Does maternal emotional arousal in disciplinary contexts moderate the relation of mothers’ disciplinary practices to children’s empathy? The role of maternal arousal in development may be multifaceted and complex.

Mot hers’ Emotional Arousal as a Moderator in the Socialization of Chi1 dren’s Empathy Paul A . Miller, Nancy Eisenberg, Richard A . Fabes, Rita Shell, Suzanne Gular

Empathy (that is, the capacity for vicarious affective responding) appears to be associated with various forms of altruistic and other prosocial behaviors for both children and adults (Eisenberg and Miller, 1987). More- over, behaviors suggesting an empathic sensitivity to others appear very early, perhaps as early as twelve to eighteen months of age (Hoffman, 1984; Radke-Yarrow and Zahn-Waxler, 1984). Fundamental questions remain, however, concerning the manner in which empathic responses are generated and constructed through social interactions with others.

This work was funded by an NIMH National Research Service Award (1 F32 MH09263-01) to Paul A. Miller. Contributions by Nancy Eisenberg were supported by the National Institute of Child Health and Development (Career Development Award 1 KMHD00717) and the National Science Foundation (BNS- 8509223, BNS-8807784). Contributions by Richard A. Fabes were supported by the National Science Foundation (BNS-8807784).

New Direciions for Child Dcvrlopmcni, no. 44. San Francisco: Josscy-Bass. Summer 1989. N. Eisenberg (ed.). Empathy and Rehlcd EmotionnI Rcspomes.

65

66

Although constitutional factors seem to play some role (Hoffman, 1981; Rushton and others, 1986), the focus of this paper is on how such respon- siveness may be learned through socialization processes in the family. Specifically, the intensity of mothers’ affective responding is examined as a potential moderator of the effectiveness of other rearing practices in organizing and structuring children’s affective capacities (however acquired) to be responsive to the feelings and needs of others. Also, an initial attempt is made to describe, from a systems theory perspective, the influence of parental socialization practices on the development and expression of children’s empathy (Bell and Harper, 1977; Ford, 1987).

Bell and Harper (1977) have described the parent-child relationship as a system of mutually constructed behaviors that are coordinated and maintained by both participants. Should either member of the dyad exhibit behaviors that are below or beyond expected limits, the other member initiates behavior that both signals that expectancies have been violated and attempts to bring the other’s behavior back into the expected range. One important aspect of the construction of this system is the establishment of a set of rules or understandings by .which the parent and child can regulate their interaction regarding the expression of intentions, feelings, and behaviors (Maccoby and Martin, 1983). The nature of this regulation involves both cognitive and affective processes (Ford, 1987).

Cognitive Methods of Self-Regulation in Social Interaction

Ford (1987) has proposed that social groups (for example, parents and children in a family) require regulatory processes to coordinate function- ing of their members. The functioning of group members is organized through a system of social-moral evaluative criteria such as social con- ventions, moral reasoning, norms, and rules that facilitate and constrain their behavior. With respect to the regulation of empathic responding, parental values or expectations for responding empathically (or not) may be seen as an important part of the child’s perceptions of his or her own successful adaptation in the family group. For example, in a family that emphasizes empathic values, behaviors that represent emotional insensi- tivity to others’ needs and feelings run the risk of less adaptive function- ing in that system (they may result in the child’s being restricted from play, experiencing parental disapproval, and so forth). Expressions of emotional sensitivity to others in this system may lead to increased adap- tive functioning (for example, parental approval, increased cohesiveness with parents, and positive self-esteem or self-efficacy appraisals). A crucial issue, however, is how parental expectations for expressing empathy, as well as the actual expression of such feelings toward others, become meaningful and important to the child.

According to Maccoby and Martin (1983), parents establish the values,

67

norms, or standards for the child’s expression of feelings and behaviors toward others through social regulatory mechanisms such as socialization practices. Maccoby and Martin have proposed that because young chil- dren are relatively unskilled in interactions with others, one of the major activities that occurs in socialization is that parents supply children with interactive competencies through various rearing practices. They have further suggested that the goal of socialization is not to totally inhibit children’s behavior but to teach them to regulate the initiation, inhibi- tion, and intensity of their behaviors and affective arousal in situationally appropriate ways.

Thus, parents may inculcate their expectations regarding the regula- tion and expression of children’s feelings and behaviors toward others by providing specific information about

Their expectations for the outcomes (goals) of the interaction between the child and others (for example, “Share the candy with your friend”) The impact of the child’s behavior on others (“You made her cry because you wouldn’t let her play with you”) Particular behaviors to be performed in a given situation (“Why don’t you take turns with the crayons”) Normative rules and expectations for behavior in the particular social situation in which the child and others are engaged (“At a party everyone gets a chance to play”).

Although useful, the mere provision of information about expected or desired behaviors appears to be an insufficient basis for the develop- ment of young children’s empathic responses toward others. A brief review of the research demonstrates that investigators have not found consistent positive relations between parents’ cognitively oriented rearing practices (for example, reasoning) and emotional empathy in young chil- dren (Barnett, King, Howard, and Dino, 1980; Feshbach, 1975; Siege1 and Barclay, 1985; Zahn-Waxler, Radke-Yarrow, and King, 1979). Thus, if cognitive information alone about preferred behaviors is not consistently associated with children’s empathic responding, other socialization prac- tices must be involved. One potential mechanism may be the parents’ expression and regulation of their own affective responsiveness during socialization encounters.

