Research ArticleParental Information, Motivation, and Behavioral SkillsCorrelate with Child Sweetened Beverage ConsumptionL. Suzanne Goodell, PhD, RD1; Michelle B. Pierce, PhD, RD2; K. Rivet Amico, PhD3;Ann M. Ferris, PhD, RD4

1Departmesity, Raleig2DepartmeConnecticu3Center for4DepartmePublic HeaAddress forPolicy, Un06108; Pho�2012 SOdoi:10.1016

240

ABSTRACT

Objective: To evaluate fit of the Information-Motivation-Behavioral Skills (IMB) model applied tosweetened beverage (SB) consumption in children.Design: Cross-sectional. Parents completed a home beverage inventory and IMB survey regarding SBconsumption.Setting: Health fairs, Special Supplemental Nutrition Program forWomen, Infants, and Children clinics.Participants: Convenience sample of 198 parents of low socioeconomic status.Main Outcome Measures: Independent variables included scores from 3 indices calculated from theIMB survey, information, motivation, and behavioral skills. The dependent variable was average childdaily caloric consumption from SB consumption calculated from the home beverage inventory.Analysis: Structural equation modeling.Results: Parental information had direct and indirect negative relationships with SB consumption. Paren-tal motivation was only indirectly associated with SB consumption mediated through behavioral skills.Parental behavioral skills had a negative correlation with SB consumption.Conclusions and Implications: In applying the IMBmodel to SB consumption, the authors found pre-liminary support for relationships between parental information, motivation, behavioral skills, and childSB consumption. Application of this model shows promise in identifying the complex interactionsbetween factors influencing consumption behaviors. As nutrition educators frequently desire to alterbehaviors, use of the IMB model may help with both targeting and assessment efforts.Key Words: poverty, beverages, parents, preschool, Information-Motivation-Behavioral Skills (IMB)(J Nutr Educ Behav. 2012;44:240-245.)

INTRODUCTION

Preschool children are consuming sig-nificantly more calories from sugar-sweetened beverages (SB) than everbefore.1 Researchers have identifiedSB as a possible contributor to exces-sive weight gain and as a likely factorin weight management problems inchildren.2,3 Preschool children fromfamilies of lower socioeconomicstatus (SES) are more likely toconsume SB between meals than arethose from higher-SES families, and

nt of Food, Bioprocessing, and Nutrih, NCnt of Medicine, Division of Public Ht Health Center, East Hartford, CTHealth, Intervention and Preventiont of Medicine, University of Conlth and Health Policy, University ofcorrespondence: Ann M. Ferris, PhDiversity of Connecticut Health Centne: (860) 208-2257; Fax: (860) 679-1CIETY FOR NUTRITION EDUC/j.jneb.2010.07.012

these children are more than 3 timesas likely to be obese at 4.5 years ofage.4 In a previous study with ethnicminority families from lower SES, theauthors found that parents did notdifferentiate between 100% juice andSB.5 To address the issues surroundingchildhood obesity and SB consump-tion, nutrition educators shouldwork with parents, the primary gate-keepers of the dietary intake of pre-school children.

Although several correlates of child-hood obesity have been identified, to

tion Sciences, North Carolina State Univer-

ealth and Population Sciences, University of

n, University of Connecticut, Storrs, CTnecticut Health Center and the Center forConnecticut, Farmington, CT, RD, Center for Public Health and Health

er, 99 Ash St, MC 7160, East Hartford, CT349; E-mail: aferris@uchc.eduATION AND BEHAVIOR

Journal of Nutrition Education and Beh

date, there continues to be a lack ofanorganizing theoryofwhat drives pa-rental behaviors that promote overex-posure to SB. To address this gap inthe literature, the current study evalu-ated a well-supported health behaviormodel, the Information-Motivation-Behavioral Skills (IMB) model,6,7 toexamine parental factors influencingtheir preschool children’s SBconsumption. Researchers havevalidated use of the IMB model withother health-related behaviors.8-10

The core constructs of the modelinclude information, motivation, andbehavioral skills.7

