Research Article

Determinants of Children’s Use of and Time Spentin Fast-food and Full-service RestaurantsAlex McIntosh, PhD1; Karen S. Kubena, PhD, RD2; Glen Tolle, MS1; Wesley Dean, PhD3;Mi-Jeong Kim, PhD4; Jie-Sheng Jan, MS1; Jenna Anding, PhD2

ABSTRACT

Objective: Identify parental and children’s determinants of children’s use of and time spent in fast-food(FF) and full-service (FS) restaurants.Design: Analysis of cross-sectional data.Setting: Parents were interviewed by phone; children were interviewed in their homes.Participants: Parents and children ages 9-11 or 13-15 from 312 families were obtained via random-digitdialing.Main Outcome Measures: Dependent variables were the use of and the time spent in FF and FS restau-rants by children. Determinants included parental work schedules, parenting style, and family meal ritualperceptions.Analysis: Logistic regression was used for multivariate analysis of use of restaurants. Least squares regres-sion was used for multivariate analysis of time spent in restaurants. Significance set at P < .05.Results: Factors related to use of and time spent in FF and FS restaurants included parental work sched-ules, fathers’ use of such restaurants, and children’s time spent in the family automobile.Conclusions and Implications: Parenting style, parental work, parental eating habits and perceptions offamily meals, and children’s other uses of their time influence children’s use of and time spent in FF and FSrestaurants.Key Words: child, parental work schedules, parenting style, fast food, restaurants, family, meal (J NutrEduc Behav. 2011;43:142-149.)

INTRODUCTION

According to data on food expendi-tures in the United States, food awayfrom home accounted for 48% of totalfood expenditures in 2008, as com-pared to 34% in 1974.1,2 The trendfrom home-prepared food to food pre-pared by others coincides with the in-creasing prevalence of overweight andobesity in this country.3 Analysis offood prepared away from home, ascompared to food prepared at home,indicates that restaurant offerings arehigher in calories and fat and lowerin nutrients.4-6 Parents of 10- to 13-year-old children reported that lack

of time and accommodation ofactivities of family members resultedin the need to obtain food awayfrom home.7 The top 3 reasons thatadults in another study chose fastfood were rapid service, convenientlocation, and good-tasting food, butavailability of nutritious food wasless important.8 However, there existother possible explanations for eatingaway from home.

The purpose of this study was toidentify determinants of children’suse of and time spent in fast-food(FF) and full-service (FS) restaurants.The relationship between children’suse of FS or FF restaurants and time

spent in them and parental work, par-enting style, perception of familymeals, usage of and time spent inthese restaurants by parents, andother uses of time by children was in-vestigated.

To explore such reasons, theoriesfrom economics (the household pro-duction model, or HPM)9,10 andsociology (role strain theory)11 drovethe perspective of the present study.Each of these complementary expla-nations was selected for its past associ-ation with outcomes in children. TheHPM predicts that outputs from fam-ily members rely on income andtime as resources for achieving goals;work outside the home by mothersor fathers is considered a constrainton time that could be devoted to fam-ily activities. Other away-from-homeuses of time by parents, such as run-ning errands, may also determine theactivities in which children engagewith their parents. Role strain theoryprovides an argument that householdmembers may fail to achieve thosegoals because of lack of time and theinterference of demands from other

1Department of Sociology, Texas A&M University, College Station, TX2Department of Nutrition and Food Science, Texas A&M University, College Station, TX3School of Rural Public Health, Texas A&M University Health Science Center, CollegeStation, TX4Department of Food and Nutrition, Silla University, Pusan, KoreaAddress for correspondence: Alex McIntosh, PhD, 4351 Texas A&M University, CollegeStation, TX 77843-4351; Phone: (979) 845-8525; Fax: (409) 862-4057; E-mail:w-mcintosh@tamu.edu�2011 SOCIETY FOR NUTRITION EDUCATIONdoi:10.1016/j.jneb.2010.04.002

142 Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011

UNDER EMBARGO UNTIL MAY 6, 2011, 12:01 AM ET

roles. The classic case of role strainconsiders the demands of householdobligations (eg, provision of healthfulmeals) in potential conflict with thedemands of work (eg, long workhours, unpredictable work schedules).Other factors of importance withinthis literature are characteristics ofwork, such as work control and workimportance, that potentially influ-ence family life.12 However, none ofthese theories captures the entiretyof factors that influence householdmembers’ behavior. So the HPM androle-strain theory models are supple-mented with (1) parental socializationvia modeling,13 (2) parenting styletheory,14 and (3) family ritual the-ory.15 With regard to socialization,fathers and mothers, for example,transmit their alcohol consumption,tobacco consumption, and eatinghabits to their children.16 The authorsargue, for example, that parents’ useof time spent in eating in FF restau-rants may serve as a model for chil-dren.

