Do cell phones, iPods/MP3 players, siblings and friends matter? Predictors of child body mass in a U.S. Southern Border City Middle School

  • Published on
    29-Nov-2016

  • View
    213

  • Download
    1

Transcript

Obesity Research & Clinical Practice (2012) 6, e39e53ORIGINAL ARTICLEDo cell phones, iPods/MP3 players, siblings andfriends matter? Predictors of child body mass in aU.S. Southern Border City Middle SchoolMarcusHoracioa DepartmTX 78041,b DepartmLaredo, Tc Canseco 78041, USReceived 1KEYWORBMI perceSiblings;Cell phoniPods/MP3 CorrespoE-mail ad1871-403X/$ doi:10.1016/ Antonius Ynalveza,, Ruby Ynalvezb, Marivic Torregosac, Palaciosc, John Kilburnaent of Behavioral Sciences, Texas A&M International University, University Boulevard, Laredo, USAent of Biology and Chemistry, Texas A&M International University, University Boulevard,X 78041, USASchool of Nursing, Texas A&M International University, University Boulevard, Laredo, TXA2 January 2011; received in revised form 28 March 2011; accepted 19 April 2011DSntile;es; playersSummaryObjective: This study examines the association of childrens (i) micro-social envi-ronment, specically siblings [kin-friends] and friends from school and neighborhood[non-kin-friends], and (ii) ownership of information and communication technologies(ICT), specically cell phones and iPod/MP3 players, with body mass index percentile(BMIp).Subjects: Fifty-ve randomly selected 6th graders with a mean age of 12 years,stratied by gender (23 boys and 32 girls), from a Texas middle school located in acity along the U.S. southern border.Methods: The linear regression of BMIp on number of siblings and of non-kin-friends,and ownership of cell phone and of iPod/MP3 player was examined using two models:M1 was based on the manual selection of predictors from a pool of potential predic-tors. M2 was derived from the predictors specied in M1 using backward eliminationtechnique. Because sample size was small, the signicance of regression coefcientswas evaluated using robust standard errors to calculate t-values. Data for predictorswere obtained through a survey. Height and weight were obtained through actualanthropometric measurements. BMIp was calculated using the on-line BMI calculatorof the Center for Disease Control and Prevention.nding author. Tel.: +1 956 326 2621; fax: +1 956 326 2474.dress: mynalvez@tamiu.edu (M.A. Ynalvez).see front matter 2011 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.j.orcp.2011.04.006e40 M.A. Ynalvez et al.Results: Findings reveal that childrens social environment and ICT ownership predictBMIp; specically, number of siblings (M2: = 0.34, p-value < .001), and ownership ofiPod/MP3 players (M2: = 0.33, p-value < .001). These results underscore the impor-tance of family in conguring, and of new personal technical devices (that encouragesolitary, and oftentimes sedentary, activities) in predicting child body mass. 2011 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd.IntroducChild obea very serStates un[1,2]. Cheffects onprone to dmental deorthopediType 2 dilosing thebody weigaccustombody. Sincon societyto the perbody of kncontributiover the (current wtion [8], scapital [1and sociaA dimecant inusocial envcated soci[1] reportpersonal comes. Prpeer suppactivity acal activiwhen thedecreasedwere within more weight kievidence ronment, particularmuch undA prevsiblings incantly asssuch nde size, which may have led to oversensitivegs g 1er dn tg tatioMP3t of theon is stuue speool, anP3 ir ag 6tho bonme widan c ado in Iodsxttudyity -trommeaso to stor stu Accf thof istical versus practical signicance is a topic of discussionAll rights reserved.tionsity is not just a cosmetic problem. It isious health concern that puts the Unitedder threat from life-threatening diseasesild obesity carries short- and long-term an individual [3,4]. Obese children areevelop cardiovascular disease, fatty liver,pression due to stigmatization, asthma,c problems, metabolic syndrome, andabetes [5,6]. Once children are obese, excess weight and reverting to an idealht are difcult as fat cells that have beened to storing excess energy remain in thee childhood obesity has a major impact in terms of health care costs in additionsonal, social, and nancial burden; a wideowledge has been built regarding factorsng towards child obesity developmentyears: inherent biological characteristicseight category) [7], genetic predisposi-edentariness [911], neighborhood social2], gene-environment interaction [13],l networks [1].nsion that has been cited as a signi-ence on child obesity development is theironment [1,14,15]. Using highly sophisti-al network analysis, Christakis and Fowler that social environment shapes not onlybehavior, but also personal health out-evious ndings indicate that parental andort is associated with increased physicalmong children [1618]. Increased physi-ty, especially among boys, was observedy were among peers [17,18]; however, physical activity was observed when they family [18]. Overweight children engagedintense physical activities than normal-ds when with peers [18]. While initiallinks child obesity and the social envi-the inuence of familial environment,ly siblings, on child body mass is veryerstudied.ious study indicates that the number ofsamplndinAnothrelatiorelatinmuniciPods/impacing tothose Thithe inment,in schpeers)iPod/Mof theamongMexicenviroties, aMexicnologylatestMethConteThis sway cis subdry sucold shomehas hihealthistics.94% opanic 1 Stat a household was negatively and signi-ociated with body weight [19]. However,ings were drawn from an overly largeamong resealinger and La very small practical sigiven the sheer size of the sample [20].imension that has been investigated ino child body mass development is thato the role of new information and com-n technologies (ICT), like cell phones and players [15,21]. Although studies on thethe Internet are replete [21], those relat- inuence of cell phones are still few, andPod/MP3 player use are still very scarce.dy adds to the paucity of literature onnce of childrens (i) micro-social environ-cically siblings (kin-friends) and friends and neighborhood (non-kin-friends ord (ii) their ownership of cell phones andplayers on body mass measured in termse-gender specic BMI percentile (BMIp), graders in a city in Texas along the U.S.