Associations between parenting behaviors and children’s Body Mass Index: The mediating effect of children’s lifestyle
Winneke A. van der Schuur University of Amsterdam
July 12th, 2013
Author Note
This second master thesis was written for the Research Master in Education Sciences at the University of Amsterdam, under supervision of dr. L. (Loes) van Gelderen and
Abstract
High prevalence estimates of childhood overweight and obesity are worrisome since
childhood obesity is related to both health and psychological problems. To counteract these problems it is necessary to improve the knowledge about determinants of children’s weight status. Previous studies that focused on the impact of parenting behaviors on childhood obesity mainly addressed isolated determinants of children’s lifestyle and/or children’s weight status. Therefore, the current study aims to investigate the mediating effect of children’s lifestyle on the associations between several parenting behaviors and children’s weight status. Data was collected using questionnaires. The sample consisted of 245 parents (219 mothers and 26 fathers; age range, 28-54 years) and 292 children (148 boys and 144 girls; age range, 8-13 years). Multiple sample path analyses were conducted to estimate a hypothesized model, controlled for parents’ BMI and parents’ educational attainment. Findings showed that
children’s lifestyle may mediate the relationship between parenting behaviors and children’s weight status. However, parameter estimates indicated that only parental control was associated with children’s lifestyle. Furthermore, inconsistent with the expectation, a healthier lifestyle was associated with a higher BMI. Although this study provides some knowledge about the mediating effect of children’s lifestyle on the association between parenting behaviors and children’s weight status, more studies are needed improve the reliability of the measurement and to better understand the causality between the constructs.
Associations between parenting behaviors and children’s Body Mass Index: The mediating effect of children’s lifestyle
Prevalence estimates indicate that at least 11.7% of the children in economically developed countries are overweight or obese (De Onis, Blössner, & Borghi, 2010). For example, in 2009-2010, 31.8% of the American children were overweight and 16.9% of the American children were obese (Ogden, Carroll, Kit, & Flegal, 2012). Another example is the prevalence of overweight and obesity in the Netherlands, the country in which the current study was conducted. In 2009, 13.0% of the Dutch children were overweight and 2.1% of the Dutch children were severely overweight or obese (CBS, 2013). Moreover, studies indicated that during 1990-2010 the worldwide prevalence of childhood overweight and obesity
increased with 2.5% (De Onis et al., 2010). Similarly, during 1980-2009 in the Netherlands the prevalence of childhood overweight and obesity increased two to three fold (Schönbeck et al., 2011).
The previous mentioned findings are problematic because review studies showed that obese children are at risk for developing health and psychological problems. For example, Deckelbaum and Williams (2012) highlighted that obesity in children may result in various health problems such as adult obesity, noninsulin-dependent diabetes mellitus (type 2 diabetes), and cardiovascular problems, like high blood pressure or elevated insulin levels. With regard to psychological problems, childhood obesity is associated with for example, low self-esteem and low quality of life (e.g., Deckelbaum & Williams, 2012; Griffiths, Parsons & Hill, 2010). To counteract these problems a better understanding of the determinants of children’s weight status is needed. Particularly, since it is easier to modify determinants that influence weight status during childhood compared to later life (Parsons, Power, & Manor, 2001; Trost, Sirard, Dowda, Pfeiffer, & Pate, 2003).
Several studies that focused on the etiology of childhood obesity found that it is likely to be multi-determined: multiple individual and environmental determinants of childhood overweight and obesity were indicated (see reviews of Monasta et al., 2010; Pate et al., 2013). With regard to individual determinants, Haworth, Plomin, Carnell, and Wardle (2008) studied same-sex twin pairs to measure the genetic influence on Body Mass Index (BMI). They found that BMI is highly heritable, which highlights an important genetic contribution to the development of childhood overweight or obesity. In addition, Rhee (2008) noted that obesity is associated with children’s lifestyle, especially the imbalance between energy intake and expenditure. This imbalance is mainly determined by children’s food intake and physical activity. Environmental determinants that contribute to the development of childhood obesity are for example educational attainment, wealth, and poverty (e.g., McCurdy, Gorman, & Metallinos‐Katsaras, 2010; McMullan & Keeney, 2013; Sabin & Juonala, 2013; Swartz & Puhl, 2003). This is shown by the review study of McMullan and Keeney (2013). They highlighted that socio-economic status is related to childhood overweight and obesity. In particular, the literature consistently provides evidence that a low socio-economic status is associated with an unhealthier lifestyle and higher levels of childhood overweight or obesity, compared to a medium or high socio-economic status.
Genes and above mentioned environmental determinants are extremely difficult to modify (Sabin & Juonala, 2013). One thing that might be easier to modify are parenting behaviors (e.g. Golan & Crow, 2008). Therefore, this study addressed associations between parenting behavior and children’s weight status (partially) mediated through children’s lifestyle. These parenting behaviors are often divided into two dimensions, parental control and parental support (e.g., Deković, Groenendaal, Noom, & Gerrits, 1996). In addition, we also addressed feeding-specific and activity-specific parenting control, which refers to the level of parental control of children’s eating and physical activity behaviors. It is especially
interesting to investigate the impact of these specific parenting behaviors since such behaviors may even have a bigger influence on children's eating behaviors or physical activity,
compared to parenting behaviors in general (Ventura & Birch, 2008). Several studies did investigate the impact of various parenting behaviors on children’s lifestyle and/or children’s weight status.
First, associations between parenting behaviors, children’s lifestyle and children’s weight status were investigated by several studies. Review studies showed that parenting behaviors were associated with children’s lifestyle and weight status (Skouteris et al., 2012, Sleddens, Gerards, Thijs, De Vries, & Kremers, 2011; Trost & Loprinzi, 2011; Ventura & Birch, 2008; Golan & Crow, 2008). Nevertheless, inconsistent results between studies were often found. To illustrate, Chen and Kennedy (2004) found a positive relation between authoritative parenting styles, which refer to high levels of parental control and high levels of parental support, and children’s BMI. In contrast, Rhee (2008) indicated that authoritarian parenting styles, referring to high levels of parental control and low levels of parental support, were more likely to increase children’s weight status, compared to authoritative parenting styles. Another example of inconsistent results is that some studies showed positive effects of authoritative parenting styles on fruit and/or vegetable intake (Kremers, Brug, De Vries, & Engels, 2003; Lytle et al., 2003, Pearson, Atkin, Biddle, Gorely, & Edwardson, 2010), while others studies found no effects (De Bourdeaudhuij et al., 2009; Vereecken, Legiest, De Bourdeaudhuij, & Maes, 2009).
Second, several studies indicated that feeding-specific parenting styles were associated with children’s eating, activity behaviors, or weight status (e.g., Skouteris et al., 2012;
Ventura & Birch, 2008; Rhee, 2008; Hurley, Cross, & Hughes, 2011; Trost & Loprinzi, 2011; Golan & Crow, 2008). For example, according to Hurley and others (2011) associations between parental feeding and weight status were frequently indicated. Especially
feeding-specific control was found to be related with child weight status. Moreover, authoritative feeding behaviors were positively associated with consumption of vegetables (Patrick,
Nicklas, Hughes, & Morales, 2005). In addition, several studies showed that activity-specific parental behaviors are related to children’s physical activity, possible via for example
stimulation and monitoring of children’s activity (e.g., Sleddens et al., 2012; Trost & Loprinzi, 2011).