Regulation of Social Interactions by Affective Processes

Ford (1987) has suggested that one function of emotional expression in social groups is to regulate interpersonal relationships and maintain the organization of the group. In the parent-child relationship, the emo- tional reactions of the parent represent a form of evaluative criteria that regulate and promote social cohesion between them. Maccoby and Martin

(1983) have suggested that, in early childhood, parental affect may con- tribute more to the meaning that the parent-child interaction has for the child than does the semantic content of parental messages. Thus, one possible reason for the inconsistent research results concerning the rela- tion between parental rearing practices and young children’s empathy is that these studies have not assessed the intensity of parents’ affective responses accompanying their other rearing practices.

From a systems approach, parental affective reactions may be effective in promoting the child’s empathic responding because they inform the child about the adaptive significance of the child’s action with respect to parental expectations. That is, the intensity of the parent’s (positive or negative) affective responses communicates the importance and value he or she places both on the child’s attending to what the parent is saying and on inhibiting or expressing appropriate feelings and behaviors when interacting with others. Moreover, the intensity of parental affective reac- tions may signal the child about potential consequences coming from the parent as a result of complying (or not) with parental expectations. For example, if the mother punitively applies verbal or physical child- rearing tactics as she reaches higher levels of emotional intensity, the child may well become fearful for his or her safety as the mother’s arousal increases. This concern for one’s own fearful state may be an early source of personal distress or self-concern in reaction to vicarious responding (Batson, 1987) and, in any event, would likely diminish the child‘s ability to process information about the needs of others (Hoffman, 1983). On the other hand, if the mother becomes highly emotional but reasons with emotional force, her increased arousal may signal the seriousness of her intent to obtain the child’s compliance with her expectations, rather than potentially painful consequences for the self. In both scenarios, the moth- er’s affective arousal regulates the child’s interpretation of the significance of her actions and messages, and guides the child’s adoption of affective and behavioral responses in the situation.

From a somewhat different perspective, Hoffman (1983, 1984) has suggested that there is a motive-arousal component accompanying par- ents’ rearing practices, which serves to arouse and motivate the child to comply with parental expectations. In power-assertive situations where the intensity of parental emotional reactions runs high, the child becomes overaroused and preoccupied with anxieties elicited by concerns about punishment, threat, and personal safety. The child is thereby unable to focus on the needs and situation of others, which is essential to eliciting the child’s vicarious affective responding (that is, empathy). With induc- tive reasoning techniques, however, parental arousal is at moderate levels, which minimizes personal safety concerns and helps the child pay atten- tion to, and cognitively process, the other-oriented components of the parents’ reasoning.

69

Unfortunately, there is little research on the interplay of parental affec- tive reactions and other rearing strategies in promoting children’s empathy. Zahn-Waxler, Radke-Yarrow, and King ( 1979) have reported data that could be considered consistent with a systems perspective. They found that mothers who used reasoning while expressing strong affect had toddlers who more frequently exhibited empathy (as measured by facial-gestural affective reactions) when responding altruistically to a distressed peer than did toddlers of mothers low in the combined use of these two techniques. Moreover, mothers’ use of reasoning without the accompanying affective arousal was not found to be related to children’s expression of altruistic behavior. Because this study focused on children’s altruistic behavior rather than empathy, however, the intensity of moth- ers’ emotional responses was not quantified; nor were mothers asked to systematically record their children’s empathic responses. In contrast, Hoffman (1963), although he did not assess mothers’ arousal directly, found that children’s consideration for others was fostered by parents’ other-oriented reasoning when such reasoning was used in rearing con- texts of low (but not high) power assertion.

Thus, it is unclear how the intensity or level of parental affect may optimize the effectiveness of parental child-rearing practices in promoting children’s empathic responding. Hoffman (1983, 1984) has suggested that the moderate level typically associated with parental reasoning is the optimal level. From a systems perspective, parental emotional reactions signal the seriousness of parental expectations for the child’s behavior, as well as impending (punitive or nonpunitive) consequences toward the child. Thus, intense parental affect could contribute positively to rela- tions between different rearing practices and children’s empathic respond- ing to others if such affect is not typically associated with punitive consequences. Finally, differences in the ages of the children may be partially responsible for these divergent effects or parental affective inten- sity; higher affective arousal may be more important for capturing the attention of toddlers than of preschool-aged children. There is clearly a need for a more refined analysis of the effect of different child-rearing strategies on children’s affective and behavioral responsiveness to others (Maccoby and Martin, 1983). In the study described in this chapter, we examined whether, and at what level, the intensity of parental affective reactions moderated the effectiveness of different rearing strategies in promoting children’s empathic responsiveness to peers in distress.

Sample and Procedures

The participants in our study were seventy-three four- to five-year-old children attending university child-care centers and their mothers (seven other subjects were lost due to equipment failure).