-rder a bicultural environment. Thatnt is characterized by tight-knit familye variety of food ranging from traditionaluisine to American fast-food, and a tech-ption behavior that is quick to utilize theCT innovation. was conducted in Laredo, Texas, a gate-to Mexico. The climate in this regionpical with long very warm (3840 C)ers, punctuated by mild and very shortns [22]. The southern border region isa segment of the U.S. population thatically been underrepresented in nationaldies and in proling health character-ording to the U.S. Census Bureau [23],e population identify themselves as His-which 29% is foreign born. Thirty-eightrchers in the behavioral sciences. For example, Ker-ee [20] contend that a very large sample will makedifference signicant, which may not necessarily ofnicance.Predictors of child body mass e41percent of the population is below 18 years of age,whereas state and national gures are at 28% and24%, respectively. Under-education and poverty arehigh. Only 13% of the population has completed abachelors degree or higher compared to the stateand national averages of 23% and 24%, respectively.Approximately 27% of the residents currently livein poverty. Household median income is $36,537comparedof $50,000householdthe entiretively.Study deThe majoface surveCentral M21.4% (3 of researcthe schoolUniversity6th gradewere obtaeach of other for obtained.English anconsent wBecauswe surve2008 and subjects uitems, thThere, wscreen, boloud and aincluded pworks, meplayer owMeasuremconductedsured usinweight weuniform bwere usedis given bheight squrate and rage-gende2 The targ171 girls) 6t20082009 (schools realVariable denitionThe outcome variable, BMIp, was derived using theCDC on-line BMI calculator [24]. Subjects gender,birth date, height, and weight were plugged intothe on-linand the nu(010) we enver oftherumbed aich rividudual straee own ofer oet. l iPo, an, ows wshiptorss (0spen cont suakfar. A Infe scr saand, wh. Soin smed or ionsds Mticas wchnutiovideard dcalcequmpa to the state and the national gures and $52,000, respectively. The average size is 3.75 while those of the state and nation stand at 2.74 and 2.59, respec-sign and samplerity of data was obtained from a face-to-y of a random sample n = 55 6th graders iniddle School.2 Non-participation rate wasgirls and 12 boys). Prior to the conducth activities, approval was obtained from district, and the Texas A&M Internationals Institutional Review Board. Rosters ofrs, by gender and in alphabetical order,ined from the counselors ofce. Fromthose rosters one for boys and thegirls a systematic random sample was Using IRB-approved consent forms ind in Spanish child assent and parentalere obtained before data collection.e of the busy schedule at the school,yed in two sessions: one in Decemberthe other in April 2009. To ensure thatnderstood instructions and questionnaireey were gathered at the school library.ith the questionnaire shown on wide-th instructions and items were read-outnswered in a synchronized manner. Itemsersonal and family characteristics, net-als and drinks, cell phone and iPod/MP3nership, Internet-, sports-, and TV-hours.ents of height (m) and weight (kg) were by the school nurse. Height was mea-g a stadiometer, and both height andre measured with subjects wearing theirut without shoes. These measurements to calculate the subjects BMI, whichy an individuals weight divided by theirared [24]. Because BMI is not an accu-eliable measure of child body mass, ther specic BMI percentile (BMIp) is used.et population comprised of N = 298 (127 boys andh graders at Central Middle School in school yearnote: to ensure anonymity and condentiality, the name is not used).socialnumbof brothe nwe usin whof indindivirecallThrused: ershipnumbInternactuaobtainHenceplayerownerpredicfriendweek Weor noat bresuppeif no.at thburgeJuice $1.00$0.50able consuselvesoccasing kiStatisResultcal tedistribTo prostandwere and frTo coe calculator [24]. The number of siblings,mber of school and neighborhood friendsre used to measure the childrens micro-ironment. To solicit information about the siblings, subjects were asked the numbers and sisters they had. In coming up wither of school and neighborhood friends, name generator and name interpreter,espondents were asked to list the namesals and provide information about theses. This technique is superior to the surveytegy.rough measures of ICT utilization wereership of cell phone (1 = yes, 0 = no), own- iPod/MP3 player (1 = yes, 0 = no), andf hours in a typical week spent on theMeasures of actual cell phone and ofd/MP3 player usage were challenging tod were prone to very unreliable results.nership of cell phones and of iPod/MP3as used as reliable measures. Essentially, was used as a proxy for utilization. Other were number of school and neighborhood10), and number of hours in a typicalt in sports activities.trolled for other factors such as whetherbjects drink water, juice, milk, or sodast, at lunch, during snack time, and atll these were coded 1 if yes, and 0ormants told us that the typical menuhool cafeteria comprised of either andwich or pizza, both served with milk. bottled water can be purchased forile an additional serving of milk costda drinks were neither sold nor avail-chool; therefore, soft drinks that were were either brought by the kids them-by their parents/guardians. On several, parents/guardians were observed bring-cDonalds meals during lunch.l analysisere derived using a variety of statisti-iques: descriptive statistics, frequencyns, correlation, and regression analyses. summary information, means, medians,eviations, minimum and maximum valuesulated for numerical variables (Table 1);encies for categorical variables (Table 2).re boys and girls on the variables ine42 M.A. Ynalvez et al.Table 1 Descriptive statistics for numerical variablesa.Variables n Mean Median SD Min MaxBMI percentileb 52 66.90 71.50 26.88 6.00 99.00Height (cm) 53 150.32 149.86 7.71 134.62 170.18Weight (kg) 53 47.89 44.91 13.22 29.48 89.36Age (years) 54 12.13 12.02 0.48 11.30 13.70No. of siblings 55 2.51 2.00 1.36 0.00 6.00No. of school and neighborhood friends 55 4.33 4.00 2.91 0.00 10.00Internet h 16.50 25.35 0.00 77.00TV hours 21.0 13.93 0.00 55.00Sports ho 12.00 12.88 0.50 52.00a Overa ariabb DeriveTable 2 VariablesSubject iLiving witBoth pareOwns a cOwns an Eats breaDrinks waDrinks juiDrinks miDrinks soDrinks waDrinks juiDrinks miDrinks soDrinks waDrinks juiDrinks miDrinks soDrinks waDrinks juiDrinks miDrinks soa Each ob OveraTables 1 used. To ccorrelatiowere empcontributiear regreemployed3 A point son correlatdichotomouours in a week 55 28.19 in a week 55 21.40 urs in a week 51 15.13 ll sample size is n = 55. Due to missing values, sample sizes per vd using CDCs on-line child BMI calculator (www.cdc.gov).Frequency distribution of categorical variablesa.b Yesn % s female 32 58 h both parents 36 65 nts working 29 53 ell phone 44 80 iPOD/MP3 player 40 73 kfast 51 93 ter at breakfast 12 22ce at breakfast 32 58 lk at breakfast 31 56 da at breakfast 7 13 ter at lunch 12 22ce at lunch 11 20lk at lunch 18 33da at lunch 33 60ter at supper 14 25ce at supper 10 18 lk at supper 8 15 da at supper 36 65 ter during snacks 28 51 ce during snacks 17 31 lk during snacks 4 7 da during snacks 31 56 f the variables listed represents a question answerable by either a ll sample size is n = 55. Due to missing values, sample sizes per variaband 2, independent samples t-test wasompute bivariate correlations, a Pearsonn and a point-biserial correlation analysisloyed [25].3 To examine the simultaneouson of predictors on BMIp, multiple lin-ssion analysis with dummy variables was.bi-serial correlation analysis is essentially a Pear-ion analysis between an interval-ratio level and as nominal level variable [25].Two regof 20 predpredictorsselected f[2527].44 As