[Skip to Navigation]
Sign In
Figure.  Predicted Body Mass Index (BMI) z Scores in Middle Childhood at Illustrative Levels of Maternal Monitoring, Family Characteristics, and Children’s Behaviors
Predicted Body Mass Index (BMI) z Scores in Middle Childhood at Illustrative Levels of Maternal Monitoring, Family Characteristics, and Children’s Behaviors

The predictor variables are those identified in model 3 (Table 2). The standardized BMI z scores were based on BMI scores calculated as weight in kilograms divided by height in meters squared.

Table 1.  Sample Descriptive Statistics by Assessment Agea
Sample Descriptive Statistics by Assessment Agea
Table 2.  Parameter Estimates for Mixed Effects Models
Parameter Estimates for Mixed Effects Models
1.
Robinson  TN.  Reducing children’s television viewing to prevent obesity: a randomized controlled trial.  JAMA. 1999;282(16):1561-1567.PubMedGoogle ScholarCrossref
2.
Crespo  CJ, Smit  E, Troiano  RP, Bartlett  SJ, Macera  CA, Andersen  RE.  Television watching, energy intake, and obesity in US children: results from the third National Health and Nutrition Examination Survey, 1988-1994.  Arch Pediatr Adolesc Med. 2001;155(3):360-365.PubMedGoogle ScholarCrossref
3.
Proctor  MH, Moore  LL, Gao  D,  et al.  Television viewing and change in body fat from preschool to early adolescence: the Framingham Children’s Study.  Int J Obes Relat Metab Disord. 2003;27(7):827-833.PubMedGoogle ScholarCrossref
4.
Dennison  BA, Erb  TA, Jenkins  PL.  Television viewing and television in bedroom associated with overweight risk among low-income preschool children.  Pediatrics. 2002;109(6):1028-1035.PubMedGoogle ScholarCrossref
5.
Wiecha  JL, Peterson  KE, Ludwig  DS, Kim  J, Sobol  A, Gortmaker  SL.  When children eat what they watch: impact of television viewing on dietary intake in youth.  Arch Pediatr Adolesc Med. 2006;160(4):436-442.PubMedGoogle ScholarCrossref
6.
Sallis  JF, Prochaska  JJ, Taylor  WC.  A review of correlates of physical activity of children and adolescents.  Med Sci Sports Exerc. 2000;32(5):963-975.PubMedGoogle ScholarCrossref
7.
Kerr  J, Rosenberg  D, Sallis  JF, Saelens  BE, Frank  LD, Conway  TL.  Active commuting to school: associations with environment and parental concerns.  Med Sci Sports Exerc. 2006;38(4):787-794.PubMedGoogle ScholarCrossref
8.
Carver  A, Timperio  A, Hesketh  K, Crawford  D.  Are children and adolescents less active if parents restrict their physical activity and active transport due to perceived risk?  Soc Sci Med. 2010;70(11):1799-1805.PubMedGoogle ScholarCrossref
9.
Whitaker  RC, Wright  JA, Pepe  MS, Seidel  KD, Dietz  WH.  Predicting obesity in young adulthood from childhood and parental obesity.  N Engl J Med. 1997;337(13):869-873.PubMedGoogle ScholarCrossref
10.
Capaldi  DM, Pears  KC, Kerr  DCR.  The Oregon Youth Study Three-Generational Study: theory, design, and findings.  ISSBD Bull. 2012;2(62):29-33.Google Scholar
11.
Barlow  SE; Expert Committee.  Expert Committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report.  Pediatrics. 2007;120(suppl 4):S164-S192.PubMedGoogle ScholarCrossref
12.
Capaldi  DM, Patterson  GR.  Psychometric Properties of Fourteen Latent Constructs from the Oregon Youth Study. New York, NY: Springer-Verlag; 1989.
13.
Kuczmarski  RJ, Ogden  CL, Grummer-Strawn  LM,  et al; National Center for Health Statistics.  CDC growth charts: United States.  Adv Data. 2000;314(314):1-27.PubMedGoogle Scholar
14.
Kuczmarski  RJ, Ogden  CL, Guo  SS,  et al.  2000 CDC growth charts for the United States: methods and development. National Center for Health Statistics.  Vital Health Stat 11.2002;(246):1-190.Google Scholar
15.
Dibley  MJ, Goldsby  JB, Staehling  NW, Trowbridge  FL.  Development of normalized curves for the international growth reference: historical and technical considerations.  Am J Clin Nutr. 1987;46(5):736-748.PubMedGoogle Scholar
16.
Dibley  MJ, Staehling  NW, Nieburg  P, Trowbridge  FL.  Interpretation of Z-score anthropometric indicators derived from the international growth reference.  Am J Clin Nutr. 1987;46(5):749-762.PubMedGoogle Scholar
17.
Moraeus  L, Lissner  L, Yngve  A, Poortvliet  E, Al-Ansari  U, Sjöberg  A.  Multi-level influences on childhood obesity in Sweden: societal factors, parental determinants and child’s lifestyle.  Int J Obes (Lond). 2012;36(7):969-976.PubMedGoogle ScholarCrossref
18.
Achenbach  TM.  Manual for the Child Behavior Checklist/4-18 and 1991 Profile. Burlington, VT: University of Vermont, Dept of Psychology; 1991.
19.
Achenbach  TM.  Manual for the Child Behavior Checklist/2-3 and 1992 Profile. Burlington, VT: University of Vermont, Dept of Psychology; 1992.
20.
American Academy of Pediatrics; Committee on Public Education.  American Academy of Pediatrics: children, adolescents, and television.  Pediatrics. 2001;107(2):423-426.PubMedGoogle ScholarCrossref
21.
Bryk  AS, Raundenbush  SW. Applications in the study of individual change. In: Bryk  AS, Raudenbush  SW, eds.  Hierarchical Linear Models. Newbury Park, CA: Sage; 1992:130-154.
22.
Satorra A, Bentler P. Scaling corrections for statistics in covariance structure analysis. Los Angeles: University of California: Dept of Statistics. http://escholarship.org/uc/item/8dv7p2hr. Published October 25, 2011. Accessed April 4, 2012.
23.
Fakhouri  TH, Hughes  JP, Brody  DJ, Kit  BK, Ogden  CL.  Physical activity and screen-time viewing among elementary school-aged children in the United States from 2009 to 2010.  JAMA Pediatr. 2013;167(3):223-229.PubMedGoogle ScholarCrossref
24.
Schmidt  ME, Haines  J, O’Brien  A,  et al.  Systematic review of effective strategies for reducing screen time among young children.  Obesity (Silver Spring). 2012;20(7):1338-1354.PubMedGoogle ScholarCrossref
25.
Steeves  JA, Thompson  DL, Bassett  DR, Fitzhugh  EC, Raynor  HA.  A review of different behavior modification strategies designed to reduce sedentary screen behaviors in children.  J Obes. 2012;2012:379215.PubMedGoogle ScholarCrossref
26.
Birken  CS, Maguire  J, Mekky  M,  et al.  Office-based randomized controlled trial to reduce screen time in preschool children.  Pediatrics. 2012;130(6):1110-1115.PubMedGoogle ScholarCrossref
Original Investigation
May 2014