These previous mentioned studies on the impact of parenting behaviors on children’s weight status or children’s lifestyle mainly addressed the isolated determinants of childhood obesity and did not consider the mediating effect of children’s eating and activity behaviors. According to Ventura and Birch (2008), investigation of the mediation effect of children’s lifestyle is essential because it is assumed that parenting does not directly affect children’s weight status. They highlighted that to examine the association between parenting behavior and children weight status including both children’s lifestyle and weight status is needed. The exception is the study of Gubbels and colleagues (2011). They did examine the mediating effect of children’s lifestyle on the association of child and parenting background
characteristics on parental control and as well as the development of children’s weight status. It was found that monitoring children eating behaviors and stimulating children’s physical activity was related to a healthy lifestyle and positive weight status development (Gubbels et al., 2011). Important limitations of this study are that these researchers only attended to the perspective of the parents and one dimension of parenting styles, parental control. Thus, less in known about the mediating effect of children’s lifestyle on the associations between different parenting behaviors and children’s weight status including both the perspective of parents and children.
To fill this knowledge gap the present cross-sectional study aims to investigate the mediating effect of children’s lifestyle on the associations between parenting styles on
children’s weight status according to parent and child report. The research question is: Is the association between parenting behaviors (general parental control, general parental support, feeding-specific parental control, activity-specific parental control) and children’s weight status (partially) mediated through children’s lifestyle (children’s eating behaviors, children’s activity behaviors). This research question will be answered by evaluating a hypothesized model (see Figure 1).
INSERT FIGURE 1 HERE.
The model assumes direct associations between parenting behaviors and children’s lifestyle. Previous studies found inconsistent results for the magnitude of these associations (Ventura & Birch, 2008; Sleddens et al., 2012). Higher levels of parental control or parental involvement could be related to a healthier lifestyle and weight status of children, compared to lower levels of parental control (Sleddens et al., 2012). However, higher levels of parental control or parental involvement could also be associated with an unhealthier lifestyle and weight status of children, compared to lower levels of parental control. For example, Faith, Scanion, Birch, Francis, and Sherry (2004) suggested that parental over control may have an adverse effect on children’s lifestyle because it may decrease children’s self-regulation. Therefore, positive or negative relations between parenting styles and children’s lifestyle or weight status could be found.
Children’s lifestyle is expected to be associated with children’s weight status. In general, it is found that unhealthy lifestyles are related to a higher weight status of children (e.g., Rhee, 2008). Children who eat unhealthy and report low levels of physical activity are expected to have an unhealthy weight status compared to children who eat healthier and report higher levels of physical activity. Previous studies did not investigate this mediation model, therefore both partial and full mediation models are evaluated.
Findings of the current study will contribute to a better understanding of the etiology and importance of the environment of childhood obesity. Eventually information about the etiology could provide guidelines to develop effective prevention and intervention programs. Prevention and intervention programs focused on parenting behaviors may improve children’s weight status, possibly mediated through a healthier lifestyle of children.
Method Recruitment
Parents and children were recruited via primary schools. In the Netherlands, sixty primary schools were approached and asked if they were willing to participate in this study. The schools were informed about the study’s content and relevance by means of an
information letter. After approximately one week, the approached schools were contacted by members of the research team. Nine schools in total were willing to participate. As soon as the schools indicated that they would participate parents were contacted by using an
information letter that was handed out to all the children in grade 3 to grade 6. The letter informed parents about the content of the present study. Parents needed to give active informed consent for their and their children’s participation using a consent form that was included in the information letter. In addition, the parents were asked to fill-in their e-mail address for their participation in the online questionnaire. Three-hundred-and-forty parents gave active consent. Approval was obtained from the ethical committee within the faculty of social and behavioral sciences of the University of Amsterdam.
Participants
The total sample consisted of 245 primary caregivers (219 mothers and 26 fathers) and 292 children (148 boys and 144 girls). Mostly there was one participating child per caregiver. In 19 cases two children of the same caregiver participated. The mean age of the primary caregivers was 41.8 (SD= 4.2; age range, 28-54 years). Ninety-one percent (n = 222) of the
caregivers indicated that they were Caucasian, the vast majority had a Dutch ethnic background. Fourteen percent (n = 35) of the parents had a low education (elementary school- lower vocational level); 54% (n = 132) had an average education (intermediate vocational level - pre-university level); and 25% (n = 60) had a high education (university of professional level – university level). One parent was underweight (BMI <18.5); 63% (n = 156) had a normal weight (BMI 18.5 to <25); 30% (n = 74) were overweight (BMI 25 to <30); 6% (n = 14) were obese (BMI = or > 30).
The mean age of the children was 10.2 (SD= 1.2; age range, 8-13 years). Fourteen percent (n = 41) of the children were in grade 3; 28% (n = 83) were in grade 4; 30% (n = 187) were in grade 5; and 28% (n = 81) were in grade 6. Ninety-eight percent (n = 285) of the children had a Dutch ethnic background. Three percent (n = 10) of the children were underweight (<5th BMI percentile); 8% (n = 22) were at risk for being underweight (5th to <15th BMI percentile); 83% (n = 242) had a normal weight (15th to <85th BMI percentile); 3% (n = 10) were overweight (85 to <95th BMI percentile); 3% (n = 8) were obese (= or > 95th BMI percentile).
Procedure
Cross-sectional quantitative data was collected by using questionnaires for parents and/or children. These questionnaires consisted mainly of standardized items about parenting behaviors and children’s lifestyle. Parents completed their questionnaires online. Members of the research team send an e-mail to ask the parents to complete the questionnaires using a direct link. The parents needed on average approximately 25 minutes to complete their questionnaires. Children filled in their paper pencil questionnaires during school within their classroom under supervision of their teacher and/or a member of the research team. The children needed plus-minus 40 minutes to fill in their questionnaires.
Parenting behaviors. According to the literature, parenting styles in general refer to attitudes and behaviors of the primary caregiver in interaction with their child. The
questionnaires of parents and children were designed to measure two dimensions of parenting styles, namely: parental control and parental support. Parental control refers to the behavioral control parents have over their child. Parental support refers to, for example, the parental responsiveness and involvement to the child (e.g., Deković, Groenendaal, Noom, & Gerrits, 1996).
Parental control. Four scales were used to measure parental control. First, we included
three scales of the parent and child version of the Nijmeegse Opvoedingsvragenlijst (NOV) (Gerrits, Vermulst, Van Boxtel, Janssens, Van Zutphen, & Felling, 1993 adapted by Gerrits, Groenendaal, Dekovi, & Noom, 1996): 8 items about rules (e.g., ‘I take care that my child
lives according to strict rules, this makes it a lot easier for the child’), 5 items about
punishment (e.g., ‘I punish my child regularly, if that is necessary’), and 7 items about
stimulation of developing autonomy (e.g., ‘I regularly encourage my child to explore things’). Second, one scale of Peeters and Woldringh (1993) was used. This scale consisted of 6 items about parental explanation (e.g., ‘I try to explain as much as possible to my child’) and was only standardized for parents; therefore, items were adapted for the children by members of the research team. Cronbach's alphas of these items in the reports of the children were .80, .67, .78, and .81; and of the parents were .78, .66, .78, and .80.
Parental support. To measure parental support three scales were included. First, we
included the 7 items about responsiveness (e.g., ‘I take care that my child lives according to
strict rules, this makes it a lot easier for the child’) of the NOV. Second, 6 items about
parental warmth (e.g., ‘I encourage my child to talk about his/her problems’) were used of the Vragenlijst Opvoedings-Dimensies (VOD) (Gerrits, Groenendaal, Deković, & Noom, 1996). Third, we used one scale about parental involvement, that consisted of 7 items (e.g., ‘If my
child is not able to do something, I will do it for him/her’), of Vergeer (1987) (adapted by
Gerrits, Deković, Groenendaal, & Noom, 1996). Cronbach's alphas of these items in the reports of the children were .89, .80, and .69; and of the parents were .89, .84, and .69.
All items of parenting styles in general were scored on a six-point-scale (1 = completely
disagree, 2 = disagree, 3 = somewhat disagree, 4 = somewhat agree, 5 = agree, 6 =
completely agree). In general, it is expected that higher scores on these scales indicate higher
levels of parental control and parental support. Additionally, qualitative questions were added to the parental questionnaire. Parents were asked what they found important in the rearing of their children and why this was important according to them.
