Relationship between Parental Behaviors and Affective Well-being in Primary School Children

The aim of this study was to examine the relationship between perceived maternal and paternal parental behaviors (support and restrictive control) and affective well-being (positive and negative affect) in chil - dren. The study involved 166 primary school children (75 girls, 91 boys), aged 8 to 12 years ( M = 10.15, SD = 0.86), who completed the Parental Behavior Questionnaire and the Positive and Negative Affect Schedule for Children. The results showed that both maternal and paternal support and restrictive control were significantly correlated with children’s positive and negative affect. Hierarchical regression analyses revealed that, after controlling for the child’s age and gender, maternal support and restrictive control significantly predicted children’s positive affect and only maternal restrictive control predicted negative affect. However, when paternal behaviors were included in the analyses, significant predictors of chil - dren’s positive affect were maternal and paternal support and paternal restrictive control, while significant predictors of negative affect were only paternal behaviors.


Introduction
Child well-being can be defined as "a dynamic process wherein a person's physical, mental, social and material situation is more commonly positive than negative" (Minkinnen, 2013, p. 3). Until recently, well-being indicators were most-ly objective focusing on a child's physical health or poverty (Raghavan & Alexandrova, 2014). There was a reluctance in asking children how they rate certain domains of their well-being that probably stemmed from the belief that children were not capable of self-assessment. However, as Nilsson et al. (2015) argued, a child's perspective should be included in re-search but with respect to the child's experiences and cognitive capacity, which allows them to understand given information. Raghavan and Alexandrova (2014) even stated that as a child matures, subjective indicators are more appropriate. Generally, using both objective and subjective indicators of well-being would be the most convenient (Pollard & Lee, 2003).
Subjective well-being refers to one's own evaluation of his/her life and is defined through cognitive and affective components. The cognitive component includes satisfaction with life and different life domains, whereas affective component refers to emotional responses to life events (Diener et al., 1999). These emotions are disposed on two dimensions: positive affect and negative affect (Kahneman & Deaton, 2010). High positive affect means a person is feeling enthusiastic, active, full of energy, focused, while low positive affect implies the state of lethargy and sadness. High negative affect implies the state of grief and dissatisfaction including other negative emotions, while low negative affect is the state in which a person is calm and serene (Watson & Clark, 1988).
Affective well-being changes as a function of age. Olino et al. (2011) found in their longitudinal research that positive affect increased from late infancy through middle/late childhood, while negative affect decreased. In adolescence the trend changes. Weinstein et al. (2007) found that positive affect significantly declined through age 14 to 16, while negative affect remained consistent. Affective well-being is of great significance from an early age in the lives of children and adolescents. Volbrecht et al. (2007) found that in children aged 12 to 25 months positive affect was related to greater empathy, both behavioral (i.e., helping) and cognitive (hypothesis testing, i.e., the amount of effort children spent trying to comprehend their mothers' distress). Jenkins et al. (2018) found that pos-itive affect is important for pain resistance, that is, children aged 6 to 18 years with cancer recovered more quickly from pain when experiencing positive affect. High positive affect in children and adolescents also helps in forming positive relationships with peers, which is a reward-related behavior especially in adolescence (Forbes & Dahl, 2005). Forbes and Dahl (2005) stated that positive affect, more so than negative affect, is a key factor in depression development. On the contrary, negative affect is positively related to depression and anxiety (Luebbe et al., 2010), as well as to externalized problems, even in two-year-old children (Buss et al., 2014).
Well-being is affected by interpersonal, intrapersonal, social, and cultural processes (Minkkinen, 2013). It is impossible to study a child's well-being separately from his or her environment. As Bronfenbrenner (1992) stated in his ecological systems theory, child development is affected by multiple interacting systems. The closest one to the child that has a direct impact on his/her development is a microsystem which includes the child's family/parents. The impact that parents have on their child's well-being is significant and a child's well-being depends on parental ability to fulfill the physical, emotional, and social needs of a child (Schor, 1995). When a family is faced with troubles, parental behaviors help in overcoming them. However, parental behaviors can also be a risk factor that impairs a child's well-being. Parental behaviors are mostly viewed through three dimensions: warmth represents emotions that parents show to their child, behavioral control represents setting rules that are used to regulate a child's behavior, and psychological control represents parent's behaviors that they use to manipulate child's emotions and functioning (Barber, 1996).