Prior research indicates that thetime mothers invest in cooking or inactivities with their children reduceschildren’s risk of overweight.10 Thus,the time entailed by maternalemployment potentially increases therisk of obesity in children.17 In addi-tion, parenting styles may predict theway that parents feed their childrenand their children’s body size. Parent-ing styles, as characterized by Rhee etal,18 include authoritative (respect forchild’s opinions, but clear bound-aries), authoritarian (strict disciplinar-ian), permissive (indulgent withoutdiscipline), and neglectful (emotion-ally uninvolved, without rules). Intheir study, children with parentswho used authoritarian, permissive,or neglectful parenting were at greaterrisk for overweight as compared tothose with more authoritative par-ents.18 Hubbs-Tait et al reported thatrestriction of some food and pressureto eat other food were practices usedby parents classified as authoritarian,whereas modeling, monitoring foodintake, and promoting children’s re-sponsibility for food selection wereused by authoritative parents.19 Fur-thermore, parental criticism and worryabout their children’s health and eat-ing habits adversely affect childrenand could encourage obesity.20 Im-pacts on children’s eating also may re-

sult from parental perceptions thatdinner represents an important familyritual.21 Work hour and work day flex-ibility benefits families, as these condi-tions often allow more parentalinvolvement in children’s activitiesand have positive health benefits.22

Orientations toward and conditionsat work, such as work control or auton-omy and the importance of work rela-tive to other roles, affect children andfamily life in general.22,23

The purpose of this study is to de-termine whether parental work sched-ules and other attributes of work,parenting style, and perceptions ofdinner as an important family ritualare related to children’s use of andtime spent in FS and FF restaurants.

METHODS

Data used in this paper come from thestudy, ‘‘Parental Time, Income, RoleStrain, Coping, and Children’s Dietand Nutrition.’’ A sample of 312 fam-ilies was obtained via random-digitdialing, with a response rate of 48%.This response rate is equal to or betterthan that of recent studies of multiplemembers of families, the rates ofwhich run from 49% to 23%.24 Bothparents (if the household was dual-headed) and 1 child between theages of 9-11 or 13-15 participated inthe sample of 2-parent households;an oversample of single-headedhouseholds made up 20% of the par-ticipating families; however, the non-response rate of children lessened thispercentage in the overall sample. Ageranges of children were selected tocapture prepubertal and postpubertalchildren; 12-year-olds were consid-ered most likely to be in a transitionalstage in this regard.

Using a questionnaire, each parentwas interviewed by telephone to ob-tain data on their work, eating habits,perceptions of dinner, and demo-graphic characteristics; these inter-views took about 32 minutes. Theyalso responded to a self-administeredincome questionnaire and main-tained a 2-day time diary. Because ofthe detailed nature of questions re-garding parents’ various sources of in-come, over 20% of the sample ofparents did not complete this ques-tionnaire. Using a questionnaire, chil-dren were interviewed for an hour in

person with questions regarding howtheir mothers and fathers parentedthem, mother’s perceptions of theireating habits and weight, and theirown perceptions of their eating habitsand other health behaviors. After thequestionnaire was finished, children’sheight and weight were measured,25

and a multipass 24-hour dietary recall(with the use of a 2-dimensional book-let containing food models) and 24-hour activity recall were conducted.The children were then instructed inmaintaining 2-day food intake and2-day activity diaries. The 24-hour ac-tivity recall and 2-day activity diariesobtained data on what children didduring each day, how much timethey spent in each activity (main ac-tivity), and where and with whomthey performed each activity. The ac-tivity recall extended from the timethe child went to bed 2 nights beforethe interview until the child went tobed the previous night. Time spentin various activities was averagedacross the 3 days. Interviewers wereinstructed in anthropometry and die-tary assessment by members of the re-search team who are registereddietitians (KK and JA). Dietary intakeand anthropometric data are not re-ported here. The study was approvedby the Institutional Review Board ofTexas A&M University after a full re-view, and written consent (adults)and assent (children) were obtained.