Parental Monitoring of Children’s Media Consumption: The Long-term Influences on Body Mass Index in Children

Author Affiliations
  • 1Oregon Social Learning Center, Eugene
  • 2School of Psychological Science, Oregon State University, Corvallis
  • 3Unit of Pediatrics, Department of Clinical Science, Intervention and Technology, Karolinska Institute, Stockholm, Sweden
JAMA Pediatr. 2014;168(5):414-421. doi:10.1001/jamapediatrics.2013.5483
Abstract

Importance  Although children’s media consumption has been one of the most robust risk factors for childhood obesity, effects of specific parenting influences, such as parental media monitoring, have not been effectively investigated.

Objectives  To examine the potential influences of maternal and paternal monitoring of child media exposure and children’s general activities on body mass index (BMI) in middle childhood.

Design, Setting, and Participants  A longitudinal study, taken from a subsample of the Three Generational Study, a predominantly white, Pacific Northwest community sample (overall participation rate, 89.6%), included assessments performed from June 1998 to September 2012. Analyses included 112 mothers, 103 fathers, and their 213 children (55.4% girls) at age 5, 7, and/or 9 years. Participation rates ranged from 66.7% to 72.0% of all eligible Three Generational Study children across the 3 assessments.

Exposures  Parents reported on their general monitoring of their children (whereabouts and activities), specific monitoring of child media exposure, children’s participation in sports and recreational activities, children’s media time (hours per week), annual income, and educational level. Parental BMI was recorded.