70

Mothers. During individual interviews in the home, mothers re- sponded to nine audiotaped vignettes (enacted by female drama students) in which a child caused a distressed or happy affective reaction in a peer. Mothers were asked to assume that the child depicted in the audiotape who caused the peers’ affective reaction was their own child. After listen- ing to each tape, mothers were asked if they would say or do anything regarding their child’s behavior. If they would do so, mothers reported how strong their emotional reaction would be to the incident using a seven-point Likert scale ranging from slight to extremely intense. They were then asked to give a first-person account of what they would say to their children (that is, “What would you say to [child‘s name]? Please talk as if she or he were standing in front of you”).

The mothers’ responses to the vignettes were transcribed and coded into six different types of child-rearing strategies. The coding scheme was established by dividing mothers’ statements into semantic units (phrases or sentences) that conveyed meaningful, yet independent, bits of information. For example, the statement “Stop that (2). You don’t need to be making fun of Kathy for losing her flower (3)” includes a negative control (code 2) and a situational definition (code 3) statement. For each incident, a mother received a score of one every time her explanation (or some segment of it) fell into one of the coding categories. Responses in each category (for example, inductive reasoning) were then summed over the six disjress incidents for a total score for that category. The coding categories and interrater reliabilities (percentage of exact agreement) are described below.

1. Inductive reasoning. Information about the impact of the child’s action on the peer or on the self as peer: “She is sad because you wouldn’t let her play with you”; “How would you feel is she did that to you?” (90 percent).

2. Negative control. Responses to the child’s behavior that included (1) withdrawal or threats thereof regarding personal-material resources: “If you can’t play together nicely, your friend will have to go home”; “Go to your room for five minutes”; (2) negative appraisals of the child’s behavior or character in the incident: “That was mean of you to do that”; and (3) verbal prohibitions of the behavior: “Stop that”; “You cannot hurt him” (87 percent).

3. Situational definitions. Descriptive information about the inci- dent including (1) psychological characteristics and motives of others: “It’s her favorite doll”; and (2) the child’s distress-inducing behavior: I‘. . . making fun of Kathy for losing her flower” (84 percent).

4. Physical control. Statements indicating the mother would spank or physically control her child: “I ought to give you a spanking for that”; “I’d take her by the arm and move her off the bike” (85 percent).

5. Prosocial suggestions. Information about behaviors the child could or should perform to resolve the problem with the peer, including nor-

71

mative rules or expectations for behavior in the incident: “Why don’t you take turns with the toy”; “This is a party, and everyone gets a chance to play” (84 percent).

6. Altruistic responding. Mothers’ enactment of behaviors or expres- sions of feelings to assist the distressed peer: “You’ll be OK” (to the peer); “Kathy, come on in here and let me see that you are not hurt” (80 percent).

ChiMren. In two sessions separated by approximately one week, chil- dren viewed two counterbalanced three-minute films in which a boy and a girl experienced a minor physical distress while playing on a play- ground (for example, both jump off a tire and appear to hurt their ankles). After each film, children nonverbally reported to a same-sexed experimenter their intensity of affective responses by pointing to one of a set of pictures of a same-sexed child expressing sad, happy, sorry, or neutral facial affect. If a sad, happy, or sorry face was chosen, the child was then asked to choose one of three pictures depicting increasing inten- sity of that affect (little bit, kind of, or very sad, happy, or sorry). A child received a score from 0 (for choosing a neutral or non-empathic affect) to 3 (accurate match at the highest level of intensity). Because the scores were positively skewed, a logarithmic transformation was used to nor- malize the distribution (Kenny, 1987).

Children’s facial affective responses were videotaped thirty seconds prior to and after the onset of the distress episodes in the films. Coding criteria for fearful, sad, and happy affects were developed based upon Ekman and Friesen’s (1975) descriptions. Criteria for scoring facial dis- tress and concerned-attention were developed and are reported in our previous research (see Eisenberg and others, 1988; Miller, 1984). The facial distress and concerned-attention indexes were developed in an attempt to identify facial responses indicative of personal distress or sympathetic responding. This was considered important because Batson (1987) has distinguished between adults’ empathic and personal distress reactions to another person in a distressful situation. According to Batson, individuals preoccupied with their own negative feelings (that is, personal distress) offer less assistance to a distressed person if given an opportunity to easily escape the situation.

For each of the five affect categories, the intensity of a child’s facial affective response was scored on a five-point rating scale every five sec- onds for the two thirty-second periods. For each film, these scores were averaged across the six intervals of each thirty-second period to obtain a baseline facial affective response score and a facial affective score in response to the distress episode. The scores were then averaged across both films to obtain one facial affective response score each for the base- line and distress episodes. Facial fear responses were dropped from the analyses because they were extremely infrequent. As with the nonverbal

72

self-report scores, children’s facial affect scores were transformed natural logarithm to improve the distribution of scores.

using a

The Moderating Role of Maternal Affect

Baron and Kenny (1986) have suggested that in tests of moderation, correlations should be low between the variable being treated as the mod- erator (mothers’ self-reported intensity of affective responses) and both the predictors (socialization practices) and dependent variables (children’s empathy measures). Intensity of mothers’ affective responses to the dis- tress vignettes was unrelated to indexes of children’s facial and nonverbal self-report of empathy and to mothers’ reported use of prosocial sugges- tions, altruistic responding, and physical control strategies. Moderate but significant relations were found between the moderator variable and mothers’ use of inductions (7(73) = .39, < .001), situational definitions (7(73) = .37, p < .001), and negative control practices (7(73) = .39, $J < .001). Although correlations between the moderator and these predictors were only moderate, their presence suggests that interpreting significant inter- actions as evidence of moderation should be done with the awareness that the effects could be influenced by separate effects from both variables on the empathy indexes.