descrition begins wselected by F-value. If told then this dropped. tted, and tle may be less than this number.Non %23 4219 3526 4711 2015 274 743 7823 4224 4448 8743 7844 8037 6722 4041 7545 8247 8519 3527 4938 6950 9123 42yes or a no.le may be less than this number.ression models were derived: M1 consistsictors selected from the list of potential in Tables 1 and 2. M2 has 14 predictorsrom M1 by called backward eliminationPredictors in M1 were selected based onbed in Neter et al. [27] on p. 353, backward elimina-ith a full model containing all potential predictorsthe analyst, and identies the one with the smallesthat F-value is less than a predetermined thresh-at predictor associated with the smallest F-valueThe model with the remaining predictors is againhe next predictor to be dropped is identied. ThePredictors of child body mass e43a review of the literature and this studys hypothe-ses. In contrast, predictors in M2 were selectedbased on their signicance probabilities under M1.Why was it necessary to generate two models?These models were generated to verify if the signif-icance of the predictors was stable across varyingnumber of predictors, and to identify predictorsthat were signicant in both models. Because ofbudgetary constraints5 and refusal rates, the nalsample sisampling psion coefHence, in(SE), robuboth modthe infereAlthougexhibits asample sieffort of regression(M1 and of BMIp rof the frthe Kolmop = 0.200;that the ttributed. M1 and M2ity amongResultsResults artains theminimum variables,and perceTable 3, tboys and fresults ofthese twoical variabthe perceto each ofthe p-valut-tests coare shownregressionprocess goesless than the5 Each subing in the stFrom Tables 1 and 2, it is clear that BMIp rangedfrom 6.00 to 99.00. Its mean is 66.9 26.9. Of the55 respondents, 58% (32) were girls and 42% (23)were boys. Average age was 12.1 0.5 years, withthe youngest about 11 (11.3 years) and the oldestalmost fouof 55) of rwhile ftyents works, 4whors; aoutoutle m of ar p.term (no8.2% snaion andand pop only11),ter,inteble con 55)ng w, soddrin bein thesomAt s5% (da,sodale 4and annt (1nt (2teen of acks werrou thaer ondsze was small. Even with a well-designedlan, the risk of obtaining unstable regres-cients comes with such a small sample.stead of using the usual standard errorsst standard errors (RSE) were used forels to calculate t-values associated withntial tests for the coefcients.h multiple linear regression results high degree of reliability even for smallzes [28], the authors made the extra(i) using RSE-based instead of SE-based results, and (ii) building two modelsM2). Finally, in checking for normalityesiduals from M1 and M2, examinationequency distributions, histograms, andgorovSmirnov tests (KS = 0.800 for M1, KS = 0.104 for M2, p = 0.200) indicatedwo sets of residuals were normally dis-Values for the variance ination factor for indicated no problems of multicollinear- predictors [25].e presented in Tables 17 . Table 1 con- means, medians, standard deviations;and maximum values for all numerical while Table 2 presents the frequenciesntages for all categorical variables. Inhe means and the standard deviations foror girls are presented, together with the independent samples t-tests comparing groups with respect to each of the numer-les. Table 4 displays the frequencies andntages for boys and for girls, with respect the categorical variables. It also presentses associated with independent samplesmparing both groups. Correlation results in Table 5. In Tables 6 and 7 M1 and M2 results are presented. on until there are no predictors that yield F-value threshold.ject received a $25.00-gift certicate for participat-udy.siblingwith matte(or ab(or ab(a litttermscellulplayerIn lunchand 9havingsumpt(21), milk, were menu20% (ing wais an availadrink (17 ofdrinkiAgainsoda theseeitherhave both. and 6and sodrink Tabboys phoneperceperceSeven0% (0ing sntheretwo ging isnumbof frierteen (13.7 years). Sixty-ve percent (36espondents were living with both parents,-three percent (29 of 55) had both par-ing. The average respondent had 2.5 1.4.3 2.9 school and neighborhood friendsm they played and discussed importantnd, spent 28.2 25.4 h in a typical week 4 h/day) on the Internet, 21.4 13.9 h 3 h/day) watching TV, and 15.1 12.9 hore than 2 h/day) engaging in sports. InICT, 80% (44 of 55) of subjects owned ahone, while 73% (40 of 55) an iPod/MP3s of meals, all subjects reported eatingt shown in Table 2), while 92.7%, 92.6%, reported eating breakfast, supper, andcks, respectively. As regards to uid con-at breakfast, 22% (12 of 55), 58% (32), 56% 13% (7) reported drinking water, juice,soda, respectively. Clearly, juice and milkular choices despite that the breakfast offered milk. At lunch, 22% (12 of 55), 33% (18), and 60% (33) reported drink- juice, milk, and soda, respectively. Thisresting observation because soda is notin school, and yet it is the most popularsumed. For snacks, 51% (28 of 55), 31%, 7% (4 of 54), and 56% (31 of 54) reportater, juice, milk, and soda, respectively.a is a popular choice. The popularity ofks at lunch and during snacks, despiteg unavailable in school could imply that kids bring soda with them to school,eone bring soda for them to school, orupper, 25% (14 of 55), 18% (10), 15% (8),36) reported drinking water, juice, milk, respectively. Clearly, majority of children at lunch, during snacks, and at supper. indicates signicant differences betweengirls with respect to owning a cellulard drinking milk during snacks. Sixty-ve5 of 23) of boys compared to ninety-one9 of 32) of girls owned a cellular phone. percent (4 of 23) of boys compared to32) for girls reported drinking milk dur-. For all other variables (Tables 3 and 4),e no signicant differences between theps. In Table 5, the most salient nd-t of the correlation between BMIp andf siblings (r = 0.458; p < 0.001). Number was positively correlated with numbere44 M.A. Ynalvez et al.Table 3 Comparison of means between boys and girlsa.Variables Boys Girls p-Valuebn Mean SD n Mean SDBMI percentilec 21 73.64 25.63 31 62.06 25.78 0.104Height (cm) 22 150.09 9.25 31 150.47 6.57 0.860Weight (kg) 22 48.89 13.35 31 47.19 13.31 0.650Age (years) 22 12.09 0.45 32 12.16 0.50 0.578No. of siblings 23 2.43 1.34 32 2.56 1.39 0.733No. of school and neighborhood friends 23 3.74 2.70 32 4.75 3.02 0.064Internet hours in a week 23 30.55 27.36 32 26.50 24.11 0.448TV hours in a week 23 20.26 10.86 32 22.23 15.90 0.494Sports hours in a week 21 16.30 13.80 28 14.28 11.43 0.560a Overall sample sizes for boys and for girls are n = 23 and n = 32, respectively. Due to missing values, sample sizes per variablemay be less than these numbers.b p-Value is associated with a two-tailed independent samples t-test.c Derived using CDCs on-line child BMI calculator (www.cdc.gov).of Internet hours (r = +0.411; p < 0.001). Numberof Internet hours was positively correlated withsports hours (r = +0.454; p < 0.001) and ownership ofiPod/MP3 player (rpb = 0.291; p < 0.05) [25]. Despitesignicant correlations, there were no problems ofmulticollinearity in