In the IMB model, information iscomposed of facts, heuristics, and im-plicit theories. The motivation con-struct includes both personal andsocial motivation,10,11 and behavioralskills refer to abilities and self-efficacyin implementing those skills invarioussituations. Specifically articulated toSB consumption, critical parental in-formation may include knowledge ofthe deleterious effects of SB overcon-sumption, what recommended

avior � Volume 44, Number 3, 2012

Journal of Nutrition Education and Behavior � Volume 44, Number 3, 2012 Goodell et al 241

exposure is, and what viable alterna-tives are available and/or recommen-ded. For motivation, personalmotivation may include parents’ atti-tudes and beliefs about the positiveand negative consequences to boththemselves and their children of al-lowing overconsumption and for im-plementing restrictions on theirchild’s consumption. Social motiva-tion includes beliefs and attitudes ofother people of importance to the par-ent and child about a child’s beveragechoices, as well as perceived negativeconsequences of implementingrestrictions on such choices. Skillsidentified as important to consump-tionmay include parents’ specific abil-ities to negotiate the child’s exposureto various kinds of beverages acrossdiverse situations and parents’ confi-dence in implementing these skills.In an IMB model of SB consumption,it would further be presumed thatparental information and motivationwould exert their effects on childbehavior primarily through behav-ioral skills, becausemanaging a child’sexposure to SB is thought to be quitecomplex and dependent on parentshaving sufficient skills.

These mediated paths essentiallymean that even well-informed andhighly motivated parents still requirecertain skills and confidence in thoseskills to ultimately influence theirchild’s exposure to SB. The modelalso includes a feedback loop to reflectcumulative processes of the modelover time, where experiences with dif-ferent skills will feed back into par-ents’ level of accurate informationand motivation through the skill andconfidence development.

This study evaluated the IMBmodel for its potential utility in char-acterizing children’s SB consumptionin a cross-sectional design in a sampleof ethnic minority parents of lowerSES. The authors hypothesized thatparents with greater informationabout SBwould also havemore behav-ioral skills related to their children’sSB consumption. The authors also hy-pothesized that parents with moremotivation toward reducing or limit-ing SB consumption would havemore behavioral skills. Finally, theauthors hypothesized that parentswith greater skills would have chil-dren who consumed fewer caloriesfrom SB. A portion of this research

was presented at the annual meetingof the American Society forNutrition.12

METHODSStudy Design and Sample

The authors collected data forthis study via an interviewer-administered IMB survey and self-reported Home Beverage Inventory(HBI) with a convenience sample ofEnglish-speaking parents of youngchildren. At local health fairs servinglow-SES populations and 2 Hartford,CT Special Supplemental NutritionProgram for Women, Infants, andChildren (WIC) offices, recruitersinitially enrolled 266 parents in thestudy. After obtaining writteninformed consent, interviewers orallyadministered a 20-minute demo-graphic and IMB survey to each partic-ipant. Researchers then scheduleda follow-up phone call with the partic-ipant for a phone inventory of all bev-erages in the participant’s home anda child beverage frequency question-naire. The first author conducted thephone interview 1-30 days after initialcontact. The University of Connecti-cut Institutional Review Board forHuman Research approved all studyprotocols by expedited review.

To be included in the study, partic-ipants were required to be the parentor legal guardian of a 1- to 5-year-oldchild who was not yet in kindergar-ten. If more than 1 child in the familyfit the criteria, parents answeredresearch questions in reference to theoldest qualifying child. Since somemeasures were collected via tele-phone, access to a telephone was re-quired. In return for completing theIMB survey, parents and their chil-dren received a small gift (eg, measur-ing spoons for parents, bubbles forchildren). For completing the HBI,researchers mailed participants a $10gift card.

Investigators estimated a targetsample size of 200 to provide suffi-cient stability in the parameterestimates and power to detectmedium-sized effects (r ¼ 0.24-0.36)13 in planned structural equationmodeling analyses.14 Anticipatinga 25% dropout rate, 2 interviewers ini-tially recruited 266 participants. Ofthe 266 original participants, 198 par-

ents completed both the IMB surveyand HBI, resulting in a 74% retentionrate.