Many studies show how different parental behaviors affect children's perceptions, emotions, and behaviors from early childhood till young adulthood. Maternal behaviors predict social behavior in preschoolers, such as conduct problems (negatively predicted by maternal inductive reasoning and positively predicted by punishment) and prosocial behavior (positively predicted by maternal inductive reasoning and warmth) (Brajša-Žganec & Hanzec, 2014). Maternal psychological control is positively associated with externalized and internalized problems in children (Olsen et al., 2002), while parental warmth is negatively correlated with externalized problems in adolescents (Eisenberg et al., 2005). Parental responsiveness negatively predicts depression, and positively predicts self-esteem, whereas psychological control positively predicts depression, and negatively predicts self-esteem (Soenens et al., 2005), as well as academic achievement (Soucy & Larose, 2000). Furthermore, parental psychological control positively predicts components of intolerance to frustration in adolescents (Filippello et al., 2018). There are differences in maternal and paternal behaviors as well as in the effect they have on a child. For example, Keresteš (2002) found that mothers show more accepting behaviors and use more behavioral and psychological control than fathers. On the contrary, Nelson and Crick (2002) found that fathers demonstrate more psychological control than mothers. They also found that paternal psychological control positively predicts relational aggression and marginally positively predicts physical aggression in girls, while it was unrelated to physical or relational aggression in boys. Maternal psychological control was also unrelated to aggression in boys, but marginally positively predicted girls' physical aggression. Smojver-Ažić and Bezinović (2011) found that maternal warmth is equally important for both boys and girls, while paternal warmth is more important for girls. Aunola et al. (2013) showed that the more mothers and fathers used psychological control the more children experienced negative emotions. Parental warmth was not correlated with negative emotions and parental behaviors in general were not correlated with positive emotions. Hankin et al. (2011) found that children who are genetically susceptible to the development of psychopathology, and whose parents use supportive parental behaviors, have higher positive affect, and vice versa. Wang et al. (2006) found that even in children at the age of two, parental behavior characterized with low psychological control is negatively associated with negative affect and is not associated with positive affect. In adolescents, parental warmth is positively, and psychological control negatively related to psychological well-being, while there is no correlation between behavioral control and well-being (Stafford et al., 2015). Moran et al. (2018) asked adults to retrospectively report their parent's behavior and found that parental warmth experienced in childhood negatively predicts negative affect, and positively predicts positive affect in adulthood. Conversely, recollections of early risky family experiences (high aggression, low warmth) predict lower well-being in emerging adulthood (Hanzec et al., 2017).
Majority of research on the relationship between affective well-being and parental behaviors concerns adolescents, often including only one parent's behaviors (usually maternal), while those involving children often neglect children's perception. Boughton and Lumley (2011) stated that parents can often be unknowing of their child's depressive symptoms, as well as of his/her positive emotional functioning. Parents can also be unwilling to honestly report their behavior towards their children. Hence, for obtaining more credible data it is important to explore the child's perspective as well. Therefore, the aim of this study was to determine the relationship between both maternal and paternal behaviors and affective well-being (i.e., positive and negative affect) in children, with parental behaviors and children's affective well-being assessed from the child's perspective. Considering the findings of previous research that addressed a link between parental behaviors and affective well-being of children, we hypothesized that maternal and paternal warmth will positively predict positive affect, and negatively predict negative affect. We also hypothesized that maternal and paternal restrictive control will negatively predict positive affect and positively predict negative affect.

Participants
A convenience sample of 166 children (75 girls and 91 boys), with a mean age of 10.15 years (SD = 0.86; TR = 8 -12), participated in this study. Participants were students in third, fourth and fifth grades of primary schools in two Croatian cities (Varaždin and Osijek). Most participants lived with both of their parents (88.6%), 6% lived with their mother and 5.4% lived with their mother and her partner.