MeasurementParental work. Based on their re-sponses to the telephone question-naire, the authors determinedwhether the parents were workingand, if they were, whether or not theirwork schedule was standard in termsof days of the week (Monday throughFriday) and time of the day (8 AM-5 PM).26 Those whose schedules in-cluded either one or both weekenddays were classified as having a non-standard work day; those whoseschedules diverged by at least 3 hoursfrom the 8 AM-5 PM time period werecharacterized as having nonstandardwork hours. The authors used 5 stan-dard items to measure perceptions ofwork control developed by Hage andAiken,27 including ‘‘There can be littleaction around here until a supervisorapproves a decision’’ and ‘‘A personwho wants to make his or her own

Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011 McIntosh et al 143

decisions would be quickly discour-aged.’’ The importance of work rela-tive to other adult roles has also beenfound to influence family life.23 Theauthors used a 5-item scale of workimportance developed and validatedby Porter et al28; example items in-clude: ‘‘My work is the most impor-tant thing in my life’’ and ‘‘Myfamily is more important to me thanmy work.’’ Each parent was also askedto rate on a 3-point scale how flexibletheir work days and work times were.

Family meal rituals. In a pilot study,the authors developed a 12-item scale,drawing on items created by Fiese andKlein.29 These items include the per-ception that ‘‘dinner is regularlyscheduled,’’ ‘‘most family membersparticipate in dinner,’’ ‘‘dinner time isa special family time,’’ and ‘‘familymembers believe it is important toeat together at least once a day.’’ Theauthors added 2 questions regardingthe eating of meals while watchingTV under the belief that watching TVduring meals with family is antitheti-cal to the idea of a ritual meal ora ‘‘proper family meal.’’ Based on theauthors’ reading of Neumark-Sztaineret al,30 they included additional itemsin the present study. These items in-cluded: ‘‘In my family, eating togetherbrings people together in enjoyableways,’’ ‘‘Mealtime is a time for talkingwith other family members,’’ and‘‘Mealtime has often been a timewhen people argue in my family.’’These items help capture at leastsome notion of the emotional andconflictual dimensions of familymeals. All of the items used to measureaspects of the family meal ritual weremeasured via a 1-to-5 Likert scale.

Parenting style and other children’svariables. The theory of parentingstyle proposed by Macoby and Martinargues that there are 4 main types ofparenting: authoritative (which con-sists of caring, clear behavioral expec-tations, punishment by privilegewithdrawal), authoritarian (less caring,attempts to control behavior, physicalpunishment and harsh criticism), per-missive (caring, few controls or behav-ioral expectations, few punishments),and neglectful (lack of caring, lack ofcontrol, few punishments).14 Parent-ing style indicators were drawn from

a scale developed by Devereux et al(‘‘My mother comforts me when Ihave problems’’; ‘‘My mother expectsme to keep my things neat’’; ‘‘Mymother worries that I cannot takecare of myself’’).31 The authors added3 items to this scale in order to measurepunishment style (eg, ‘‘My motherpunishes me by not letting me to bewith my friends’’; ‘‘My mother pun-ishes me by yelling, nagging, or scold-ing’’). The same questions were thenasked about fathers’ parenting style.

Maternal worries and concerns re-garding their children’s weight havebeen found to affect children’s eatinghabits and body weight.32 Thus, chil-dren were asked about the degree towhich their mothers worried abouttheir weight and believed that theirchildren should lose weight, eat less,and exercise more (eg, ‘‘My motherthinks I weigh too much’’) (Cronbacha ¼ 0.68). Given the influence peershave over children,16 the childrenwere asked about the degree to whichthey believed they had greater free-dom of choice when eating withfriends. Time children spent in restau-rants was obtained from the time di-ary and activity recall questionnaires.

Statistical Analysis

The children’s activity recall33 and 2-day time diaries used coding proceduresdeveloped by Robinson and Godbey.34

These diaries included codes for eatingbreakfast, lunch, dinner, snacks, foodpreparation, and cleanup after meals;and eating at or away from home. In ad-dition, variables for time use in exercise,watching, television, and numerousothers were created. However, in somecases the authors coded with greaterspecificity than their coding allows,namely, because the children listed thename of the restaurant when they indi-cated they had eaten out. The authorscreated 2 FF variables, beginning withadummyvariable representingwhetherthey had eaten at an FF restaurant at anytime during the 3-day activity period. Asecond variable reflected the actual timein minutes spent in the FF restaurant; allwho had not eaten fast food were givena score of 0. Fast-food entries includedMcDonald’s, Chick Fil-A, Burger King,and others. Full-service restaurants in-cluded regional and national chainssuch as Golden Corral, as well as localrestaurants. Again, a dummy variable

for having eaten at least once or morein FS restaurant(s) was created, as wasa time spent eating at an FS restaurantvariable. The authors used the samecoding scheme for parents’ time use. Fa-thers’ and mothers’ time spent in min-utes in each kind of restaurant wereconsidered as measures of socializationvia modeling.

