Mothers’ and Fathers’ Reports of their Supportive Responses to their Children’s Negative Emotions over Time

SYNOPSIS

Objective

Parents’ emotion socialization practices are thought to be moderately stable over time; however, a partner’s socialization practices could initiate change.

Design

We examined mothers’ and fathers’ reports of their supportive responses to their children’s negative emotions when the target child was 7 years old and again at age 10. We tested a dyadic, longitudinal path model with 111 mother-father pairs.

Results

Significant actor and partner effects emerged. Parents’ age 7 responses predicted their own age 10 responses and their partners’ later responses.

Conclusions

Parents’ reported responses to children’s negative emotions during middle childhood are predicted by their own earlier responses and by their partners’ responses.

INTRODUCTION

Parents’ socialization of children’s emotion is a multifaceted, complex process that facilitates social development and aids children in understanding, expressing, and regulating emotion (Eisenberg, Cumberland, & Spinrad, 1998). Compared to positive emotions, negative emotions are harder for children to cope with, and there are more rigid expectations regarding the expression of negative emotion (Malatesta & Haviland, 1982; Ramsden & Hubbard, 2002). Thus, the way in which parents respond to their children’s negative emotions in the moment constitutes an important method of emotion socialization (Eisenberg et al., 1998). Theoretical and empirical work has suggested that supportive parental responses, such as encouraging children to express themselves or teaching strategies that help to alleviate distress, are associated with less difficulty regulating emotional arousal (Eisenberg et al., 1998; Tao, Zhou, Wang, 2010). Supportive responses provide children with tools they can utilize when facing challenging situations independently (Denham, 1997). Indeed, empirical work has shown supportive emotion socialization techniques to be related to the number of coping strategies generated to regulate negative arousal, as well their effectiveness, in addition to promoting overall social competence (Cole, Dennis, Smith-Simon, & Cohen, 2009).

Parent Development Theory suggests that parents’ practices and perceptions of their parenting role change over time in response to their developing child (Mowder, 2005). However, longitudinal investigations of parents’ responses to their children’s negative emotions have focused on how responses at one point in time relate to later child outcomes (e.g., Tao et al., 2010). Although it is plausible that these behaviors remain consistent over time due to mean-level stability in parents’ philosophies about emotions (Michalik et al., 2007), there is emerging evidence that emotion socialization practices change as children become more independent, have a more complex understanding of emotions, and develop effective regulatory strategies. Eisenberg et al. (1999) reported moderate to high mean-level stability for mothers’ responses to children’s negative emotions during middle childhood, although their expectations for children appeared to shift at the start of elementary school and again during early adolescence. Thus, change in emotion socialization is not likely to be linear, but responsive to contextual and developmental changes. One aim of this study was to explore the mean-level stability of parents’ reports of their supportive responses to negative emotions from age 7 to age 10 to better understand these influential parenting practices over time.

According to Family Systems Theory, mother-child and father-child dyads are interdependent subsystems in the family (Cox & Paley, 1997). Thus, changes in emotion socialization practices are likely influenced by one’s partner. Mothers and fathers jointly create a socialization context for their children and influence each other’s emotions, cognitions, and behaviors. Emotion socialization researchers have given increased attention to fathers’ responses to children’s negative emotions, demonstrating that fathers tend to provide fewer supportive responses than mothers (Cassano, Perry-Parrish, & Zeman, 2007). Still, few studies recognize the dyadic nature of mothers’ and fathers’ responses, and none to date has explored how parents’ emotion socialization practices are predicted by a partner’s practices using a longitudinal framework. Not only is a dyadic design truer to the dynamic nature of family life, but a better understanding of partner effects increases our knowledge of how adaptive (and maladaptive) practices permeate families. Therefore, the second aim of the current report was to explore whether mothers’ and fathers’ reports of their supportive responses to children’s negative emotions at the start of elementary school predicted each other’s responses 3 years later in middle childhood using a longitudinal dyadic design accounting for the interdependence between parents.

METHODS

Participants

The current sample utilized data from two of three cohorts of children who are part of a larger, ongoing longitudinal study conducted in Greensboro, North Carolina, USA. We excluded cohort 1 because father data were not available for these participants. All cohorts were recruited through child care centers, the County Health Department, and the local Women, Infants, and Children program. Participants in cohort 2 were recruited at 2.5 years during 2000–2001 (n=130), and participants in cohort 3 were recruited in 1998 when children were 6 months old (n=140). There were no significant demographic differences between cohorts with regard to gender, ethnicity, or socioeconomic status (SES). Of the 270 original participants in cohorts 2 and 3, 234 families (86.7%) participated at 7 years. Again, there were no significant differences between families that did and did not participate in terms of child gender or ethnicity. Families with lower SES when children were 2.5 years old were less likely to continue participation at the 7-year assessment, t(432)=−2.31, p<.05.