Instruments
Parental Behavior Questionnaire (cro. Upitnik roditeljskog ponašanja, URP29; Keresteš et al., 2012) measures children's perception of parenting behaviors, separately for mothers and fathers. It consists of 29 items divided into seven subscales: Warmth (4 items, e.g., "Shows me that she/he loves me."), Autonomy (4 items, e.g., "Teaches me how to fight for myself and my ideas."), Parental Knowledge (4 items, e.g., "She/he usually knows when I am having test in school."), Inductive Reason-ing (5 items, e.g., "She/he often tells me how I am supposed to behave and what I should be like."), Permissiveness (3 items, e.g., "I easily persuade her/him to do what I want."), Punishment (5 items, e.g., "She/he punishes me by banning something (watching TV, going out, etc."), and Intrusiveness (4 items, e.g., "She/he interferes too much in my life."). The subscales are grouped into three global dimensions of parental behavior: Parental Support (Warmth, Autonomy, Parental Knowledge, and Inductive Reasoning), Restrictive Control (Punishment and Intrusiveness), and Permissiveness. The participants' task is to estimate the agreement with the items using a 4-point Likert scale ranging from 1 -not at all true, to 4 -entirely true. The total score is calculated for each dimension as the mean of the responses to the corresponding items.
In this study, children's assessments of both their mothers' and fathers' parenting behaviors were used (without the Inductive Reasoning subscale). The Cronbach's alphas for maternal and paternal Parental Support and maternal and paternal Restrictive Control in this study were .66, .79, .73, and .80, respectively. Reliability of the Permissiveness subscale was below the acceptable level, so we decided to omit this dimension from further analyses.
The Positive and Negative Affect Schedule for Children (PANAS-C, Ebesutani et al., 2012) is a 10-item self-report measure assessing positive (joyful, cheerful, happy, lively, proud) and negative (miserable, mad, afraid, scared, sad) affect in children and youth. The participants' task is to rate how often they felt in the described way in the past few weeks using a 5-point Likert scale ranging from 1 -very slightly or not at all, to 5 -extremely. The total score is calculated for the Positive Affect (PA) and the Negative Affect (NA) as the mean of the responses to the corresponding items. In this study, the Cronbach's alpha for the PA was .71 (after removing the item "lively" due to low saturation with the factor and reduction of the reliability of the PA scale) and .63 for the NA.
The instruments used were pretested. Small convenience samples of students in third, fourth and fifth grades of one primary school participated in cognitive interviews while filling out the questionnaires. In this way, we evaluated the age suitability of the content of the questionnaire and the method of testing for all age groups. These cognitive interviews showed that students had no problem reading instructions, words, phrases, or text questions. They mostly understood the words and questions in the questionnaires. Items and expressions that some students had difficulties understanding were revised to make them more understandable to students of different ages.

Procedure
Prior to data collection, approvals from the Ethical board of the Institute of Social Sciences Ivo Pilar, Croatian Ministry of Science and Education, school principals, and parents were obtained. Students completed questionnaires in schools, during class. For the youngest age group (3rd-grade students), items were read out loud to facilitate item understanding and accuracy of responding. This work has been fully supported by Croatian Science Foundation under the project "Child Well-being in Family Context (CHILD-WELL)" (IP-2019-04-6198). Data analyzed and presented in this paper are a part of a pilot-study conducted within the mentioned project.

Results
The results showed that children, in average, experience high positive affect and low negative affect. They perceived both maternal and paternal support as high, and parental restrictive control as low (Table 1).
The Kolmogorov-Smirnov test showed that distributions of results in all variables significantly differed from normal. The distribution of results of positive affect, maternal and paternal support were significantly negatively skewed, while the distribution of results of negative affect, maternal and paternal restrictive control were positively skewed. However, in absence of extreme values, and with large enough samples, regression analysis is robust against violation of the normality assumption (Schmidt & Finan, 2018). Therefore, we removed univariate and multivariate outliers and analyzed the data using regression analyses.
Correlations presented in Table 1 show that children who experienced positive affect more often tended to experience less nega- Note. 1 coded as 1 = boy, 2 = girl; * p < .05, ** p < .01 tive affect. The child's gender was significantly correlated only with negative affect, indicating that boys tended to experience more negative affect. On the other hand, the child's age was significantly negatively correlated only with positive affect, indicating that older children experienced less positive affect. All correlations between parental behaviors and indicators of children's affective well-being were significant. Maternal and paternal support were positively related to positive affect and negatively to negative affect, while maternal and paternal restrictive control were positively related to negative affect and negatively related to positive affect.
Regarding predictor intercorrelations, all but the one between paternal support and restrictive control were significant. Given that the correlation between maternal and paternal restrictive control was high, we checked for the possible issue of multicolinearity using VIF (variance inflation factor) and tolerance as indicators, with VIF larger than 10 (or 2.5 in weaker models), and tolerance less than .10 indicating multicolinearity (Senaviratna & Cooray, 2019). In this study both indicators were acceptable, tolerance larger than .44 and VIF smaller than 2.26 for all the variables used, which suggested there was no problem with multicolinearity.
