Actor and Partner Effects of Parenting Stress and Co-parenting of Parents of Children with Atopic Dermatitis on Marital Conflict: Multiple-group Analysis Based on Mother’s Employment Status

Objective This study aims to determine the effect of parenting stress and coparenting on marital conflict and the actor and partner effects of parental variables and to identify the control effect based on the mother’s employment status. Method In this study, among all children who participated in the panel study and health questionnaire survey, 161 fathers and 161 mothers raising seven-year-old children recently treated for atopic dermatitis for 12 months were selected as the final study participants. Results The father’s parenting stress had an actor and a partner effect on the co-parenting, and the mother’s parenting stress had an actor and a partner effect on the co-parenting. The father’s parenting stress only had an actor effect on the marital conflict(β=.32, p<.001). The father’s co-parenting had an actor(β=-.29, p<.001) and a partner effect(β=-.22, p<.001) on the marital conflict, and the mother’s co-parenting had an actor(β=-.39, p<.001) and a partner(β=-.19, p<.001) effect on the marital conflict. There were statistically significant differences between two groups in terms of the path coefficient of the father’s parenting stress affecting father’s marital conflict, path coefficient of the father’s co-parenting affecting father’s marital conflict, and path coefficient of the mother’s co-parenting affecting father’s marital conflict Conclusions This study is meaningful in providing basic data for the development of an atopic dermatitis family management program by investigating the actor and partner effects of parenting stress and co-parenting on marital conflict as well as the control effect based on the mother’s employment status.

to long-term care as with other chronic illnesses, which negatively affects family functioning [10].
In a society in which childcare is mostly carried out by mothers, the physical symptoms of the child are a cause of condemnation and guilt for the mother, and these symptoms of the child and the parenting stress of the mother induce conflict between the husband and wife, which destroys the family function and has a negative effect on the recovery of the child's health [11]. Additionally, if the mother also works and has to balance work and family, then the level of stress is higher than for a stay-at-home mom, and the increase of the mother's parenting stress is associated with the increase of the father's parenting stress, also influencing marital conflict [12]. However, if both parents recognize parenting as a collective responsibility and actively participate, the parenting stress can be moderated with co-parenting and have a positive effect on the marital relationship in terms of raising children [13].
Additionally, in the co-parenting model proposed by Feinberg [14], it was reported that parenting stress affects co-parenting, and the degree of co-parenting affects the child's adaptation as well as marital adaptation. To help with management of children with AD, it is important to be interested in the child's health problems, but maintaining the function of the family, particularly between the husband and wife, by keeping an interest in the parenting stress and co-parenting could be an effective way to increase the recovery of children with AD. However, previous studies related to children with AD examined the effect of self-efficacy of the parents and satisfaction of the marital relationship on children's behavior [15], the effect of self-efficacy of the mother on the family management [16], and sleep disorder in parents raising children with AD [17], so there is limited research that comprehensively approaches children with AD and parents. In particular, considering that children are highly dependent on parents due to the nature of the disease, studies that investigate the parental variables at the same time are necessary for the management of atopic children.
Because parents are in an interdependent relationship, the Actor-Partner Interdependent Model (APIM) suggested by Kenny [18] is recommended for analyzing the interrelation between parent-related variables, and this study also applied the APIM model to identify the effects of parenting stress and co-parenting on marital conflict among parents with children with AD. For the couple data, the mutual dynamics of the couple was not reviewed when using individual data. Even if the data were collected from both members of the couple, analyzing such interdependent data individually as independent data, it violates the main hypothesis of reasoning statistics, resulting in low calculation of standard error and a possibility of committing a Type 1 error. Thus, such interdependent couple data must be analyzed by applying APIM [18]. In particular, marital conflict is a result of interaction between parents, so it is necessary to determine the effects of parenting stress and co-parenting instead of analyzing the mother and father individually. Therefore, the present study aims to determine the effect of parenting stress and co-parenting on marital conflict and the actor and partner effects of parental variables and to identify the control effect based on the mother's employment status to provide basic data for the development of an AD family management program.

Study design
This is a cross-sectional descriptive study to investigate the actor and partner effects of parenting stress and co-parenting on marital conflicts and the differences between groups according to the mothers' employment status utilizing the 8th Panel Study on Korean Children [ Figure 1].

Study Participants
The present study included parents above the age of 19  raising seven-year-old children recently treated for AD for 12 months were selected as the final study participants. In the structural equation model, the minimum recommendation for the sample size is 10 times the free parameter, and the ideal size is 150-400 participants, so 161 participants in this study constituted a sufficient sample size to analyze actor and partner effects using the structural equation model.

Measurement
In the study, the validity of the tool was confirmed through confirmatory factor analysis. Convergent validity was confirmed to be greater than .50 for each factor loading, greater than .70 for construct reliability, and greater than .50 for average variance extracted, and discriminant validity was found to be valid when the AVE (averaged variance extracted) values of the different latent variables were greater than the square of the correlation coefficient between the latent variables.