When both the predictor and moderator variables are continuous (as they were in this study), the moderator is constructed by multiplying the presumed moderator variable by the predictor variable to create a multi- plicative interaction term (Baron and Kenny, 1986; Cohen and Cohen, 1975). For each empathy index, we constructed a series of hierarchial regression equations in which the moderator variable (mothers’ level of affective intensity) was entered first, followed by a given predictor (one of the socialization variables) and then by the interaction term (moder- ator multiplied by socialization practice). For children’s facial affective responses, baseline facial affect scores (reactions prior to the distress epi- sodes in the films) were entered into the equation before the moderator. If the interaction vector accounted for a significant amount of variance in predicting a given empathy measure (suggesting the presence of modera- tion), correlations were computed between the given socialization pre- dictor and empathy index, using either two-or three-way splits of the distribution of the moderator. For facial indexes of empathy, partial cor- relations were used to control for the effect of baseline facial responding. In addition, regression models to test for the interaction between gender and the moderator and predictor variables were computed by entering sex (and the appropriate second- and third-order interaction terms) into the model. If significant interaction effects were found for a given predictor variable (none were obtained for the moderator), regression models as described above were run separately by sex for that variable.

73

The results suggested that the intensity of mothers’ emotional reac- tions influenced the impact of child-rearing practices on children’s empathic responsiveness, somewhat more so for girls than boys. More- over, certain rearing practices appeared to have direct effects on empathic expression for both boys and girls.

Maternal Inductions and Altruistic Responding

Mothers’ reports of inductions and altruistic responding significantly, positively predicted children’s sad facial responses (partial q 7 1 ) = .33 and 2 3 , p s < .003 and .05, respectively. Moreover, mothers’ affective responses moderated the relation of their altruistic responding to children’s sad facial responses to their distressed peers (Table 1). Children’s sad facial responses became increasingly positively related to mothers’ altruistic responding as mothers’ reported intensity of emotional reactions in- creased (Table 2). Also, the impact of mothers’ inductions and altruistic responses on children’s facial distress reactions was moderated by moth- ers’ emotional reactions to the distress vignettes (Table 1). At low levels of mothers’ emotional reactions, children’s distressed facial reactions were positively associated with mothers’ inductions and (for girls only) signif- icantly, positively related to maternal altruistic responding. At the highest IeveIs of intensity of mothers’ emotional reactions, however, the relation became negative for both practices; for inductions, this relation was marginally significant (Table 2).

In addition, physical control strategies were significantly, negatively predictive of children’s facial distress reactions and positively associated with their facial sad responses (see results and discussion below). Thus, physical control, inductions, and altruistic responding were positively associated with children’s facial sad responses, which are considered to indicate an empathic or sympathetic orientation to distressed others. Moreover, physical control practices were negatively associated with children’s distressed facial reactions, and (at high levels of mothers’ emotional reactions) inductive reasoning and altruistic responding were negatively related to children’s distressed facial responses. These findings are consistent with theory if distressed facial reactions are viewed as a possible index of personal distress or anxiety (that is, a focus on one’s own negative feelings or personal needs) rather than an index of empathic responding (Batson, 1987). Thus, these findings suggest that mothers’ use of inductive reasoning and altruistic responding at high levels of affective arousal may actually decrease children’s distressed facial reactions (that is, persona1 distress responses) when witnessing peers in distress.

Finally, unexpectedly, girls’ self-reports of sympathetic (sorry) responses were negatively related to mothers’ altruistic responding

Tab

le 1

. R

elat

ions

Bet

wee

n M

othe

rs' S

ocia

lizat

ion

Mes

sage

s and

Chi

ldre

n's E

mpa

thic

Res

pond

ing

~~

~~

~ ~

Soci

al izn

t ion

Dire

ctio

n M

essa

ge T

ype

Empa

thy

Inde

x ? C

hang

e' F

DF

P se

x an

d Ef

fecld

~~

~~

~ ~

+ M

ain

Indu

ctio

ns

FS

AD

~ .l

l 10

.25

1.69

.002

M

and

F

Altr

uist

ic

SR-S

OR

RY

'" .1

0 3.9

0 1,

36

.05

Fem

ales

- M

ain

resp

ondi

ng

FSA

D

.05

3.91

1,

69

.05

M a

nd F

+

Mai

n

FDIS

T~

.05

4.60

1.6

8 .04

M

and

F

- Int

.04

3.88

1,

68

.05

M a

nd F

+

Int

FDIS

T .1

6 7.8

2 13

4 .01

Fe

mal

es

- Int

N

orm

ativ

e/pr

osoc

ial

FSA

D

.09

7.73

1.68

.01

M a

nd F

- I

nt

sugg

estio

ns

.22

10.5

3 1.3

4 .003

Fem

ales

- I

nt

FDIS

T .I

3 5.8

8 1.3

4 .0

2 Fe

mal

es

+ In

t

Situ

atio

nal d

efin

ition

FD

IST

.08

8.60

1,69

.0

05

M a

nd F

+

Mai

n .0

9 5.