Specific parenting behaviors. In the current study specific parenting behaviors are divided in feeding-specific and activity-specific parental control. Feeding-specific parental control refers to parental behavioral control of their children’s eating behaviors. Activity-specific parental control refers to the parental behavioral control of their children’s physical activity.
Feeding-specific parental control. Feeding-specific parental control was measured with
three scales of the standardized parental and child version of the Dutch Child-Feeding Questionnaire (CFQ) (Johnson & Birch, 1994 adapted and translated by Jansen, Mulkens & Jansen, 2007): 8 items about parental restrictions in their children’s eating behaviors (e.g., ‘I
have to be sure that my child does not eat to many sweets’), 4 items about parental pressure to
eat (e.g., ‘My child should always eat all the food on his/her plate’), and 3 items about monitoring their children’s eating behaviors (e.g., ‘How much do you keep track of the snack
2 food that the child eats?’). Cronbach's alphas of these items in the reports of the children
were .66, .59, and .88; and of the parents were .71, .64, and .96.
Activity-specific parental control. Activity-specific parental control was measured with
Jansen et al., 2007). The adapted parental items used by Gubbels et al. (2011) were used as a reference during this process. Eventually, 8 items about parental restrictions in their
children’s physical inactivity (e.g., ‘I have to be sure that my child does not sit behind the
computer too much’), 4 items about parental pressure to be physically active (e.g., ‘If I do not control the physical activity of my child than he/she would be less physical active than that he/she needs’), and 3 items about monitoring their children’s physical activity (e.g., ‘How much do you keep track of the amount of time your child is physical active?’) were
constructed. Cronbach's alphas of these items in the reports of the children were .69, .58, and .78; and of the parents were .69, .59, and .87.
All previous mentioned items were scored on a five-point-scale (1 = never, 2 =
seldom, 3 = half of the times, 4 = mostly, 5 = always). It is assumed that higher scores on
these scales indicate higher levels of feeding-specific and activity-specific parental control. Additionally, a qualitative question was added to the parental questionnaire: parents were asked what their role was regarding their child’s eating behaviors and physical activity.
Children’s lifestyle. Children’s lifestyle was divided into children’s eating behaviors and children’s physical activity. Children’s eating behaviors refer to the dietary intake of children, especially actual food intake and dietary patterns. Children’s physical activity was defined as any bodily movement that requires energy expenditure (Sleddens et al., 2012). Children’s eating behavior and activity were measured using standardized items of the STC nutrition and physical activity assessment and counseling instrument (e.g., Perrin et al., 2010) and the Washington State Healthy Youth Survey (Healthy Youth Survey 2008, 2010).
Children’s eating behaviors. Twelve items were focused on healthy (e.g., ‘On average how often does your child eat fruits per day?’) or unhealthy (e.g., ‘How many time per week does your child eat snacks?’) eating behaviors of children. Cronbach's alphas of
Children’s physical activity. Ten items concerned the level of physical activity (e.g., ‘How many times per week did your child exercise or participate in physical activity for at least 20 minutes that made your child sweat and breathe hard, such as basketball, soccer, running, swimming, fast dancing, or similar activities?’) and level of sedentary behaviors
(e.g., ‘On average how many hours a day does your child watch television?’) of children. Cronbach's alphas of these items in the reports of the children were .63 and .68; and of the parents were .53 and .49.
These items were scored on a five-point-scale: (1 = never, 2 = 1 or 2 times in a week, 3 =
3 or 4 times in a week, 4 = 5 or 6 times in a week, 5 = every day or 1 = zero, 2 = one, 3 = two,
4 = three, 5 = four or more). Additionally, qualitative questions were added to the parental questionnaire. The parents were asked what they found important regarding their child’s eating behavior and physical activity and why this was important to them.
Children’s weight status. Parents were asked about their children’s weight and height, these answers were used to determine children’s BMI. Children’s BMI and the corresponding BMI percentiles were calculated using the children’s BMI-percentile-for-age Calculator, USDA/ ARS Children’s Nutrition Research Center. The BMI percentiles were used in the analyses.
Analyses
Structural equation modeling (Kline, 2011) was applied to analyze associations between parenting behaviors and children’s weight status, (partially) mediated through children’s lifestyle. Multi-group path analyses were conducted with a computer program called OpenMX (Boker et al., 2011). The hypothesized model is recursive and therefore identified. Identification of this model allowed for the estimation of all model parameters with the maximum likelihood estimation method. Before interpreting the fit and parameters the converged solution was checked for inadmissibility. All solutions were admissible.
Model fit was evaluated with the chi-square measure of exact fit, at a significance level of α=.05. The model has exact fit to the data if the chi-square value is not significant (p > .05). In addition, the approximate fit index root mean square error of approximation (RMSEA) and its 90% confidence interval were evaluated. RMSEA values below .08 or .05 indicate respectively satisfactory or close fit (Browne & Cudeck, 1993, cited by Kline, 2011). If the upper bound value of the 90% confidence interval was above .08, correlation residuals were analyzed and if absolute values exceeded .10 the highest correlation residual was taken into account for confirmatory model modification. An advantage of multiple sample models is the possibility to simultaneously estimate models, which provides opportunities to test the equality of parameters across groups. Therefore, the equality of direct effects across groups was tested.
The chi-square difference and Expected-Cross Validation Index (ECVI) difference were assessed to analyze if the model fit significantly improved or significantly deteriorated from the previous model. For the chi-square difference test a significance level of α = .05 was used and the ECVI difference was significant if the 90% confidence interval did not include zero. In addition, the Root Deterioration per Restriction (RDR) was assessed, a RDR below .05 indicated essentially equivalent fit and a RDR above .08 indicated not essentially
equivalent fit. The significance of the parameter estimates was calculated with the standard errors, using a significance level of α = .05. The total and indirect effects were significant if the 95% confidence intervals did not include zero.
Results Preliminary analyses
Before estimating the hypothesized model preliminary analyses were conducted. First, missing value analyses were performed. First, assumption were checked and found to be satisfactory. However, missing values were indicated. The little’s MCAR tests showed that
the missing values were completely at random for both groups (Parent reports: χ2 (603) = 635.05, p = .177, n = 245; child reports: χ2 (664) = 658.40, p = .554, n = 292), so further analyses with these estimated missing values is plausible. Second, a multivariate analysis of variances was applied to examine parent-child differences in their perceptions of parenting behaviors and children’s lifestyle. The variables children’s BMI, parents’ BMI, and parents’ educational attainment were not included in the analyses, because these variables were only measured via parent reports. Multivariate analysis of variance showed a significant difference between parent and child reports, F (6, 530) = 27.99, p <.001, ηp2 = .24. Therefore, univariate analyses of variance were performed to investigate specific differences between perceptions of primary caregivers and children.
INSERT TABLE 1 HERE.
Although all effect sizes were small it was found that parents reported significantly higher levels of parental control (F (1, 535) = 97.29, p < .001, ηp2= .15), parental support (F (1, 535) = 21.48, p <.001, ηp2 = .04), feeding-specific parental control (F (1, 535) = 3.89, p =.05, ηp2= .01), activity-specific parental control (F (1, 535) = 61.64, p < .001, ηp2 = .10), and children’s healthy eating behaviors (F (1, 535) = 25.73, p < .001, ηp2= .05). The means and standard deviations for the study variables are presented in Table 1. Pearson’s correlations among the study variables and covariates for parents (above diagonal) and children (below diagonal) are displayed in Table 2.
