71) = - .17, p < .10.

Age and emotional insight.—In contrastto emotional expressiveness, emotional in-sight was unrelated to age. Contrary to ex-pectation, denial of negative emotions dui-ing the stimulus videos did not decreasewith age, r{71) = ~ .03, nor did congruencebetween facial expressions and verbally re-ported emotions increase, r(71) = .06.

Age and role taking.—As expected,role-taking ability, as assessed by Selman'sprocedure, increased strongly with age, r(71)= .82, p < .0001. Thus, over this age range,cognitive changes were more striking thanchanges in emotional expressiveness andemotional insight.

Relations across Domains: EmpathyEmotional expressiveness and empa-

thy.—As expected, expressiveness was fre-quently related to empathy. As shown in Ta-ble 4, 10 of 35 comparisons were significant;the omnibus null hypothesis was rejected,p < .0001. The contrasting pattern that weexpected was evident: laboratory measuresof expressiveness were positively related toempathy, whereas the expression of angershowed negative associations. These lattercorrelations only reached significance forparent ratings of children's anger, sug-gesting that the expression of anger In familycontexts may be particularly important forchildren's empathy with peers—a findingconsistent with attachment theory (e.g..,Bretherton, 1990).

Emotional insight and empathy.—Asindicated in Figure 1, we expected emo-tional insight to be related to empathy androle taking. As shown in Table 5, this expec-tation was supported for empathy (omnibusp < .005) but not for role taking. Accuracywas most strongly associated with higher fa-

460 Child DevelopmentTABLE 4

CORBELATIONS BETWEEN EMOTIONAL EXPRESSIVENESS AND EMPATHY

EMOTIONAL EXPRESSIVENESS

Facial Interview Ratings of Anger By;

EMPATHY Neg Pos Po.s Neg Parent Teacher Self

EmCon 28** -19^ .26* .30** -.38*** - .05 .13Facial 36*** .39*** - .06 .03 .16^ - .17 - .01Teacher 07 .03 .ir .09 -.26* -.14 -.11Self 01 .24' .29* .40** - .23 .08 .14Friend 05 - .09 .07 - . 2 1 ' - . 3 1 * - ,20 - ,09

NOTE.—Emcon: mean Empathy Continuum score. Facial Neg: Observer-rated intensity of negative valence facialemotions. Faeial Pos; Observer-rated intensity of positive valence facial emotions. Intview Pos: Vignette interview,self-reported positive emotional intensity. Intview Neg; Vignette intervievv, self-reported negative emotional in-tensity.

^For boys, r ^ ,54**; for girls, r = -.20; the correlations differ significantly, z = 2,99**.

* p < .05.**p < .01,' **p< .001.Omnibus null hypothesis; p < .0001.

cial empathy scores and denial with lowerEmpathy Continuum scores. As well, denialshowed marginally significant relations withempathy across methods. These correla-tions, considered as a set, suggest that ourmeasures of emotional insight are tappingsomething general, not just context specific.

Role taking and empathy.—As ex-pected, greater role taking was associatedwith greater empathy. As shown in Table 6,two of five comparisons were significant,omnibus p < .001. Not surprisingly, the rela-

TABLE 5

CORRELATIONS BETWEEN EMOTIONAL INSIGHT,ROLE TAKING, AND EMPATHY

Ac-curacy Denial

Role taking;Selman 16^

Empathy;EmCon 19-Facial 49***Teacher - .09Self 18=

. Friend 12

.00

- ..38***.10

-.17^- . 2 3 +- .14

tion was strongest for Empathy Continuumscores, which integrate cognitive and af-fective factors. Nevertheless, it also held forempathy measures in which role-takingcomponents are not obvious {i.e., best friendratings and, for girls, self reports).

Empathy and prosocial behaviors.—Em-pathy was positively associated with proso-cial behaviors. Unexpectedly, the evidencewas much clearer for boys than girls, Asshown in Table 7, six of 55 comparisons forgirls and 15 of 55 comparisons for boys weresignificant. Thus, although the omnibus nullhypothesis was rejected for both (indicatingan overall relation between these variable

TABLE 6

CORBELATIONS BETWEEN ROLE TAKING ANUEMPATHY

Empathy Role-Taking Scores

EmConFacial ...TeacherSelfFriend ..

,54*-- .04- .17

NOTE,—EmCon; mean Empathy Continuum score.Self; self-reported empathy (Bryant questionnaire); N =40.

"For girls, r - .38**; boys, r -- - ,35; the correla-tions differ significantly, z = 2.23*.

+ p < .10.* ))< ,05.**p< .01.*** p < .001.Omnibus null hypothesis; p < .005.

NOTE,—EmCon; mean Empathy Continuum score.Self; self-reported empathy (Bryant questionnaire); N ~40.

^ For girls, r = .58**; boys, r = - .11 ; the correla-tions differ significantly, i = 2.25*.

• p < .10,* p < ,05.

