BACKGROUND: Research suggests that general parenting dimensions and styles are associated with children’s BMI, but directionality in this relationship remains unknown. Moreover, there has been little attention to the influences of both mothers’ and fathers’ parenting. We aimed to examine reciprocal relationships between maternal and paternal parenting consistency and child BMI.
METHODS: Participants were 4002 children and their parents in the population-based Longitudinal Study of Australian Children. Mothers and fathers self-reported parenting consistency, and children’s BMI was measured at 4 biennial waves starting at age 4 to 5 years in 2004. Bidirectionality between parenting and child BMI was examined by using regression analyses in cross-lagged models.
RESULTS: The best-fitting models indicated a modest influence from parenting to child BMI, whereas no support was found for bidirectional influences. For mothers, higher levels of parenting consistency predicted lower BMI in children from Waves 1 to 2 and 3 to 4; for example, for every SD increase in mothers’ parenting consistency at Wave 1, child BMIz fell by 0.025 in Wave 2 (95% confidence interval: −0.05 to −0.003). For fathers, higher levels of parenting consistency were associated with lower child BMI from Waves 1 to 2 and 2 to 3.
CONCLUSIONS: Parenting inconsistency of mothers and fathers prospectively predicted small increases in offspring BMI over 2-year periods across middle childhood. However, child BMI did not appear to influence parenting behavior. These findings support recent calls for expanding childhood overweight interventions to address the broad parenting context while involving both mothers and fathers.
- BMIz —
- body mass index z-score
- CFI —
- comparative fit index
- CI —
- confidence interval
- LSAC —
- Longitudinal Study of Australian Children
- MLR —
- maximum likelihood estimation with robust standard errors
- PC —
- parenting consistency
- RMSEA —
- root mean square error of approximation
What’s Known on This Subject:
Parents influence their child’s overweight development through lifestyle-related parenting practices. Although broader parenting dimensions may also affect children’s BMI, reverse causality is possible and there have been calls to examine the possible impacts of fathers.
What This Study Adds:
More consistent parenting prospectively predicted lower child BMI with effects equally strong for fathers and mothers. There was little evidence of child BMI influencing parenting. Improved child BMI could be among the benefits of promoting parenting consistency of both parents.
Overweight affects ∼25% of children in Western countries,1,2 tends to be chronic,3 and conveys an increased risk to future health and well-being.4 The rapid increase in rates of childhood obesity cannot be explained by genetics and biology alone, focusing attention on potentially modifiable environmental factors, including the role of parents in the development of their children’s overweight.5,6 The family environment is the most important context for children, within which parents influence child development through providing rules, experiences, and resources, and by acting as role models.7
To date, parenting research in the childhood obesity field has primarily focused on behaviors of mothers only, and predominantly in the context of children’s feeding and physical activity.6 There is now a small body of research suggesting that parents may influence their children’s BMI through their broader approaches to parenting.6 In a recent cross-sectional analysis of 4- to 5-year-old children participating in the Longitudinal Study of Australian Children (LSAC), Wake et al8 reported that low parental consistency was associated with higher BMI in children. Intriguingly, this association held only for paternal (and not maternal) parenting consistency once accounting for possible confounding factors. The parenting dimensions of warmth and irritability were unrelated to child weight. Wake et al8 also examined Baumrind’s widely used typology of 4 parenting styles,9 and found that paternal permissive and disengaged parenting styles (both characterized by low levels of consistency) were associated with higher odds for children to be overweight as compared with authoritative parenting. However, inference on causality and direction of these associations was restricted owing to the cross-sectional data available at that time, a limitation shared by the majority of research in this area.6 Despite recent calls to examine the possible impacts of fathers as well as mothers,6,10 the few longitudinal studies conducted on this topic have only examined whether maternal parenting predicted subsequent child weight, and none have investigated the possibility of reverse causation.11–14
Parenting consistency reflects the degree to which parents set and ensure compliance with age-appropriate instructions, rules, and expectations.15 Inconsistent parenting negatively affects children’s general behavior,16,17 and potentially their lifestyle behaviors. Children whose parents are high in consistency may be more likely to live in households where there are clear expectations in general, and also around healthy behaviors (eg, rules regarding television viewing, screen time, and physical activity, bed-time routines, and the timing and type of foods consumed). These influences could be expected to support healthy weight patterns.