Parenting stress
For the parenting stress survey, "burden and distress from carrying out parents' role" among the subfactors of the parenting stress scale developed by Kim and Kang [19] was extracted by the Panel Study on Korean Children's research team, and a tool with 11 questions confirmed through a preliminary survey from 2007 was used.
A total of 11 questions were based on a five-point scale, and higher scores signify high parenting stress. For the reliability of the tool in the study by Kim and Kang [19], Cronbach's alpha was .86. For the reliability of the tool in this study, Cronbach's alpha was .88 for the father and .90 for the mother.

Co-parenting
For the co-parenting survey, the measurement tool developed by Mchale [20] was translated by the Panel Study on Korean Children's research team, and a total of 16 questions (four subcategories: family unity, discipline, criticism, conflict), which underwent the preliminary survey, were selected based on a seven-point scale.
Higher sum of scores signifies a high level of co-parenting. In Mchale's study [20], the reliability of the tool was Cronbach's alpha .59-.82, and, in this study, Cronbach's alpha was .88 for the father and .86 for the mother.

Marital conflict
For marital conflict, the measurement tool developed by Markman et al. [21]

Ethical considerations
The 8th Panel Study on Korean Children was approved by the institutional review board of the Korea Institute of Child Care and Education (IRB No. KICCEIRB-2015-03). The current work was also conducted after review by the Institutional Review Board of C University.

Data collection and analysis
The data were obtained from the website of the Panel Study on Korean Children

General participant characteristics
In terms of the residential areas of the participants, 62 lived in large cities (38.5%), 69 lived in towns (42.9%), and 30 lived in small and medium-size cities. The mean age of the fathers was 40.5 years old while that of the mothers was 37.9 years old.

Descriptive statistics of measurement variables
The mean parenting stress score of the fathers was 1.9 points, and that of the mothers was 2.3 points. The mean co-parenting score of the fathers was 5.2 points, and that of the mothers was 5.4 points. The mean marital conflict score of the fathers was 1.9 points, and that of the mothers was 2.1 points. For each measurement variable, the absolute value of skewness (−0.65 to 0.81) did not exceed 2, and the absolute value of the kurtosis (−0.17 to 1.15) did not exceed 4 [ Table 1].

Correlation between measurement variables
Each measurement variable showed a statistically significant correlation at the significance level of .05, and the absolute value of the correlation between the variables did not exceed .8, confirming that there was no problem of multicollinearity [ Table 1].

Verification of measurement invariance of measurement variables
To verify the homogeneity of the father and mother's data within one measurement tool, four competing models were compared. The first model is the baseline model, the second constrains the factor loading, the third constrains the covariance of the error, and the fourth constrains the factor loading and covariance of the error. In this study, the results of χ 2 and TLI, CFI, RMSEA, which are not sensitive to the number of cases, confirmed measurement invariance [ Table 2].

Actor and partner effect of parenting stress and co-parenting on marital conflict
To verify the effects of parenting stress and co-parenting on marital conflict, the normality of the measurement variables was investigated. The univariate normality of each measurement variable satisfied the normal distribution condition by showing the absolute value of the skewness and the kurtosis ranging less than 2, but multivariate normality was not satisfied at the significance level of .05 with multivariate index = 4.10 and CR = 6.10. If multivariate normality is not satisfied, there may be a problem of upward biasing the threshold when estimating the parameters. However, even if the multivariate normality is not assumed, it is reported that the estimated parameter is reliable when using the maximum likelihood method and when the sample size is 120 or greater. Therefore, the model was estimated without converting the data. The goodness of fit of the hypothetical model was evaluated with GFI, AGF, CFI, NFI, TLI, RMSEA, and SRMR, where the goodness of fit of the model is interpreted to be favorable when GFI, AGFI, CFI, NFI, and TLI are greater than 0.9. The fit of the model is considered good if RMSEA and SRMR are less than 0.05, fair when less than 0.10, and low when greater than 0.10.
Among the demographic characteristics reported as the factors affecting parenting stress using the maximum likelihood method, age, educational background, household income, sex of the child, and AD diagnosis period presented in this study were used as the control variables to conduct a hypothetical model test, which was confirmed as the model with appropriate goodness of fit with χ 2 = 15.59, df = 10, RMSEA = .02, SRMR = .04, GFI = .95, AGFI = .94, CFI = .97, NFI = .97, TLI = .96.
Nine out of a total of 12 hypotheses were selected [ Table 3]. The father's parenting stress had an actor effect (β = −.46, p < .001) on co-parenting and a partner effect  Table 3]. In addition, the father's parenting stress (β = .17, p = .004) had an indirect effect on the father's marital conflict, and the father's parenting stress had an indirect effect on the mother's marital conflict (β = .04, p = .005), but the total effect (β = .16, p = .269) of the father's parenting stress on the mother's marital conflict was not statistically significant. The mother's parenting stress had an indirect effect on the father's (β = .07, p = .005) and mother's (β = .18, p = .003) marital conflict.