***p < .001.Omnibus null hypothesis; p < .001.


TABLE 7

CORRELATIONS BETWEEN EMPATHY AND PROSOCIAL BEHAVIOBS BY SEX OF CHILD

PROSOCIAL

Laboratory:Helps - .11Shares 23+Coop - .11

Teacher:Helpful - .27Coop - .20Shares - .17Comforts - .13

Parent:Helpful -.14Coop 25*Comforts 04

Friend:Prosocial .24"''

Laboratory;Helps 16Shares 27+Coop .14

Teacher:Helpful 33*Coop 21Shares 19Comforts 33*

Parent;Helpful 25+Coop 38*Comforts 31*

Friend:Prosocial 42*

EMPATHY

EmCon Facial Teacher Self Friend

.03

.12

.02

.09

.02

.23

.04

.13

.15

.05

.05

.26

Girls

.06

.17

.44**

,87***,61***.69***.85***

.11

.11

.14

,20

Boys

- .08-,15

.05

.04-,07-.26

.09

,24- .05

.02

- ,34

.34^ - . 1 7

NOTE.—EmCon; mean Empathy Continuum score.^ p < .10.• p < ,05.** p < .01,•**p< .001,Omnibus null hypothesis; for girls, p < ,0001; for boys, p < .00001,

.05,08.06

,04,11.10.14

,12.26'.17

,52*

060130*

1341*1810

031329 +

.35*

.11

.23

Y3***

,41*.59***.69***

.27+

.20- , 0 1

- . 2 5- . 0 8

.19

.04- . 3 5- . 2 5- . 1 2

.34 +- . 0 1

.33 +

.33*

.12- . 0 2

.26

.46*

.21

.31 +

- . 1 4.00

- . 0 8

.62***

sets for both genders), significant correla-tions occurred more frequently for boys thangirls, x'(l) = 4.77, p < .05.

For both genders, all within-source cor-relations were significant (i.e., the correla-tions between teaeher-rated empathy andteacher-rated prosocial behaviors, and best-friend rated empathy and prosocial behav-ior), suggesting that teachers and friends seeempathy as part of a coherent constellationof prosocial qualities.

Correlations across sources constitutemuch clearer evidence for the causal role ofempathy in prosocial behaviors. Of the 50such correlations in Table 7, 10 were sig-

nificant for boys (omnibus p < .001), butonly one for girls (omnibus p > .35). If weadopted this more rigorous across-source cri-terion, then, evidence would indicate thatboys' empathy (but not girls') was related totheir prosocial behavior. Thus, consistentwith reports that girls are more empathicthan boys but not more prosocial, this corre-lational evidence indicates that empathy isan important determinant of boys' prosocialbehavior, but has less effect on the prosocialbehavior of girls.

Relations across Domains: A LatentVariables Path Analysis

In order to clarify this rather complexset of findings, data were subjected to a la-

462 Child Developmenttent variable path analysis (LohmoUer,1984). According to Falk and Miller (1992),this method is particulary appropriate whenrelations between theoretical constructs can-not be specified exactly, when empiricalmeasures have some degree of unreliability,when there are many manifest and latentvariables, and when sample sizes are moder-ate or small. Thus this type of analysis isappropriate under conditions found in thepresent study (and found as well as in psy-chological research in general). In addition,by creating composite variables based onshared variance, latent variables path analy-sis offers a view of emotional expressive-ness, empathy, and prosocial behavior thatcuts across sources and methods. This is es-pecially useful in the present case, in whichwithin-source correlations between empa-thy and prosocial behaviors are high.

Because teacher scales of prosocial be-havior were strongly correlated with one an-other (median r - .60, p < .0001; see Strayer& Roberts, 1994), they were averaged to-gether prior to the latent variable analysis toform a single teacher measure of prosocialbehavior. The same was done for parentscales of prosocial behavior (median r = ,58,p < .0001).

When domains did not reduce to a sin-gle latent variable (as in the case of emo-tional expressiveness), they were divided onthe basis of independent principal compo-nent analyses. Analyses within the pathmodeling program confirmed that variableshad been grouped appropriately.

Three analyses were done, one for theentire sample (N = 73), and one for eachgender. The entire sample was used for themain analysis because we neither expectednor found functional differences betweengenders in the relations between age, roletaking, emotional expressiveness, and empa-thy. In contrast, our correlational analysesindicated important gender differences inthe relation between empathy and prosocialbehaviors, and so for this portion of tiiemodel, path analyses were carried out sepa-rately for girls and boys. To ensure compara-bility, all components ofthe model (Fig. 1)were included in all analyses. For the rela-tions between age, role taking, emotional ex-pressiveness, and empathy, the analyses bygender can be considered replicating sub-samples with Ns of 38 and 35; they thereforeprovide us with infonnation about the stabil-ity of the path coefficients generated in theanalysis ofthe full sample.

In order to control Type I error, ex-pected paths were deleted if they failed toaccount for at least 10% ofthe variance. Thisvalue corresponds to an alpha level of .01with a sample size of 73. All paths found inthe main analysis were included in the sub-sample analyses for purposes of comparabil-ity and interpretation. Because of the con-ceptual overlap between these analyses, itwill be convenient to describe them jointly.

Age-related trends.—As reported ear-lier, age strongly predicted role taking, ac-counting for 68% of the variance in Selmanscores. The path coefficient (.82) was stablein the two subsamples, differing by onlyplus or minus .02 (see Figs. 3 and 4), indicat-ing an equivalent relation for girls and boys.In contrast, and contrary to e)q)ectation, agewas only weakly correlated wiA latent emo-tional expressiveness (r = -.04), latent ex-pressed anger (r = - .19), and emotional in-sight (r = .04).