All scales were subjected to com-mon factor analysis to determinewhether each scale could be repre-sented by 1 or more underlying factors.Common factor analysis is preferred tothe more familiar principal compo-nents because the former, unlike thelatter, excludes unique variance in thecalculation of common variance.35 Inthe case of the parenting style vari-ables, because each parent’s style wasmeasured by 30 items, a second-ordercommon factor analysis was run toprovide greater clarity to the factorsproduced by the first-order analysis.

Logistic regression was used in themultivariate analysis of whether chil-dren had eaten in either an FS or FF res-taurant during the 3-day period.Ordinary least squares was used to per-form the multivariate analyses of ac-tual time spent in either an FS or FFrestaurant. Given that time diaries pro-duce a high number of zeros (no timespent in a given activity), results werechecked using Tobit analysis. Controlvariables included children’s risk ofobesity (body mass index), sex, race/ethnicity, and age group. Initial analy-ses included parents’ income. Neithermothers’ nor fathers’ income nortime spent at work was related to res-taurant use, and a considerable num-ber of parents failed to fill out theincome questionnaire provided tothem. As a consequence, parents’ in-come and work time were droppedfrom the analyses. Analyses were per-formed using SAS (version 9.2, SASInstitute, Inc., Chicago, IL, 2008), andrelationships were considered statisti-cally significant at the .05 level or less.

RESULTS

Descriptive statistics are presented inTable 1. In addition, modal educationlevels were ‘‘some college’’ for mothersand a ‘‘bachelors degree’’ for fathers.Coefficients from logistic regressionand ordinary least squares regression,shown in Tables 2-5, indicate that

144 McIntosh et al Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011

several factors of interest were relatedto (1) whether children had visitedrestaurants, both FF and FS, and (2)the time children spent in them.

Results of the Common FactorAnalyses and ReliabilityAssessment

The 5 indicators of work control weresubjected to a common factor analy-

sis, resulting in a single factor foreach working parent (Cronbach a ¼0.68 for mothers, 0.69 for fathers).Common factor analysis of the workimportance scale of 5 items produceda single factor for each working parent(Cronbach a ¼ 0.74 for mothers, 0.73for fathers). The 30 indicators fromthe parenting style scale developedby Devereux et al,29 plus items devel-oped to measure punishment style,underwent first-order and second-

order common factor analysis.35 Fac-tors produced by the analyses reflectthe general types of parenting pro-posed by Macoby and Martin,14,36

producing factors that reflectmaternal authoritative parenting andpaternal authoritarian parenting, aswell as the other types of parentingstyles (Cronbach a ranged from 0.71to 0.86 for the subscales produced bythe factor analyses). Common factoranalysis of maternal worries abouttheir children’s eating habits,exercise habits, and weight produceda single factor. The 2 questionsmeasuring children’s perception oftheir eating habits when with friendswere found to be significantlycorrelated, and thus the authorscreated a new variable by averagingtheir responses. Finally, commonfactor analysis of the family mealritual items produced 2 factors fromthe mothers’ responses and 2 factorsfrom the fathers’ responses. The firstfactor represents dinner as a specialfamily ritual, and the second reflectsa special family night/special Sundaylunch (Cronbach a for the items thatmade up the subscales produced bythe factor analysis ranged from 0.77to 0.87).

Logistic and OLS RegressionResults

Eating in FF and in FS restaurants andtime spent doing so by fathers were di-rectly related to those variables inchildren (Tables 2-5). The more timefathers spent eating in FS restaurants,the more likely their children were toeat and spend time in an FSrestaurant. The more time fathersspent eating in FF restaurants, themore likely their children were to eatand spend time in FF restaurants.However, no significant relationshipswere observed between thesevariables for mothers and theirchildren.

The work schedule of parents caninfluence the location of meals.When the work schedule of both par-ents could be classified as standard,the likelihood of the child eating inFS restaurants increased (Table 2). Asimilar relationship was shownwhen fathers reported their workschedules were flexible, an indicationof a sense of autonomy at work and

Table 1. Sociodemographic Background Variables of Children and Their Parents

Variable Proportion (%) Mean SDChild’s race/ethnicity (n ¼ 312)

White 71.6Black 13.4Hispanic 13.5Asian 1.5

Child’s age group (n ¼ 312)9-11 y 54.313-15 y 45.7

Child’s sex (n ¼ 312)Female 51.0Child’s BMI (n ¼ 311) 20.8 5.0Mother’s age, y (n ¼ 320) 41.9 5.8Father’s age, y (n ¼ 247) 44.8 5.6

BMI indicates body mass index.