Main Outcomes and Measures  Predictions to level and change in child BMI z scores were tested.

Results  Linear mixed-effects modeling indicated that more maternal, but not paternal, monitoring of child media exposure predicted lower child BMI z scores at age 7 years (95% CI, −0.39 to −0.07) and less steeply increasing child BMI z scores from 5 to 9 years (95% CI, −0.11 to −0.01). These effects held when more general parental monitoring, and parent BMI, annual income, and educational level were controlled for. The significant negative effect of maternal media monitoring on children’s BMI z scores at age 7 years was marginally accounted for by the effect of child media time. The maternal media monitoring effect on children’s BMI z score slopes remained significant after adjustment for children’s media time and sports and recreational activity.

Conclusions and Relevance  These findings suggest that parental behaviors related to children’s media consumption may have long-term effects on children’s BMI in middle childhood. They underscore the importance of targeting parental media monitoring in efforts to prevent childhood obesity.

Children’s media consumption has been consistently linked to childhood obesity.1-5 Research on children’s media use has focused on identifying factors of direct relevance to weight, such as the number of hours children spend watching a television or computer,2,3 the presence of television screens in children’s bedrooms,4 and whether eating while watching television increases caloric intake.5 Although studies have revealed parents’ role in structuring children’s physical activities,6-8 prior work has not effectively addressed how parents monitor and influence sedentary activities and the links to children’s obesity. Better understanding of the role that parents may play in the monitoring of their children’s media consumption with an examination of more general forms of monitoring is critical to the development of targeted, preventive interventions for obesity. In the current study, we sought to disentangle the potential influences of multiple parenting behaviors and parent and family characteristics on body mass index (BMI) in children to identify specific parental behaviors that may protect or put children at risk for unhealthy weight development.

Because obesity by middle childhood portends lifelong risk,9 identifying modifiable parental influences in this period, such as lack of media monitoring, may have implications for health in childhood and adulthood. We used longitudinal data from middle childhood,10 when parents still maintain primary responsibility for children’s health behaviors. We first distinguished among parental monitoring behaviors thought to confer specific risk for obesity (media monitoring) and more general direct monitoring (supervision) and indirect monitoring (communication and time spent with child) that are broadly relevant to developmental risk. We also controlled for potentially confounding family characteristics (parental BMI, annual income, and educational levels). Finally, we explored whether these associations would be attenuated by child behaviors thought to confer specific risk for obesity (child media time and participation in sports and recreational activities).11

We hypothesized that parental monitoring behaviors (direct, indirect, and media monitoring) would be inversely related to intraindividual and interindividual variation in child BMI beyond prediction from family characteristics (parental BMI, annual income, and educational level). Given the sedentary behavior associated with media time and exposure to potential food advertisement, monitoring of child media exposure was hypothesized to be especially relevant. Lower monitoring of a child’s whereabouts and activities (direct) and less communication and time spent with child (indirect) were hypothesized to affect child BMI through parental absence and unawareness of children’s diets and/or behaviors. Stronger monitoring effects were expected for mothers, who are more often children’s primary caregivers, than for fathers. We also tested whether associations between parental monitoring behaviors and child BMI, especially associations related to media monitoring, would be attenuated by child behaviors hypothesized to be inversely (participation in sports and recreational activities) or directly (media time) related to child BMI.

Methods
Participants

After the Oregon Social Learning Center institutional review board approval and written parental informed consent, 213 children (55.4% girls) and their 103 fathers and 112 mothers were assessed across childhood during the Three Generational Study (3GS),10 which originally examined the intergenerational transmission of risk for antisocial and maladaptive behaviors and substance abuse. Fathers had been originally recruited at age 9 or 10 years owing to elevated neighborhood risk for delinquency (Oregon Youth Study12 [n = 206]) and assessed nearly annually to age 37 years. The 3GS is a predominantly white, Pacific Northwest community sample. Current data included assessments performed from June 1998 to September 2012. Eligible 3GS children (≤2 per partner of the men) participated at the age 5-, 7-, and 9-year assessments (89.4%, 92.9%, and 93.0%, respectively). Children were considered for the present analyses if (1) their heights and weights were measured at least once across the assessments (92.6%, 96.6%, and 97.5% of participating 3GS children, respectively, at the 3 age assessments) and (2) their data were complete for other study predictors (86.0%, 87.3%, and 82.1% of those meeting criterion 1).