Age was correlated with latent empathyand latent prosocial behavior (rs = .56 and.32, respectively), indicating that older chil-dren were both more empathic and moreprosocial, consistent with reports for othersamples. However, in keeping with thepaths diagrammed in Figure 1, analysis ofresiduals indicated that age was not directlylinked to either factor. In the full-samplemodel, including age as a direct predictor ofempathy and prosocial behavior only ac-counted for an additional 5% ofthe variancein latent empathy (increasing R^ from .60 to.65) and failed to account for any additionalvariance in prosocial behavior (fl decreasedfrom .26 to .25). Similar pattems were seenin analyses done separately by gender. Thusage-related increases in empathy appear tobe largely due to increasing ability to under-stand the plight of others, with greater empa-thy leading, as expected, to greater prosocialbehavior.

Predictors of latent empflt/n/.—As ex-pected, role taking was strongly linked to la-tent empathy (primarily defined by EmpathyContinuum scores and best friend ratings),accounting for approximately 23% of its vari-ance, F(l, 71) = 21.21, p < .0001. This rela-tion was moderately stable across sex; sub-sample estimates ofthe path coefficient (.49)differed by plus or minus .07, and R^ variedfrom .12 (boys) to .35 (girls).

Confirming expectations, emotional in-s i ^ directly contributed to latent empathy,accounting for 10% ofthe variance, F(l, 71)= 7.89, p < .01. The path coefficient (.26)


was stable across sex; K^ varied from .08(girls) to .15 (boys), indicating that absenceof denial and recognition of one's own emo-tional experience contribute to greater em-pathy across gender.

Contrary to expectation, the indirectlink from emotional insight to empathy viarole taking was not found. Emotional insightwas only weakly linked to role taking, r =.03, suggesting that this measure of role tak-ing (Selman & Jaquette, 1977) is tapping aprimarily cognitive mode of understandingothers.

As expected, children who were moreemotionally expressive on our laboratorymeasures were also more empathic. Latentemotional expressiveness was positively re-lated to latent empathy, accounting for 16%of its variance, F(l, 71) = 13.52, p < .001.Stability was not high across sex; in the sub-sample estimates, the path coefficient (.40)varied by plus or minus .12, and R s from.06 (girls) to .25 (boys), suggesting that thisrelation may be somewhat more importantfor boys.

Also as expected, expression of angerwas negatively related to latent empathy, ac-counting for 12% of its variance, F(l, 71) =9.68, p < .01. The path coefficient (-.30)

was stable across sex; R^ varied from .12(boys) to .19 (girls), indicating that for bothgirls and boys, reports of anger, especiallyby parents, were associated with lower lev-els of empathy across methods and sources.

In all, 60% ofthe variance in latent em-pathy was accounted for by role-takingscores, latent expressiveness, expression ofanger, and emotional insight, F(4, 68) =25.50, p < .00001. R s for girls (.68) and boys(.64) were similar. Thus, as expected, bothcognitive and emotional factors made impor-tant contributions to empathy, consistentwith the dual cognitive-affective nature ofempathy.

Empathy as a predictor of prosocial be-havior.—In the main analysis, latent empa-thy was an important predictor of latent pro-social behaviors (defined by best friend,teacher, laboratory, and parent measures, asshown in Fig. 2), accounting for 26% of itsvariance, F(l, 71) = 24.95, p < .00001. How-ever, this combined analysis masked impor-tant gender differences.

As shown in Figure 3, the analysis forboys was similar to the main analysis buteven stronger, as indicated by path coeffi-cients and R^s. For boys, empathy accountedfor 55% of the variance, F(l, 33) = 40.33,

1.00 age

1.00 Role Taking

(Express iveness)

.71 Interview Neg

.66 Interview Pos

.57 Facial Neg

.33 Facial Pos

(Expraasion of anger)

.90 Parent ^ ^

.48 Teacher ""'^

.34 Self report

(Insight) / ^

.98 Mo denial^/

.57 Accuracy

.40

^ ^-.30

.26

(Enâthy)

^ .86 BmCon

.72 Friend

.30 Self

.30 Teacher

.16 Facial

.. ^

//

.51

(Prosocial)

.70 Friend

.63 Teacher

.36 Parent

.52 Shares

.42 Coop

.24 Helps

FIG, 2.—Empathy and prosocial behavior: an empirical latent variable model (N = 73). Notes:Coefficients for measured variables are their factor loadings; coefficients for arrows between latentvariables are standardized path coefficients or beta weights. For the overall model, RMS Cov(e,u) =.10, indicating a moderate fit with the data. The latent variables expressiveness and expression of angercorrelated at .09. EmCon = Empathy Continuum.

464 Child Development1.00 ase

1.00 Bole Taking

{Expressiveness)

.66 Interview Nag

.ao Interview Poa

.45 facial Nag

.27 Facial Pos

(Expression of anger)

.64 P&rent ^ ^

.64 Teacher ^

,67 Self report

(insight) /

.93 No denial/

.19 Accuracy

.43

\

.52

-.33 ,

/

/

.34

(Qm

• ^ .81

.64

.01

.70

^ .07

/

:iathy)

SmCon

Priend

Self

Teacher

Facial

.74

{prosocial)

.62 Friand

.79 Teacher

.25 Parent

.53 Shares

.39 Coop

.54 Helps

FIG, 3.—Empathy and prosocial behavior: an empirical latent variable model foi boys (N = 35).Notes: Coefficients for measured variables are their factor loadings; coefficients for arrows betweenlatent variables are standardized path coefBcients or beta weights. For the overall model, RMS Cov(e,u)= ,13, indicating a moderate fit with the data. The latent variables expressiveness and expression ofanger correlated at .06. EmCon = Empathy Continuum.

p < .00001, in a latent variable defined byteacher, best friend, and laboratory mea-sures of prosocial behavior. Thus for boys,empathy accounted for a substantial propor-tion ofthe variance in a wide range of proso-cial behaviors.