Alternatively, this relationship may flow in the opposite direction with children’s lifestyle behaviors and weight influencing parenting consistency. Children’s unhealthy lifestyle behaviors and overweight could contribute to parents’ stress, undermining their capacity for maintaining parenting consistency. Perhaps more plausibly, child overweight could increase parenting consistency, with some parents responding by setting rules and becoming more controlling and consistent around diet and activity. This might be particularly true for girls, whose parents have been shown to be more aware of a high BMI in their child than parents of boys.18 In contrast, parents who are overweight themselves may underestimate their child’s overweight and may be less likely than normal-weight parents to adjust their parenting in response to child overweight.19,20
Disentangling associations between parenting and child BMI, and understanding how these develop over time, requires repeated measures of parenting and child BMI. With 4 waves of data now available, the nationally representative LSAC21 provides an ideal opportunity to address these questions, building on our previous cross-sectional research from the first wave only. Therefore, our first aim was to explore the mutual influences of maternal and paternal parenting consistency and child BMI across the pre- and primary school years, using a cross-lagged modeling approach.22 We hypothesized that parenting consistency not only predicts but is also a consequence of child BMI. A second aim was to assess the moderating influence of child gender and parental weight status on these associations. We hypothesized that a higher child BMI would be prospectively associated with more parenting consistency among girls and normal-weight parents.
Design and Study Population
This study was conducted by using data from Waves 1 to 4 of the nationally representative Growing up in Australia: Longitudinal Study of Australian Children. The LSAC sampling design and field methods have been described extensively elsewhere.21 Participants (n = 4983) were aged 4 to 5 years when recruited in 2004 and 10 to 11 years (n = 4169, 84% retention) at Wave 4 data collection. Data were collected every 2 years via interviews, questionnaires, and direct anthropometric measurements. Written informed consent was obtained, and the Australian Institute of Family Studies Ethics Committee approved the study.
Of the 4983 participating children, we excluded those whose BMI data were missing at all 4 waves (n = 16) and without any responses to the parenting consistency items from both mothers and fathers (n = 965), resulting in a sample of 4002 child-parent triads. Excluded children (n = 981) had a similar BMI at baseline (P = .09), but were more likely to be in the lowest tertile of socioeconomic position, or to have a non-English speaking or Indigenous background (P < .001), than children without missing data (n = 4002).
In each biennial wave of data collection, parents gave detailed information about various aspects of their parenting behaviors. To assess the frequency with which parents set and enforced clear expectations and limits, LSAC’s 5-item measure of parenting consistency was used.23 The items are: 1) How often does your child get away with things that you feel should have been punished? 2) How often is your child able to get out of punishment when she(/he) really sets her mind to it? 3) When you discipline your child, how often does she ignore the punishment? 4) When you give your child an instruction or make a request to do something, how often do you make sure that she does it? and 5) If you tell your child she will get punished if she doesn't stop doing something, but she keeps doing it, how often will you punish her? Responses were on 5-point Likert scales (0 = never/almost never to 4 = all the time) and reverse coded where appropriate to indicate higher consistency, with sum scores ranging from 0 to 20. Internal consistency was acceptable for mothers and fathers at all 4 waves (α 0.68–0.72). Previous validations have demonstrated the expected relationships with other constructs (eg, lower scores associated with socioeconomic disadvantage and poorer child development).23
At each wave, study staff measured children’s weight and height (dressed in light clothing) using standardized equipment and procedures. BMI was calculated as kg/m2 and converted into age- and gender-specific z-scores (BMIz) based on the 2000 Centers for Disease Control growth charts.24
Covariates included child gender, indigenous status, language other than English spoken at home, family socioeconomic position (composite variable based on annual family income and parental education and occupational status),25 and parental BMI derived from self-reported height and weight. For the moderation analyses, parental BMI was categorized as non-overweight (<25 kg/m2) and overweight/obese (≥25 kg/m2).
Analyses were performed by using SPSS version 17.0 and Mplus version 6.0. For Aim 1, cross-sectional and longitudinal correlations between maternal/paternal parenting consistency and child BMIz were estimated with Pearson’s correlation coefficients (r), while bidirectional influences were examined by using a cross-lagged modeling approach.22 Models were estimated by using maximum likelihood estimation with robust standard errors (MLR) to account for non-normality of the data.26 Survey weights were applied to account for differential non-response at Wave 1.27 Missing values in child BMI, parenting consistency and covariates were accounted for by full information maximum likelihood procedures available in Mplus. This method estimates model parameters and standard errors using all available data while adjusting for the uncertainty associated with missing data.28 Of the 4002 included children, 3312 children (82.8%) had complete BMI data across all 4 waves, and 419 (10.5%), 154 (3.8%), and 117 (2.9%) children had missing BMI data in 1, 2, or 3 waves, respectively. Across all waves, parenting consistency scores were complete for 2964 mothers (74.1%) and 1788 fathers (44.7%); 638, 238, and 162 mothers, and 834, 668, and 712 fathers had missing scores in 1, 2, and 3 waves respectively.