INSERT TABLE 2 HERE. Multi-group path analyses
Model modification. To test if parenting behaviors were associated with children’s weight status, mediated through children’s lifestyle, the hypothesized model was estimated (Model 1). The findings indicated no exact fit of the hypothesized model, χ2(14)= 44.27, p < .001. The RMSEA did indicate approximate satisfactory fit, RMSEA= .06 (90% CI [.04 ;
.09]). However, the upper bound of the 90% confidence interval of the RMSEA indicated neither close nor satisfactory approximate fit. This finding showed that model modification was needed.
The correlation residuals were analyzed separately for parents and children to provide information for possible model modification. The highest correlation residual (.16) was found between the variables children’s healthy eating and activity behaviors for both parent and child reports. Based on the theoretical background a covariance between healthy eating and activity behaviors was added to the hypothesized model (Model 2). The estimates of the model with this covariance indicated no exact fit, χ2(12)= 26.73, p = .008, but did indicate approximate satisfactory fit, RMSEA= .05 (90% CI [.02 ; .07]). Moreover, difference tests showed a significant difference or not essentially equivalent fit between the models,
χ2
D(2)=17.53, p <.001, ECVID= .03 (90% CI [.01 ; .06]), RDR= .12 (90% CI [.07 ; .18]). This indicated that the model fit improved with addeding the covariance between healthy eating and activity behaviors. Therefore, this model was accepted as the full mediation model.
To test whether the assocations between parenting behaviors and children’s weights status are fully mediated or partially mediated through children’s lifestyle, direct effects between parenting behaviors and children’s BMI were added to the model (see the dotted lines in Figure 1). This model is referred to as the partial mediation model (Model 3). The partial mediation model indicated no exact fit, χ2(8)= 23.31, p = .002. The RMSEA indicated approximate satisfactory fit, RMSEA= .06 (90% CI [.03 ; .09]). Addition of the direct effects to Model 2 did not significantly improve the model fit or indicated essentially equivalent fit,
χ2
D(4)= 3.42, p =.490, ECVID= -.01 (90% CI [-.01 ; .01]), RDR= .00 (90% CI [.00 ; .06]). In conclusion, the full mediation model was preferred over the partial mediation model. Further analyses were based on the full mediation model.
For the full mediation model it was tested if the direct effects across groups were similar (Model 4). Specifically, equality constraints on the direct effects were imposed. This model did not fit exactly, χ2(34)= 68.35, p < .001, but did indicate approximate close fit, RMSEA = .04 (90% CI [.03 ; .06]). However, adding the equality constraints significantly deteriorated the model fit, χ2D(22)= 41.61, p = .006, ECVID = -.00 (90% CI [-.03 ; .04]), RDR = .04 (90% CI [.02 ; .06]). This implies that not all the direct effects across groups were equal. Therefore, one of the constraints on the direct effects was removed.
Removal of the constraint of the direct effect of parental control on children’s healthy eating behaviors (Model 5) resulted in the most acceptable model fit, χ2(33)= 58.93, p = .003, RMSEA = .04 (90% CI [.02 ; .05]). Difference tests indicated that the model fit did not significantly deteriorated anymore or indicated essentially equivalent fit, χ2D(21)= 32.20, p = .056, ECVID = -.02 (90% CI [-.04 ; .02]), RDR = .03 (90% CI [.00 ; .05]). This model was selected as the final model. Important fit statistics of the model selection proces are displayed in Table 3.
INSERT TABLE 3 HERE
Parameter estimates. Inconsistent with the expectations, no significant effects were found for the direct effects of parental support, feeding-specific parental control, and activity-specific parental control on children’s healthy lifestyle behaviors. In contrast, the direct effect of parental control on children’s healthy activity behaviors was significant across groups. Higher levels of parental control were related to healthier activity behaviors of children according to both parent (.08) and child reports (.12). Furthermore, for the parental reports the effect of parental control on children’s healthy eating behaviors was significant. Higher levels of parental control were associated (-.14) with less healthy eating behaviors of children. For child reports the effect of parental control on children’s healthy eating behaviors was not significant. Lastly, in both groups the direct effect of healthy eating behaviors on children’s
BMI was significant. Inconsistent with the expectations, healthier eating behaviors were related to higher levels of children’s BMI in both parent (.09) and child reports (.12).
INSERT TABLE 4 HERE.
Results of the analyses of indirect effects between parenting behaviors and children’s weight status confirmed the previous findings. Only the indirect effect (-.05) of parental control on children’s BMI was significant according to reports of the parents. In total, higher levels of parental control resulted in a lower BMI of children. The final multiple sample path model explained 15% of the variance in children’s BMI according to parent reports and 17% of the variance in children’s BMI according to child reports. Table 4 provides an overview of all important parameter estimates. Standardized parameter esitmates of the effects in the models of parents and childeren are provided in Figure 2a and Figure 2b.
INSERT FIGURE 2a HERE. INSERT FIGURE 2b HERE.
Previous mentioned results seem to provide some indications of the importance of the mediation of children’s lifestyle in the relation between parenting behaviors and children’s weight status, based on the model fit and explained variance of children’s BMI. However, only a few significant direct and indirect effects were found. The impact of the covariates, parents’ BMI and parents’educational attainment, on the model could explain these results. First, parents’ BMI had a significant effect on activity-specficic parental control (.16/.12) and children’s BMI (.360/.374). A higher BMI of parents was related to higher levels activity-specific parent control and children’s BMI. Second, parents’ educational attainment was significantly related to parental control (-.09/-.17), parental support (-.09/-.19), children’s eating behaviors (.29/.21), children’s activity behaviors (.12/.08) and children’s BMI (-.11/-.11). Higher educational attainment of parents was related to healthier eating and activity behaviors of children. In addition, higher educational attainment was associated with lower
levels of parental contol, parental support, and children’s BMI. Moreover, a significant indirect effect of parents’ educational attainment on children’s BMI (0.02/0.20) was found. Higher levels of parents’ educational attainment were related to a higher BMI of children. The important impact of these covariates on the constructs in the model were consistent with the expectations and could explain the model fit and explained variance of children’s BMI. Table 5 provides an overview the associations between the covariates and constructs in the hypothesized model.
INSERT TABLE 5 HERE. Qualitative analyses
Qualitative questions that were included in the parental questionnaire about the importance of children’s eating behaviors, children’s activity behaviors, and their role in children’s lifestyle were aswered by 75% (n = 184) of the parents. Exploratory analyses of these reports were performed. The results of this analyses about children’s eating and activity behaviors will be discussed consecutively.
Children’s eating behaviors. Several aspects of children’s eating behaviors were important according to reports of the parents. The majority (74%, n = 136) of the parents reported that healthy eating behaviors of their children are essential–“It is important that my
child eats healthy”. For example, 41 (22%) parents acknowlegdge that their children must eat
fruits and/or vegetables– “I find it important that my child eats more vegetables”. In addition, 45% (n = 85) of the parents mentioned the importance of variation and/or discovering
unknown food–“We find balanced foods very important”. Furthermore, 32% (n = 59) of the parents emphasized the essential contribution of the structure or eating patterns of their
children to develop healthy eating behaviors–“That he eats regularly”. Lastly, eating a snacks or candy once in a while (38%, n = 70) or having family meals (7%, n = 13) were also
reported as important by parents– “Enough vegetables and fruits and not too much in
between, but something nice must also be possible”.
Parents found previous mentioned aspects, such as healthy, divers and structured eating behaviors, important for two main reasons. First, most parents (91%, n = 168) noted that these aspects are important for their children’s health–“because it is healthy”. To illustrate, they mentioned that children need divers or healthy food to grow, to have enough energy to function, and to diminish the risk of being overweight or obese– “Having breakfast
gets the motor going”. Second, 28% (n = 52) of the parents reported that structured meals,
eating heatlhy, and/or eating divers will provide guidelines for their children to adopt healthy eating behaviors in later life–“Children are developing and lay a foundation for their health
in the future”.