In contrast, the analysis for girls (Fig. 4)indicated a more modest relation betweenlatent empathy and prosocial behaviors withbest friends, R^ = .13, F(l, 36) = 5.38, p <.05. Girls' empathy was unrelated to a sec-ond latent prosocial variable reflecting coop-erative behavior with peers (r = .15), Thusgirls' empathy, in contrast to boys', was re-lated more weakly and to a more restrictedrange of prosocial behaviors.

Given this pattern of gender differ-ences, we re-examined data previously re-ported in Strayer and Roberts (1989), doinga latent variable path analysis for each gen-der. (In the original report, data were aggre-gated across gender.) A pattern similar tothat just described emerged in this smallersample of 7-year-olds. For boys (N - 25),latent empathy (defined by high loadings onthe Bryant Inventory and parent ratings) ac-counted for 21% of the variance in parent-rated prosocial behavior, F(l, 23) = 6.11,p < .025, whereas for girls (N ^ 19), latent

empathy accounted for only 2% of the vari-ance in prosocial behavior, F(l, 17) - .35,p > .55. Thus in this independent samplealso, the relation between empathy and pro-social behavior was stronger for boys thanfor girls.

Mean Gender DifferencesGiven the functional gender differences

that emerged in the correlational and latentvariables analyses, we next report mean dif-ferenc-es across gender. As will be seen,mean differences in this sample are similarto those reported in other samples, sug-gesting that the functional differences de-scribed here may exist elsewhere.

Gender and emotional expressive-ness.—Gender differences in expressive-ness were only marginally significant, multi-variate F(7, 61) = 2.04, p < .07, and wereconfined to laboratory measures. Gonsistentwith gender role display rules (Brody, 1985;Lewis & Michalson, 1983), girls describedtheir negative emotions while watching thestimulus videos as more intense: meanswere 1.3 and 1.1 for girls and boys, respec-tively, F(l, 67) - 4.92, p < .05. In contrast,ratings of emotional expressiveness by par-ents, teachers, and children failed to showgender differences, suggesting that gender


1.00 age

.56

.28

(Empathy)

.88 BnCon

.77 Friend

.43 Self

.08 Facial

(Prosocial:

.92

.36 .60

.18

Frif"

Friend

Share

parent

(Pirosocial: Peers)

.88 Coop U.

.86 Teacher

.24 Heiv:.

1.00 Role Taking.

(Bxpreasiveneas)

.63 interview Neg

.57 interview POS

.56 Facial Neg

-.26 Facial Pos

(Expression o£ anger)

1.00 Parent

.18 Teacher

- .06 self report

(Insight)

.90 No denial

.66 Accuracy

FIG, 4.—Empathy and prosocial behavior: an empirical latent variable model for girls, N = 38.Notes: Coefficients for measured variables are their factor loadings; coefficients for arrows betweenlatent variables are standardized path coefBcients or beta weights. For the overall model, RMS Cov(e,ii)= .10, indicating a moderate fit with the data. The latent variables expressiveness and expression ofanger correlated at — ,29. Teacher-rated empathy deleted from latent empathy because of a substantialnegative loading (-.55). EmCon = Empathy Continunm.

may form part of the implicit frame of refer-ence used when ratiug boys' or girls' expres-siveness.

Gender and insight.—There were con-sistent gender differences in emotional in-sight, multivariate F(2, 66) - 4.95, p < .02.Consistent with a socialization model ofemotion, girls displayed higher levels of ac-curacy, univariate F(l, 67) = 7.49, p < .01(means = 1.3 and .7 for girls and boys, re-spectively), and lower levels of denial, uni-variate F(l, 67) = 5.56, p < .025 (means =1.5 and 2.3). Present data are thus consistentwith views that geuder socialization pres-sures may serve to distance boys from theiremotional experiences (Brody, 1985; Lewis& Michalson, 1983).

In contrast to understanding their emo-tional experiences, girls and boys had equiv-alent scores on their cognitive understand-ing of others' experience. Means for roletaking were 4.4 and 4.5, respectively, f(71)= .13, p > .85.

Gender and empathy.—As in other sam-ples (Lennon & Eisenberg, 1987), girls oftenhad higher empathy scores than boys, multi-variate F(6, 62) - 3.57, p < .005. Girls hadhigher scores than boys on the Empathy

Continuum, univariate F(l, 67) = 4.91, p <.05 (means = 6.4 and 5.1, respectively), andthey described themselves as more em-pathic on the Bryant Empathy Question-naire, univariate F(l, 36) = 8.29, p < .01(means = 13.2 and 10.2, respectively). Cirlswere also more often facially empathic thanboys, univariate F(l, 67) = 6.57, p < .025(means = 1.7 and 1.1, respectively), a proce-dure that often minimizes gender differ-ences.

Gender and prosocial behaviors.—It isnotable (given gender differences in empa-thy) that there were no gender differencesin prosocial behaviors, multivariate F( l l , 57)- .41, p > .90.

