This study examined the associations between parenting style and health related behavior of children. Overall, significant correlations were found between the parenting constructs and parental education, BMI of the mother, emotional eating, external eating, consumption frequency of sweet foods, vegetables and soft drinks and sleep duration. Nevertheless, most of the correlations between parental constructs and health related behavior of the children were low (r < 0.3), indicating that general parenting probably operates as a more distal predictor of childhood weight-related outcomes than more proximal behavior-specific parenting practices, e.g. feeding practices. In contrast with previous research, we did not classify the parenting style into one of the four broad categories (authoritarian, authoritative, permissive, and uninvolved or neglectful), but we described different aspects of the parenting style in more detail by using five key constructs for which a score was calculated. By using for each parent–child pair the scores for the five constructs as assessed by the CGPQ, we tried to respond to most of the criticism given on previous research in the parenting area, namely, the poorly described definition of the different parenting characteristics. Moreover, using the CGPQ made it possible to distinguish between “behavioral control”, “overprotection” and “coercive control”, which allowed us to differentiate the positive aspects of “control” from the more negative aspects. As an example, we found a positive correlation between the construct “coercive control” and the consumption frequency of sweet food whereas a negative correlation was found between the construct “behavior control” and emotional eating. Although the questionnaire that we used did not classify parents in one parenting style we can state that the authoritative parenting style corresponds with the parenting constructs “nurturance”, “behavioral control” and “structure”. In the western European population the authoritative parenting style is the most prevalent parenting style , which is in line with our study, in which most of the parents scored high on the aforementioned parenting constructs.
In this study, parental education was significantly associated with two parenting constructs, i.e. “overprotection” and “structure”. Higher educated parents tend to overprotect their children more and gave them more structure than lower educated parents. This is in line with the results of previous research showing that higher parental education is positively associated with more monitoring, control and restriction ,–. No other studies were found that reported a correlation between parental education and providing structure to their children. A possible explanation for this association (parental education versus structure) is maybe the awareness of higher educated parents about the importance of structure in a child’s life and their consistency in acting predictable upon the child.
This study also indicated that mothers with a higher BMI were more controlling and overprotective towards their children. This is in contrast with the study of Wardle et al.  where obese mothers were less controlling than normal-weight mothers. It should however be noted that the study of Wardle et al.  only considered the aspect ‘control’ in the context of eating. Mothers with a higher BMI could be more controlling and overprotective in response to the fear that their children will have the same weight-related problems as themselves.
We did not find an association between the parenting constructs and the BMI nor body fat percentage of the children. The results in the existing literature on this subject are mixed. Some researchers found an association between the parenting style or some parenting constructs and the weight of the children ,,, while others did not find an association ,. One possible explanation for this is the distal relation between parenting and BMI. There are many factors that influence this relation and we were not able to consider all these factors. Furthermore, the findings of some researchers indicate that the impact of general parenting on children’s weight status depends on characteristics of both the child and the parents .
Concerning eating behavior, the results indicated that the parenting constructs of “structure” and “behavioral control” were negatively correlated with emotional eating. This is in line with the results of previous research, where more monitoring and authoritative parenting induced less emotional eating ,. In this study, there was no significant association between parenting constructs and restraint eating. However, we found a small negative correlation between the parenting construct of “structure” and “external eating”. This can possibly be explained by the fact that parents scoring high on the parenting construct “structure,” provide their children rules and boundaries. If parents impose rules about when and what children can eat, children have less possibilities to eat following an external clue. No other studies investigating the relation between general parenting and external eating were found.
In this study there was no association between the parenting constructs and physical activity. Other studies have shown that parents can influence the activity of their child by their own activity and by logistic support. But most studies did not find an associations between physical activity or sedentary behavior and general parenting –. Next, we investigated whether there was a relation between parenting and the children’s food consumption. In the current literature mixed results have been reported for specific food related parenting practices and children’s eating behavior ,,. A possible explanation is the heterogeneity of the definition of the parenting practices. In our study, parents that scored higher on “coercive control” had children that ate more frequently sweet foods. When parents score high on “coercive control”, this may possibly induce a negative atmosphere in the family. This can cause more stress for the children, leading to higher cortisol levels. Recent research showed a positive interaction between high cortisol levels and the intake of sweet foods . In contrast, the children of parents that scored high on “overprotection” ate less frequent fruit and vegetables. When we considered fruit and vegetables separately, the relation only remained for the vegetables. We did not find other studies that reported an association between excessive involvement and excessive monitoring and fruit and vegetables intake. A possible explanation for our finding is that, when parents excessively monitor their children while eating, they also put pressure on them to eat more. After all, pressure to eat is in most studies associated with fewer intakes of foods ,–. The parenting construct “structure” was in this study negatively associated with the frequency of soft drink consumption (total and light), while the parenting construct “overprotection” was positively associated with the frequency of soft drink consumption. Two kinds of soft drinks were considered: light and non-light soft drinks. When we considered the two categories of soft drinks separately the light soft drinks showed a positive association with “behavioral control”. Most studies investigating the relationship between parenting and the consumption of soft drinks only included adolescents. Only a few studies included children of 12 years or younger. These studies found a positive relation between the permissive parenting style and soft drink consumption. “Structure” can be viewed as a parenting construct that is lacking in a permissive parenting style.