the regression analyses.M1 and M2 regression results are presented inTables 6 and 7, respectively. M1 indicated a verygood t between observed and predicted BMIp val-ues (R2 = 82.3%; adj-R2 = 68.8%). In other words, M1is able to account for 82.3% of the variation in BMIp.In general, results using usual standard errors (SE)Table 4 Comparison of percentages for categorical variables between boys and girlsa.Variables Boys Girls p-ValuebYes No Yes Non % n % n % n %Living with both parents 17 73.9 6.0 26.1 19.00 59.4 13.00 40.6 0.264Both parents working 15 65.2 8.0 34.8 14.00 43.8 18.00 56.3 0.120Owns a cell phone* 15 65.2 8.0 34.8 29.00 90.6 3.00 9.4 0.033Owns an IPOD/MP3 Player 16 69.6 7.0 30.4 24.00 75.0 8.00 25.0 0.662Eats brea 30.0Drinks wa 7.0Drinks jui 22.0Drinks mi 15.0Drinks so 4.0Drinks wa 6.0Drinks jui 5.0Drinks miDrinks soDrinks waDrinks juiDrinks miDrinks soDrinks waDrinks juiDrinks miDrinks soa Overamay be lesb p-Valudifferencekfast 21 91.3 2.0 8.7 ter at breakfast 5 21.7 18.0 78.3 ce at breakfast 10 43.5 13.0 56.5 lk at breakfast 16 69.6 7.0 30.4 da at breakfast 3 13.0 20.0 87.0 ter at lunch 6 26.1 17.0 73.9 ce at lunch 6 26.1 17.0 73.9 lk at lunch 6 26.1 17.0 73.9 12.0da at lunch 16 69.6 7.0 30.4 17.0ter at supper 7 30.4 16.0 69.6 7.0ce at supper 3 13.0 20.0 87.0 7.0lk at supper 4 17.4 19.0 82.6 4.0da at supper 15 65.2 8.0 34.8 21.0ter during snacks 9 39.1 14.0 60.9 19.0ce during snacks 4 17.4 19.0 82.6 13.0lk during snacks* 4 17.4 19.0 82.6 0.0da during snacks 15 65.2 8.0 34.8 16.0ll sample sizes for boys and for girls are n = 23 and n = 32, respectivelys than these numbers.e is associated with a two-tailed independent samples t-test. A var at the 5%. 93.8 2.00 6.3 0.7360 58.3 5.00 41.7 0.9910 68.8 10.00 31.3 0.0630 46.9 17.00 53.1 0.0940 12.5 28.00 87.5 0.9540 18.8 26.00 81.3 0.5250 15.6 27.00 84.4 0.3480 37.5 20.00 62.5 0.3830 53.1 15.00 46.9 0.2220 21.9 25.00 78.1 0.4810 21.9 25.00 78.1 0.4120 12.5 28.00 87.5 0.6200 65.6 11.00 34.4 0.9760 59.4 13.00 40.6 0.1440 40.6 19.00 59.4 0.0570 0.0 31.00 100.0 0.0430 51.6 15.00 48.4 0.327. Due to missing values, sample sizes per variableiable tagged by an asterisk (*) denotes signicantPredictors of child body mass e45Table 5 Correlation matrix.Variables Y X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 X17 X18 X19 X20 X21 X22 X23 X24BMI percentile Y 1.00 No. of siblings X1 .458** 1.00 No. of schooland neigh-borhoodfriendsX2 0.08 0.07 1.00 Internet hoursin a weekX3 0.15 0.13 .411** 1.00 TV hours in aweekX4 0.10 0.13 0.25 0.26 1.00 Sports hours ina weekX5 0.24 0.18 0.16 .454** 0.17 1.00 Owns a cellphone(1 = yes;0 = no)X6 0.07 0.18 0.20 0.07 0.04 0.01 1.00 Owns an IPOD(1 = yes;0 = no)X7 0.03 0.23 0.14 .291* 0.18 0.07 0.00 1.00 Eats breakfast(1 = yes;0 = no)X8 0.20 0.16 0.24 0.23 0.10 0.21 0.14 0.17 1.00 Drinks water atbreakfast(1 = yes;0 = no)X9 0.18 0.04 0.17 0.05 0.25 0.11 0.18 0.23 0.02 1.00 Drinks juice atbreakfast(1 = yes;0 = no)X10 0.11 0.01 0.14 0.14 0.06 0.01 0.13 0.14 0.10 0.00 1.00 Drinks milk atbreakfast(1 = yes;0 = no)X11 0.23 0.01 0.08 0.13 0.04 0.01 0.11 0.05 0.04 0.02 .300* 1.00 Drinks soda atbreakfast(1 = yes;0 = no)X12 0.26 0.10 0.01 0.11 0.12 0.23 0.06 0.01 0.11 0.06 0.23 0.01 1.00 Drinks water atlunch(1 = yes;0 = no)X13 0.17 0.17 0.11 0.12 0.20 0.01 0.07 0.03 0.02 .360** .269* 0.11 0.06 1.00 e46 M.A. Ynalvez et al.Table 5 (Continued)Variables Y X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 X17 X18 X19 X20 X21 X22 X23 X24Drinks juice atlunch(1 = yes;0 = no)X14 0.25 0.01 0.10 0.13 0.01 0.10 0.14 0.00 0.04 0.07 0.15 .348** 0.08 0.18 1.00 Drinks milk atlunch(1 = yes;0 = no)X15 0.08 0.05 0.24 .317* 0.18 0.02 0.04 0.10 0.05 0.01 0.12 0.09 0.15 0.18 0.25 1.00 Drinks soda atlunch(1 = yes;0 = no)X16 0.15 0.08 0.18 .287* 0.14 0.21 0.04 0.17 0.23 0.11 0.02 0.11 0.09 0.07 0.06 .538**1.00 Drinks water atsupper(1 = yes;0 = no)X17 0.03 0.13 0.14 0.08 0.01 0.14 0.13 .358** 0.00 .298* 0.16 0.09 0.10 0.20 0.02 0.13 0.14 1.00 Drinks juice atsupper(1 = yes;0 = no)X18 0.01 0.04 0.00 0.10 0.03 0.02 0.24 0.08 0.13 0.09 .304* 0.04 0.04 0.21 0.24 0.03 0.10 .374** 1.00 Drinks milk atsupper(1 = yes;0 = no)X19 0.15 0.04 0.10 0.05 0.07 0.04 0.08 0.02 0.12 0.03 0.14 0.16 0.00 0.16 0.18 0.18 0.08 0.12 0.06 1.00 Drinks soda atsupper(1 = yes;0 = no)X20 0.23 0.13 0.11 0.11 0.20 0.07 0.08 0.10 0.20 0.08 0.15 0.06 0.05 0.01 0.12 0.02 0.11 .278* 0.25 .568** 1.00 Predictors of child body mass e47Table 5 (Continued)Variables Y X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 X17 X18 X19 X20 X21 X22 X23 X24Drinks waterduringsnacks(1 = yes;0 = no)X21 0.25 0.08 0.18 0.18 .287* 0.09 0.07 .299* 0.01 0.22 0.22 .319* 0.07 .303* 0.02 0.11 0.04 0.20 0.15 0.02 0.00 1.00 Drinks juiceduringsnacks(1 = yes;0 = no)X22 0.07 0.17 0.03 0.14 0.08 0.12 0.14 0.12 0.19 0.16 0.09 0.19 0.02 0.03 .354** 0.12 0.18 0.24 .399**0.16 0.18 0.15 1.00 Drinks milkduringsnacks(1 = yes;0 = no)X23 0.14 0.22 0.20 0.08 0.21 0.06 0.03 0.14 0.19 0.02 0.04 0.25 0.11 0.21 0.21 0.04 0.23 0.16 0.05 .479** 0.24 0.00 0.03 1.00 Drinks sodaduringsnacks(1 = yes;0 = no)X24 0.06 0.07 0.18 0.25 0.01 .300* 0.03 0.14 0.10 0.19 0.09 0.09 .332* 0.16 0.03 0.06 .312* 0.08 0.03 0.06 0.07 0.11 0.10 0.04 1.00* Signicant rank correlation at the 0.05 level.** Signicant rank correlation at the 0.01 level.e48 M.A. Ynalvez et al.Table 6 Multiple Linear Regression Model (M1) for BMIpa.M1 predictors SE-based resultsb RSE-based resultsc S.E. c R.S.E.No. of siblings 0.35** 2.08 0.35** 1.88No. of school and neighborhood friends 0.05 0.91 0.05 0.89Internet hours in a week 0.25 0.13 0.25 0.10Sports hours in a week 0.34** 0.23 0.34*** 0.14Owns a cell phone (1 = yes; 0 = no) 0.26* 7.32 0.26 8.89Owns an iPOD/MP3 player (1 = yes; 0 = no) 0.42** 7.36 0.42*** 5.82Drinks juice at breakfast (1 = yes; 0 = no) 0.13 7.06 0.13 6.50Drinks milk at breakfast (1 = yes; 0 = no) 0.34* 7.43 0.34** 5.69Drinks water at lunch (1 = yes; 0 = no) 0.12 8.19 0.12 8.00Drinks juice at lunch (1 = yes; 0 = no) 0.14 9.09 0.14 7.00Drinks milk at lunch (1 = yes; 0 = no) 0.40* 9.85 0.40** 6.33Drinks soda at lunch (1 = yes; 0 = no) 0.35* 9.12 0.35** 6.13Drinks water at supper (1 = yes; 0 = no) 0.48** 8.01 0.47** 9.57Drinks juice at supper (1 = yes; 0 = no) 0.23 8.20 0.23 9.43Drinks milk at supper (1 = yes; 0 = no) 0.83*** 12.95 0.83*** 13.63Drinks soda at supper (1 = yes; 0 = no) 0.84*** 7.85 0.84*** 9.51Drinks water during snacks (1 = yes; 0 = no) 0.53*** 6.62 0.53*** 5.53Drinks juice during snacks (1 = yes; 0 = no) 0.00 6.96 0.00 6.50Drinks milk during snacks (1 = yes; 0 = no) 0.31* 13.40 0.31* 11.81Drinks soda during snacks (1 = yes; 0 = no) 0.14 5.81 0.14 5.26Coefcient of determination (R2) 82.30 82.30 Adjusted coefcient of determination (adj. R2) 68.80 68.80 Sample size (n) available for analysisd 47 47 a denotes the standardized regression coefcients.b *, **, ***Denote signicance at the 0.05, 0.01, and 0.001 levels, respectively.c Regression coefcients for SE- and RSE-based results are the same, but not the standard errors.d Overall sample size is n = 55 but due to missing values, effective sample size