In summary, although parents andteachers did not describe girls as more ex-pressive than boys, girls reported more in-tense negative emotions during our labora-tory procedures, and they showedsignificantly higher levels of emotional in-sight and empathy. (Based on their latentvariables path model, higher levels of empa-thy for girls could be attributed in part totheir higher scores on emotional insight) Incontrast, girls and boys showed equivalentlevels of prosocial behavior. As noted ear-lier, this pattern of mean differences (girls


higher on empathy but not higher on proso-cial behavior) has been found in other sam-ples (Lennon & Eisenberg, 1987; Radke-Yarrow et al., 1983), and the absence of adose-response relation implies that empathyis not a strong determinant of girls' prosocialbehavior. Thus, the analysis of mean genderdifferences is consistent with the conclu-sions drawn from the path analyses and sug-gests that similar functional differences maybe present in other samples.

DiscussionThese results make several important

contributions to the literature on emotionalexpressiveness, empathy, and prosocial be-havior. Our findings present clear evidencein support ofthe hypothesis that empathy isan important contributor to prosocial behav-ior. They also provide evidence of a func-tional gender difference in the relation ofempathy to prosocial behavior, indicatingthat this relation is much stronger for boysthan for girls. Just as importantly, they sup-port the contention that there are importantlinks between emotional factors and empa-thy, drawing attention to the distinct posi-tive contributions made by emotional in-sight, expressiveness when witnessingothers in emotionally charged situations,and the negative effects of expressed angerin natural contexts in which the child is in-volved as a participant.

Empathy and Emotional ExpressivenessAs expected, given the dual cognitive-

emotional nature of empathy, the presentstudy found that empathy was predicted byboth a cognitive measure (role taking) audby emotional factors. It thus establisheslinks between emotional expressiveness andempathy, suggesting that children who ex-perience emotions of sadness, happiness,and fear more keenly, who recognize and ac-cept these feeiings, but who also manage tomoderate their own anger in social interac-tions, tend to be more empathic.

Factor loadings for laboratory measuresof expressiveness indicate that greater em-pathy is assoeiated with greater ability to ex*perience and express positive emotions asweil as negative emotions. Empathy as a mo-tivator of prosocial behavior may indeed re-quire both abilities. The ability to share pos-itive affect facilitates positive, cooperativesocial interactions (Sroufe, Schork, Motti,Lawroski, & LaFreniere, 1984), and the abil-ity to respond with concern motivates shar-ing, helping, and comforting. Thus currentresults support the contention that positive

and negative affect are both important com-ponents of empathy.

Our results also indicate the need to dis-tinguish types of negative affect. As shownin Figure 2, expressiveness during our labo-ratory procedures was unrelated to ratingsof expression of anger in real-life settings:measures fell into two distinct, essentiallyorthogonal (r = - .09) factors that had differ-ent functional relations with empathy. Mea-sures that assessed intensity of happiness,sadness, and fear were positively associatedwith empathy, and those that assessed angerwere negatively associated with empathy.(Whereas parent, teacher, and self report rat-ings focused on the expression of anger, ourstimulus videos [Table 1] seldom elicitedanger from children: less than 10% of all ver-bal responses and 5% of all facial expres-sions; see Strayer & Roberts, in press). Thuscurrent results support the contention thatexpression of anger should be consideredseparately from expressions of sadness andfear.

We believe that current findings mayhave implications for related areas of re-search. For example, it is plausibLe to thinkthat links found between parents' responsesto children's emotional distress and chil-dren's prosocial behaviors (Roberts, 1995;Roberts & Strayer, 1987) may be mediatedin part by the consequences such practiceshave for their empathy (Zahn-Waxler et a!..1979). The important role of emotional ex-pression in the current sample suggests thatresearch on the socialization of empathyshould include measures assessing the so-cialization of emotions as well.

Gender and Empathy-Prosocial RelationsLatent variable path analyses provided

clear evidence for the importance of empa-thy for a broad range of boys' prosocial be-haviors and for an important functional dif-ference between genders in the relation ofempathy and prosocia! behavior. We believethis is unusually clear evidence because la-tent variables for empathy and prosocial be-havior agêgated across methods andsources. In doing so, they leave behind er-ror, method, and source variance unique toeach ofthe manifest (or measured) variables,and therefore provide unusually good mea-sures ofthe underlying constructs.

For girls, the relation between empathyand prosocial behaviors was modest in sizeand restricted to prosocial behaviors withfriends, as assessed by best friend ratingsand a laboratory task involving sharing withbest friends. That empathy was less impor-


tant for girls' prosocial behavior was also ap-parent in raw correlations between empathyand prosocial behaviors (Table 7) and in theanalysis of mean differences, in which girlswere more empathic than boys (multivariatep < .01), but not more prosocial (multivariatep > .90). Finally, this same functional differ-ence was found in a reanalysis of previouslypublished data on a separate sample of 7-year-olds (Strayer & Roberts, 1989).

We do not know the reasons for func-tional variations in the relations of empathyand prosocial behavior, but differences ingender socialization are a plausible candi-date. One possibility is that social norms re-quire girls to be prosocial whether they feelempathic or not. This appears to be the casefor nurturant, care-giving behavior, an areain which girls are often described as moreprosocial than boys, with differences usuallyascribed to gender-role socialization pres-sures (Radke-Yarrow et al., 1983; Wylie,1964; Zahn-Waxler, Cole, & Barrett, 1991).To the extent that such socialization pres-sures are important, the role of empathy asa motivator of prosocial behavior would bediminished.