To our knowledge this is the first study that investigated the relation between general parenting and children’s sleep duration. We found a significant negative association between the parenting construct of “coercive control” and the child’s average sleep duration. We could not formulate a possible explanation for this result. Further investigation to confirm and explain this result is necessary.
After correcting for the confounding factors that we chose on the basis of the existing literature and our results, some relations were no longer significant. We found a significant negative association between the parenting constructs “behavioral control” and “structure” and the amount of light soft drinks a child consumes daily as well as a significant negative association between the total soft drinks (light and regular) and the parenting construct “structure”. An association that we did not find in the univariate analysis is the positive association between the parenting constructs “nurturance” and “coercive control” with sedentary behavior in the multivariate model. Furthermore, the association between sleep duration and “coercive control” remained.
Study strengths and limitations
The current study has several strengths and limitations. By using the air displacement plethysmography technology, we could precisely measure the fat percentage of the children. Moreover, many different aspect of children’s health related behaviors were measured: dietary consumption, eating behavior, sleep duration, physical activity, sedentary behavior and body composition. This allowed us to look at many different relationships. Several significant associations were found, however, most of the correlations and regression coefficients were small, indicating that also other factors besides those investigated in this study play a role in the relation between the parenting style and the health behavior of the children, e.g. more proximal behavior-specific parenting practices such as feeding practices. More focused studies are needed to get further insight in the role of the parenting style.
As a result of this broad scope taking on board different parameters related to health behavior, we performed a vast number of statistical analyses to test many different hypotheses. Table 3 shows the results of 125 correlation analyses (5 parental constructs multiplied with 25 child characteristics). Using a p-value of 0.05, we have a high chance (99%) that we have observed a significant result just by chance (false positive). Applying a Bonferroni correction, we would need to use a p-value of 0.0004 (0.05/125); resulting in the fact that none of correlations described in Table 3 would be considered significant. However, the Bonferroni correction is known to be very conservative and increases the risk on false negative results (correlations considered to be non-significant while they are significant). Applying this correction would lead to a result in which some corrections that are significant are not considered like that. Because of the exploratory nature of this study, it was decided not to adjust for multiple testing. The main purpose of the paper was to find out whether correlations exist between the different parental constructs as measured by the newly developed CPGQ and children’s health behavior. Significant correlations between different parameters were found, indicating that it is relevant to use this instrument further on in investigations studying in depth the role of parenting style on children’s health behavior. Further investigations to confirm the results found in this study need to include studies with a longitudinal design.
Because the parenting constructs were only investigated in a cross-sectional setting we cannot be sure they are invariable. Two earlier studies investigated whether parenting style is stable. They both found that parenting style was stable for at least two years ,. Recently, the CGPQ was applied in a longitudinal design to investigate the moderating role of general parenting on the relationship between food parenting practices and children’s dietary behavior . However, given the cross-sectional design of this study, we should be careful with interpreting causality. Next, the internal consistencies (Cronbach’s alphas) for the CGPQ were quite low – particularly for structure and behavioral control. This could be due to possible heterogeneity of these constructs in this sample. Furthermore, the questionnaires were mainly completed by mothers. However, there is no evidence that this has introduced any bias. Another limitation is the self-reported parental BMI, parenting behaviors, the self-reported food consumption frequency and the self-reported physical activity and sedentary behavior. Therefore, it is likely that the present study yielded underestimates of associations between scale scores of the CGPQ and child health outcomes. Indeed, as Shiely et al.  demonstrated in their study, people tend to underreport their weight. Moreover, when reporting parenting style, food consumption and physical activity, people can have the tendency to give social desirable answers. Finally, we cannot generalize these results to other groups since the study population was mostly middle or high social class. This is a study with a rather exploratory nature and more studies are needed to establish scientific evidence on the relation between parenting and children’s health related behavior. If these results are confirmed, parents should be advised to apply the more positive parenting construct “structure”, e.g. by supervising and managing its activities and “behavioral control”, e.g. by helping their child to achieve certain goals.