Table 2. Children’s Use of Full-service Restaurants Regressed on Parental andControl Variables (logistic regression; n ¼ 138)

VariableOddsRatio

LowerConfidence

Interval

UpperConfidence

IntervalMother worried about child’s weight 1.75* 1.05 2.94Child can eat what he/she wants

with friends0.57* 0.34 0.95

Father’s work schedule flexible 3.20*** 1.67 5.81Both parents standard work schedules 3.26*** 1.11 9.59Father’s use of full-service restaurants 4.08*** 1.59 10.20Sex 1.71 0.69 4.25Age group 1.40 0.53 3.69non-Hispanic white 0.55 0.14 1.79Child BMI 1.52 1.09 3.39Likelihood ratio c2 40.49***Pseudo R2 0.24

BMI indicates body mass index.*Significant at the .05 level; **Significant at the .01 level; ***Significant at the .001level.Note: Likelihood ratio c2 is the ratio of a null-model vs an alternative model inwhich independent variables are thought to provide information about the out-come (dependent variable). Pseudo R2 is a counterpart to a proportion of vari-ance explained coefficient from an ordinary least squares regression analysisand is based on the likelihood ratio.

Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011 McIntosh et al 145

the ability to juggle time in order toeat dinner out. Furthermore, themore fathers perceived theycontrolled their work, the more timetheir children spent eating in FSrestaurants (Table 4). Perceptions ofwork were also important; thosemothers who considered their em-ployment role to be more importantthan their role as a mother weremore apt to have children whoreported spending more time in FFrestaurants than other children(Table 5).

Other perceptions also seemed toaffect restaurant use. Fathers who con-sidered meals, specifically family din-ner, to be an important family ritualwere found to have children who re-ported less use of and time spent inFF restaurants (Tables 3 and 5).Concern by mothers about children’sweight was related to greaterlikelihood of use of FS restaurants(Table 2).

Parenting style was another factorthat was involved in choosing meallocations. When both fathers and

mothers were categorized as being ne-glectful, the time spent by their chil-dren in FF restaurants increased(Table 5). On the other hand, childrenof mothers who exhibited authorita-tive parenting characteristics reportedless use of FF restaurants (Table 3).

Two other factors were related torestaurant eating in this study. Thefirst was the children’s perceptionthat they could eat what they wantwhen with friends. As expected withsubjects in this age range, the morethese subjects perceived they hadthis kind on autonomy when withfriends, the less time they reportedeating in FS restaurants (Table 2).The other variable that was positivelyrelated to the frequency of use of FFrestaurants and time spent in bothFS and FF restaurants was time spenttraveling by children (Tables 3-5). Astime spent in the car with theirparents increased, the likelihood ofeating at both types of restaurantsand the time spent in each type ofrestaurant increased.

DISCUSSION

The results of this study demonstratethe importance of a multi-theoreticalapproach to the study of this type offood behavior of children. Use ofand time spent in restaurants of vari-ous types are differentially affectedby conditions associated with thework of parents, the parenting styleemployed by mothers and fathers, pa-rental perceptions of the importanceof family meals, parental socializationvia modeling of eating behavior, timechildren spend in a car with their par-ents, and peer influences over foodchoices. However, variables from thehousehold production model (in-come, time spent working) did notpredict either use of or time spent inthese restaurants.

In the present study, a more stan-dard work schedule or the ability tocontrol one’s work schedule (variablesthat reflect role conflict) was condu-cive to having meals in FS restaurantsthat require more time than FF restau-rants. The term ‘‘fast food’’ results inan expectation of rapid service andshort time to eat. Miller and Han sug-gest a consequence of nonstandardwork schedules of parents could be de-velopment of overweight children.37

Table 3. Children’s Use of Fast-food Restaurants Regressed on Parental andControl Variables (n ¼ 187)

Variable Odds Ratio

LowerConfidence

Interval

UpperConfidence

IntervalAuthoritative parenting by mother 0.23** 0.09 0.62Father perceived dinner as family ritual 0.39*** 0.24 0.63Father’s use of fast-food restaurants 1.26** 1.12 1.34Time child spent traveling 1.04 1.01 1.07Sex 0.78 0.31 1.96Age group 0.49 0.19 1.28non-Hispanic white 4.18* 1.06 16.39Child BMI 1.10 0.30 3.20Likelihood ratio c2 44.81***Pseudo R2 0.21

BMI indicates body mass index.*Significant at the .05 level; **Significant at the .01 level; ***Significant at the .001level.Note: Likelihood ratio c2 is the ratio of a null-model vs an alternative model inwhich independent variables are thought to provide information about the out-come (dependent variable). Pseudo R2 is a counterpart to a proportion of vari-ance explained coefficient from an ordinary least squares regression analysisand is based on the likelihood ratio.