Boys, in contrast, appear to be underless pressure to behave prosocially (Radke-Yarrow et al., 1983; Wylie, 1964). Boys arepermitted to express more anger than girls(Brody, 1985; Lewis & Michalson, 1983) andthey do so more often (Strayer & Roberts,in press). Aggression and mock-fighting aremore frequent and more tolerated in boys(Parke & Slaby, 1983). There are sociologicalreports that in adolescence boys may evenbe expected to engage in irresponsible be-havior that is not permitted in girls (e.g., Wy-lie, 1964). Under these less constrained cir-cumstances, the role of empathy as amotivator of cooperative, helpful, responsi-ble behavior may be more apparent.

Whatever their exact nature, one wouldexpect differential socialization pressures togive rise to differential functional relationsbetween variables. Such differential pro-cesses may help to explain the sometimesinconsistent links between empathy andprosocial behaviors reported by other inves-tigators (e.g., Bamett, 1987). Our results sug-gest that amlysis by gender is important inthis area of research.

Developmental TrendsIn our path analysis, age was strongly

linked to increasing cognitive ability to un-derstand others, but not to emotional expres-siveness or to the recognition and accep-tance of one's own emotional experiences.

This may be due in part to the substantialgains in understanding emotions that occurin toddlerhood and preschool (Denham etal., 1994; Saarni, 1990). However, the weakassociations between age and our laboratorymeasures of expressiveness and emotionalinsight may also be partly artifactual. Thestimulus materials were chosen to be under-standable by even the youngest childrenstudied here (Strayer, 1993), and, in the caseof our laboratory measures, latent expres-siveness aggregated across variables forwhich we expected differing age relations.

In contrast, correlational evidence indi-cated that modest trends of the type ex-pected did occur. Older children experi-enced more intense negative emotionswhile viewing our stimulus materials (as in-dicated by their facial expressions), but de-scribed their feelings as less intense. Consis-tent with this, teacher and parent ratingsdescribed older children as less intense intheir expressions of anger. Thus our resultsare consistent with socialization models inwhich school-age children face increasingpressure to regulate or moderate their ex-pressions of negative affect, particularlyanger. It is certainly possible that bettermeasures of the expression of anger wouldprovide clearer evidence of increased regu-lation over this age range.

As expected, age was not related to pro-social behaviors independently of empathy.However, like our emotionally laden stim-uli, our laboratory measures of prosocial be-havior were chosen to be age appropriateacross our sample, that is, comprehensibleby the youngest and reasonably engaging forthe oldest In addition, it is plausible to as-sume that ratings of prosocial behaviors byparents, best friends, and teachers incorpo-rated age into their frames of reference, thatis, that children were tacitly rated in compar-ison to similar children on tasks that wereage appropriate to them. Thus current mea-sures were not designed to elicit differencesin prosocial behavior due to increasing mo-tor or cognitive skills related to particularprosocial behaviors as such. Rather, ourfindings suggest that once an ability entersa child's prosocial behavioral repertoire, em-pathy will have a bearing on whether it isperformed—especially if the child is male.

ReferencesBamett, M. (1982). Empathy and prosocial behav-

ior in children. In T. Field, A. Huston, H.Qnay, L. Troll, & G. Finley (Eds.), Review ofhuman development. New York: Wiley.


Barnett, M. (1987). Empathy and related re-sponses in children. In N. Eisenberg & J.Strayer (Eds,), Empathy and its development(pp. 146-162). New York: Cambridge Univer-sity Press.

Batson, C. D., Fnltz, J., & Schoenrade, R. A.(1987). Adults' emotional reaction to the dis-tress of others. In N. Eisenberg & J. Strayer(Eds.), Empathy and its development (pp.163-184). New York: Cambridge UniversityPress.

Blanck, P., Rosenthal, R., Snodgrass, S., DePanlo,B., & Zuckerman, M. (1981). Sex differencesin eavesdropping on nonverbal cues: Devel-opmental changes./oumoZ of Personality andSocial Psychology, 41, 391-396.

Bowlby, J. (1982). Attachment and loss: Vol. 1.Attachment. New York: Basic.

Bretherton, I. (1990). Open communication andinternal working models: Their role in the de-velopment of attachment relationships. In R.Thompson (Ed.), Nebraska symposium onmotivation 1988: Socioemotional develop-ment (pp. 57-113). Lincoln: University of Ne-braska Press.

Brody, L. (1985). Gender differences in emotionaldevelopment: A review of theories and re-search. Journo? of Personality, 53, 102-149.

Bryant, B. K. (1982). An index of empathy for chil-dren and adolescents. Child Development,53, 413-425.

Buck, R. (1977). Nonverbal communication accu-racy in preschool children: Relationshipswith personality and skin conductance, jour-nal of Personality and Social Psychology, 35,225-236.

Campos, J., Campos, R., & Barrett, K. (1989).Emergent themes in the study of emotionaldevelopment and emotion regulation. Devel-opmental Psychology, 25, 394-402.

Chisholm, K., & Strayer, J. (1995). Verhal and fa-cial measures of children's emotion and em-pathy. Journal of Experimental Child Psy-chology, 59, 299-316.

Cole, P. (1986). Children's spontaneous control offacial expression. Child Development, 57,1309-1321.

Connolly, K., & Brnner, J. (1974). Competence:Its nature and nurture. In K. Connolly & J.Bruner (Eds.), The growth of competence (pp.3-10). New York: Academic Press.