In order to examine the contribution of maternal and paternal parenting behaviors to children's positive and negative affect, hierarchical regression analyses were conducted. The child's age and gender were entered in the first step of the regression as control variables. In the second step, maternal support and restrictive control were entered, while paternal support and restrictive control were entered in the third step (Tables 2 and 3). After controlling for the child's age and gender (which did not significantly contribute to the positive affect in the first step of the regression), maternal support and restrictive control explained 19% of the positive affect variance, with both parenting behaviors as significant predictors. In the third step of the regression, paternal behaviors explained additional 7% of the positive affect variance. When paternal behaviors were entered in the analysis, maternal restrictive control ceased to be a significant predictor of the positive affect, while the child's gender became a significant predictor. Considering that the gender was not significantly correlated with the positive affect, it represents a suppressor variable in this model. In the final model a total of 29% of the positive affect variance was explained, with paternal restrictive control as the best predictor, followed by maternal support, paternal support, the child's age and gender.
Regarding negative affect (Table 3), the results showed that socio-demographic variables entered in the first step explained small but statistically significant 4% of the negative affect variance, with the child's gender as the significant predictor. After controlling for the child's age and gender, maternal behaviors in the second step of the regression explained additional 18% of the negative affect variance, with both maternal behaviors as significant predictors. When maternal behaviors were entered in the analysis, gender ceased to be a significant predictor. In the third step of the regression, paternal behaviors explained additional 5% of the negative affect variance. When paternal behaviors were entered in the analysis, maternal behaviors ceased to be significant predictors of the child's negative affect. In the final model a total of 23% of the negative affect variance was explained, with paternal restrictive control as the best predictor, followed by paternal support.

Discussion
The aim of this study was to determine the relationship between maternal and paternal be- haviors (support and restrictive control) and affective well-being (positive and negative affect) in primary school children. The results only partially confirmed our hypotheses. Although both maternal and paternal behaviors were significantly correlated with children's positive and negative affect, paternal behaviors proved to be somewhat more important in predicting children's affective well-being.
As expected, individual contribution of maternal behaviors in predicting positive affect was significant. Children who perceived more support and less restrictive control from their mothers experienced more positive affect. After controlling for maternal behaviors, paternal behaviors significantly added to positive affect prediction, with children perceiving more paternal support and less restrictive control experiencing more positive affect. In the final regression model, among other parental behaviors, maternal restrictive control had no contribution in predicting the child's positive affect, while paternal restrictive control was the best predictor.
Previous studies also showed that parental warmth and support experienced in childhood is associated with higher positive affect in childhood (Hankin et al., 2011), and higher psychological well-being in adolescence and adulthood (Moran et al., 2018;Stafford et al., 2015). This relationship can be explained in relation to the self-determination theory (Soenens & Vansteenkiste, 2010). Supportive relationships with parents enable children to fulfill their need for connectedness, one of the three basic psychological needs which, when fulfilled, leads to well-being, including higher positive affect. On the other hand, parental psychological control prevents the fulfilment of the need for autonomy, also one of the basic psychological needs, which decreases the child's inner feeling for independent and active participation in the environment and leads to experiencing low positive affect.
The difference in the importance of maternal (nonsignificant) and paternal (the most significant) restrictive control in predicting children's positive affect some authors explain by the fact that fathers play a different role in the child's life, encouraging children to engage in interactive activities more than mothers (Aunola et al., 2013). Fathers encourage children more to be active and engage in fun activities, and children perceive such interactions with their fathers as emotionally supportive. That is, the relationship with the father is perceived positive if he allows them to participate independently and undisturbed in different activities. Therefore, if fathers use restrictive control, which violates children's autonomy and hinders their free exploration of the world, it can have more effect on children's experience of positive affect. Thus, the greater importance of paternal restrictive control for positive affect in children over maternal control may stem from the way fathers normally spend time with their children. Also, it can stem from the traditional parental roles nurtured by (Croatian) culture, with fathers more often taking the role of a disciplining parent, while mothers are nurturers expressing more care and warmth. Because the child is used to being disciplined by the father, experiencing paternal restrictive control leaves greater consequences on the child's affective well-being.