Multiple group analysis according to the mother's employment status
To confirm the significant difference of the intergroup path coefficients, the critical ratio for difference of free and constrained models between 12 paths in the study model was confirmed. As a result, there were statistically significant differences between two groups in terms of the path coefficient (critical ratio for difference = −2.408) of the father's parenting stress affecting father's marital conflict, path coefficient (critical ratio for difference = 2.753) of the father's co-parenting affecting father's marital conflict, and path coefficient (critical ratio for difference = 2.952) of the mother's co-parenting affecting father's marital conflict [ Table 4].

Discussion
This study aimed to identify the actor and partner effects of parenting stress and co-parenting on marital conflicts in parents raising children with AD and to further discuss the differences between groups based on the mother's employment status.
First, the parenting stress of the fathers and mothers of children with AD had actor and partner effects on both parents' co-parenting. The results of this study are similar to the findings of May et al. [13] and Feinberg [14] who reported parenting stress to affect co-parenting based on the co-parenting model conducted in parents of children with autism.
Such results show that co-parenting is a process in which couples discuss the principles of child-raising and sharing of the burden of child-raising and cooperate with each other [22], where, as the husband and wife are more affectionate and supportive in relation to each other, parents are not only confident in their parental role but cooperate well together while the couple's intimacy and reliability decrease as blame and conflict increase when parenting stress is exchanged between the parents [20], resulting in decreased motivation for co-parenting. Therefore, considering that the parents of children with AD have a higher level of parenting stress than the parents of children with other chronic diseases, medical professionals might seek and apply interventions that can reduce parenting stress in addition to improving physical symptoms of the children as an effective way to achieve physical and emotional stability of the child.
Second, it was found that marital conflicts perceived by the fathers of children with AD were influenced by the actor effect of father's parenting stress and co-parenting and partner effect of mother's co-parenting. As reported in a previous study on marital conflict and parenting [23], marital conflicts increase when a couple perceives that they have not been able to get help and cooperation from their spouses while raising their children; the degree of the father's perceived marital conflict increases if the father of the child with AD believes that he did not get much help from the mother for parenting. As it was confirmed in previous studies that marital conflicts are associated with child's internal and external problem behaviors [23,24] and affect the restoration of health in a child with AD, there is a need for medical professionals to approach with more interest in the marital conflicts that are perceived by the father of children with AD, and it is necessary to approach the degree of mother's participation in parenting and the attitude toward co-parenting at the same time to reduce the degree of marital conflict experienced by the father.
Third, the marital conflict perceived by the mothers of children with AD was affected by the actor effect of the mother's co-parenting and the partner effect of the father's co-parenting. Such results are similar to the study results by Feinberg [14], who reported that the co-parenting of the parents had a positive effect on the adaptation between husband and wife. For parents who raise children, marital relationship is closely related to the process of raising children, and, in particular, the results of a study [24] reporting that children are more likely to be affected by their father's emotional and behavioral status than their mother's suggest the importance of the father's co-parenting in raising children with AD. In addition, the mother could become dependent on the father in the process of raising the child; during this process, mothers tend to underestimate the quality of their marital relationship if they perceive the father's level of parental involvement to be low. Therefore, to lower the degree of marital conflict perceived by the mothers of children with AD, it is necessary to confirm the attitude and degree of father's coparenting in addition to positively increasing the attitude and degree of mother's co-parenting.
Fourth, parents' parenting stress had an indirect effect on the marital conflict perceived by the father while mother's parenting stress had an indirect effect on the marital conflict perceived by the mother. Mothers are usually the primary caregiver of children with AD, and, because they feel great burden [9], active intervention for the parenting stress of the mothers of children with AD could be an important factor in reducing marital conflict. In addition, in the path analysis according to the employment status of the mothers, there were significant differences among the groups in the path of father's parenting stress to marital conflict of the father, the path of father's co-parenting to marital conflict of the father, and the path of mother's co-parenting to marital conflict of the father. Such study results indicate that, when the mother is employed, the proportion of father's role in parenting is relatively high, where the degree of parenting stress perceived by the father affects co-parenting, resulting in establishing the relationship of affecting marital conflict. Therefore, there is a need for planning and approaching a distinctive intervention plan for the father's parenting stress and co-parenting among parents of children with AD with employed mothers. Meanwhile, if the mother is a stay-at-home mother, the key factor that affects father's marital conflict was confirmed to be the mother's co-parenting, and, thus, it is also important to approach with a lot of interest in the intervention plan based on this.

Conclusion
The study is meaningful in providing basic data for the development of an AD family management program by investigating the actor and partner effects of parenting stress and co-parenting on marital conflict as well as the control effect based on the mother's employment status. A follow-up study is suggested to develop a family management program for children with AD considering actor and partner effects of parenting stress, co-parenting, and marital conflicts and to further verify the effects. In addition, parenting stress may appear distinctively according to the severity of symptoms in children with AD, so it is necessary to grasp these relationships, and follow-up studies are needed to identify various factors to alleviate stress.    Figure 1 The dyadic model of parenting stress, co-parenting and marital conflict: the moderating effec