Child participation rates ranged from 66.7% to 72.0% of all eligible and from 71.7% to 77.5% of all participating 3GS children across assessments. Averaged over assessments, independent sample t tests indicated that the included children, compared with excluded participating children, had fathers who engaged in more media monitoring (95% CI, 0.14 to 0.98), mothers who engaged in less direct monitoring (95% CI, −0.05 to −0.001), and parents with higher annual incomes (95% CI, $1300 to $12 700). There were no other significant differences between the groups for other study predictors or child BMI. Of the children’s mothers and fathers, 99.5% and 88.6%, respectively, were the children’s biological parents (others were stepparents). Information on children’s living situations is presented in Table 1.

Procedures

Parents and children were assessed when children were aged 5, 7, and 9 years, using questionnaires, interviews, and physical measurements. Children participated at 1, 2, or 3 assessments (33, 97, and 83 children, respectively). For 82.1% to 94.4% of children and 87.0% to 94.5% of parents, both parents’ reports were available for all study predictors except for parents’ BMI (measured in both parents for 56.8% of children).

Measures

For parental general monitoring, media monitoring, and children’s activities, items, response ranges, internal consistencies, percentage of total variance explained, and correlations between parents’ scores are provided in the eTable in the Supplement. Items had to demonstrate adequate associations with their corresponding scales (individual item-to-total correlations of ≥0.20). Averaged across assessments and scales, item-to-total correlation ranged from 0.28 to 0.78. All scales were unidimensional. Maternal and paternal monitoring measures were used at each assessment to form time-variant (intraindividual) variables. Other children and family variables served as controls and were averaged across parents to form time-variant aggregate variables. Cross-time averages of these time-variant measures and of those collected at only 1 or 2 assessments served as time-invariant (interindividual) variables.

Dependent Variable: Children’s BMI z Scores

Heights and weights were obtained via physical measurements for 91.5% of the children (195 of 213) and 89.0% of the time-by-person observations (422 of 474) and by parent reports for the remaining children using the Centers for Disease Control and Prevention growth charts.13,14 The BMI scores (calculated as weight in kilograms divided by height in meters squared) were converted to standardized BMI-for-age-and-sex z scores. Thus, the dependent variable represents deviation from a national mean BMI. Two biologically implausible BMI z scores were excluded from the sample.15,16

Time-Invariant Predictors

Three family characteristics (parental annual income and educational level, measured at all 3 assessments, and parent BMI, measured at 1 or 2 assessments) were averaged and entered into the analyses as time-invariant control predictors. Maternal and paternal monitoring of media exposure and children’s media time were assessed only at the age 5- and 7-year assessments; these time-invariant independent variables were computed as means of these 2 assessments.

Parental Annual Income

Annual household income was the mean of the mothers’ and fathers’ reports. One outlier was set to the second-highest value of $182 400. Scores were divided by 10 000 so that regression coefficients are interpretable as interindividual differences associated with a $10 000 difference in annual income.

Parental Educational Level

Parents selected their highest level of education completed from 7 categories, ranging from “less than 7th grade” to “graduate degree.” Based on past research17 and for parsimony, parental educational level was categorized to denote that 0, 1, or 2 parent(s) had more than 12 years of education at 1 assessment or more. The effects of parental educational level on child BMI were tested for 0 vs 2 parents and 1 vs 2 parent(s) with some form of postsecondary education.

Parental BMI

Parents’ BMI scores were calculated from physical measurements for all fathers and 44.6% of mothers (others contributed self-reports). When both parents’ BMI scores were available (for 95 parents of 121 children), they were averaged; otherwise, the parental BMI variable denoted only paternal BMI (19 fathers of 88 children) or maternal BMI (1 mother of  4 children).

Parental Monitoring of Media Exposure

Parents answered 3 questions regarding their limiting of their child’s media exposure (D.M.C., K.C.P., J. Wilson, MS, and L. Bruckner, BS, Parent Interview [unpublished instrument], Oregon Social Learning Center, 1998). Response scales ranged from 1 (never or almost never) to 5 (always or almost always), in addition to an option for “never ever watched television or videos or played video games,” which was recoded to 5. For children aged 5 and 7 years, respectively, 19.5% and 7.1% of mothers and 13.2% and 4.9% of fathers restricted their children from playing video games.