is less than this number.Table 7 Multiple Linear Regression Model by Backward Elimination Method (M2) for BMIpa.M2 predictors SE-based resultsb RSE-based resultsc S.E. c R.S.E.No. of siblings 0.34** 1.97 0.34*** 1.73Internet hours in a week 0.29* 0.12 0.29* 0.13Sports hours in a week 0.37** 0.22 0.37*** 0.17Owns a cell phone (1 = yes; 0 = no) 0.24* 6.39 0.24 7.86Owns an iPOD/MP3 player (1 = yes; 0 = no) 0.33** 6.12 0.33*** 4.99Drinks milk at breakfast (1 = yes; 0 = no) 0.23* 5.81 0.23** 4.70Drinks milk at lunch (1 = yes; 0 = no) 0.44** 7.13 0.44*** 5.67Drinks soda at lunch (1 = yes; 0 = no) 0.39** 6.52 0.39*** 4.70Drinks water at supper (1 = yes; 0 = no) 0.41** 6.76 0.41** 8.30Drinks juice at supper (1 = yes; 0 = no) 0.18 6.48 0.18 7.02Drinks milk at supper (1 = yes; 0 = no) 0.79*** 11.17 0.79*** 10.63Drinks soda at supper (1 = yes; 0 = no) 0.76*** 6.91 0.76*** 7.52Drinks water during snacks (1 = yes; 0 = no) 0.46*** 5.88 0.46** 6.58Drinks milk during snacks (1 = yes; 0 = no) 0.30* 11.31 0.30** 9.78Coefcient of determination (R2) 79.60 79.60 Adjusted coefcient of determination (adj. R2) 70.70 70.70 Sample size (n) available for analysisd 47 47 a denotes the standardized regression coefcients.b *, **, ***Denote signicance at the 0.05, 0.01, and 0.001 levels, respectively.c Regression coefcients for SE- and RSE-based results are the same, but not the standard errors.d Overall sample size is n = 55 but due to missing values, effective sample size is less than this number.Predictors of child body mass e49and robust standard errors (RSE) give very simi-lar pattern of signicant variables, except for cellphone ownership, which was signicant in the SE-,but not in the RSE-based results. All other signi-cant predictors in the SE- are also signicant in theRSE-basedrienced shBecausresults arenumber ohours in an iPod/Ming milk amilk at luat lunch supper (per ( = ( = 0.84( = 0.53snack ( associatedIn othehours; dridrinking wlow BMIp ing soda alunch, anBMIp kids.are presensignicantilar, with added to T( = 0.29results inspent in aIn the possible eresearch both M1 results) aand are dDiscussiSiblings aAn indiviindividualresourcesand healtmicro-socthat relatwith schooTable 6 (showed that of those two aspects, number of sib-lings was highly signicantly associated with bodymass. Specically, results indicated that greaternumber of siblings within the household was signi-cantly associated with children who have low BMIp.esuls (sl and, thilardergiblinto ]. Iiblinir m enve nactivncretivitposnt whooer of thvelyith ctoudyen ay ly rthe.g. ard,eralingshadnmetiviby eed o aties.ctivs to sing vidi phy of te ed ch tfamodel results, although some predictors expe-ifts in their degrees of signicance.e of the small sample size, the RSE-based appropriate. These results indicate thatf siblings ( = 0.35; p < 0.010), sport-a week ( = 0.34; p < 0.001), owningP3 player ( = +0.42; p < 0.001), drink-t breakfast ( = +0.34; p < 0.010), drinkingnch ( = +0.40; p < 0.001), drinking soda( = +0.35; p < 0.01), drinking water at = 0.47; p < 0.01), drinking milk at sup-0.83; p < 0.001), drinking soda at supper; p < 0.001), drinking water during snacks; p < 0.001), and drinking milk during= +0.31; p < 0.05) were all signicantly with BMIp.r words, having more siblings, more sportsnking water, milk, and soda at supper; andater during snacks are all associated withkids. In contrast, owning an iPod, drink-t lunch, and drinking milk at breakfast,d during snacks are associated with high Like Table 6, SE- and RSE-based resultsted in Table 7. Obviously, the pattern of variables in Tables 6 and 7 is very sim-the exception of one signicant variableable 7, namely: Internet hours in a week; p < 0.05). In other words, RSE-based Table 6 indicated that Internet hours week is associated with high BMIp kids.next section, links to the literature,xplanations, and hypotheses for futureare forwarded. Signicant predictors in(RSE-based results) and M2 (RSE-basedre considered stable predictors of BMIpiscussed in the following section.onnd friendsduals social environment shapes thats access to material and non-material; as well as that individuals life chancesh status. This study focused on childrensial environment, specically, its aspectse to their interaction with siblings, andl and neighborhood friends. Results fromM1 RSE) and 7 (M2 RSE) consistentlySuch rfriendschooHenceSimon kinwith slikely [19,29that sby thesocialand thtion, that ical acnot imronme(or scnumbone opositiity, wrisk fathis stbetwethis mter onthus oties (ebackyconsidSibhave envirocal acthereproviding tactiviical aacceswatchbe proto belihoodevaluareportof suence for mts appear to highlight the salience of kin-iblings) over non-kin-friends (friends in neighborhood) in predicting body mass.e question: why siblings and not friends? results are reported in studies conductedarten-level age-groups in which childrengs had lower BMI, and hence were lessbe obese than children without siblingst stands to reason that it is very likelygs in the same household as subjects,ere co-presence, may have provided theironment that allowed ready access toeeded stimulus for child-to-child interac-e cooperative play, and other activitiesase the time a child allocates to physi-y. All these would have been difcult, ifsible, to come by if a childs social envi-as mainly comprised of non-kin-friendsl and neighborhood friends) [29]. In af studies [3032], number of siblings wase variables found to be signicantly and associated with levels of physical activ-a small number of siblings serving as ar for obesity. Although the results from did not show any signicant correlationnumber of siblings and physical activity,be attributable to the fact that the lat-eferred to number of sports hours, andr forms of non-structured physical activi-running around the house, playing in the or leisure biking) were not taken intotion. living in the same household may either (i) created an immediate socialnt conducive to non-structured physi-ties (e.g. running around or rough play)xpending calories in the process, or (ii)constant availability of playmates result-lmost limitless hours of a variety of One of the challenges to increasing phys-ity among obese children is their easyedentary alternatives [33], which includeTV and playing video-games. Siblings mayng the needed motivation to allocate timesically active, which decreases the like-sedentary activities. In a study done tofamily-based obesity treatment, it wasthat one of the potential advantagesreatment is the opportunity to inu-ily members. With greater opportunitiesing and supporting behavioral change ine50 M.A. Ynalvez et al.larger families, these children showed signicantpositive response to family-based obesity treat-ment [33].The results of this study also suggest that greaternumber of siblings in the household may haveforced decreasinsharing tyhouseholdneighborhamong indin turn inhmay haveated withalso be thsame housharing osubjects computerporary timThe lityouth whoity includoutdoor psafety cothe perceneighborhphysical asafe spacphysicallyHaldemancan be hysame