With regard to the parental role their children’s eating behaviors it was found that al parents feel resposible for their children’s eating behaviors. First, 76% (n = 141) of the parents highlighted that they decide what children eat and what food they are exposed to–“I
decide what food my child gets at home”. Second, parents (29%, n = 53) mentioned that they
are important because they provide guidelines or function as examples for their children’s eating behaviors–“It is my role to give a good example”. Third, 25% (n = 47) of the parents reported that they observe children’s eating behaviors and controle the intake of high fat or high sugar food to control the energy intake of their children–“Keeping an eye on my child
with regard to candy or other snacks”. Lastly, 5% (n = 10) of the parents noted that they did
not experience difficulties in guiding their children’s eating behavior because their children were willing to eat healthy autonomically–“Fortunately, my daughter loves fruits and raw
vegetables so I do not experience much trouble with guiding her in that”.
Children’s activity behaviors. It was found that all parents seem aware of the
active, and/or play outside– “I find it very important that my children work out”. In addition, they reported that activity is essential for their children’s health and/or physical condition– “Being active is healthy”. Moreover, 32% (n = 59) of the parents mentioned that it enhances their children’s development mentally or physically. To illustrate, parents often reported that for example teamsports or playing outside improves children’s social skills and/or motor control–“Being active is a good way to improve social contacts”. Furthermore, 15% (n = 8) of the parents reported that exercise is an important outlet and/or that it improves their sleep– “They sleep better when they are more active”. Surprisingly, only 14% (n = 26) of the parents explicitly mentioned that children need to exersice to control their weight status– “because it
is not healthy to be fat”. Lastly, altough all parents addressed that children need to be
physically acitive to improve their health and physical condition, 30% (n = 56) of the parents emphasized the importance of having fun, balancing activity versus rest, and/or developing their imagination–“I find it important that she has fun in being active”.
Concerning the role of the parents in their children’s activity behaviors parents were mainly focused on stimulating children’s activity instread of controlling. In particular, 54% (n = 99) reported that they mainly stimulate their children’s activity–“He needs
encouragement, otherwise he will sit too long behind a computer screen”, while 13% (n = 13)
mentioned a more controlling role–. Furthermore, parents (24%, n = 44) emphazised that they are important in stimulating their childnren’s eating behaviors by providing guidelines, function as example, and being active together– “We encourage to join a sport and to play
outside when the weather is good enough”. Finally, parents mentioned frequently (40%, n =
59) that their children’s enjoyed being active. Therefore they did not experience diffuculties in activating their children– “I do not worry about that [his activity] he is always out to cycle
or play soccer”.
The aim of the current study was to investigate the mediating effect of children’s eating and activity behaviors on the associations between various parenting behaviors and children’s weight status. Examination of this mediation effect was needed because previous studies focused on parenting behaviors mainly addressed isolated determinants of children’s lifestyle and/or childhood obesity. Results indicated that children’s lifestyle could be fully mediated the relation between several parenting behaviors (parental control, parental support, feeding-specific parental control, and activity-specific parental control) and children’s weight status, controlled for the effect of parents’ BMI and educational attainment. This finding is consistent with our hypothesis and the assumptions of Ventura and Birch (2008).
Specifically, it confirms that parenting behaviors do not directly affect children’s weight status and it highlights the importance of including children’s lifestyle when investigating the associations between parenting behaviors and children’s weight status. However, it is
important to mention that we only found a few significant associations within this model. Concerning the associations between parenting behaviors and children’s lifestyle, we indicated significant relations between parental control and children’s lifestyle. First, we found that parental control was positively associated with children’s healthy activity
behaviors according to both parent and child reports. In contrast, previous studies did not find such association (see review of Sleddens et al., 2011). This inconsistency might be caused by the age of the children, since the other studies focused on older children and adolescents. Parents may have less impact on children’s lifestyle behaviors when age increases (e.g., Pearson, Biddle, & Gorely, 2009). Second, it was found that only according to parent reports parental control was negatively associated with children’s healthy eating behaviors. Perhaps parental control has an adverse effect on children’s lifestyle due to a less developed self-regulation of these children (Faith et al., 2004). Another explanation might be that parents of children who eat less healthy are more likely to show higher levels of parental control, which
emphasizes the expected bidirectional effects between children’s lifestyle and parenting behaviors (Ventura & Birch, 2008).
It must be said that other parenting behaviors were not significantly related to children’s lifestyle. For example, parental support was not related to children’s lifestyle, which is consistent with previous studies (see review of Trost & Loprinzi, 2011). In addition, we did not find associations between feeding-specific or activity-specific parental control and children’s lifestyle. This was unexpected since previous studies indicated that feeding- or activity specific parenting behaviors have more impact on children’s lifestyle compared to general parenting behaviors (e.g. Gubbels et al., 2011; Faith et al., 2004; Trost & Loprinzi, 2011; Ventura & Birch, 2008). However, these studies did already show that the impact of eating-specific and activity-specific parenting behaviors might be limited. To illustrate, Faith and colleagues (2004) found that only parental feeding restrictions were associated with increased eating behaviors of children and children’s weight status. Furthermore, according to the review of Sleddens and others (2012) several studies did not find a relation between activity-specific parental control and children’s activity behaviors as well. Thus, the contribution of parental support, feeding-specific parental control, and activity-specific parental control to children’s lifestyle might be limited.
With regard to the direct effects between children’s lifestyle and weight status
unexpected results were found. First, children’s healthy eating behaviors were associated with children’s BMI according to both parent and child reports. However, inconsistent with our hypothesis, healthier eating behaviors were related to a higher BMI of children.
Misconceptions of children’s weight by parents could explain this finding, since previous studies showed that parents of overweight children often underestimate their children’s weight (e.g. McMullan & Keeney, 2013). These misconceptions might also clarify the
activity behaviors were not related to children’s BMI. Similar results were found in several other studies (see review of Pate et al., 2013). In sum, although there is consensus about the important impact of children’s lifestyle on their weight status, review studies found mixed results between studies. The high amount of discrepancies between studies on the association between children’s lifestyle and children’s weigh status may emphasize the methodological challenges in investigating this relation (e.g., Pate et al., 2013; Bleich, Ku, & Wang, 2010; Must & Tybor, 2005).
As expected, the covariates parents BMI and parents’ educational attainment have an important impact on several variables in the hypothesized model. Similar to the findings of Haworth and others (2008), it was found that parents’ BMI was positively associated with children’s BMI. Moreover, we found that higher levels of educational attainment were related to a healthier lifestyle and lower weight status of children. This is consistent with the results of McMullan and Keeney (2013). The impact of the educational attainment of parents on children’s lifestyle could be due to high economic stress. This highlights the role of society in reducing childhood obesity. Parents with lower educational attainment might struggle to provide healthy food and activity possibilities for their children due financial and/or social constraints. Increasing the economic resources may enhance possibilities of parents to adopt a healthy lifestyle (McCurdy et al., 2010). In addition, the stress that these parents experience may also be amplified by the constant exposure in Western society to unhealthy foods and sedentary activities (e.g., Swartz & Puhl, 2003).
An important strength of this study is that we used reports of both parents and children. Including multiple perceptions could increase the reliability of the measurements. Another strength is that we included multiple parenting behaviors, covariates, and the
parenting behaviors on children weight status, did not included children’s lifestyle, or neglected the importance of covariates (e.g. Ventura & Birch, 2008).