MeasuresInformation-Motivation-BehavioralSkills survey. Based on focus groupand interview data from the partici-pant population,5,15 an extensivesearch of the literature, and researchexperience in the community,this research team developeda preliminary IMB-based SB consump-tion survey targeting 3 parental con-structs related to children’s SBconsumption: information, motiva-tion, and behavioral skills. An expertpanel developed initial items, whichwere reviewed by members of the par-ticipant population for understand-ing, readability, and content validity.Survey items were intended to sampleacross diverse aspects of information,motivation, and behavioral skills andas such, demonstrated low internalconsistency within each index. Sur-vey items were therefore summedinto the 3 indices determined a prioriin terms of the intended, face validconstruct the item was developedto represent. Subsequently, thesesummed scales (information, motiva-tion, and behavioral skills) were ana-lyzed through structured equationmodeling using a measured variableapproach in further analyses to repre-sent the 3 SB constructs. Five itemswere used to assess aspects of informa-tion, 4 items for personal (2 items)and social (2 items) motivation, and7 items for behavioral skills. Higherscale scores reflected higher amountsof accurate information, higheramounts of total motivation, andhigher levels of behavioral skills.

Home Beverage Inventory. Inter-viewers trained parents to completethe HBI in their home, includinghow to measure their child’s typicalserving sizes using 4 differently sizedand colored cups, during the initialstudy encounter. At the first visit, par-ticipants received a HBI journal, self-addressed return envelope, and4 cups to take home. During thefollow-up phone call, participantsorally reported their HBI, listing allnonalcoholic beverages in the house-hold including liquids, powders, con-centrates, tea, coffee, sweeteners, and

242 Goodell et al Journal of Nutrition Education and Behavior � Volume 44, Number 3, 2012

syrups. Participants reported the bev-erage brand, type, flavor, size, con-tainer, and number of containers foreach drink. The interviewer recordedthe participants’ responses and usedprobes to ensure clarity and complete-ness. Participants estimated howoften the study child drank each bev-erage listed in their HBI and howmuch they drank at 1 time followinga beverage frequency methodology.The interviewer asked participantshow often the study child consumedeach beverage listed in the HBI, fol-lowed by the probes, ‘‘Does (child’sname) drink (name of beverage) everyday, every week, every month, or ev-ery year?’’ and then ‘‘Howmany timesa (day, week, month or year) does(child’s name) drink (name of bever-age)?’’ Participants used the providedcups as a reference, indicating thecolor of cup from which the child typ-ically consumed each beverage and towhat level the parent filled the cupwith each beverage. The HBI wasdeveloped for the current study.

Calories from SB consumption. Us-ing the results from the HBI, researchassistants entered each beverage intoNutritionist Pro Diet Analysis soft-ware (version 2.5, Axxya Systems,Stafford, TX, 2005). For those itemsnot included in the software database,assistants contacted the manufac-turer, procured the item’s nutritionalvalues, and entered the specificationsinto the program’s customizable data-base. Serving sizes were estimated bycalculating the volume of beveragesconsumed on a daily basis. The inves-tigators then calculated average dailycalories consumed from SB. Totaldaily calories consumed from 4 SB cat-egories (juice drinks and juice-flavored beverages, flavored milk, hotbeverages, and sodas) were summedto create a measure of average dailycalories consumed from SB. For sim-plification, this variable was termedSB consumption.

Initial review of the intake dataindicated a non-normal distribution(M ¼ 282 kcal; SD ¼ 374; skew ¼ 2.4;kurtosis ¼ 7.5). Inspection of this dis-tribution revealed 2 substantially out-lying values. Analysts ‘‘capped’’ thedistribution such that these 2 valueswere reset to the closest reported valuenot exceeding 3 standard deviationunits from the mean in the original

distribution. In this case, the cap forcalories was set to 1,402 kcal, whichproduced a more normal, though notcompletely normal, distribution (M ¼274 kcal; SD ¼ 338; skew ¼ 1.7; kurto-sis ¼ 2.6). Thus, the adjusted variablecould range from 0 to 1,402 kcal, andmaintained its immediate interpret-ability.Although this strategymay sac-rifice some variability in the sampledistribution, it is important to notethat only 2 scores were directlyaffected, and the mean and standarddeviation were only mildly altered.