For mothers and fathers separately, 4 competing models were tested, each assuming different associations between parenting consistency and BMI (Fig 1). In all models, covariates were regressed on the continuous variables of child BMI and parenting consistency at Wave 1, a correlation was included between child BMI and parenting consistency at Wave 1, as well as cross-sectional residual correlations between these 2 measures for Waves 2 to 4. Model 1 was the stability model, which only allowed associations between child BMI across waves and between parenting consistency across waves. In Model 2, lagged effects of child BMI at each wave on parenting consistency at each subsequent wave were added to Model 1. In Model 3, lagged effects of parenting consistency on child BMI in the subsequent wave were added to Model 1. Finally, in Model 4 the cross-lagged effects of BMI on parenting and of parenting on BMI were simultaneously added to Model 1.
Acceptable-to-good fit was determined by a comparative fit index (CFI) >0.90, and a root mean square error of approximation (RMSEA) <0.08.29 Model improvement was tested using the Satorra-Bentler χ2-difference test for MLR estimation methods30 with significance set at P < .05. We tested whether Models 2 and 3 were a significant improvement to Model 1. To formally test for bidirectional associations, we examined whether Model 4 was a significant improvement over the best fitting of Models 1 to 3. The standardized regression coefficients of the best-fitting model are presented.
For Aim 2, for the best fitting mothers’ and fathers’ models, we tested moderation by child gender and parental weight status. We examined whether cross-lagged paths differed between boys and girls, and between normal weight and overweight parents. Using multiple group analyses, a model with all parameters free to vary between the 2 groups (boys and girls; normal weight and overweight parents) was compared with a model in which lagged paths were constrained to be equal for the 2 groups, with a lack of significant improvement indicating equivalence in the model parameters for the 2 groups.
Mean age of children at Wave 1 (n = 4002) was 4.2 years, and 51% were boys (see Table 1). Forty-two percent of mothers and 66.8% of fathers were overweight. Correlations between maternal and paternal parenting consistency within each wave were modest (r 0.22–0.36).
Table 2 shows that maternal and paternal parenting consistency were negatively correlated with child BMI, both within and between waves, indicating that higher levels of parenting consistency were associated with slightly lower child BMI. In all but 2 instances, correlations were larger for fathers than for mothers.
Table 3 gives an overview of the model fit indices for the alternative cross-lagged models. Patterns of results were similar for mothers and fathers. The stability models (Model 1) were a good fit to the data for both maternal and paternal parenting consistency. The addition of lagged pathways from child BMI to parenting consistency (Model 2) did not improve model fit. In contrast, the models with lagged pathways from parenting consistency to child BMI (Model 3) had a significantly better fit to the data than the stability models. The final models including bidirectional influences (Model 4) were not a significant improvement over Model 3, indicating the addition of bidirectional paths was not an improvement to the unidirectional model.
The lagged parenting to child BMI models (Model 3) were thus accepted as the best-fitting and most parsimonious models for mothers and fathers. Figure 2 shows that the strongest associations were between BMI across waves (eg, standardized β Wave 1 to 2 for maternal and paternal models = 0.77, 95% confidence interval [CI]: 0.75 to 0.79) and between parenting consistency across waves (eg, standardized β Wave 1 to 2 for maternal parenting = 0.60, 95% CI: 0.58 to 0.63). In addition, higher levels of parenting consistency predicted lower child BMI in the next wave at most points. For mothers, this held from Waves 1 to 2 and 3 to 4, and for fathers from Waves 1 to 2 and 2 to 3. However, associations were small; for example, each 1.0 SD increase in mothers’ parenting consistency at Wave 1 was associated with a 0.025 reduction in child BMIz at Wave 2 (95% CI: −0.05 to −0.003).