62

1,30

.0

2 M

ales

+

Mai

n SR

-SO

RR

Y"

.08

3.29

1,

35

.08

Fem

ales

- I

nt

Phys

ical

cqn

trol

FSA

D

.085

7.7

5 1,

68

.01

M a

nd F

+

Int

FDIS

T .0

4 3.8

0 1.6

8 .0

6 M

and

F

- M

ain

Neg

ativ

e co

ntro

l SR

-SO

RR

Y

.08

6.38

1,70

.0

1 M

and

F

- Mai

n

FCO

NC

ER

N~

.06

5.35

1.6

9 .0

2 M

and

F

- M

ain

.07

3.14

1,35

.0

8 Fe

mal

es

- In

t

'Sel

f-rep

ort o

f in

tens

ity of

son

y fe

elin

gs.

bSad

, dist

ress

ed, a

nd c

once

rned

-att

entio

n fac

ial r

espo

nses

, res

pect

ivel

y.

'Am

ount

of

vari

ance

in e

mpa

thy

inde

x ac

coun

ted

for

by t

he s

ocia

lizat

ion

vari

able

. dP

ositi

ve o

r ne

gativ

e re

latio

n of

th

e so

cial

izat

ion

vari

able

with

the

em

path

y in

dex

as a

mai

n ef

fect

(M

ain)

or

in i

nter

actio

n (I

nt)

with

the

m

oder

a tor

.

75

Table 2. Relations Between Mothers’ Messages and Children’s Empathy at Levels of Mothers’ Emotional Responses

Levels of Mothers’ Emotional Reactions

Total Sample Female Subjects Socialization Variable and Empathy Index Low Mid &.Ch Low High

Inductions FDIST~ .20 3 4 -.38‘

Altruistic responding FSAD~ .21d .17 .44d FDIST .43 -.08 -.21 .35 -.031

Normative/prosocial suggestions

FSAD .41d .06 -.42 . I 1 -.53 FDIST .02 -.01 .28 -.04 .08

Situational definitions” SR-SORRY” .09 -.18 -.32 .03 -.35

Physical control

Negative control’ FSAD

SR-SORRY

-.09 .01 .42‘

-.02 -.I9 -.46d -.26 -.52d Note: Table 2 shows the relation of a given socialization variable (for example, physical control) to an empathy index (for example, SR-SORRY) at high and low levels of mothers’ emotional responses for female subjects and at three levels of affective intensity for the total sample. Correlations are included only for patterns that were significant at p < .10 in the regression analysis.

A nonsignificant correlation for either a two-way or a three-way split suggests that the groups were not split at the point of interaction suggested by the significant effect for moderation. In this case, the significant interaction for the regression is interpreted. ‘Self-reports of intensity of sorry affective responses. bSad and distressed facial responses, respectively.

dp 5 .05. ‘ p 5 .lo.

(r(39) = -32, fi < .04). That is, mothers who reported using altruistic responses tended to have daughters who reported feeling less sympathetic to the peers in the distress films. One possible explanation is that in our vignettes, the peers’ distresses were often physical. Thus, mothers’ altru- istic responses typically involved statements ensuring that no harm was done to the peer rather than statements about their feelings about the peer (for example, “Kathy, come here and let me see you are OK”). From an observational-learning perspective, the child is thus more likely to witness comforting, physical caretaking behaviors than verbal statements directly expressing sorrow or sadness or the other child (which occur

76

more frequently with loss or disappointment). Thus, a pattern of altruis- tic responding that includes mothers’ expectations for their children’s own affective responding may be necessaq in promoting children’s verbal expression of sympathetic feelings toward others.

Normative/Prosocial Suggestions

Girls’ distressed facial reactions increased when mothers used norma- tive prosocial suggestions (henceforth called prosocial suggestions) at high levels of emotional intensity (Tables 1 and 2). If facial distress is, in fact, an index of personal distress rather than of empathy or sympathy, then this practice appears to be associated, at least for girls, with decreases in empathic orientation. Similarly, mothers’ prosocial suggestions were associated with less sad facial affect as mothers’ reported intensity of emotional reactions increased, and this effect was especially strong for girls (Table 1). Typically, prosocial suggestions direct the child to perform some positive social behavior to resolve the problem with the peer (for example, “This is a party and everyone gets a chance to play”). If spoken at high levels of emotional intensity, however, expectations for changes in behavior may be perceived as nonnegotiable or coercive demands, thereby inducing reactance rather than empathy or sympathy for the peer (Hoffman, 1963). Thus, we might expect to find increases in personal distress and decreases in sad facial affect in response to the distressed peers as mothers’ intensity of emotional reactions increased.

This finding also suggests the possibility that the effectiveness of mothers’ affective responding as a moderator may increase as a function of the similarity of content of the child-rearing practice and the nature of the response expected from the child. For example, mothers’ prosocial suggestions (as well as their modeling of altruistic responses) instruct children on the performance of prosocial behaviors that the children might adopt to improve their interactive competencies with peers in conflict situations. Thus, mothers’ affective responding may act more powerfully to moderate the relation to these practices to children’s actual prosocial behaviors to a distressed peer than to moderate children’s affec- tive (empathic or sympathetic) responses. Analyses are currently undenvay to examine this possibility.