Table 4. Time Children Spend in Full-service Restaurants Regressed on Parentaland Control Variables (OLS) (n ¼ 186)

VariableUnstandardized

CoefficientStandardizedCoefficient

Father’s control at work 2.55** 0.17Father’s time spent

in full-service restaurants0.07*** 0.31

Time children spent traveling 0.07*** 0.20Sex -1.40 -0.05Age group 3.98* 0.14non-Hispanic white -0.33 -0.01Child BMI 3.37 0.09R2 0.17***

BMI indicates body mass index.*Significant at the .05 level; **Significant at the .01 level; ***Significant at the .001level.

146 McIntosh et al Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011

However, perceptions of work playa role as well. Mothers in this studywho considered their work more im-portant than other roles, includingparenthood, had children who spentmore time eating in FF restaurants.Perhaps a high priority is given bythese mothers to the demands of thejob, whereas needs of the family, in-cluding family meals, are consideredlater. In the case of either nonstandardschedules and/or perceptions of workimportance, fast food is expedient, al-though not necessarily a good optionfor families’ future health.

The authors found children whospent more time traveling, usually inthe family car, used FF restaurantsmore and spent more time in both FFand FS restaurants. This associationsuggests either that eating out whilerunning errands results from othertime constraints or that families havea preference for eating out when theyspend more time running errands. Ifmore time is spent riding in the car,less time might be available to the childand others in the car for other activi-ties, including family meals at home.

Results of the current study indicatethat children used FF restaurants less iftheir mothers were authoritative par-ents, who were more likely to monitorand be concerned about nutritionalquality of their child’s food intake.Children spent more time in FF restau-rants if both parents were uninvolvedwith them (a neglectful parenting

style). An earlier report from the pres-ent study by Kim et al noted thatmothers who were predominantlynurturing (authoritative) as comparedto controlling were more likely tohave adolescents of healthy weight.36

It is clear, however, that parentingstyle becomes important only whenFF restaurants are a consideration.

An authoritarian parenting style,including the use of food restrictionto control intake, is associated withmaternal concern about a child’sweight.38 Perception by children thattheir mothers were worried abouttheir weight was related to more useof FS restaurants. Portion sizes and cal-orie content of meals from FS restau-rants, as well as FF restaurants, are ofconcern.5,6 Food away from homemeans that responsibility for allaspects of meals is delegated to others.

Children who responded that theyusually eat more with friends fre-quentedFS restaurants lessoften. Eatingmore with friends increases the chanceof peer modeling affecting intake,39 al-though these children may have feltfreer to eat when they were away fromparents who attempt to control food in-take (authoritarian parents).16

A striking finding in the presentstudy was the strong association be-tween the use of and time spent inboth FS and FF restaurants by childrenand use of and time in restaurants bytheir fathers. Modeling of dietary be-haviors by parents is a key factor in

the development of food preferencesand food behaviors of infants and chil-dren.40 These results suggest that fa-thers often spend time with theirchildren at FF or FS restaurants fora snack or a meal. Since dietary behav-iors, like relying on food away fromhome and eating fast food, have beenshown to track from childhoodthrough adolescence into youngadulthood, fathers should be encour-aged to model healthful food choiceswhen they obtain food and to eatwith children at home. After all, fa-thers who believe that dinner is an im-portant family ritual reduce children’suse of fast food; this perception shouldalso be encouraged among fathers.

The present sample was limited tochildren aged 9-11 and 13-15 who re-side in the Houston Metropolitan Sta-tistical Area. In addition, this samplewas not fully representative of Hous-ton families with children in that thissample contained fewer low-incomefamilies as well as fewer minority fami-lies than suggested by the United StatesPopulation Census. In addition, al-though these data include the amountof time spent obtaining takeout food,the data on eating takeout food donot differentiate eating takeout foodat home or elsewhere. Unlike a stan-dard time diary approach, secondaryor simultaneous activities of childrenwere not measured. It is also importantto note that the dates on whichmothers and fathers maintained theirtime diaries almost never coincidedwith the days for which children re-ported their own activities, althoughparents had been requested to recordthe same days for themselves as fortheir children. Thus, for example, theamount of time fathers spent in full ser-vice restaurants is not for the same timeperiod about which the child has re-ported on her/his time expenditure. Fi-nally, to increase response rates, futurework in this area should consider fi-nancial incentives to motivate familymembers to cooperate.