Children’s Media Time

Parents reported the typical number of hours their child spent (1) watching television or movies and (2) playing video games during school-year weekdays and weekends (D.M.C., K.C.P., J. Wilson, and L. Bruckner, Parent Interview). A weighted mean (5/7 × weekday hours + 2/7 × weekend hours) was calculated to denote the typical number of hours of media time per day averaged for both television/movies and video games. Mothers’ and fathers’ reports were averaged into a composite child media time score.

Time-Varying Predictors

Separate scores for general monitoring and children’s activities were created at each of the assessments (at ages 5, 7, and 9 years) to determine time-specific associations with child BMI z scores.

General Parental Monitoring

Parents completed 3 items describing direct parental monitoring practices (supervision, awareness, and control of child whereabouts and associates). Response scales ranged from 1 (very often/every day) to 5 (never). Identical items were used at ages 5 and 7 years but were modified at age 9 years to be more developmentally appropriate (D.M.C., K.C.P., J. Wilson, and L. Bruckner, Parent Interview; D.M.C. and J. Wilson, Monitor and Parent-Child Relationship [unpublished instrument], Oregon Social Learning Center, 1998). Five items regarding more indirect forms of parental monitoring (conversation about child’s day and time spent with child; response scale: 0-7 d/wk) assessed at each age and reported by parents were not significantly associated with direct monitoring practices and thus were considered separate predictors of child BMI (D.M.C. and J. Wilson, Monitor and Parent-Child Relationship).

Children’s Sports and Recreational Activity

Parents reported on their child’s participation in sports (2 items)18,19 and family recreational activities (2 items; G.R. Patterson, PhD, Family Activities Checklist [unpublished instrument], Oregon Social Learning Center, 1982). Item response scales were recoded to range from 0 to 3, and mothers’ and fathers’ reports were averaged at each time point to create a composite score (hereafter referred to as activities).

Data Analytic Strategy

Dependence among children’s BMI z scores across middle childhood and siblings’ BMI z scores were accounted for by fitting 3-level linear mixed-effects models. Children’s ages were grand-mean centered at the middle assessment (mean age, 7.3 years), and children’s BMI z scores at age 7.3 years were free to vary within the sample (ie, random child intercept). Models 1 and 2 for mothers and fathers addressed the first set of hypotheses examining the simultaneous influences of (1) direct and indirect general monitoring as both intraindividual and interindividual predictors and (2) media monitoring as an interindividual predictor of children’s BMI z scores while also controlling for interindividual differences in parents’ BMI, annual income, and educational level.

Model 3 addressed the second set of hypotheses examining whether maternal monitoring effects on children’s BMI z scores were attenuated by children’s activities (as an intraindividual and interindividual predictor) and media time (as an interindividual predictor). Averaging mothers’ and fathers’ BMI scores was necessary to retain sufficient sample size, but this results in a loss of information and prevents individual examination of effects, which may vary for children with only 1 overweight or obese parent rather than 2. We thus included both a main effect of mean parent BMI and an interaction term between parent BMI and a contrast coefficient denoting whether the parental BMI effect on child BMI varied for those children with 2 (coded as 0.5) vs only 1 overweight or obese parent (coded as −0.5). This effect applies only to children with complete data on mothers’ and fathers’ BMI scores; children with only 1 parent BMI measure (coded as 0) were excluded from the mean comparison and included in the main effect of parent BMI.

Results

Descriptive statistics (Table 1) indicated that children’s mean BMI z scores increased from 0.61 to 0.82 across middle childhood. Overweight and obesity prevalence across all assessments ranged from 37.1% to 51.8% for children. Of the 121 children with complete maternal and paternal BMI scores, 69, 40, and 12, respectively, had 2, 1, or 0 overweight or obese parent(s) at 1 assessment or more; of the 92 children with 1 parent report, 55 mothers or fathers were overweight or obese. Mothers’ and fathers’ general direct monitoring showed slight increases over time, whereas all other variables appeared relatively stable. Mothers and fathers reported similar mean levels of direct monitoring, whereas mothers reported more indirect and media monitoring than fathers. Averaged across assessments, children spent 1.74 h/d watching television and/or playing video games, and 19.6% of children spent more than the American Academy of Pediatrics’ recommended maximum of 2 h/d.20