Situational Definitions

Mothers’ statements describing characteristics of the peer’s or the child’s behavior relevant to the incident were positively associated with children’s facial distress reactions (partial ~ ( 7 0 ) = .29, p < .Ol), and this relation was especially evident for boys (partial q 3 0 ) = .40, p < .02; Table 1). Furthermore, there was a marginally significant finding in which

77

mothers’ intensity of affective responding negatively moderated the rela- tion between their use of situational definitions and their daughters’ self- reports of sympathetic (sorry) affective responses (Table 1). Girls’ sympa- thetic responses to the distressed peers decreased somewhat as mothers used situational definitions at increasing levels of emotional intensity. These findings suggest that this type of information lessens children’s sympathetic orientation toward others. This information, however, is thought by some theorists to increase children’s awareness of others’ needs and situations, which would be expected to enhance empathic responding (Hoffman, 1983). One possible explanation for this dis- crepancy is that situational definitions may induce nonsympathetic ori- entations if the mother delivers so much information that the child cannot extract key elements relevant to the peer’s plight. In fact, the relation between mothers’ situational definitions and the total number of explanations they used (the sum of all coding categories) was high (~(73) = .90, p < .001). Furthermore, there is some evidence to suggest that adults may create reactance in children by providing them with state- ments designed to elicit prosocial orientations (Staub, 1971). In brief, mothers’ who explain too much may be producing effects opposite of what they intend.

Physical Control

Boys’ and girls’ sad facial affective responses became positively related to mothers’ stated use of physical control at high levels of mothers’ emo- tional reactions to the distress incidents (Tables 1 and 2). There also was a main effect in which mothers’ physical control statements negatively predicted children’s reactions of facial distress to the distressed peers in the films. These findings are consistent with Roe’s (1980) finding that girls in Greek families remained considerate of others’ feelings if mothers who used physical punishment did so within a supportive parent-child relationship. Moreover, Sears, Maccoby, and Levin (1976) found that physical punishment combined with either warmth or reasoning (defined here as inductions) was positively rather than negatively associated with children’s positive social behavior.

Because of the Sears, Maccoby, and Levin (1976) finding regarding reasoning, an additional regression was run to determine whether moth- ers’ inductions (reasoning about the impact of the child’s behavior on others) also acted as a moderator of the relation between mothers’ physi- cal control practices and the children’s sad facial affective responses. A significant effect for moderation was found (3 change = .08, F(I, a) = 7.57, p < .Ol); children’s sad facial affective responses were positively asso- ciated with mothers’ reported physical control statements at only the highest level of inductive reasoning (partial ~ ( 2 2 ) = .40, p < .05). Thus, it

appears that physical control strategies may play a role in promoting children’s emotional responsiveness to others, but that the influence of these strategies must be viewed within the context of the quality of other parental practices, as well as the quality of the overall parent-child rela- tionship (Baumrind, 1973; Radke-Yarrow, Zahn- Waxler, and Chapman, 1983).

Negative Control

Mothers’ use of negative control practices was negatively related to children’s empathic and sympathetic responses. Specifically, negative control practices were significantly, negatively related to self-reports of sympathetic (sorry) responses ( ~ ( 7 0 ) = 24, p < .05; Table 1). Moreover, for girls, self-reported sympathy became increasingly negatively related to these practices as mothers’ intensity of emotional reactions increased (Table 2). Also, children’s facial reactions of concerned attention were negatively related to mothers’ use of negative control practices. In this particular case, the zero-order correlation between this practice and chil- dren’s concerned-attention responses was nonsignificant (r(70) = -. 16), but once the effect of the intensity of mothers’ emotional reactions was par- tialed out, the relation became significant (partial q m ) = -.27, p < .02). Thus, mothers’ affective responses appeared to be acting as a suppressor variable, such that the negative effects of mothers’ control practices on children’s concerned-attention responses did not appear until the var- iance attributable to their affective responses was removed from the correlation.

These findings suggest that negative control practices hinder chil- dren’s empathic responding toward others. This may be due to several factors. First, negative control statements often may serve to emotionally reject or isolate the child, or communicate negative attributions about the child’s behavior or character. Such practices may induce children to focus on their own feelings of fear, anxiety, or rejection, or on negative self-evaluations (for example, self-recrimination), rather than on the needs, feelings, or situation of the other in distress. Furthermore, such statements do not provide the child with information relevant to devel- oping appropriate affective and behavioral responses to the peer and thus may inhibit the development of self-efficacy or control beliefs, at least with respect to responding to others in distress (Skinner, 1985). Finally, because such techniques may produce an association between children’s own distress and the distress of others close to them (they are made to feel “bad” when they cause someone else to feel “bad”), the use of negative control may be associated with children’s development of strategies designed to avoid others’ distresses.

Summary

Overall, our findings suggest that the moderating effects of maternal affective arousal in the socialization of children’s empathic responding are complex. On the one hand, at low levels of mothers’ affective inten- sity, positive relations were found between children’s sad facial reactions and mothers’ prosocial suggestions. At the highest levels of affective responding, however, mothers’ altruistic responding was significantly and positively related to children’s sad facial responses, and mothers’ altruism and inductive reasoning was negatively related to children’s distressed facial reactions (Table 2).