For Aim 2, we conducted multi-group analyses on the best-fitting mothers’ and fathers’ models (Model 3) to examine moderation by gender and parental weight. Again results were similar for both parents and indicated that the cross-lagged paths from parenting consistency to child BMI did not differ in strength and direction between boys and girls (mothers: χ2-difference = 0.1, P = .99; fathers: χ2-difference = 3.1, P = .38), or between parents who were normal weight versus overweight (mothers: χ2-difference = 4.9, P = .18; fathers: χ2-difference = 0.8, P = .84).
Findings from this large contemporary cohort of Australian children suggest that higher levels of parenting consistency precede slightly lower BMI in children across a 6-year period from ages 4–5 to 10–11 years. Unlike the earlier cross-sectional findings at age 4 to 5 years in this same cohort (which implicated only fathers),8 effects of parenting consistency were very similar for mothers and fathers. Contrary to our hypothesis, child BMI did not appear to influence parenting behavior, nor were associations influenced by child gender or parent weight status.
Associations of parenting consistency predicting lower child BMI were small and, although evident between every wave, these were not always statistically significant. This may reflect difficulty in achieving statistical significance when assessing a modest association over a 2-year time interval, coupled with high short-term stability in children’s BMI and in parenting consistency, suggesting that both are largely set early in life. In the context of the array of factors known to influence children’s weight (including genetic inheritance and the energy intake-expenditure balance),5 the findings indicate that the broader parenting environment may also contribute. As parenting inconsistency is quite a distal factor from child BMI, relationships were expected to be weak. Nevertheless, both mothers’ and fathers’ inconsistency in parenting may generalize to greater difficulty in instituting and enforcing rules, household organization, and mealtime routines that could be protective against children’s weight gain.
Although the findings for fathers are entirely novel, these modest prospective consistency-BMI associations are in line with previous longitudinal studies. Most have examined parenting styles (ie, authoritarian, permissive, or neglectful) and showed that children raised by authoritative mothers (characterized by warmth as well as consistency) had lower mean BMIs than children who were raised with other styles.6,11–14 Maternal parenting consistency as a stand-alone construct has been investigated less often and, to our knowledge, only once using a longitudinal design. In an unpublished thesis, Hejazi31 showed that lower parental consistency (assessed for the parent most knowledgeable about the child, probably often the mother) among 2.5-year-old Canadian children predicted a trajectory leading to a relatively high BMI at age 9 years. Similar to our study, effects were small. Interesting next steps in this field would be to examine whether effects of mothers' and fathers' parenting on child BMI vary across countries that differ in terms of social norms regarding mothers' and fathers' involvement in child rearing.
Contrary to our hypothesis, no evidence was found for child BMI predicting parenting consistency. This contrasts with research on children’s psychosocial well-being, showing bidirectional associations between parenting and children’s behavior or temperamental characteristics.32 Unlike behavioral problems, child overweight appears to have little direct impact on children’s health,33 family activities and routines, and objective health care costs (S. Clifford, personal communication, 2013) in the first decade of life at the population level. Thus, the only real driver to change parenting would be recognition that their child is at risk for future poor health, yet it is well documented that most parents do not consider their overweight children to be overweight.34,35 This study does not rule out the possibility that child BMI could affect more specific parenting dimensions or practices related to lifestyle behaviors like physical activity and eating behavior. In fact, some recent studies found that children’s weight status acts as an antecedent to parent’s food-related restrictions.36,37 Together, our and previous findings suggest that domain-specific parenting may be partly context driven, whereas more general parenting is a quite stable construct that typically influences child development, but not the reverse.
A key strength of the current study was our longitudinal examination of both mothers’ and fathers’ parenting in relation to child BMI. Despite our initial report of a cross-sectional association between fathers’ parenting consistency and child BMI in 2007,8 research regarding fathers’ influences on children’s overweight has remained sparse.6,10 In keeping with our earlier analyses, fathers’ parenting consistency prospectively predicted small decreases in offspring BMI over 2-year periods across middle childhood. Other strengths of the current study were its large, population-based sample and repeated assessments of both parenting consistency and objective BMI. Our study also had some limitations. While accounting for selective participation by using sampling weights, we did not account for selective loss to follow-up. Data were more often missing in children from disadvantaged families. Although we recently showed that such children have a particularly high risk for overweight,38 children excluded from the present analyses because of missing data did not have a higher BMI than included children. Furthermore, given the availability of data from 4 waves only, it was not possible to identify age trends in the parenting-BMI relationship. Associations may change in adolescence, a period characterized by changes in parenting and other environmental influences on eating and activity.