housand sociato be physmental faGiven close famappears timportantactivity anside envirdecrease cal activias importing the tr[31]. Thuslic spacesan urban ow of trknow eachical regioit stands friends wco-presenschool anwho are less accessible in the engagement of phys-ical activities because of their not being readilyavailable).Cell phones and iPods/MP3 playersilizaarce an [38ole ship of aP3 of . Bative notf iPoted gs soporhoney isredtams: a owties,atedesuties ing. erda com outt is ed s coly [4vidieertivithatt anIp, mion atiocont utictiviing. to obe-sedy trethe sharing of food resources therebyg the availability of food per child. Suchpically occurs with siblings in the same more than it would with friends, evenood friends. This food-resource sharingividuals living in the same household mayibit or limit excessive calorie in-take that contributed to lower body weight associ- subjects who have many siblings. It coulde case that having many siblings in thesehold not only results to the forced-f food, but could have also diminishedopportunities to access and use homes and Internet facilities which, in contem-es, are the platform for games.erature on child obesity suggests that are at risk of decreased physical activ-e children living in neighborhoods wherehysical activity is restricted by climate,ncerns, or lack of facilities [34]. Evenption of environmental factors such asood safety has been noted as barriers toctivity [35,36]. Hence, the need to createe to allow families to exercise, and to be active has been suggested in Gruber and [35] and Chatterjee et al. [37]. Thus, itpothesized that siblings living within theehold create a readily accessible physicall environment that allows these childrenically active without reliance on environ-ctors that promote physical activity.the conservative orientation of and theily ties among Hispanics-Americans, ithat the home environment is a more factor in shaping childrens physicald body mass development than their out-onment [32]. Environmental stimuli thatsedentary behaviors and increase physi-ty within the home have been identiedant targets in preventing weight gain dur-ansition from childhood to adolescence, in an environment (i) with limited pub- for physical activities, (ii) situated withinarea located by the border with a highansient migrants wherein people hardly other, and (iii) embedded in a geograph-n that is one of the hottest in the U.S.;to reason that the role of siblings (kin-ho are more accessible because of theirce and co-location) takes salience overd neighborhood friends (non-kin-friendsICT utjob seleisurstatustive rownerershipiPod/Mdictorin M2predictentlyrole oindicandinwho recell pWhtant pis tantion iplayeractiviassocimay ractivisnackLeathphoneels of[41]. Iobservphonequentby prowith pical acgiven detecon BMdirectassociIn playertary asnacklengesamongto nonobesittion impacts many aspects of daily life:h, productivity in work and in school,d sociability, and even personal health,39]. This study investigated the predic-of cell phone and of iPod/MP3 player on body mass distribution. In M1, own- cell phone was not, but ownership of anplayer was shown to be a signicant pre-child BMIp. These same results are seensed on these results, the hypothesized role of cell phone ownership was consis- observed in both models. In contrast, thed/MP3 player ownership was consistentlyby these two regression results. Thesemehow deviate from Lajunen et al. [40]t a weak but positive correlation between usage and BMI. iPod/MP3 player ownership an impor-ictor of BMIp? Assuming that ownershipount to usage, then a possible explana-lthough both cell phone and IPod/MP3nership are linked to seemingly solitary cell phone use is more likely to be with person-to-person interaction whichlt in planning and executing organizedwith peers, and decreasing tendency forSupporting evidence can be adduced fromle [41] who reports that increased cellmunication is associated with high lev-door physical activities among childrenalso reported that higher self-esteem wasamong children who frequently use cellmpared to those who use these less fre-2]. In this sense, cell phone ownership,ng enhanced opportunity for interactionss, may indirectly result in increased phys-ty and decreased sedentary behavior. But results from both M1 and M2 failed toy predictive role of cell phone ownershipore research is needed to establish theand magnitude of cell phone ownershipsn with body mass.rast, the authors argue that iPod/MP3lization is a predominantly solitary seden-ty, which may be conducive to eating andAs mentioned earlier, one of the chal-increasing physical activity, particularlyse children, is how to provide easy accessentary activities [33]. Thus, a number ofatment programs is focused on strategiesPredictors of child body mass e51that combine ways to reduce sedentary behaviors,induce physical activity to increase energy expen-diture, and decrease opportunities for unnecessaryenergy in-take [33]. In rening those strategies,analysts will need to recognize and consider therole of iPowith otheopment.DespiteLaredo, r(80%; 44 of 55) owbehind thhypothesetral Middleof the genLaredo is U.S. Highnorth is famodern, householdof childreareas seecans, whoenjoying tthat cell pable in ththese novers and laA t-tesownershipresults indership andrate mighthey are interactiotions withthemselveboys as w a claim while bare more cell phoneting, and ado with BMwith BMIpDrinks atRSE-basedesting patsnacks in rat breakfassociatedat supperdrinking msupper are all associated with low BMIp. The nd-ing that milk intake at breakfast is associated withhigher BMIp is contrary to most ndings that milkconsumption specically at breakfast contributedto lower body mass. It is possible that the schoolria uch ajoilk.he tlatenot e an sece a vegburg eamplds w by thouirov stuIp wnd datedch. swered d asmednchmpt pred t dums oageshoierg andlustudynt s prtancopmat hducer od/MP3 player usage, and its interactionr factors that inuence body mass devel- the low socioeconomic situation inesults indicate high rates of cell phoneof 55) and of iPod/MP3 player (73%; 40nership. What could be the explanationese seemingly paradoxical ndings? Threes are in order: (i) it is possible that Cen- Schools students are not representativeeral population. According to informants,roughly divided into north and south withway 59 as the anecdotal boundary. Therther from the U.S.