Parents’ General and Media Monitoring

The first set of hypotheses relating parental monitoring behaviors to children’s BMI z scores while controlling for family characteristics were partially supported for mothers only (Table 2; models 1 and 2). Maternal media monitoring predicted level and change in children’s BMI z scores; maternal and paternal general monitoring did not. Children whose mothers engaged in less media monitoring had higher BMI z scores at age 7 years and more steeply increasing BMI z scores from age 5 to age 9 years. The years in which parents engaged in more direct and indirect monitoring, however, were not predictive of concurrent decreases in children’s BMI z scores across middle childhood. Parent BMI scores predicted children’s BMI z scores at age 7 years but not changes in BMI z scores from 5 to 9 years. The parental BMI effect at age 7 years was amplified for children with 2 overweight or obese parents vs 1. Children from homes where only 1 parent (rather than 2) had more than 12 years of education had higher BMI z scores at age 7 years, but there were no significant differences in the prediction of child BMI z score slopes. Finally, neither children's BMI z scores at age 7 years nor their BMI z score slopes significantly differed for children from homes where neither (rather than both) parents had more than 12 years of education, nor was parental annual income a significant predictor of either outcome. The maternal and paternal media monitoring models, respectively, accounted for 2.7% and 2.3% of variation in children's BMI z scores across time and 23.7% and 18.8% of variation between children's BMI z scores.21

Parents’ General and Media Monitoring Controlling for Children’s Behaviors

Finally, we considered whether children’s activities and media time might attenuate the associations between parental BMI and educational level and between maternal media monitoring and child BMI z scores (Table 2; model 3). Intraindividual increases in children’s activities across middle childhood were marginally associated with concurrent decreases in children’s BMI z scores; however, interindividual variability in children’s activities did not relate to lower BMI z scores at age 7 years or less steeply increasing BMI z scores from age 5 to age 9 years. Next, results suggested that the negative effect of maternal media monitoring on BMI z scores at age 7 years was marginally accounted for by the effect of child media time, whereas the effect of maternal media monitoring on children’s BMI z score slopes remained significant after adjustment for children’s media time and activities. The significance of all family characteristics effects remained unchanged from the prior maternal monitoring model, model 1. The addition of children’s activities and media time yielded improvement in model fit over model 1 (scaled χ25 difference, 11.43; P = .04),22 explaining 4.1% of children’s variability in BMI z scores across time and 27.3% of the variability in BMI z scores between children.21 The Figure depicts children’s predicted BMI z scores, given their ages and the predictor variables identified in model 3.

Discussion

This study highlights the importance of parenting behavior in children’s weight development across middle childhood. Less maternal monitoring of media exposure predicted higher BMI at age 7 years and increasing deviance from child BMI norms from ages 5 to 9 years. Several competing explanations for these effects were ruled out. First, whereas lower parental educational level, higher parental BMI, and having 2 vs 1 overweight or obese parent(s) were risk factors for child obesity, they did not account for effects of maternal media monitoring on child BMI z scores. Second, maternal media monitoring, but not direct or indirect general monitoring, was associated with child BMI. Thus, low maternal media monitoring does not seem to reflect more general parental disengagement or lack of awareness regarding children’s behaviors and whereabouts. The association between lower maternal media monitoring and higher child BMI was primarily explained by a tendency for these children to spend more hours per week watching television and playing video games. This supports the validity of our interpretation that child media time has direct effects on BMI, is under substantial control by parents, and therefore is a prime target for family intervention.

The link between children’s media time and obesity is not new.1-5 To our knowledge, however, the link has not been established in longitudinal studies that sufficiently control for the competing influences of parents’ BMI, income, educational level, and other forms of parenting (ie, general monitoring or supervision). The American Academy of Pediatrics recommends that child media time should be limited to 2 h/d,20 but many children spend more time in front of a screen than is recommended.23 The results of interventions aimed at decreasing television time have not been adequate,24,25 perhaps because of their low intensity,26 but strengthening parental practices regarding limiting child media use is an important focus in family-based interventions.

Our findings also indicated that child BMI was marginally responsive to changes in children’s participation in sports and recreational activities over time such that deviations from age and sex norms were greater in years of decreased activity. Future research should explore these associations by using more comprehensive measures of physical activity and explore dietary options and child eating patterns.