Of the 234 families that participated in the 7-year visit, 184 mothers reported being in a two-parent family. Only two-parent families in which both parents reported on their responses to child negative emotions at the 7-year assessment were included in the present analyses (n=111 couples). Fathers who completed questionnaires at the 7-year visit were more educated on average than fathers who did not participate, t(178)=−2.67, p<.01. All fathers included were living in the home with the mother and child, 99% were married to the mother, and 97% were the biological father of the child. In the final sample, 47% (n=52) of the children were female, 82% (n=91) were European American, 13% (n=14) were African American, and 5% (n=6) were of mixed or other ethnicities. The composition of the families also varied; 16% (n=18) of the families had 1 child, 49% (n=53) had 2 children, 23% (n=26) had 3 children, and 12% (n=14) had 4 or more children. Mothers’ and fathers’ median education level was a college degree.

Procedures

When the children were 7 and 10 years old, families came to the study site for two visits. At the second visit, research assistants gave mothers who resided with fathers a packet of questionnaires for the father to complete; 60% and 81% of fathers living in the home completed these packets at the 7-year and 10-year assessments, respectively. Mothers were compensated $125 total for participating in both visits and completing questionnaire packets. Fathers were compensated an additional $10 for completing questionnaires.

Responses to Child’s Negative Emotions

Parents independently completed the Coping with Children’s Negative Emotions Scale (CCNES; Fabes, Eisenberg, & Bernzweig, 1990) when children were 7 and 10 years old. The CCNES is a self-report measure in which mothers and fathers respond to 12 hypothetical situations in which their child expresses distress. Parents indicate the likelihood of each of six possible responses ranging from 1 (very unlikely) to 7 (very likely). The measure yields 6 subscales, 3 of which were averaged to indicate supportive responses: problem-focused reactions, emotion-focused reactions, and expressive encouragement. Cronbach’s alpha for the supportive aggregate was .90 for mothers and .93 for fathers when children were 7 years old and .92 for mothers and .95 for fathers when children were 10 years old.

RESULTS

Reports of supportive responses averaged 5.65 (SD=.63, range: 3.92−6.83) at age 7 and 5.59 (SD=.67, range: 3.72−6.75) at age 10 among mothers, and 5.22 (SD=.80, range: 3.11−7) at age 7 and 5.24 (SD=.85, range: 1.47−6.83) at age 10 among fathers. Socio-economic status was included as a covariate due to its association with attrition. We also examined child gender, child ethnicity, and parental education as potential covariates. Child gender was associated with study variables and was controlled in the analyses. Fathers of girls reported more supportive responses than fathers of boys at age 7, t(109)=−2.17, p=.032.

Of the 111 couples with complete mother and father CCNES data at age 7, 85% (n=94) of mothers and 77% (n=86) of fathers completed the CCNES measure at the 10-year wave. This resulted in 7% missingness. To account for missing data, which were missing completely at random according to Little’s MCAR test, χ²(66)=60.90, p=.65, we used full information maximum likelihood (FIML) estimation.

We tested intra-class correlation coefficients (ICCs) to examine interdependence between mothers’ and fathers’ reports of their supportive responses. ICCs, interpreted in the same way as Pearson correlations, take into account rank-order stability as well as the mean difference between dyad members to estimate dependence in dyadic data (Kenny, Kashy, & Cook, 2006). Correlations showed that mothers’ and fathers’ responses were significantly related at age 7, ICC=.17, p=.033, and age 10, ICC=.37, p<.01.

An omnibus test of distinguishability was performed to determine the extent to which mothers’ and fathers’ reported responses differed in means, variances, and correlations. Constraints were added using structural equation modeling in Mplus v.7.11 (Muthén & Muthén, 1998–2012) to compare the fit of models with these characteristics freely estimated to those set to be equal across parents. Mothers were more supportive than fathers at age 7, χ²Δ(1)=16.83, p<.01, and age 10, χ²Δ(1)=11.77, p<.01; there was greater variation across fathers in the sample than mothers at age 7, χ²Δ(1)=6.41, p<.05, and age 10, χ²Δ(1)=5.60, p<.05; across time, fathers’ earlier responses were more predictive of their later responses than mothers’, χ²Δ(1)=4.37, p<.05; and parents’ reports of their responses were more strongly related to their partners’ at age 10 than age 7, χ²Δ(1)=3.82, p=.05.