-Mexico border, moreand is populated by more high-incomes than the south; (ii) it may be the casen from a minority population in peripheralking to identify with mainstream Ameri- are typically depicted by mass media ashe latest fad in new ICT; or (iii) it may behones and iPods are much more afford-e U.S. that even low-income parents ndel devices cheaper than personal comput-ptops to give kids.t revealed signicantly lower cell phone among boys than girls, but regressionicated no association between such own- BMIp. Girls high cell phone ownershipt be explained by the hypothesis thatmore sociable and into conversationaln (i.e., chatting, texting). From conversa- subjects (and with informants, who weres students at Central), girls describedasting money and time on video-games consistent with the ndings of Park [43]oys viewed girls as busy talking. If girlsinto cell phones than boys, and if much of use among girls is for chatting and tex-lso if cell phone ownership has nothing toIp then it is gender that might associate. meals and during snacks results from M1 and M2 reveal two inter-terns regarding drinks at meals and duringelation to child BMIp. First, drinking milkast, at lunch, or during snacks are all with higher BMIp; while drinking milk is associated with lower BMIp. Second,ilk, drinking soda, and drinking water atcafeteings sthe mwas mand tchocowere clusivto theat homsh orthan maybeand siare kiforcedeatingother they pfutureBMmilk aassociat lunsugar-prepalimiteconsuing luconsumealsreportsumedin terbeverfood cthe endrinksConcThis sronmeplayerimpordevelsis thand renumbdid not have other breakfast food offer-as a variety of grains and fruits, and thusr source for caloric energy for subjects In this study, the quantity of milk intakeype of milk preferred (i.e., plain versus-avored milk; or skim versus whole milk)measured. Hence, this nding is incon-d requires further research. With regardond observation, it could be that foodsre those (e.g. traditional Mexican foods,etables) less palatable to childrens tasteer sandwiches and pizzas. As such, kidsting less of these real foods at supper,y stuff themselves with uids (e.g. thereho hate eating vegetables, but whenparents, they would ease the difculty ofse foods by taking them with water, ords). These assertions are anecdotal, butide a basis for generating hypotheses fordies.as negatively associated with drinkingrinking soda at supper, but was positively with drinking milk and drinking sodaA possible explanation is that intake ofetened beverages and milk during family-meals (typically at supper) may be more to intake than that of these beverages in the presence of friends at school dur- time. Woodruff et al. [44] found lowion of sugar-sweetened beverages withepared and consumed at home. Theyhat although large portion sizes are con-ring home meals, such meals are lowestf caloric count and intake of sweetened. In this sense, the presence of healthyces during family meal times attenuatesy and caloric impact of sugar-sweetened milk on BMI.ion found that childrens micro-social envi-of siblings and ownership of iPod/MP3edict BMIp. These results underscore thee of family in conguring body massent among children, and support the the-aving siblings induce physical activitye sedentary behavior [45]. Although thef siblings is not a modiable factor ine52 M.A. Ynalvez et al.inuencing obesity, the nding that more siblingsin the household is signicantly associated withlower BMI suggests: (i) that health care providersfocus on interventions and preventive strategiesthat promote collaborative activities and siblinginteractioput more among ch(iii) that pinto consicies to prIn constion studiof child bAmericanmight be in other Uplayer owcal devicenon-physichild bodyactivities have no sare muchtime are such thatmembers mates.Finally,dren who those whofast, at luwith kids the timinnot this win BMIp? Oesis about(i.e., juiclinked to to keep indata. Hentionships may, at behypotheseConictThe authoAcknowThis reseInternatioaward to the rst author. The authors as a teamowe special thanks to the following: Cecilia Briones,Rosie Salazar, Roxanne Hernandez, David Canales,and Patricia Keck for helping plan and conductthe surveys; Isabel Perez and Cristina Rosales,nteata 9 Nesea Ynarovg fond lrs anntedrenristakcial ndicinang Tbrief dicinghesting the f08;16rokawbutio the esityilly JJy 200mundy in c08;13us Krs in Sou07;96rooqiesityd Medersehanssoolcurnalen M/comhaviosearcurson, Waeen urnalanzin, Grcial ntivity09;99n, (ii) that school-based efforts need toemphasis in promoting physical activitiesildren who do not have siblings [29], androximal social networks should be takenderation in developing programs and poli-omote physical activity among the youth.ideration of the location and the popula-ed, this is not to say that family inuencesody mass are exclusive to the Hispanic- population, but perhaps these inuencesstronger among Hispanic-Americans than.S. ethnic groups. In addition, iPod/MP3nership highlights how personal techni-s that encourage solitary and oftentimescal activities might strongly inuence mass development. Solitary non-physicalmay be very real threats for children whoiblings or for those who have siblings that older or younger; and, who at the samein a location with heavy migration ows these children largely depend on familyto assume the role of buddies and play- it is also the nding of this study that chil-drink milk at supper had lower BMIp than did not. However, drinking milk at break-nch, or during snacks was all associatedwith higher BMI. Might this suggest thatg of drinking milk inuences whether orill contribute to an increase or a decreaser generally, might this point to a hypoth- how the timing of and the types of uide, milk, soda or water) consumed arechild weight. Ultimately, it is important mind that this study used cross-sectionalce, no attempts to derive causal rela-should be made. Instead, these resultsst, be taken as a springboard to generates that can guide future research. of interestrs declare no conict of interest.ledgementsarch was funded by a 2008 Texas A&Mnal University, University Research Grantwho enary da 200mer RAmielto impNursintory, agradeconseRefe[1] ChsoMe[2] Hua Me[3] Hugein 20[4] KutriofOb[5] Reet[6] Edsit20[7] Hutoin 20[8] Faoban[9] AnJoschJo[10] ChTVbeRe[11] LaDAscrJo[12] FrMJsoac20red, validated, and performed prelimi-analysis for this research project throughIH-TAMIU PUENTES Undergraduate Sum-rch Experience; Susan Aguilar and Marcuslvez, who provided insightful commentse the manuscript; the Canseco School ofr granting access to its computer labora-ast but certainly not least to the 6thd their parents/guardians who assented,, and participated in this research.cesis NA, Fowler JH. The spread of obesity in a largeetwork over 32 years. New England Journal ofe 2007;357(4):3709.T, Horlick MN. Trends in childhood obesity research:analysis of NIH-supported efforts, Journal of Law.e and Ethics 2007;35(1):14853. AR, Reilly JJ. Disease management programs tar-obesity in children: setting the scene for wellnessuture. Disease Management and Health Outcomes(4):25566.a N, Satoh H. Recent trends of body mass index dis-n among school children in Sendai, Japan: decreaseprevalence of overweight and obesity, 20032009. Research and Clinical Practice 2011;5(1):e18.. Childhood obesity: an overview. Children and Soci-7;21(5):3906.s LD. Social implications of overweight and obe-hildren. Journal for Specialists in Pediatric Nursing(3):191200., Ludvigsson JF, Enskar K, Ludvigsson J. Risk fac-childhood obesity- ndings from the all babiestheast Sweden (ABIS) cohort. Acta Paediatrica(9):131520. IS. Genetic and hereditary aspects of childhood: best practice and research. Clinical Endocrinologytabolism 2005;19(3):35974.n LF, Lillegaard ITL, verby N, Lytle L, Klepp K,on L. Overweight and obesity among Norwegianhildren: changes from 1993 to 2000. Scandinavian of Public Health 2005;33(2):99106., Liou Y, Wu J. The relationship betweenputer time and adolescents health-promotingr: a secondary data analysis. Journal of Nursingh 2008;16(1):7585. KR, Eisenmann JC, Welk GJ, Wickel EE, Gentilelsh DA. Combined inuence of physical activity andtime recommendations on childhood overweight. of Pediatrics 2008;153(2):20914.i L, Elliott MN, Cuccaro P, Schuster M, Gillilandunbaum JA, et al. Inuences of physical andeighborhood environments on childrens physical and obesity. American Journal of Public Health(2):2718.Predictors of child body mass e53[13] Nieters A, Becker N, Linseisen J. Polymorphisms in candi-date obesity genes and their interaction with dietary intakeon n-6 polyunsaturated fatty acids affect obesity risk in asub-sample of the EPIC-Heidelberg cohort. European Jour-nal of Nutrition 2002;41(5):21021.