Next to these strengths this study has several limitations, which provide important directions for future research. First, the reliability of some measurements is questionable. To diminish social desirability bias future studies could include more reliable measurements to measure children’s lifestyle. For example, pedometers or accelerometers can be used to measure children’s activity and diaries to measure children’s eating behaviors (e.g. Sleddens et al., 2012). Furthermore, the measurement of children’s weight status could be improved by measurements of weight and height by trained researcher and ideally should also include measurements of the waist circumference and skinfold thickness (Freedman & Sherry, 2009). Second, it is important to mention that the results only addressed associations between
parenting behaviors, children’s lifestyle, and children’s weight status and not the causality between these constructs due to the cross-sectional design of this study. Ventura and Birch (2008) highlighted the bidirectional effects of parenting behaviors and child characteristics. Previous studies did indicate that children’s weight status might influence parenting
behaviors. Therefore, experimental and longitudinal studies are needed to address the causality between these constructs.
Next to these methodological changes in future research, these results provide indications to focus more on the importance of the broader societal system instead of
parenting behaviors. Although qualitative analyses showed that parents have the knowledge about aspects that are important in children’s healthy eating and activity behaviors, they still could struggle with their own lifestyle and weight and/or not be able to provide healthy lifestyle opportunities for their children. In particular, variables like educational attainment and weight status of the parents seem essential in the development of childhood obesity. One essential explanation could be that the Western society is preoccupied with unhealthy food
which makes it extremely challenging for parents and children to adopt a healthy lifestyle. Logically, parents seem not able to solve this societal problem on their own (Swartz & Puhl, 2003). Future research needs to focus on the impact of various determinants in the
obesogenic environment. Practically, parent’s might be supported by for example providing financially attractive food and activity possibilities, reducing children’s exposure to
advertising of unhealthy food, and learning them how to implement their knowledge about healthy eating and activity behaviors.
In conclusion, findings of the current study contribute to a better understanding of the mediating effect of children’s lifestyle on the association between parenting behaviors and children’s weight status. However, this study provides limited evidence of the importance of parenting behaviors on children’s lifestyle and weight status. Only parental control was associated with children’s lifestyle. It does indicate that it is essential to include multiple perspectives, children’s lifestyle, and covariates, such as parental BMI and socio-economic status in the investigation of determinants of childhood obesity. Future studies using reliable measurements are needed to improve our understanding of the importance broader societal system, and the causality between parenting behaviors, children’s lifestyle, and children’s weight status. Moreover, since the impact of parenting behaviors on children’s lifestyle and eventually children’s weight status might be limited, the parental role in reducing childhood obesity may be overestimated. Therefore, more attention is needed for other determinants in the obesogenic environment.
References
Bleich, S. N., Ku, R., & Wang, Y. C. (2010). Relative contribution of energy intake and energy expenditure to childhood obesity: a review of the literature and directions for
future research. International journal of obesity, 35(1), 1-15. doi: 10.1038/ijo.2010.252
Boker, S. (2011). OpenMx: An Open Source Extended Structural Equation Modeling Framework. Psychometrika, 76, 306-317. doi:10.1007/s11336-010-9200-6 CBS (2013). StatLine: Onder- en overgewicht, lengte en gewicht; vanaf 1981 [StatLine:
Under- and Overweight, length, and weight, since 1981]. Den Haag/Heerlen. Chen, J. L. & Kennedy, C. (2004). Family Functioning, Parenting Style and Chinese Children's Weight Status. Journal of Family Nursing, 109(2), 262–279. doi: 10.1177/1074840704264021
De Bourdeaudhuij, I., Te Velde, S. J., Maes, L., Perez-Rodrigo, C., De Almeida, M. D. V., & Brug, J. (2009). General parenting styles are not strongly associated with fruit and vegetable intake and social-environmental correlates among 11-year-old children in four countries in Europe. Public health nutrition, 12(2), 259–266. doi:
10.1017/S1368980008002930
Deckelbaum, R. J. & Williams, C. L. (2001). Childhood Obesity: The Health Issue. Obesity
Research, 9(S11), 239S–243S. doi: 10.1038/oby.2001.125
De Onis, M., Blössner, M., & Borghi, E. (2010). Global prevalence and trends of overweight and obesity among preschool children. The American journal of clinical nutrition,
92(5), 1257–1264. doi: 10.3945/ajcn.2010.29786
Deković , M., Groenendaal, J. H. A., Noom, M. J., & Gerrits, L. A. W. (1996). Theoretisch kader ern opzet van het onderzoek [Theoretical background and design of the study]. Ripens, J., Hermanns, J. M. A., & Meeus, W. H. J. (Red.), Opvoeden in Nederland
Faith, M. S., Scanion, K. S., Birch, L. L., Francis, L. A., & Sherry B. (2004). Parent-Child Feeding Strategies and Their Relationships to Child Eating and Weight Status. Obesity
Research, 12(11), 1711–1722. doi:10.1038/oby.2004.212
Freedman, D. S., & Sherry, B. (2009). The validity of BMI as an indicator of body fatness and risk among children. Pediatrics, 124(1), S23-S34. doi: 10.1542/peds.2008-3586E Gerrits, L. A. W., Deković , M., Groenendaal, J. H. A., & Noom, M. J. (1996).
Opvoedingsgedrag [Parenting behavior]. Ripens, J., Hermanns, J. M. A., & Meeus, W. H. J. (Red.), Opvoeden in Nederland [Parenting in the Netherlands], (pp. 41-69). Assen: Van Gorcum.
Gerrits, L. A. W., Groenendaal, J. H. A., Deković, M., & Noom, M. J. (1996). Interne
rapportage over een Nederlandse bewerking van de Parenting Dimensions Inventory (PDI) [Internal report about a Dutch adaptation of the Parenting Dimensions Inventory (PDI)]. Utrecht: Universiteit Utrecht.
Gerrits, J. R. M., Vermulst, A. A.,Van Boxtel, D. A. A. M., Janssens, J. M. A. M., Van Zutphen, R. A. H., & Felling, A. J. A. (1993). Parenting in Dutch families: A
representative description of Dutch family life in terms of validated concepts
representing characteristics of parents, children, the family as a system and parental socio-cultural value orientations. Nijmegen: University of Nijmegen, Institute of Family
Studies.
Golan, M., & Crow, S. (2004). Parents are key players in the prevention and treatment of weight‐related problems. Nutrition reviews, 62(1), 39-50. doi: 10.1111/j.1753-4887.2004.tb00005.x
Gubbels, J. S., Kremers, S. P. J., Stafleu, A., De Vries, S. I., Goldbohm, R. A., Dagnelie, P. C., De Vries, N. K., Van Buuren S., & Thijs, C. (2011). Association between
body mass index: the KOALA Birth Cohort Study. International Journal of
Behavioral Nutrition and Physical Activity 2011, 8(18), 1–13. doi:
10.1186/1479-5868-8-18
Griffiths, L. J., Parsons, T. J., & Hill, A. J. (2010). Self‐esteem and quality of life in obese children and adolescents: A systematic review. International Journal of Pediatric
Obesity, 5(4), 282–304. doi: 10.3109/17477160903473697
Haworth, C. M. A., Plomin, R. Carnell, S., & Wardle, J. (2008). Childhood Obesity: Genetic and Environmental Overlap With Normal-range BMI. Obesity, 16(7), 1585–1590. doi: 10.1038/oby.2008.240
Healthy Youth Survey 2008 (2010) Healthy Youth Survey 2008: Analytic Report. Washington State Department of Health, Office of the Superintendent of Public Instruction,
Department of Social and Health Services, Department of Commerce, Family Policy Council and Liquor Control Board.
Hurley, K. M., Cross, M. B., & Hughes, S. O. (2011). A systematic review of responsive feeding and child obesity in high-income countries. The Journal of Nutrition, 141(3), 495–501. doi: 10.3945/jn.110.130047
Jansen, E., Mulkens, S., Sanders, E. & Jansen, A. (2011). How do children experience their parents' feeding practices? A study using the Dutch Child Feeding Questionnaire - child version. Netherlands Journal of Psychology, 66(1), 26–32.
Johnson, S. L., & Birch, L. L. (1994). Parents' and children's adiposity and eating style.