Regarding the prediction of the child's negative affect, individual contribution of maternal behaviors was significant. Children who perceived less support and more restrictive control from their mothers experienced more negative affect. After controlling for the maternal behaviors, paternal behaviors significantly added to the prediction, with children perceiving less paternal support and more restrictive control experiencing more negative affect. In the final regression model only paternal behaviors significantly predicted chil-dren's negative affect, with paternal restrictive control being the best predictor.
These results are in line with previous studies showing negative relations between both maternal and paternal warmth and support and children's negative affect (Boughton & Lumley, 2011;Moran et al., 2018). Previous studies also found that both maternal and paternal psychological control positively predicted children's negative emotions, but with stronger effects of paternal control (Aunola et al., 2013), while some showed that maternal, but not paternal control, positively predicted depressive symptoms (Boughton & Lumley, 2011). Aunola et al. (2013) pointed out that, although both maternal and paternal control significantly positively predicted negative emotions, maternal control was more stable, while paternal changed from day to day, with fathers adjusting their use of restrictive control according to their child's negative emotions. It is possible that this fluctuation of the paternal control contributes more to the child's negative affect because the father's behavior is not consistent, the child does not know what reactions to expect from the father, which increases frustration and confusion. On the other hand, as for the positive affect, paternal control could be more important for the child's negative affect because fathers usually support children and contribute to their development through play through which they encourage autonomy, while mothers contribute more through care (Cabrera et al., 2007). Therefore, because of the child's expectations from the relationship with the father, when fathers use restrictive control which discourages the child's autonomy it is more reflected in the child's negative affect.
Although previous studies mostly examined maternal behaviors and pointed to the importance of maternal parenting for the child's development and well-being, the importance of paternal behaviors in addition to mater-nal behaviors in this study show that fathers have an important role in their children's affective well-being. Cabrera et al. (2007) emphasized that paternal behaviors should be viewed separately from maternal behaviors and that models of paternal behaviors should be developed in a way that will assume some specific mechanisms by which fathers affect children's development. Paternal behaviors should also be viewed within the cultural context. It is possible that by changing cultural norms about what kind of parental behavior is expected of mothers and fathers, children's perceptions and expectation of parental behaviors would change too, as well as their individual and combined significance for children's development and well-being.
The present study used children's assessments of both their mothers' and fathers' parenting behaviors as well as their self-rating of affective well-being, while many previous studies mainly used parental reports. However, although child reports are an advantage, they are subjective and can reflect distorted memories, social desirability, or be under influence of other factors and situations in the child's life. Therefore, future studies should include both parental and child reports to compare the results on the relationship between parental behaviors and child well-being depending on the source of the data on both. Given the significant correlations between gender and negative affect, as well as gender and paternal restrictive control, which indicate that boys tend to experience more negative affect and perceive their fathers as more restrictive and controlling, future studies with large enough samples of boys and girls should examine gender differences or gender as a moderator in the relationship between parental behaviors and child well-being outcomes. Parental practices of mothers and fathers often differ according to the child's gender, but boys and girls could also per-ceive certain behaviors of their mothers and fathers differently (Smojver-Ažić & Bezinović, 2011). Therefore, it is possible for maternal behavior to be more important for affective well-being in girls, and paternal behavior for affective well-being in boys, or as suggested by Nelson and Crick (2002), and Smojver-Ažić and Bezinović (2011), paternal practices could be more important for girls than for boys. Correlational and cross-sectional study design used in the present study does not allow causal conclusions. It is also possible, and previously shown, that child affect affects parental behaviors. It is important to study the mechanisms that could mediate the relationship between certain parental behaviors and positive and negative affect. Variables such as meeting the child's needs or the child's perception of parental acceptance/rejection, which is a result of certain parental behaviors, should be included as suggested by some previous studies (e.g., Van der Kaap-Deeder et al., 2017). Furthermore, longitudinal studies would be beneficial in providing information on the change in the relationship of maternal and paternal behaviors with child well-being over time and in different developmental periods, which is important to consider since the affective well-being changes as a function of age (Olino et al., 2011;Weinsten et al., 2007).
To conclude, the results of the present study showed that both maternal and paternal behaviors are important for the child's affective well-being, with paternal restrictive control being the most important predictor of both positive and negative affect of children. These results could be used as a basis for different interventions aimed at encouraging desirable parental behaviors (warmth and support) and reducing those that could impair the child's affective well-being (psychological/restrictive control), with an emphasis on the active involvement of fathers in all interventions and programs.