Statistical Analysis

Key demographic variables (ethnicity,education, marital status, WIC andFood Stamp participation, employ-ment, parent age, and child age)were examined for association withthe 3 main IMB-model–based con-structs (information, motivation, andbehavioral skills) and the dependentvariable (SB consumption). Pearsoncorrelations and chi-square analysesrevealed no significant relations, andthus no covariates were included inthe model evaluation.

Evaluation of the IMB model of SBconsumption was performed via struc-tural equationmodeling, where SB con-sumption was explained by the 3constructs: information, motivation,and behavioral skills. All associated di-rect and indirect paths were specifiedamong the core constructs, creatinga just-identifiedmodel. Each of the pre-dictedassociationswasevaluatedon thebasis of magnitude and direction of in-dividual path estimates. The results ofthe just-identified model were thenused to explore a parsimonious model,where any non-influential direct pathfrom information or motivation to SBconsumption in the first model was setto 0. Overall fit of the parsimoniousmodelwas examined via standardfit in-dices (eg, chi-square, Root Mean SquareError of Approximation [RMSEA], Com-parative Fit Index [CFI]) with values of0.90 or greater for CFI, less than 0.10for RMSEA, and nonsignificant chi-square values indicative of good fit.Model comparisons via changes in chi-square, Akaike Information Criterioncomparison (large values indicating im-proved fit), and relative (such as Tucker-Lewis Index [TLI] and Normal Fit Index[NFI]) and absolute Goodness of Fit In-

dex (GFI) fit indices (NFI andGFI valuesof0.95 andTLI values of approachingorgreater than 1 indicating good fit) wereconsidered. For all analyses, AMOSstructural equation modeling software(version 5.0.1, Smallwaters Corpora-tion,Chicago, IL,2003)wasused,whichnecessitated obtaining some of the rela-tive fit indices under the conditions ofno missing data. Values for TLI andGFI were derived from an n of 197 (1participant was missing 2 items fromthose making up the behavioral skillsconstruct). The outcome variable of in-terest, SB consumption, was nearly nor-mally distributed,14 which permitteduse of the unbiased Maximum Likeli-hood estimator with all measured vari-ables considered continuous measures.

RESULTSDemographics

A total of 198 participants completedIMB and HBI measures. The sampleprovided appropriate power for themodel evaluated assuming mediumeffects.14 As indicated in the Table,participants were primarily of ethnicminority backgrounds and lower SES.Almost one quarter of participantsdid not have a high school diplomaor its equivalent, whereas approxi-mately half had a high school or gen-eral equivalency diploma. Over twothirds of the children in the samplelived in a single-parent home. Al-though 1 parent participant selectedtheir 11-month-old infant as a referentin completing the measures, 99.5% ofthe sample provided responses in ref-erence to their 1- to 5-year-old child.

The present sample missing mea-sures contributed to the exclusion of26% of the original sample in the finalanalyses. Exploration of ‘retention’ re-vealed that missingness was likely ran-dom across most study variables, but itdid appear associated with the infor-mation subscale. Those retained inthefinal analysesappeared slightlybet-ter informed than those excluded be-cause of missing data, although thisdifference was driven primarily bya single item (older children can haveaccess to a wider variety of drinks).