Pediatrics, 94(5), 653-661.
Kline, R. B. (2011). Principles and Practice of Structural Equation Modeling, 3rd ed. New York: The Guilford Press.
Kremers, S. P., Brug, J., De Vries, H., & Engels, R. C. (2003). Parenting style and adolescent fruit consumption. Appetite, 41(1), 43–50. doi: 10.1016/S0195-6663(03)00038-2
Lytle, L. A., Varnell, S., Murray, D. M., Story, M., Perry, C., Birnbaum, A. S., & Kubik, M. Y. (2003). Predicting adolescents' intake of fruits and vegetables. Journal of Nutrition
Education and Behavior, 35(4), 170–178. doi: 10.1016/S1499-4046(06)60331-X
McCurdy, K., Gorman, K. S., & Metallinos‐Katsaras, E. (2010). From poverty to food insecurity and child overweight: a family stress approach. Child Development
Perspectives, 4(2), 144-151. doi: 10.1111/j.1750-8606.2010.00133.x
McMullan, J. & Keeney, S. (2013). A review of the literature on the social and environmental factors which influence children (aged 3–5 years) to be obese/ overweight and the accuracy of parental perceptions . Health Education Journal February, 0(0), 1–7. doi: 10.1177/0017896912471034
Monasta, L., Batty, G. D., Cattaneo, A., Lutje, V. Ronfani, L., Van Lenthe, F. J., & Brug, J. (2010). Early-life determinants of overweight and obesity: a review of systematic reviews. Obesity reviews, 11(10), 695–708. doi: 10.1111/j.1467-789X.2010.00735. Must, A., & Tybor, D. J. (2005). Physical activity and sedentary behavior: a review of
longitudinal studies of weight and adiposity in youth. International Journal of Obesity,
29, S84–S96. doi: 10.1038/sj.ijo.0803064
Ogden, C. L., Carroll, M. D., Kit, B. K., & Flegal, K. M. (2012). Prevalence of obesity and trends in body mass index among US children and adolescents, 1999-2010. JAMA: the
journal of the American Medical Association, 307(5), 483–490. doi:
10.1001/jama.2012.40
Pearson, N., Atkin, A. J., Biddle, S. J., Gorely, T., & Edwardson, C. (2010). Parenting styles, family structure and adolescent dietary behaviour. Public Health Nutrition, 13(8), 1245–1253. doi: 10.1017/S1368980009992217
Pearson, N., Biddle, S. J., & Gorely, T. (2009). Family correlates of fruit and vegetable consumption in children and adolescents: a systematic review. Public health
nutrition, 12(2), 267–283. doi: 10.1017/S1368980008002589
Parsons, T. J., Power, C., & Manor, O. (2001). Fetal and early life growth and body mass index from birth to early adulthood in a 1958 British cohort: Longitudinal study.
British Medical Journal, 323, 1331–1335. doi: 10.1136/bmj.323.7325.1331
Pate, R. R., O'Neill, J. R., Liese, A. D., Janz, K. F., Granberg, E. M., Colabianchi, N., & Taverno Ross, S. E. (2013). Factors associated with development of excessive fatness in children and adolescents: a review of prospective studies. Obesity Reviews. doi: 10.1111/obr.12035
Patrick, H., Nicklas, T. A., Hughes, S. O., & Morales, M. (2005). The benefits of authoritative feeding style: caregiver feeding styles and children's food consumption patterns.
Appetite, 44(2), 243–250. doi: 10.1016/j.appet.2002.07.001
Peeters, J, & Woldringh, C. (1993). Leefsituaties van kinderen tot 12 jaar in Nederland [Living
situations of children till 12 years in the Netherlands]. Nijmegen: Instituut voor
Toegepaste Sociale Wetenschappen (ITS).
Perrin, E. M., Jacobson Vann, J. C., Benjamin, J. T., Cockrell Skinner, A., Wegner, S., & Ammerman, A. S. (2010). Use of a Pediatrician Toolkit to Address Parental
Perception of Children’s Weight Status, Nutrition, and Activity Behaviors. Academic
Pediatrics, 10(4), 274–281. doi: 10.1016/j.acap.2010.03.006
Rhee, K. (2008). Childhood overweight and the relationship between parent behaviors, parenting style, and family functioning. The Annals of the American Academy of
Political and Social Science, 615(1), 12–37. doi: 10.1177/0002716207308400
Sabin, M. A. & Juonala, M (2013). Obesity: Predicting risk of later obesity from the first day of life. Nature Reviews Endocrinology, 9, 136–138. doi:10.1038/nrendo.2013.16
Schönbeck, Y., Talma, H., van Dommelen, P., Bakker, B., Buitendijk, S. E., Hirasing, R. A., & Van Buuren S. (2011). Increase in prevalence of overweight in Dutch children and adolescents: a comparison of nationwide growth studies in 1980, 1997 and 2009. PloS
one, 6(11), e27608. doi:10.1371/journal.pone.0027608
Skouteris, H., McCabe, M., Ricciardelli, L. A., Milgrom, J., Baur, L. A., Aksan, N., &
Dell’Aquila, D. (2012). Parent–child interactions and obesity prevention: a systematic review of the literature. Early Child Development and Care, 182(2), 153–174. doi: 10.1080/03004430.2010.548606
Sleddens, E. F., Gerards, S. M., Thijs, C., De Vries, N. K., & Kremers, S. P. (2011). General parenting, childhood overweight and obesity‐inducing behaviors: A
review. International Journal of Pediatric Obesity, 6(2), e12–e27. doi: 10.3109/17477166.2011.566339
Sleddens, E. F. C., Kremers, S. P. J., Hughes, S. O., Cross, M. B., Thijs, C., De Vries, N. K., & O’Connor, T. M. (2012). Physical activity parenting: a systematic review of questionnaires and their associations with child activity levels. Obesity reviews,
13(11), 1–19. doi: 10.1111/j.1467-789X.2012.01018.x
Schwartz, M. B., & Puhl, R. (2003). Childhood obesity: a societal problem to solve. Obesity
reviews, 4(1), 57–71. doi: 10.1046/j.1467-789X.2003.00093.x
Trost, S. G. & Loprinzi, P. D. (2011). Parental influences on physical activity behavior in children and adolescents: A brief review. American Journal of Lifestyle Medicine,
5(2), 171–181. doi: 10.1177/1559827610387236
Trost, S. G., Sirard, J. R, Dowda, M., Pfeiffer, K. A., & Pate, R. R. (2003). Physical activity in overweight and nonoverweight preschool children. International Journal of Obesity
Ventura, A. K. & Birch, L. L. (2008). Does parenting affect children’s eating and weight status? International Journal of Behavioral Nutrition and Physical Activity, 5(15), 1– 39. doi: 10.1186/1479-5868-5-15
Vereecken, C., Legiest, E., De Bourdeaudhuij, I., & Maes, L. (2009). Associations between general parenting styles and specific food-related parenting practices and children's food consumption. American Journal of Health Promotion, 23(4), 233–240. doi: 10.4278/ajhp.07061355
Vergeer, M. M. (1987). Effecten van de geboorte van een tweede kind. Twee onderzoeken naar
de interacties tussen eerstgeborenen en ouders. [Effects of the birth of a second child. To studies on the interactions between first-borns and parents]. Lisse: Swets en
Appendix
Table 1
Descriptive statistics of the included study variables.