[14] Marsden P. Models and methods in social network analysis.In: Carrington PJ, Scott J, Wasserman S, editors. Recentdevelopments in network measurements. New York, NY:Cambridge University Press; 2005. p. 830.[15] Ahn Y, Kim K-W. Study on utilization status of internetand needs assessment for developing nutrition educationprograms among elementary school children. NutritionResearch Practice 2007;1(4):3418.[16] Lown DA, Braunschweig CL. Determinants of physical activ-ity in Americ[17] Beets MSocial provide2006;30[18] Salvy SE, PaluactivityPsychol[19] Bhargavbehavioweightsearly chJournal[20] Kerlinge4th ed.[21] Larwin while i2008;32[22] Escape com/es[23] US Cenhttp://[24] CDC. Agov/hechildren[25] Field AOaks, C[26] Draper York, N[27] Neter Jlinear m[28] McGuinhighly slogical [29] Chen AYearly ePrevent[30] Mazur ATenderachildrenand Env[31] Timperet al. Fand weight change in children. International Journal ofPediatric Obesity 2008;3(3):1607.[32] Crawford D, Cleland V, Timperio A, Salmon J, Andri-anopoulos N, Roberts R, et al. The longitudinal inuenceof home and neighbourhood environments on childrensbody mass index and physical activity over 5 years: theCLAN study. International Journal of Obesity 2010;34(7):117787.[33] Epstein LH, Paluch RA, Raynor HA. Sex differences in obesechildren and siblings in family-based obesity treatment.Obesity Research 2001;9(12):74653.[34] American Academy of Pediatrics. Council on SportsMedicine and Fitness and Council on School Health. Activehealthy living: prevention of childhood obesity throughreas06;11uber ildho09;6(ble Sencyset aed ch07;10atterd barrving Comque ity 09;11alvez. Wns o05;63junenn A, Kth boong athersed se phohavioalth cksonernetors ademmputrk H., seolesc07.odruysicaake ants. Jmuzznes (edirelow-income, overweight African American girls.an Journal of Health Behavior 2008;32(3):2539.W, Vogel R, Forlaw L, Pitetti KH, Cardinal BJ.support and youth physical activity: the role ofr and type. American Journal of Health Behavior(3):27889.J, Bowker JW, Roemmich JN, Romero N, Kiefferch R, et al. Peer inuence on childrens physical: an experience sampling study. Journal of Pediatricogy 2008;33(1):3949.a A, Jolliffe D, Howard LL. Socio-economic,ural and environmental factors predicted body and household food insecurity scores in theildhood longitudinal study-kindergarten. The British of Nutrition 2008;100(2):43844.r FN, Lee HB. Foundations of behavioral research. Belmont, CA: Wadsworth Thomson Learning; 2000.KH, Larwin DA. Decreasing excessive media usagencreasing physical activity. Behavior Modication(6):93856.to Laredo Texas. n.d. http://www.thecityoaredo.cape/.sus Bureau. State and county quickfacts; 2010.quickfacts.census.gov/qfd/states/48000.html.bout BMI for children and teens. http://www.cdc.althyweight/assessing/bmi/childrens bmi/abouts bmi.html.. Discovery statistics using SPSS. 3rd ed. ThousandA: Sage; 2009.N, Smith H. Applied regression analysis. 2nd ed. NewY: John Wiley & Sons; 1981., Kutner MH, Nachtsheim CJ, Wasserman W. Appliedodels. 4th ed. Boston, MA: McGraw-Hill; 1996.ness D, Bennett S, Riley E. Statistical analysis ofkewed immune response data. Journal of Immuno-Methods 1997;201(1):99114., Escarce JJ. Family structure and childhood obesity,hildhood longitudinal study kindergarten cohort.ing Chronic Disease 2010;7(3):A5060., Klimek K, Telega G, Hejda G, Wdowiak L, Maecka- E. Risk factors for obesity development in school from south-eastern Poland. Annals of Agriculturalironmental Medicine: AAEM 2008;15(2):2815.io A, Salmon J, Ball K, Baur LA, Telford A, Jackson M,amily physical activity and sedentary environmentsinc20[35] Grch20[36] Gaquonag20[37] Chanseof[38] Dubil20[39] YnWMsio20[40] Lanewiam[41] LebathbeHe[42] JaIntdicacCo[43] Paityad20[44] Wophintde[45] CogeAvailable online at www.scienced physical activity, policy statement. Pediatrics7(5):183442.KJ, Haldeman LA. Using the family to combatod and adult obesity. Preventing Chronic Disease3):A10610., Chang Y, Krull JL. Television watching and fre- of family meals are predictive of overweightnd persistence in a national sample of school-ildren. Journal of the American Dietetic Association7(1):5361.jee N, Blakely DE, Barton C. Perspectives on obesityriers to control from workers at a community center low-income Hispanic children and families. Journalmunity Health Nursing 2005;22(1):2336.RB, Ynalvez MAH. Internet practice and socia-in South Louisiana. New Media and Society(4):487507. MA, Duque R, Sooryamoorthy R, Mbatia P, Shrumhen do scientists adopt the internet? dimen-f connectivity in developing areas. Scientometrics(1):3967. H-R, Keski-Rahkonen A, Pulkkinen L, Rose RJ, Rissa-aprio J. Are computer and cell phone use associateddy mass index and overweight? A population studytwin adolescents. BMC Public Health 2007;7:24.dale ST. Factors associated with communication-edentary behaviors among youth: are talking onne, texting, and instant messaging new sedentaryrs to be concerned about? Journal of Adolescent2010;47(3):3158. LA, von Eye A, Fitzgerald HE, Witt EA, Zhao Y.t use, videogame playing and cell phone use as pre-of childrens body mass index (BMI), body weight,ic performance, and social and overall self-esteem.ers in Human Behavior 2011;27(1):599604. Longitudinal relationships between physical activ-dentary behaviors and obesity in children andents. University of North Carolina at Chapel Hill;ff SJ, Hanning RM, McGoldrick K. The inuence ofl and social contexts of eating on lunch-time foodmong Southern Ontario, Canada, middle school stu-ournal of School Health 2010;80(9):4218.ie AG, Allison DB. The search for human obesitycover story). Science 1998;280(5368):1374.ct.comDo cell phones, iPods/MP3 players, siblings and friends matter? Predictors of child body mass in a U.S. Southern Border Ci...IntroductionMethodsContextStudy design and sampleVariable definitionStatistical analysisResultsDiscussionSiblings and friendsCell phones and iPods/MP3 playersDrinks at meals and during snacksConclusionConflict of interestAcknowledgementsReferences