IMB Model of SB Consumption

As indicated in Figure 1, parentalinformation about SB consumption

Table. Participant Characteristics

Frequency

n %Parent characteristicsEthnicityAfrican American 72 36Latino/Hispanic 67 34West Indian 35 18Other 24 12

SexFemale 185 93

Marital statusSingle 125 63Partnered/married 62 31Separated/divorced 11 6

EducationLess than high school diploma 54 27High school diploma or equivalent 90 46Some college and/or technical school 54 27

Employment statusUnemployed 102 52Part-time 43 22Full-time 53 27

Participation in the Special Supplemental NutritionProgram for Women, Infants, and ChildrenCurrent 185 93Previous 9 5Never/unknown 4 2

Age, y (� SD) 30.3 (� 8.2)Range, y 18–64

Child characteristicsEthnicityAfrican American 71 36Latino/Hispanic 57 29West Indian 23 12Multiple ethnicities 29 15Other 18 9

SexFemale 111 56

Age, y (� SD)a 3.1 (� 2.4)Range, y 0.9-5.911-23 mo. 44 2224-35 mo. 47 2436-47 mo. 49 2448-59 mo. 46 2360-70 mo. 12 6

aThe relationship between age of child and calories consumed via sweetenedbeverages was not significant (P ¼ .18).

Journal of Nutrition Education and Behavior � Volume 44, Number 3, 2012 Goodell et al 243

was related to greater skills in regulat-ing SB consumption exposure; it wasalso directly related to fewer caloriesconsumed by children via SB con-sumption. Greater motivation to limitSB consumption related to greaterbehavioral skills and skills weredirectly related to lower SB consump-

tion in their children. In full, thismodel accounted for approximately14% of the variance in the child SBconsumption.

Based on the full model results, theauthors explored model fit for a parsi-monious model, where the nonmedi-ated path from motivation to SB

consumption was set to 0. Theauthors explored model fit for thisparsimonious model, because theassociation between motivation andSB consumption was not significantin the full model. This parsimoniousmodel, depicted in Figure 2, againsupported the direction and magni-tude of the structural hypotheses ofthe model; it also provided a generallyacceptable fit of the model to the sam-ple data (CFI ¼ 0.99; RMSEA ¼ 0.05).Direct comparison between the parsi-monious model and the full model in-dicated that the parsimonious modeldid not significantly deteriorate fitand was a preferable model (chi-square change [n ¼ 198; df ¼ 1] ¼1.52; P ¼ .22; Akaike Information Cri-terion ¼ 27.5 [vs 28.0 for the fullmodel]; TLI ¼ 1.0; GFI ¼ 0.998; NFI¼ 0.987). The parsimonious model ac-counted for 13% of the variance inchild SB consumption.

DISCUSSION

By using the IMB model, the authorswere able to evaluate direct and medi-ating relationships between modelconstructs (information, motivation,and behavioral skills) and the targetbehavior (SB consumption). Greaterparental information was signifi-cantly correlated with reduced SB con-sumption (direct pathway). Thisfinding suggests that the traditionalfocus of nutrition education, improv-ing parental knowledge, may lowerchildren’s SB consumption. For exam-ple, providing parents with factsabout excessive SB consumption andweight gain may alter parental knowl-edge and decrease child SB consump-tion. The relationship betweeninformation and SB consumptionwas even stronger when the directpathway was combined with the indi-rect pathway via behavioral skills.Therefore, in addition to the informa-tion provided on the relationshipbetween excessive SB consumptionand weight gain, teaching parentshow to look for 100% juice on thenutrition facts label may furtherincrease desired outcomes.

Children’s SB consumption waslower when parents had both highmotivation to limit intake and theskills required to do so. Simply beingmotivated, in the absence of the skills

Figure 1. Information-Motivation-Behavioral model of sweetened beverage consump-tion with standardized path coefficients (just-identified model). *P < .05; **P < .001.

244 Goodell et al Journal of Nutrition Education and Behavior � Volume 44, Number 3, 2012

and efficacy needed to actually imple-ment limitations to SB consumption,did not directly correlate with SB con-sumption. The impact of behavioralskills on SB consumption has receivedsupport in other research as well. Pre-vious researchers found that childrenwhose parents consciously limit theirsugar intake consume fewer caloriesfrom SB.16 Although not yet studiedin relationship to SB consumption,parental role modeling and encour-agement skills have been found topositively influence other desiredfood behaviors.17,18

Parents’ information and motiva-tion are important to the IMB model,but these 2 constructs are moststrongly correlated with children’sconsumption of SB through their rela-tionship with behavioral skills. Apply-ing these outcomes, parents ona limited budget may need to acquireskills to select affordable, healthfuldrink alternatives over lower-pricedSB. Important parental skills linkedto child SB consumption includemodeling,19 the development and

Figure 2. Information-Motivation-Behavioraltion with standardized path coefficients (parNote: The model run with no missing observaof Fit Index (.998).

implementation of rules regarding SBconsumption,16 providing childrenwith alternative beverages,20 andinformation-gathering techniques.21

Nutrition educators are in a primaryposition to address these topics.