The negative relation of mothers’ inductions and altruistic respond- ing to children’s facial distress reactions at high levels of mothers’ affec- tive responding could be interpreted to mean that these practices are associated with decreases in children’s personal distress reactions to dis- tressed peers (that is, with decreases in one’s focus on one’s own negative feelings). However, the positive relations between mothers’ inductions and altruistic responding and children’s facial distress reactions at low levels of mothers’ emotional responding could support the interpretation of facial distress reactions as vicarious emotional responses. The affective response depicted by the children in the empathy films involved physical distress, so a purely vicarious affective (that is, empathic) response might be expected to be expressed facially as a distressed reaction. Thus, the variability in the results regarding children’s facial distress reactions sug- gests that to interpret them as an index of either empathy or personal distress requires considering the nature of the empathy stimulus and may ultimately require assessing additional criterion variables (for exam- ple, examining the relation of facial distress reactions to actual prosocial behavior).

Results not found in previous research were that high-intensity paren- tal affect combined with negative control practices was associated with a lessening of children’s sympathetic orientations, whereas situational def- initions were positively associated with children’s facial distress reactions to peer distress. The variability in these findings may be explained, in part, by the interpretation that parental affect may potentiate the impact of the semantic content of parental messages to the child. That is, if the content of the message is inductive, the mother’s intense affect may heighten the meaningfulness of the relation between the child’s behavior and its consequences for the peer’s situation or feelings (unless, of course, the parent overwhelms the child with information). In contrast, if the content of the message is emotionally or psychologically punitive (for example, coercive or negative control), the parent’s intensity may exacer- bate fearful or anxious feelings, or negative self-perceptions in the child,

80

leading to a focus on the self rather than others. Thus, the issue may not be one of determining some optimal intensity of parental affective responses but one of identifying the meaning and patterns of parents’ emotional reactions associated with different contents of their socializa- tion messages. Furthermore, searching for an optimal level of parental affective intensity ignores the fact that there is enormous variation in emotional expression in parent-child relationships, as well as in the mean- ings of emotional arousal within that relationship, as understood by the parent and the child.

If this line of reasoning is correct, it may be advantageous to deter- mine the meaning of the intensity of parental emotional reactions with respect to different socialization practices in parent-child relationships, an approach that is consistent with systems theory formulations (Bell and Chapman, 1986; Ford, 1987). For example, punishment in and of itself is typically associated with negative social outcomes for children (Maccoby and Martin, 1983). In this research, however, mothers’ stated use of physical control was found to be positively associated with chil- dren’s empathic responding toward others when such physical control occurred in the context of high levels of mothers’ affective responding and reasoning but not when used without these practices. Similarly, Radke-Yarrow, Zahn-Waxler, and Chapman (1983) have pointed out that power-assertive methods may not unilaterally produce negative effects on children’s sociomoral development if these methods are used within an overall positive and responsive parent-child relationship. Our under- standing of the socialization of empathy may be enhanced if investigative methodologies allow for idiographic assessments of the meaning of the parents’ emotional reactions that accompany other rearing practices in the family system. In that regard, researchers interested in the socializa- tion of children’s empathic responding have typically measured different rearing practices as separate, independent variables. The existence of moderator variables suggests that a more effective approach may be to examine relations between different patterns of socialization practices and children’s development of empathy. A promising avenue for future research may be in identifying particular patterns of socialization prac- tices that especially enhance children’s development of empathy.

To conclude, this research represents a first attempt at investigating the role of maternal affective arousal as a moderator of the effect of mothers’ child-rearing practices on children’s empathic responding. Moreover, the use of a multimethod approach to the assessment of empathy, as well as of differentiations among empathy, sympathy, and personal distress, expands our knowledge of the various ways in which children’s affective responsiveness may be linked to the affective states of others. Our understanding of the socialization of empathy may be enhanced still further by naturalistic studies (versus simulated vignettes)

81

and more objective assessments of the intensity, as well as the positive or negative quality, of mothers’ self-reported emotional reactions in these socialization encounters.

References

Batson, C. D. “Prosocial Motivation: Is It Ever Truly Altruistic?” In L. Berkowitz (ed.), Advances in Experimental Social Psychology.” Vol. 20. Orlando, Fla.: Academic Press, 1987.

Barnett, M. A., King, L. M., Howard, J. A., and Dino, G. A. “Empathy in Young Children: Relation to Parents’ Empathy, Affection, and Emphasis on the Feel- ings of Others.” Deuelopmental Psychology, 1980,14, 243-244.

Baron, R. M., and Kenny, D. A. “The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Con- siderations.” Journal of Personality and Social Psychology, 1986, 51, 1173-1 182.

Baumrind, D. “The Development of Instrumental Competence Through Sociali- zation.” In A. Pick (ed.), Minnesota Symposia on Child Psychology. Vol. 7. Minneapolis: University of Minnesota Press, 1973.

Bell, R. Q., and Chapman, M. “Child Effects in Studies Using Experimental or Brief Longitudinal Approaches to Socialization.” Developmental Psychology,

Bell, R. Q., and Harper, L. V. Child Effects un Adults. Hillsdale, N.J.: Erlbaum, 1977.