There were some study limitations. The sample was not racially or ethnically representative of the US population and was relatively small, thus not permitting comparisons between boys and girls. Other limitations include subjective measures of children’s activities and media time and parental behavior and our inability to rule out potential social desirability biases. Incomplete data precluded the ability to test intraindividual effects of media monitoring and independent maternal and paternal BMI effects on child BMI. Furthermore, although the study used a longitudinal design, it was not possible to distinguish the temporal ordering of predictors and outcomes, and the design did not allow for causal inferences. In addition, we did not consider how parental monitoring may differ between overweight, less active children and more active children who become overweight. Nor did we consider how parental monitoring and child activities may be predicted by child BMI because monitoring for an overweight or obese child may differ from monitoring for a normal-weight child. Finally, models explained relatively little variation in children’s BMI z scores across middle childhood and approximately one-fourth of the variation between children in BMI z scores at age 7 years. Thus, future research must identify additional explanatory variables.

Conclusions

The current study examined the influences that parents’ behaviors may have on children’s weight across middle childhood. Our results suggest that interventions aimed at parental supervision and control of child media exposure may promote healthy child weight development during middle childhood.

Back to top
Article Information

Accepted for Publication: December 12, 2013.

Corresponding Author: Paulina Nowicka, PhD, Unit of Pediatrics, Department of Clinical Science, Intervention and Technology, Karolinska Institute, B62, 115 35, Stockholm, Sweden (paulina.nowicka@ki.se).

Published Online: March 17, 2014. doi:10.1001/jamapediatrics.2013.5483.

Author Contributions: Dr Tiberio had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Tiberio, Kerr, Capaldi, Nowicka.

Drafting of the manuscript: Tiberio, Kerr, Nowicka.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Tiberio, Kerr.

Obtained funding: Kerr, Capaldi, Pears.

Study supervision: Capaldi, Nowicka.

Conflict of Interest Disclosures: None reported.

Funding/Support: This project was supported by the National Institute of Drug Abuse (grant R01 DA 015485 to Dr Capaldi), the National Institute on Alcohol Abuse and Alcoholism (grant 1R01AA018669 to Dr Capaldi), the National Institute of Child Health and Development (grant HD 46364 to Dr Capaldi), and the Sweden-America Foundation (Dr Nowicka), the Swedish Society for Medical Research (Dr Nowicka), and the Vinnmer Marie Curie International Qualification (Dr Nowicka).

Role of the Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, the National Institute on Alcohol Abuse and Alcoholism, the National Institute of Drug Abuse, or the National Institute of Child Health and Development.