This study is unique in exploring dynamics between child BMI and parenting behavior of both parents. Findings supported a protective influence of parenting consistency on child BMI, with no evidence of an influence from child BMI to parenting. Although not large, these associations were only slightly smaller than the effects of other factors repeatedly cited as contributing to or being protective for children’s overweight development, like breast feeding.39,40 Thus, our results support recent calls for expanding childhood overweight interventions to address the broad parenting context41; these should involve both mothers and fathers.
This paper uses confidentialized unit record files from Growing Up in Australia, the Longitudinal Study of Australian Children (LSAC). The study is conducted in partnership between the Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA), the Australian Institute of Family Studies (AIFS), and the Australian Bureau of Statistics (ABS). The findings and views reported are those of the authors and should not be attributed to FaHCSIA, AIFS, or the ABS. We thank all the parents and children who took part in Wave 1 until 4 of LSAC. We acknowledge the peer review provided by the LSAC analysis group comprising staff from the Parenting Research Centre and Murdoch Childrens Research Institute.
- Accepted September 12, 2013.
- Address correspondence to Pauline Jansen, PhD, Murdoch Childrens Research Institute, Flemington Rd, Parkville, Victoria 3053, Australia. E-mail:
Dr Jansen conducted the literature review and analyses and drafted the manuscript; Dr Giallo contributed to the design, data analyses, interpretation of results, and drafting of sections of the Methods and Results; Dr Westrupp contributed to the data analyses, interpretation of results, and drafting of sections of the Methods and Results; Dr Wake was co-senior author and contributed to the design of the study, interpretation of results, and drafting of the manuscript; Dr Nicholson was senior author and supervised the design of the study, data analyses; and drafting of the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Dr Jansen was supported by a Rubicon grant 446-11-010 from the Netherlands Organisation for Scientific Research (NWO) and the Marie Cofund Action. Dr Wake and Dr Nicholson were supported by National Health and Medical Research Council Career Development Awards (Dr Wake: 546405; Dr Nicholson: 390136) and Research Fellowships (Dr Wake: 1046518). Murdoch Childrens Research Institute research is supported by the Victorian Government's Operational Infrastructure Support Program, and the Parenting Research Centre receives funding from the Victorian Government Department of Education and Early Child Development.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Freedman DS,
- Sherry B
- Reilly JJ,
- Methven E,
- McDowell ZC,
- et al
- ↵Sleddens EF, Gerards SM, Thijs C, de Vries NK, Kremers SP. General parenting, childhood overweight and obesity-inducing behaviors: a review. Int J Pediatr Obes. 2011;6(2-2):e12–e27
- Bandura A
- ↵Wake M, Nicholson JM, Hardy P, Smith K. Preschooler obesity and parenting styles of mothers and fathers: Australian national population study. Pediatrics. 2007;120(6). Available at: www.pediatrics.org/cgi/content/full/120/6/e1520
- ↵Baumrind D. The development of instrumental competence through socialization. In: Pick D, ed. Minnesota Symposium on Child Psychology. Minneapolis, MN: University of Minnesota Press; 1973:3-46
- Rhee KE,
- Lumeng JC,
- Appugliese DP,
- Kaciroti N,
- Bradley RH
- Olvera N,
- Power TG
- Gray M,
- Smart D
- ↵Zubrick SR, Smith GJ, Nicholson JM, Sanson AV, Jackiewicz TA; for the LSAC Research Consortium. Parenting and families in Australia [Social Policy Research Paper]. Canberra, Australia: Australian Government, Department of Family, Community Services and Indigenous Affairs; 2007
- Blackmore T,
- Strazdins L,
- Gibbings J
- Muthén LK,
- Muthén BO
- ↵Soloff C, Lawrence D, Misson S, Johnstone R. Wave 1 weighting and non-response (LSAC Technical Paper No. 3). Melbourne, Australia: Australian Institute of Family Studies; 2006
- Brown M,
- Cudeck R
- Satorra A
- ↵Hejazi S. Temperament, parenting, and the development of childhood obesity. PhD thesis: University of British Columbia, Vancouver, Canada; 2007
- Webber L,
- Cooke L,
- Hill C,
- Wardle J
- Harder T,
- Bergmann R,
- Kallischnigg G,
- Plagemann A
- Owen CG,
- Martin RM,
- Whincup PH,
- Davey-Smith G,
- Gillman MW,
- Cook DG
- Copyright © 2013 by the American Academy of Pediatrics