Cohen, J., and Cohen, P. Applied Multiple RegressionlCorrelation Analysis for the Behavioral Sciences. Hillsdale, N. J.: Erlbaum, 1975.

Eisenberg, N., and Miller, P. A. “The Relation of Empathy to Prosocial and Related Behaviors.” Psychological Bulletin, 1987,101, 91-1 19.

Eisenberg, N., Shaller, M., Fabes, R. R., Bustamante, D., Mathy, R. A., Shell, R., and Rhodes, K. “Differentiation of Personal Distress and Sympathy in Children and Adults.” Developmental Psychology, 1988,24, 766-775.

Ekman, P., and Friesen, W. V. Unmasking the Face. Englewood Cliffs, N.J.: Prentice-Hall, 1975.

Feshbach, N. D. “The Relationship of Child-Rearing Factors to Children’s Aggression, Empathy, and Related Positive and Negative Social Behavior.” In J. DeWit and W. W. Hartup (eds.), Determinants and Origins of Aggressive Behaviors. The Hague, Netherlands: Mouton, 1975.

Ford, M. Humans as Self-Constnrcting Living System: A Developmental Perspective on Behavior and Personality. Hillsdale, N. J.: Erlbaum, 1987.

Hoffman, M. L. “Parent Discipline and the Child’s Consideration of Others.” Child Development, 1963,34, 572-588.

Hoffman, M. L. “Is Altruism a Part of Human Nature?” Journal of Personality and Social Psychology, 1981, 40, 121-137.

Hoffman, M. L. “Affective and Cognitive Processes in Moral Internalization.” In E. Tory Higgins, D. N. Ruble, and W. W. Hartup (eds.), Social Cognition and Social Development. Cambridge, England: Cambridge University Press, 1983.

Hoffman, M. L. “Empathy, Its Limitations, and Its Role in Moral Intemaliza- tion.” In W. M. Kurtines and J. L. Gewirtz (eds.), Morality, Moral Behavior, and Moral Development. New York: Wiley, 1984.

Kenny, D. A. Statistics for the Social and Behavioral Sciences. Boston: Little, Brown, 1987.

Maccoby, E. E., and Martin, J. A. “Socialization in the Context of the Family:

1986,22, 595-603.

82

Parent-Child Interaction.” In P, H. Mussen (ed.), Handbook of Child Psychol- ogy. Vol. 4: Socialization, Personality, and Social Development. New York: Wiley, 1983.

Miller, P. A. “Maternal Child Rearing Practices and Daughters’ Empathic Responses to Peer Distress.” Unpublished doctoral dissertation, Department of Educational Psychology, University of Texas, Austin. (Disse7tation Abstracts International, 1984, 46B, 676, Order No. 85-08316.)

Radke-Yarrow, M., and Zahn-Waxler, C. “Roots, Motives, and Patterns in Chil- dren’s ProsociaI Behavior.” In E. Staub, D. Bar-Tal. J. Karlyowski, and J. Reykowski (eds.), Development and Maintenance of Prosocial Behavior: Intema- tional Perspectives on Positive Morality. New York: Plenum, 1984,

Radke-Yarrow, M., Zahn-Waxler, C., and Chapman, M. “Children’s Prosocial Dispositions and Behavior.” In P. H. Mussen (ed.), Handbook of Child Psychol- ogy. Vol. 4: Socialization, Personality, and Social Development. New York: Wiley, 1983.

Roe, K. V. “Early Empathy Development in Children and the Subsequent Inter- nalization of Moral Values.” Journal of Social Psychology, 1980,110, 147-148.

Rushton, J. P., Fulker, D. W., Neale, M. C, Nias, D.K.B., and Eysenck, H. “Altru- ism and Aggression: The Heritability of Individual Differences.” Journal of Personality and Social Aychology, 1986,50, 1 192- 1 198.

Sears, R. R., Maccoby, E. E., and Levin, H. Patterns of Childrearing. Stanford, Calif.: Stanford University Press, 1976.

Siegel, M., and Barclay, M. S. “Children’s Evaluations of Fathers’ Socialization Behavior.” Developmental Psychology, 1985, 21, 1090-1096.

Skinner, E. A. “Determinants of Mother Sensitive and Contingent-Responsive Behavior: The Role of Childrearing Beliefs and Socioeconomic Status.” In I. E. Sigel (ed.), Parental Belief System: The Psychological Consequences for Children. Hillsdale, N. J.: Erlbaum, 1985.

Staub. E. “The Use of Role Playing and Induction in Children’s Learning of Helping and Sharing Behavior.” Child Development, 1971, 42, 805-816.

Zahn-Waxler, C., Radke-Yarrow, M., and King, R. A. “Child Rearing and Chil- dren’s Prosocial Initiations Toward Victims of Distress.” Child Developnent, 1979,50, 319-330.

83

Paul A. Miller is an assistant projessor of psychology at the West Campus of Ariwna State University.

Nancy Eisenberg is a professor of psychology at Arizona State University.

Richard A. Fabes is an associate professor in the Department of Family Resources and Human Development at Arizona State University.

Rita Shell is a graduate student in social developmental fisychology at Arizona State University.

Suzanne Gular is a graduate student in educational psychology at Ariwna State University.