A longitudinal distinguishable Actor-Partner Interdependence Model (APIM; Kashy & Kenny, 2000) was used to examine whether mothers’ or fathers’ scores predicted those of their partners over time. APIM models account for interdependence in dyadic data. Actor effects in the longitudinal APIM models estimated the extent to which parents’ reports of their responses at age 7 were predictive of their own later responses at age 10. Partner effects estimated the extent to which parents’ reported responses at age 7 were predictive of their partners’ responses at age 10. We tested paths in a structural equation model using path analysis. The model controlled for child gender and SES.

Figure 1 shows the structural model with standardized coefficients (covariates not pictured). We evaluated model fit using the root mean square error of approximation (RMSEA), the comparative fit index (CFI), the standardized root mean square residual (SRMR), and the chi-square statistic (χ²). The model had good fit, RMSEA=.02 [90% c.i.=.00–.15], CFI=1.00, SRMR=.03, χ²(4)=4.14, p=.39. The actor and partner effects tested in the model were significant. Parents’ reports of their supportive responses at age 7 were highly predictive of their own supportive responses at age 10 and were also predictive of their partners’ reported supportive responses at age 10. Adding equality constraints revealed that mothers’ and fathers’ partner effects were not statistically different from each other, χ²Δ(1)=1.77.

Figure 1.

Longitudinal APIM model for supportive parental responses controlling for child gender and SES.

DISCUSSION

Middle childhood is a developmental period that requires children to appropriately express and effectively manage negative emotional arousal in order to adapt to rigorous academic demands and complex social situations. Parents’ reactions to children’s anger, fear, and sadness play a crucial role in their ability to regulate emotion and behave in a socially appropriate manner (Eisenberg et al., 1998). Mothers and fathers are both important socializing agents and are likely to influence each other’s emotion socialization practices. Thus, the current report incorporated mothers’ and fathers’ reports of their responses to children’s negative emotions using dyadic data in a longitudinal design across a 3-year period during middle childhood. This dyadic approach contributes to our understanding of family processes by allowing us to model a more complete picture of parenting practices and predictors of these behaviors while statistically accounting for interdependence in mothers’ and fathers’ behaviors.

A significant degree of dependence was found between mothers’ and fathers’ reported supportive responses within the same families at both time points; parents’ responses were more closely related at age 10 than age 7, consistent with research demonstrating that mothers and fathers tend to report more similarity in parenting practices over time (Feinberg, 2003). Mothers reported providing significantly more supportive responses to their children’s negative emotions than fathers at both time points, and fathers’ reported responses varied across the sample more than mothers’. After accounting for the established dependence within dyads, we found that both mothers’ and fathers’ reports of their supportive responses were predicted by their partners’ earlier reported responses. Despite the fact that mothers tend to carry more responsibility for attending to family members’ emotions than fathers (Erickson, 2005), these findings are consistent with Family Systems Theory and suggest that parents may influence each other’s practices. Parents may make shifts in their expectations for children’s emotional displays as children mature, and adjust to each other accordingly.

However, it is important to note that parents’ reported responses to children were not markedly different at age 7 and age 10. Significant actor effects demonstrated strong mean-level stability in parents’ reports over time, suggesting that any change in parents’ supportive responses predicted by partners was modest. This is consistent with previous research demonstrating moderate to high stability coefficients for emotion socialization practices (Eisenberg et al., 1999). Parents appear to hold beliefs or philosophies about emotions that persist across time.

There are some noteworthy limitations to this study. First, data from mothers and fathers were only available at two time points, and due to a non-experimental design, causality cannot be determined. Additional waves of data could shed light on transactional processes within dyads. Second, this report focused on the parents’ relationship, but future research including larger samples should examine interactions that include child effects. Child gender, pubertal status, and temperament are factors that may contribute to changes in emotion socialization practices. Third, the sample was limited to mothers and fathers living in the same home. We are not able to generalize these findings to parents living separately whose partner effects may vary from the present results. Additionally, participating fathers had higher education levels than those who lived with mothers but chose not to participate, which could have resulted in an advantaged group that may differ in their behaviors from less advantaged fathers.

IMPLICATIONS FOR PRACTICE, APPLICATION, THEORY, AND POLICY

The emotion socialization messages parents provide to children are dynamic. In this study, parents’ reports of their responses were positively associated over time, as parents appeared to initiate changes in their partners’ practices with older children. These findings enhance our understanding of the processes of emotion socialization within families. For example, these findings suggest that efforts to increase a parent’s tendency to encourage children to appropriately express and regulate their negative feelings may lead to positive changes in a partner’s responses, as well, regardless of parent gender. Thus, understanding how one parent may respond to changes in the other parent’s behaviors reveals important information about the indirect socialization effects children may experience.