Davis, M., Hull, J., Young, R., & Warren, G.(1987). Emotional reactions to dramatic filmstimnli: The influence of cognitive and emo-tional empathy. Journal of Personality andSocial Psychology, 52, 126-133.

Denham, S., Zoller, D., & Couchoud, E. (1994).Socialization of preschoolers' emotion under-standing. Developmental Psychology, 30,928-936.

Dunn, J., & Brown, J. (1991). Relationships, talkabout feelings, and the development of affectregulation in early childhood. In j . Garher &K. Dodge (Eds.), The development of emotionregulation and dysregulation (pp. 89-108).Gamhridge: Cambridge University Press.

Eisenherg, N., Fabes, R., Bustamante, D., Mathy,R., Miller, P., & Lindholm, E. (1988). Differ-entiation of vicariously induced emotional re-actions in children. Developmental Psychol-ogy, 24, 237-246.

Eisenberg, N., Fabes, R., Miller, P., Fnltz, j . .Shell, R., Mathy, R,, & Reno, R. (1989). Rela-tion of sympathy and personal distress to pro-social behavior: A multimethod stndy. Jour-nal of Personality 6- Social Psychology, 57,55-66.

Eisenberg, N., Fabes, R., Murphy, B., Karbon, M.,Maszk, P., Smith, M., O'Boyle, G., & Suh, K.(1994). The relations of emotionality and reg-ulation to dispositional and situational empa-thy-related responding. Journal of Personal-ity and Social Psychology, 66, 776-797.

Eisenberg, N., & Miller, P. A. (1987a). The rela-tion of empathy to prosocial and related be-haviors. Psychological Bulletin, 101, 91-119.

Eisenberg, N., & Miller, P. A. (1987b). Empathy,sympathy, and altruism: Empirical and con-ceptual links. In N. Eisenberg & J. Strayer(Eds.), Empathy and its development (pp.292-316). New York: Cambridge UniversityPress.

Falk, R., & Miller, N. (1992). A primer for softmodeling. Akron, OH: University of AkronPress.

Feshbach, N. (1975). Empathy in children: Sometheoretical and empirical considerations.Counselling Psychologist, 5, 25-30.

Feshbach, N., & Roe, K. (1968). Empathy in six-and seven-year-olds. Child Development, 34,133-145.

Flapan, D. (1968). Children's understanding ofsocial interaction. New York: Teachers Col-lege Press.

Flavell, j . (1985). Cognitive development. Engle-wood Cliffs, NJ: Prentice-Hall,

Gilbert, D. (1969). The young child's awarenessof affect. Child Development, 40, 629-640.

Graves, N., & Graves, T. (1983). The cultural con-text of prosocial development: an ecologicalmodel. In D. Bridgeman (Ed.), The nature ofprosocial development (pp. 243-264). NewYork: Academic Press.

Hoffman, M, (1975). Developmental synthesis ofaffect and cognition and its implications foraltruistic motivation. Developmental Psychol-ogy, 2, 607-622.

Hoffman, M. (1982). Development of prosocialmotivation: empathy and guilt. In N. Eisen-berg (Ed.), The development of prosocial be-


havior (pp. 281-314). New York: AcademicPress.

Hughes, R., Tingle, B., & Sawin, D. (1981). Devel-opment of empathic understanding in chil-dren. Child Development, 52, 122-128.

Iannotti, J. (1978). The effect of role-taking experi-ences on role taking, empathy, altruism, andaggression. Developmental Psychology, 14,119-124.

Izard, G., & Dougherty, L. (1982). Two comple-mentary systems for measuring facial expres-sions in infants and children. In C. Izard(Ed.), Measuring emotions in infants andchildren (pp. 97-126). New York: GambridgeUniversity Press.

Izard, C, Dougherty, L., & Hembree, E. (1983).A system for identifying affect expressions byholistic judgement (Affex). Newark: Univer-sity of Delaware Press.

Krebs, D,, & Russell, G, (1981). Role-taking andaltruism: When you put yourself in the sboesof another, will they take you to their owner'said? In J. Rnshton & R. Sorrentino (Eds.), Al-truism and helping behavior (pp. 137-166).Hillsdale, NJ: Erlbaum.

Lazarus, R. (1991). Emotion and adaptation. NewYork: Oxford University Press.

Lennon, R., & Eisenberg, N. (1987). Cender andage differences in empathy and sympathy. InN. Eisenberg & ]. Strayer (Eds.), Empathyand its development (pp. 195-217). NewYork: Cambridge University Press.

Lewis, M., & Michalson, L. (1983). Children'semotions and moods: Developmental theoryand measurement. New York: Plenum.

LohmoUer, J-B. (1984). LVPLS 1.6 program man-ual: Latent variables path analysis with par-tial least-squares estimation. Berlin: Zentral-archiv.

Madsen, M., & Connor, G. (1973). Gooperativeand competitive behavior of retarded andnonretarded children at two ages. Child De-velopment, 44, 175-178.

Mehrabian, A., & Epstein, N. (1972). A measureof emotional empathy, youma/ of Personality,40, 525-543.

Miller, C,, Galanter, E., & Pribram, K. (1960).Plans and the structure of behavior. NewYork: Holt, Rinehart & Winston.