Variables Parents (n=245) Children (n=292)
M SD Range M SD Range Parenting behaviorsa Control 4.61 0.42 3.54 - 5.75 4.10 0.71 1.81 - 5.90 Support 5.02 0.44 3.90 - 5 .95 4.76 0.78 1.57 - 5.89 Feeding-specific control 3.16 0.53 1.33 - 4.33 3.05 0.70 1.04 - 4.89 Activity-specific control 2.96 0.59 1.00 - 4.32 2.51 0.70 1.00 - 4.58 Children’s lifestylea
Healthy eating behaviors 3.66 0.31 2.33 - 4.50 3.50 0.41 2.33 - 4.50 Healthy activity behaviors 3.06 0.40 1.90 - 4.00 3.13 0.49 2.10 - 4.00
Children’s weight statusa
Children’s BMI percentile 45.25 26.56 0.00 - 99.30 47.39 23.11 0.00 - 99.30
Covariatesa
Parent’s BMI 24.11 3.61 18.20 - 45.00 24.10 3.22 18.20 - 45.00 Parent’s educational level 3.78 1.50 1.00 - 7.00 3.80 1.41 1.00 - 7.00
Table 2
Zero-order Pearson’s correlations between study variables for parents (above diagonal) and children (below diagonal).
Variables 1 2 3 4 5 6 7 8 9 Parenting behaviors 1 Control .45*** .06 .06 -.17** .11 .12 -.10 -.15* 2 Support .53*** .05 -.10 -.12 .08 -.08 -.10 -.20** 3 FS controla .33*** .20** .39*** .05 .01 .10 .12 .02 4 AS controla .32*** .10 .63*** .13 -.09 .11 .07 .13* Children´s lifestyle 5 Healthy eating .13* .13* .18** .20** .15* .08 .09 .20** 6 Healthy activity .10 .11 .08 .01 .23*** -.06 .05 .03
Children´s weight status
7 Children’s BMI -.03 -.02 -.02 .02 .11 -.02 .34*** -.08
Covariates
8 Parent’s BMI -.02 -.03 .06 .20** .05 -.13* .40*** .02 9 Parent’s education -.13* -.09 -.01 -.05 .32*** .12* -.06 .03
Note: * p<.05, ** p<.01, ***p<.001, a FS control = Feeding-specific control, AS control = activity specific control
Table 3
Values of selected fit statistics for the path models with (partial) mediation effects of lifestyle (children’s healthy eating and activity behaviors) on the impact of parenting behaviors: parental control, parental support, feeding-specific parental control, and activity- specific parental control, on children’s BMI. Analyzed across samples of parents (n = 245) and children (n = 292).
90% CI
Models χ2(df) p RMSEA χ2D(df) p ECVID Lower Upper RDR Model 1 44.27(14) <.001 .06
Model 2 26.73(12) .008 .05 17.53(2) <.001 .03 .01 .06 .12 Model 3 23.31(8) .002 .06 3.42(4) .490 -.01 -.01 .01 .00 Model 4 68.35(34) <.001 .04 41.62(22) .006 -.00 -.03 .04 .04 Model 5 58.93(33) .003 .04 32.20(21) .056 -.02 -.04 .02 .03
Table 4
Maximum likelihood parameter estimates for the final model, including the effect of parenting behaviors on children’s BMI mediated through children’s lifestyle, controlled for the effects of parents’ BMI and parents’educational attainment, analyzed across samples of parents and children.
Parents (n = 245) Children (n = 292)
Parameters Unst. SE St. Unst. SE St.
Direct effects Control Eatingb -0.11* 0.05 -.14* 0.06 0.04 .11 Control Activityb 0.08* 0.04 .08* 0.08* 0.04 .12* Support Eatingb 0.03 0.03 .04 0.03 0.03 .06 Support Activityb 0.04 0.04 .04 0.04 0.04 .07 FS controla Eatingb 0.05 0.02 .08 0.05 0.02 .08 AS controla Activityb -0.05 0.03 -.08 -0.05 0.03 -.08
Eatingb child BMI 0.36* 0.14 .09* 0.36* 0.14 .13*
Activityb child BMI -0.06 0.11 -.02 -0.06 0.11 -.02
Disturbance Variances
Eating behaviors 0.09* 0.01 .88* 0.14* 0.01 .92*
Activity behaviors 0.16* 0.01 .97* 0.23* 0.09 .97*
Children’s BMI 1.53* 0.14 .85* 1.10* 0.09 .83*
Indirect effect Unst. 95% CI Unst. 95% CI
Control Child BMI -0.045* -0.107 ; -0.004 0.017 -0.015 ; 0.058 Support Child BMI 0.009 -0.014 ; 0.037 0.009 -0.014 ; 0.037 FS controla Child BMI 0.017 -0.001 ; 0.044 0.017 -0.001 ; 0.044 AS controla Child BMI 0.003 -0.011 ; 0.021 0 .003 -0.011 ; 0.021
Note: * p<.05, a FS control = Feeding-specific control, AS control = activity specific control, b Eating = children’s eating behaviors, Activity = children’s activity behaviors
Table 5
Maximum likelihood parameter estimates for the covariates, parents’ BMI and parents’ educational attainment, in final model, analyzed across samples of parents and children.
Parents (n = 245) Children (n = 292)
Parameters Unst. SE St. Unst. SE St.
Direct effects
Parents’ BMI Control -0.10 0.06 -.08 -0.10 0.06 -.04 Parents’ BMI Support -0.12 0.07 -.10 -0.12 0.07 -.05 Parents’ BMI FS controla 0.14 0.07 .09 0.14 0.07 .06 Parents’ BMI AS controla 0.25* 0.08 .15* 0.25* 0.08 .12* Parents’ BMI Eatingb 0.06 0.04 .06 0.06 0.04 .05 Parents’ BMI Activityb -0.02 0.06 -.02 -0.02 0.06 -.01 Parents’ BMI Child BMI 1.33* 0.14 .36* 1.33* 0.14 .37* Parents’ education Control -0.05* 0.02 -.17* -0.05* 0.02 -.09* Parents’ education Support -0.05* 0.02 -.19* -0.05* 0.02 -.09* Parents’ education FS controla 0.01 0.02 .02 0.01 0.02 .01 Parents’ education AS controla 0.02 0.02 .04 0.02 0.02 .03 Parents’ education Eatingb 0.06* 0.01 .29* 0.06* 0.01 .21* Parents’ education Activityb 0.03* 0.01 .11* 0.03* 0.01 .08* Parents’ education Child BMI -0.10* 0.04 -.11* -0.10* 0.04 -.11*
Indirect effect Unst. 95% CI Unst. 95% CI
Parents’ BMI Eatingb 0.013 -0.002 ; 0.036 -0.003 -0.022 ; 0.014 Parents’ BMI Activityb -0.027* -0.054 ; -0.006 -0.027* -0.054 ; -0.006 Parents’ BMI Child BMI 0.028 -0.005 ; 0.076 0.021 -0.011 ; 0.067 Parents’ education Eatingb 0.004 -0.001 ; 0.010 -0.004 -0.010 ; 0.000 Parents’ education Activityb -0.007* -0.013 ; -0.002 -0.007* -0.013 ; -0.002 Parents’ education Child BMI 0.023* 0.005 ; 0.044 0.020* 0.004 ; 0.039
Total affects
Parents’ BMI Eatingb 0.069 -0.016 ; 0.153 0.052 -0.033 ; 0.136 Parents’ BMI Activityb -0.047 -0.157 ; 0.062 -0.047 -0.157 ; 0.062 Parents’ BMI Child BMI 1.363* 1.077 ; 1.648 1.357* 1.071 ; 1.642 Parents’ education Eatingb 0.066* 0.045 ; 0.087 0.058* 0.037 ; 0.079 Parents’ education Activityb 0.023 -0.003 ; 0.050 0.023 -0.003 ; 0.049 Parents’ education Child BMI -0.07* -0.143 ; -0.005 -0.077* -0.146 ; -0.008
Note: * p<.05, a FS control = Feeding-specific control, AS control = activity specific control, b Eating = children’s eating behaviors, Activity = children’s activity behaviors
Figure 1. A hypothesized model, which highlights the (partial) mediation of children’s