Limitations of the current studyinclude use of a cross-sectional designand a preliminary measure of the IMBmodel constructs that did not offerthe opportunity to construct orevaluate more sophisticated modelevaluations. Only a longitudinaldesign would permit assessment ofthe causal implications of the IMBmodel and its articulation to SBconsumption, and thus the cross-sectional design of the present studydoes not permit conclusions regard-ing directionality. This is an emergingmodel of SB consumption with nowell-validated measures of theproposed model constructs; the mea-sures crafted for the present researchprovided representation of only cer-tain aspects of the larger theoreticalmodel, and psychometric propertiesof the measure limited the analytic

model of sweetened beverage consump-simonious model). *P < .05; **P < .001.tions produces an acceptable Goodness

approach. The current results are bestpositioned as a ‘‘proof of concept’’ inproviding an initial exploration of anIMB model that can be further evalu-ated withmore complex study designsand measurement strategies.

Adopting comprehensive measure-ment strategies and incorporatingmore objective assessments of SB con-sumption may be desirable, but cau-tion should be exercised to avoidincreasing participant burden viaintensive measures, as in the presentsample missing measures contributedto the exclusion of 26% of the originalsample in the final analysis. Althoughthose retained in the final analyses ap-peared slightly better informed thanthose excluded because of missingdata, the increase in information inthose included in the final analysisdid not appear to alter the outcomesof the study, because results directlysupported the role of information ininfluencing behavioral skills and SBconsumption. Finally, the measuredvariable approach tomodel evaluationlimits the otherwise available advan-tages to latent variable modeling, pri-marily in terms of estimating error.When adapting the IMB survey for fu-ture use, researchers can strengthenthe instrument by including addi-tional items that measure a widerrange of information, motivation,and behavioral skills related to SB con-sumption. There is a need for addi-tional research targeting thedevelopment of comprehensive scalesof SB-related information,motivation,andbehavioral skills that provide suffi-cient internal validity and also directlyestablish strong external validity.

IMPLICATIONS FORRESEARCH ANDPRACTICE

Application of this model showspromise in identifying the complex in-teractions between factors influencingpreschool children’s beverage con-sumption, particularly those relatedto parent-child feeding interactions.As nutrition educators frequently de-sire to alter the behavior of parentsand their preschool children, theymayuse the IMBmodel todevelop, im-plement, and assess interventions tar-geting specific items found toinfluence preschool children’s SB

Journal of Nutrition Education and Behavior � Volume 44, Number 3, 2012 Goodell et al 245

consumption, including parental in-formation, motivation, and behav-ioral skills. Further development ofcomprehensive models of SB con-sumption is essential to optimizeinterventions created to improve pre-school children’s beverage consump-tion patterns.

ACKNOWLEDGMENTS

This researchwasconductedasapartofDr. Goodell’s doctoral dissertation ob-tained at the University of Connecti-cut. Dr. Goodell was supported bya National Needs Fellowship providedby the US Department of Agriculture.We thank WIC clinic staff and partici-pating families for so generously pro-viding their time and resources insupport of our research endeavors.We also thank the other members ofthe Husky Nutrition Research Group,particularly StephanieWei andConnieCantor, withoutwhomnoneof this re-search would have been completed.

SUPPLEMENTARY DATA

Supplementary data associated withthis article can be found in the onlineversion at doi:10.1016/j.jneb.2010.07.012.

REFERENCES

1. Wang YC, Bleich SN, Gortmaker SL.Increasing caloric contribution fromsugar-sweetened beverages and 100%fruit juices amongUS children and ado-lescents, 1988-2004. Pediatrics. 2008;121:e1604-e1614.