CONCLUSIONS

This study examined data from 312families with a child between theages of 9-11 or 13-15 about use oftime and eating out as related to par-enting style, parental work, and fam-ily meal rituals. The authors are

Table 5. Time Children Spent in Fast-food Restaurants Regressed on Parental andControl Variables (OLS) (n ¼ 135)

VariableUnstandardized

CoefficientStandardizedCoefficient

Both parents neglectful 2.36*** 0.27Father perceived dinner as family ritual -1.07* -0.15Mother perceived work as more

important than other roles1.76* 0.14

Father’s time spent in fast-foodrestaurants

0.25*** 0.28

Time child spent traveling 0.06*** 0.37Sex -0.93 -0.04Age group -1.02 -0.06non-Hispanic white 2.67* 0.15Child BMI -1.88 -0.10R2 0.32***

BMI indicates body mass index.*Significant at the .05 level; **Significant at the .01 level; ***Significant at the .001level.

Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011 McIntosh et al 147

unaware of other studies that assessthis combination of variables. Find-ings support the concept that par-ents’ work influences family meals.When both parents have a standardschedule, father has a flexible sched-ule or he feels that he has control atwork, children use FS restaurantsmore. Also work-related is the sugges-tion that children whose mothersrank their job role as more importantthan their other roles, including thatof parent, reported spending moretime in FF restaurants. Time spentby children traveling in the car wasdirectly related to use of FF restau-rants and time spent in both FF andFS restaurants.

This may be the first report of influ-ence of family dinner being consid-ered a family ritual. Fathers who heldthis belief had children who fre-quented FF restaurants less. Childrenwith authoritative mothers used FFrestaurants less. On the other hand,children with neglectful parents spentmore time in FF restaurants. Childrenwho stated they ate more with theirfriends than with family, a possiblesign of food restriction and authoritar-ian parenting, were less likely toreport eating in a FS restaurant.

Although no association was seenbetween time spent and use of FF orFS restaurants by mothers and theirchildren, such relationships were seenfor fathers and their children. Dietaryquality of children is influenced bythe manner in which parents interactwith their children (parenting style),time available for family meals, andthe role restaurants play in their lives.

ACKNOWLEDGMENTS

Datapresented inthispaper weredrawnfrom the project ‘‘Parental Time, In-come, Role Strain, Coping, and Child-ren’s Diet and Nutrition’’ (43-3AEM-0-80075) funded through a grant fromthe United States Department of Agri-culture-ERS Food Assistance and Nutri-tion Research Program. The viewsexpressed in this paper do not necessar-ily reflect those of the United States De-partment of Agriculture.

REFERENCES

1. ClausonA,LeibtagE.FoodCPIandExpen-ditures Briefing Room. Table 12. US De-

partmentofAgriculture.www.ers.usda.gov/Briefing/CPIFoodAndExpenditures/Data/Expenditures_tables/table12.htm. AccessedAugust 5, 2010.

2. Stewart H, Blisard N, Jolliffe D. Let’seat out: Americans weigh taste, conve-nience, and nutrition. Economic Infor-mation Bulletin Number 19. http://www.ers.usda.gov/publications/eib19.Published October 2006. Accessed July27, 2010.

3. Ogden CL, Carroll MD, Curtin LR,McDowell MA, Tabak CJ,Flegal KM. Prevalence of overweightand obesity in the United States, 1999-2004. JAMA. 2006;295:1549-1555.

4. Paeratakul S, Ferdinand DP,Champagne CM, Ryan DH,Bray GA. Fast-food consumptionamong US adults and children: dietaryand nutrient intake profile. J Am DietAssoc. 2003;103:1332-1338.

5. Binkley JK. Calorie and gram differ-ences between meals at fast food and ta-ble service restaurants. Rev Agr Econ.2008;30:750-763.

6. Mancino L, Todd J, Lin B. Separatingwhat we eat from where: measuringthe effect of food away from home ondiet quality. Food Policy. 2009;34:557-562.

