Children’s Emotion Expression in the Home When Interacting With Their Parents

Jacqueline Sperling, M.A., C.Phil. and Rena Repetti, Ph.D.

University of California, Los Angeles

Introduction We know very little about the types, frequencies, and intensities of emotions that children

express in daily life. One study observed pre-schoolers’ emotion expression when interacting

with their mothers in a laboratory setting and found that the dyads predominantly expressed

positive and neutral emotions (Denham & Grout, 1993). Researchers proposed that this

overall positive/neutral behavior would be expected in a laboratory setting due to the social

desirability bias. Another study in a laboratory setting recorded pre-schoolers’ emotion

expression during a frustrating “waiting task” that required participants to wait eight minutes

with their mothers until they were allowed to open a gift (Cole, Teti, & Zahn-Waxler, 2003).

Children expressed emotion an average of 29 times over the course of 8 minutes, and they

expressed positive emotion as frequently as they did anger. However, laboratory settings are

unnatural environments, so some researchers have had participants complete daily reports to

get a better sense of emotion expression in everyday life. For example, using Experience

Sampling Methods (ESM) to collect adolescent and parent reported emotion experiences,

Larson & Richards (1994) found that adolescents endorsed the “extreme” or intense end of the

emotion experience response scale more frequently than their parents: they reported that they

felt “very positive” five times more often and “very negative” two times more often than their

parents did. However, even though emotion experiences are typically triggered by social

interactions, questionnaire measures usually ignore the relevant social context. This study

observed children’s emotion expression when interacting freely with their parents at home.

Participants • 31 families recruited by the Center for Everyday Lives of Families (CELF) for a

study of dual-earner middle-class families which was funded by the Alfred P. Sloan

foundation

• Inclusion criteria:

•Both parents worked at least 35 hours per week

•Two-to-three children: one of which, the target child, was aged between 8 and 12

years old

•Had a mortgage on the home

• Of the 31 target children, three were in second grade, thirteen in third grade, nine in

fourth grade, three in fifth grade, and three in sixth grade. There were 17 male and

14 female target children.

•Two families were headed by male same-sex couples. Because this study focused on

mother- and father-child interactions, one father from each couple was randomly

selected to be included in the analyses.

Method The data used for this project is part of a larger study which included videotaped

naturalistic observation, self-report measures, semi-structured interviews, and

salivary cortisol sampling. The present study focused on a portion of the video

recorded data which included filming the 31 families in their homes over the course of

one weekday and one weekend day. On the weekday, filming began when children and

their parents awoke until they left home and then resumed when they returned from

school and work. On the weekend day, filming began when the children awoke and

continued until the children went to sleep. A camera was assigned to follow each

parent; therefore, the target child was not always captured on screen with the parent.

The videographers followed parents when they left home to visit community settings

and used wireless microphones to capture social interactions without having to hover.

In order to systematically sample the video recordings of parent-child interactions,

all 30-second clips of mother-child and father-child interactions on each weekday and

weekend day were coded. If there were more than 200 30-second clips for a dyad one

day, coders only observed the first 200 clips when footage was ranked starting from

the longest video segments of dyadic interactions. Coders identified whether the child

expressed an emotion at least once during each video clip. Coding is discussed in more

detail above.

Proportion of Clips in Which Specific Emotions were Expressed by the Daughters (D) and Sons (S) Across Both Days When Co-

Present with Target Parent in the 29 Families with Both Mothers and Fathers

M(D) M(S) Median(D) Median(S) SD(D) SD(S) Range(D) Range(S)

Mother-Child Dyads

Happiness .36 .28 .35 .27 .12 .13 .18-.57 .09-.56

Sadness .04 .06 .03 .04 .05 .04 .00-.04 .00-.18

Anger .08 .08 .06 .05 .06 .08 .01-.19 .00-.25

Surprise .01 .01 .01 .01 .02 .01 .00-.08 .00-.04

Fear .00 .00 .00 .00 .01 .00 .00-.02 .00-.01

Disgust .01 .01 .01 .00 .00 .01 .00-.02 .00-.02

Father-Child Dyads

Happiness .37 .26 .39 .24 .11 .13 .19-.56 .01-.54

Sadness .04 .04 .02 .03 .04 .04 .00-.14 .00-.12

Anger .08 .07 .05 .06 .07 .06 .01-.24 .00-.23

Surprise .01 .01 .01 .01 .01 .01 .00-.02 .00-.05

Fear .00 .00 .00 .00 .00 .00 .00-.00 .00-.00

Disgust .01 .00 .00 .00 .01 .00 .00-.04 .00-.02

Observational Behavior Coding System Child Expression or Discussion of Emotion: A team of 19 coders was trained on pilot video clips in which a parent and non-target child

were co-present. Emotions included both nonverbal and verbal expressions of Ekman’s (1992) six universal emotions: anger (e.g., furrowed

eyebrows or stomping one’s feet), fear (e.g., raised eyebrows in conjunction with widened eyes and the stretching of one’s lips horizontally),

sadness (e.g., tearfulness and pouting), happiness (e.g., smiles and laughter), disgust (e.g., wrinkling of the nose or a gaping mouth with a tongue

extrusion), and surprise (e.g., a dropped jaw and open mouth). Dichotomous scoring was used to determine whether any of these emotions was

expressed by the target child. Coders identified the specific emotion expressed and the time the display began.

Inter-rater reliability was maintained by having two coders separately rate each of the four days of clips in which the target parent and child were co-

present on film (two days for mother-child and two days for father-child dyads) for a family and then resolve any coding disagreements. Coders

had to agree on the start time of each expressed child emotion before submitting final data. The 19 coders were paired with a different team

member for each family assignment, and reliability was measured by aggregating data across all coder pairs. There was a total of 15,071 30-

second clips across dyads, days, and families.

Total Number of 30-Second Parent-Child Clips Per Day for Each Parent-Child Dyad

n of Dyads M Median SD Range

Mother-Child Dyads

Weekday 29 133 136 51 35-200

Weekend 29 128 143 63 18-200

Father-Child Dyads

Weekday 31 108 108 58 11-200

Weekend 31 134 161 74 4-200

Child Sex Differences (n = 29 families) With both parents, daughters (M = .37) expressed happiness in a greater proportion of

clips than sons (M = .28, t(27) = 2.23, p ≤ .05). With both parents, daughters also

expressed more fear (M = .003) and disgust (M = .008) in a greater proportion of clips

than sons (fear: M = .001, t(27) = 2.15, p ≤ .05; disgust: M = .004, t(27) = 2.62, p ≤ .01).

Future Directions

• In the process of coding the intensity of children’s negative emotions, parents’

behavioral and emotional responses to children’s negative emotions, and children’s

emotional reactions to parents’ responses.

• Multilevel modeling will be implemented to test hypotheses of how parents’ responses

are linked with children’s emotion expression.