2. Gillis LJ, Bar-Or O. Food away fromhome, sugar-sweetened drink con-sumption and juvenile obesity. J AmColl Nutr. 2003;22:539-545.

3. Welsh JA, Cogswell ME, Rogers S,Rockett H, Mei Z, Grummer-

Strawn LM. Overweight among low-income preschool children associatedwith the consumption of sweet drinks:Missouri, 1999-2002. Pediatrics.2005;115:e223-e229.

4. Dubois L, Farmer A, Girard M,Peterson K. Regular sugar-sweetenedbeverage consumption between mealsincreases risk of overweight amongpreschool-aged children. J Am DietAssoc. 2007;107:924-934.

5. DrakeR,PierceM,Goodell LS, FerrisA.Selection of beverages for young chil-dren: the decision process of low resourceurban parents. FASEB J. 2007;21:16.

6. Fisher JD, FisherWA.ChangingAIDS-risk behavior. Psychol Bull. 1992;111:455-474.

7. Fisher WA, Fisher JD, Harman J, Suls J,Wallston KA. The information-motivation-behavioral skills model:a general social psychological approachto understanding and promoting healthbehavior. In: Suls JM, Wallston KA,eds. Social Psychological Foundations ofHealth and Illness. Malden, MA: Black-well Publishing; 2003:82-106.

8. Singh S. Study of the effect of infor-mation, motivation and behaviouralskills (IMB) intervention in changingAIDS risk behaviour in female univer-sity students. AIDS Care. 2003;15:71-76.

9. Anderson ES, Wagstaff DA,Heckman TG, et al. Information-Moti-vation-Behavioral Skills (IMB) model:testing direct and mediated treatmenteffects on condom use among womenin low-income housing. Ann BehavMed. 2006;31:70-79.

10. Fisher JD, Fisher WA, Amico KR,Harman JJ. An Information-Motivation-Behavioral Skills model ofadherence to antiretroviral therapy.Health Psychol. 2006;25:462-473.

11. AmicoKR,BartaW,Konkle-ParkerDJ,et al. The Information-Motivation-Behavioral Skills model of ART

adherence in a Deep South HIVþ clinicsample. AIDS Behav. 2009;13:66-75.

12. Goodell LS, Amico K, Pierce M,Ferris A. An Information-Motivation-Behavioral Skills model for childsweetened beverage consumption.FASEB J. 2008;22:2.

13. Cohen JA. power primer. Psychol Bull.1992;112:155-159.

14. Kline RB. Principles and Practice of Struc-tural Equation Modeling. 2nd ed. NewYork, NY: Guilford Press; 2005.

15. Goodell LS, Pierce MB, Bravo CM,Ferris AM. Parental perceptions ofoverweight during early childhood.Qual Health Res. 2008;18:1548-1555.

16. Liem DG, Mars M, De Graaf C. Sweetpreferences and sugar consumption of4- and 5-year-old children: role of par-ents. Appetite. 2004;43:235-245.

17. Cullen KW, Baranowski T,Rittenberry L, Olvera N. Social-environmental influences on children’sdiets: results from focus groups withAfrican-, Euro- and Mexican-American children and their parents.Health Educ Res. 2000;15:581-590.

18. Iannotti RJ, O’Brien RW,Spillman DM. Parental and peer influ-ences on food consumption of pre-school African-American children.Percept Mot Skills. 1994;79:747-752.

19. Brown R, Ogden J. Children’s eatingattitudes and behaviour: a study of themodeling and control theories of paren-tal influence. Health Educ Res.2004;19:261-271.

20. Ebbeling CB, Feldman HA,Osganian SK, Chomitz VR,Ellenbogen SJ, Ludwig DS. Effects ofdecreasing sugar-sweetened beverageconsumption on body weight in adoles-cents: a randomized, controlled pilotstudy. Pediatrics. 2006;117:673-680.

21. Krueger R. Focus Groups: A PracticalGuide for Applied Research. 2nd ed.Thousand Oaks, CA: Sage Publica-tions; 1994.