References
1.
Robinson  TN.  Reducing children’s television viewing to prevent obesity: a randomized controlled trial.  JAMA. 1999;282(16):1561-1567.PubMedGoogle ScholarCrossref
2.
Crespo  CJ, Smit  E, Troiano  RP, Bartlett  SJ, Macera  CA, Andersen  RE.  Television watching, energy intake, and obesity in US children: results from the third National Health and Nutrition Examination Survey, 1988-1994.  Arch Pediatr Adolesc Med. 2001;155(3):360-365.PubMedGoogle ScholarCrossref
3.
Proctor  MH, Moore  LL, Gao  D,  et al.  Television viewing and change in body fat from preschool to early adolescence: the Framingham Children’s Study.  Int J Obes Relat Metab Disord. 2003;27(7):827-833.PubMedGoogle ScholarCrossref
4.
Dennison  BA, Erb  TA, Jenkins  PL.  Television viewing and television in bedroom associated with overweight risk among low-income preschool children.  Pediatrics. 2002;109(6):1028-1035.PubMedGoogle ScholarCrossref
5.
Wiecha  JL, Peterson  KE, Ludwig  DS, Kim  J, Sobol  A, Gortmaker  SL.  When children eat what they watch: impact of television viewing on dietary intake in youth.  Arch Pediatr Adolesc Med. 2006;160(4):436-442.PubMedGoogle ScholarCrossref
6.
Sallis  JF, Prochaska  JJ, Taylor  WC.  A review of correlates of physical activity of children and adolescents.  Med Sci Sports Exerc. 2000;32(5):963-975.PubMedGoogle ScholarCrossref
7.
Kerr  J, Rosenberg  D, Sallis  JF, Saelens  BE, Frank  LD, Conway  TL.  Active commuting to school: associations with environment and parental concerns.  Med Sci Sports Exerc. 2006;38(4):787-794.PubMedGoogle ScholarCrossref
8.
Carver  A, Timperio  A, Hesketh  K, Crawford  D.  Are children and adolescents less active if parents restrict their physical activity and active transport due to perceived risk?  Soc Sci Med. 2010;70(11):1799-1805.PubMedGoogle ScholarCrossref
9.
Whitaker  RC, Wright  JA, Pepe  MS, Seidel  KD, Dietz  WH.  Predicting obesity in young adulthood from childhood and parental obesity.  N Engl J Med. 1997;337(13):869-873.PubMedGoogle ScholarCrossref
10.
Capaldi  DM, Pears  KC, Kerr  DCR.  The Oregon Youth Study Three-Generational Study: theory, design, and findings.  ISSBD Bull. 2012;2(62):29-33.Google Scholar
11.
Barlow  SE; Expert Committee.  Expert Committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report.  Pediatrics. 2007;120(suppl 4):S164-S192.PubMedGoogle ScholarCrossref
12.
Capaldi  DM, Patterson  GR.  Psychometric Properties of Fourteen Latent Constructs from the Oregon Youth Study. New York, NY: Springer-Verlag; 1989.
13.
Kuczmarski  RJ, Ogden  CL, Grummer-Strawn  LM,  et al; National Center for Health Statistics.  CDC growth charts: United States.  Adv Data. 2000;314(314):1-27.PubMedGoogle Scholar
14.
Kuczmarski  RJ, Ogden  CL, Guo  SS,  et al.  2000 CDC growth charts for the United States: methods and development. National Center for Health Statistics.  Vital Health Stat 11.2002;(246):1-190.Google Scholar
15.
Dibley  MJ, Goldsby  JB, Staehling  NW, Trowbridge  FL.  Development of normalized curves for the international growth reference: historical and technical considerations.  Am J Clin Nutr. 1987;46(5):736-748.PubMedGoogle Scholar
16.
Dibley  MJ, Staehling  NW, Nieburg  P, Trowbridge  FL.  Interpretation of Z-score anthropometric indicators derived from the international growth reference.  Am J Clin Nutr. 1987;46(5):749-762.PubMedGoogle Scholar
17.
Moraeus  L, Lissner  L, Yngve  A, Poortvliet  E, Al-Ansari  U, Sjöberg  A.  Multi-level influences on childhood obesity in Sweden: societal factors, parental determinants and child’s lifestyle.  Int J Obes (Lond). 2012;36(7):969-976.PubMedGoogle ScholarCrossref
18.
Achenbach  TM.  Manual for the Child Behavior Checklist/4-18 and 1991 Profile. Burlington, VT: University of Vermont, Dept of Psychology; 1991.
19.
Achenbach  TM.  Manual for the Child Behavior Checklist/2-3 and 1992 Profile. Burlington, VT: University of Vermont, Dept of Psychology; 1992.
20.
American Academy of Pediatrics; Committee on Public Education.  American Academy of Pediatrics: children, adolescents, and television.  Pediatrics. 2001;107(2):423-426.PubMedGoogle ScholarCrossref
21.
Bryk  AS, Raundenbush  SW. Applications in the study of individual change. In: Bryk  AS, Raudenbush  SW, eds.  Hierarchical Linear Models. Newbury Park, CA: Sage; 1992:130-154.
22.
Satorra A, Bentler P. Scaling corrections for statistics in covariance structure analysis. Los Angeles: University of California: Dept of Statistics. http://escholarship.org/uc/item/8dv7p2hr. Published October 25, 2011. Accessed April 4, 2012.
23.
Fakhouri  TH, Hughes  JP, Brody  DJ, Kit  BK, Ogden  CL.  Physical activity and screen-time viewing among elementary school-aged children in the United States from 2009 to 2010.  JAMA Pediatr. 2013;167(3):223-229.PubMedGoogle ScholarCrossref
24.
Schmidt  ME, Haines  J, O’Brien  A,  et al.  Systematic review of effective strategies for reducing screen time among young children.  Obesity (Silver Spring). 2012;20(7):1338-1354.PubMedGoogle ScholarCrossref
25.
Steeves  JA, Thompson  DL, Bassett  DR, Fitzhugh  EC, Raynor  HA.  A review of different behavior modification strategies designed to reduce sedentary screen behaviors in children.  J Obes. 2012;2012:379215.PubMedGoogle ScholarCrossref
26.
Birken  CS, Maguire  J, Mekky  M,  et al.  Office-based randomized controlled trial to reduce screen time in preschool children.  Pediatrics. 2012;130(6):1110-1115.PubMedGoogle ScholarCrossref
×