Parke, R., & Slaby, R. (1983). The development ofaggression. In E. M. Hetherington (Ed.), P. H,Mussen (Series Ed.), Handbook of child psy-chology: Vol. 4. Socialization, personality,and social development (pp. 547-642). NewYork: Wiley.

Piaget, J., & Inhelder, B. (1969). The psychologyof the child (H. Weaver, trans.). New York:Basic. (Original work published 1966).

Radke-Yarrow, M., Zahn-Waxler, G., & Chapman,M. (1983). Ghildren's prosocial dispositions

and behavior. In E. M. Hetherington (Ed,),P. H. Mussen (Series Ed.), Handbook of childpsychology: Vol. 4. Socialization, personal-ity, and social development (pp. 469-546),New York: Wiley.

Roberts, W. (1984). Family interactions and childcompetence in a preschool setting. Unpub-lished doctoral dissertation, Simon FraserUniversity, Burnaby.

Roberts, W. (1995). The socialization of emotionexpression: Relations with prosocial behav-ior and competence in five samples. Manu-script submitted for publication.

Roberts, W., & Strayer, J. (1987). Parents' re-sponses to the emotional distress of their chil-dren: Relations with children's competence.Developmental Psychology, 23, 415-422.

Saarni, C. (1979), Children's understanding of dis-play rules for expressive behavior. Develop-mental Psychology, 15, 424-429.

Saarni, C. (1989), Ghildren's understanding ofstrategic control of emotional expression insocial transactions. In G. Saarni & P, Harris(Eds.), Children's understanding of emotions(pp, 181-208), Gambridge: Cambridge Uni-versity Press.

Saarni, G. (1990). Emotional competence: Howemotions and relationships become inte-grated. In R, Thompson (Ed.), Nebraska sym-posium on motivation 1988: Socioemotionaldevelopment (pp. 115-182), Lincoln: Univer-sity of Nebraska Press,

Schachter, S., & Singer, J, (1962). Cognitive, so-cial, and physiological determinants of emo-tional state. Psychological Review, 69,379-399,

Selman, R., & Jaquette, D. (1977). The develop-ment of interpersonal awareness: A manualconstructed by the Harvard-Judge Baker So-cial Reasoning Project. Harvard University,Cambridge, MA.

Shantz, C. (1983). Social cognition. In J. H, Flavelland E. M, Markman (Eds.), P. H. Mussen (Se-ries Ed,), Handbook of child psychology: Vol.3. Cognitive development. New York: Wiley.

Sroufe, L., Schork, E., Motti, F,, Lawroski, N., &LaFreniere, P. (1984), The role of affect insocial competence. In C, Izard, J. Kagan, &R. Zajonc (Eds.), Emotions, cognition, and be-havior (pp. 289-319). Cambridge: CambridgeUniversity Press.

Strayer, F. (1981), The nature and organization ofaltruistic behavior among preschool children.In J. Rushton & R. Sorrentino (Eds.), Altruismand helping behavior (pp. 331-348), Hills-dale, NJ; Erlbaum.

Strayer, J, (1985, August), Children's affect andempathy in response to TV dramas. Paperpresented at the meeting of the AmericanPsychological Association, Los Angeles,


Strayer, J. (1987). Affective and cognitive perspec-tives on empathy. In N. Eisenberg & J.Strayer (Eds.), Empathy and its development(pp. 218-244). New York: Cambridge Univer-sity Press.

Strayer, J. (1993). Children's concordant emotionsand cognitions in response to observed emo-tions. Child Development, 64, 188-201.

Strayer, J., & Roberts, W. (1989). Children's empa-thy and role taking: Child and parental fac-tors, and relations to prosocial behavior. Jour-nal of Applied Developmental Psychology,10, 227-239.

Strayer, J., & Roberts, W. (1994, June). Empathy,expressiveness, and prosocial behavior. Pa-per presented at meetings of the CanadianPsychological Association, Penticton, BC.

Strayer, J,, & Roberts, W. (in press). Facial andverbal measures of children's emotions andempathy. International Journal of BehavioralDevelopment.

Trivers, R. (1983). The evolution of cooperation.In D. Bridgeman (Ed.), The nature of proso-cial development (pp. 43-60). New York: Aca-demic Press.

Underwood, B., fit Moore, B. (1982). Perspectivetaking and altruism. Psychological Bulletin,91, 143-173.

Weir, K., Stevenson, J., & Graham, P. (1980). Be-havioral deviance and teacher ratings of pro-social behavior. Journal of the AmericanAcademy of Child Psychiatry, IQ, 68-77.

Wylie, L. (1964). Village in the Vancluse. NewVork: Harper & Row.

Yarrow, M., & Waxier, C. (1976). Dimensions andcorrelates of prosocial behavior in young chil-dren. Child Development, 47, 118-125.

Zahn-Waxler, C , Cole, P., & Barrett, K. (1991).Guilt and empathy: Sex differences and im-plications for the development of depression.In J. Garber & K. Dodge (Eds.), The develop-ment of emotion regulation and dysregula-tion (pp. 243-272). New York: CambridgeUniversity Press.

Zahn-Waxler, C , Radke-Yarrow, M., & King, R.(1979). Child rearing and children's prosocialinitiations towards victims of distress, ChildDevelopment, 50, 319-330.

Empathy, Emotional Expressiveness, and Prosocial …faculty.tru.ca/wlroberts/prosocial.pdfEmpathy, Emotional Expressiveness, and Prosocial Behavior William Roberts Cariboo College

Documents