7. Cluskey M, Edlefsen M, Olson B, et al.At-home and away-from-home eatingpatterns influencing preadolescents’ in-take of calcium-rich food as perceivedby Asian, Hispanic and non-Hispanicwhite parents. J Nutr Educ Behav.2008;40:72-79.

8. Rydell SA, Harnack LJ, Oakes JM,Story M, Jeffery RW, French SA. Whyeat at fast-food restaurants: reported rea-sons among frequent consumers. J AmDiet Assoc. 2008;108:2066-2070.

9. Becker GS. Towards a theory of timeallocation. The Economic J. 1965;76:493-517.

10. You W. Time Well Spent: How Watch-ing Time Can Reduce Having OverweightChildren. Saarbrucken, Germany:VDM Verlag Dr. Muller Akteinge-schellschaft & Co.; 2008.

11. Michel JS, Michelson JK, Kotrba LM,LeBreton JM,Baltes BB.A comparativetest of work-family conflict models andcritical examination of work-familylinkages. JVocat Behav. 2009;74:199-218.

12. Voydanoff P.Work role characteristics,family structure demands, and work-family conflict. JMarriage Fam. 1992;50:749-761.

13. BanduraA.Social LearningTheory. Eng-lewood Cliffs, NJ: Prentice-Hall; 1977.

14. Macoby EE, Martin J. Socialization inthe context of family: parent-child in-teractions. In: Mussen PH,Hetherington EM, eds. Handbook ofChild Psychology. New York, NY: Wi-ley; 1983. 1-101.

15. Fiese BH. Fifty years of research onfamily rituals: cause for celebration? JFam Psychol. 2002;16:381-390.

16. Peters LWH, Widfferink CH,Hoekstra F, Buijs GJ, ten Dam GTM,Paulussen TGWM. A review of simi-larities between domain-specific deter-minants of four health behaviorsamong adolescents. Health Educ Res.2009;24:198-223.

17. Anderson PM, Butcher KF, Levine PB.Maternal employment and overweightchildren. JHealth Econ. 2004;22:477-504.

18. Rhee KE, Lumeng JC, Appugliese DP,Kaciroti N, Bradley RH. Parentingstyles and overweight status in firstgrade. Pediatrics. 2006;117:2047-2054.

19. Hubbs-Tais L, Kennedy TS, Page MC,Topham GL. Parental feeding practicespredict authoritative, authoritarian, andpermissive parenting styles. J Am DietAssoc. 2008;108:1154-1161.

20. Eckstein KC, Mikhall LM, Ariza AJ,Thompson JS, Millard SC, Binns HJ.Parents perceptions of their child’sweight and health. Pediatrics.2006;117:681-691.

21. Almas R. Domestic dinner: representa-tions and practices of a proper familymeal among young suburbanites. JConsumer Culture. 2006;6:203-228.

22. Fenwik R, Tausig M. The health andfamily-social consequences of shiftwork and schedule control. In:Epstein CF, Kallenberg AL, eds. Fight-ing for Time: Shifting Boundaries ofWork and Social Life. New York, NY:Russell Sage; 2004:77-110.

23. Parcel TL, Menegham EG. Parents’ Jobsand Children’s Lives. New York, NY:Aldine-DeGruyter; 1994.

24. Hendy HM, Williams KE, Camise TS,Eckman N, Hedemann A. The ParentMealtimeActionScale (PMAS):develop-ment and association with children’s dietand weight. Appetite. 2009;52:328-339.

25. Lohman TG, Roche AF, Martorell R.Anthropometric Standardization ReferenceManual. Champaign, IL: Human Ki-netics Books; 1988.

26. Presser HB.Working in a 24/7 Economy:Challenges for Families. New York, NY:Russell Sage Foundation; 2004.

27. Hage J, AikenM. 1969. Relationship ofcentralization to other structural prop-erties. Admin Sci Q. 1969;12:72–91.

148 McIntosh et al Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011

28. Porter LW, Steers M, Monday RT,Borilian PV. Organizational commit-ment, job satisfaction, and turn-overamong psychiatric technicians. J ApplPsychol. 1974;59:603-609.

29. Fiese BH, Kline CA. Development ofthe Family Ritual Questionnaire: initialreliability and validation studies. J FamPsychol. 1993;6:290-301.

30. Neumark-Sztainer D, Story M,AckardD,Moe J, PerryC. Familymealsamong adolescents: findings from a pilotstudy. J Nutr Ed. 2000;32:335-340.

31. Devereux EC Jr, Brofenbrenner U,Suci GJ. Patterns of parent behavior inthe United States and the Federal Re-public of Germany. Int Soc Sci J.1962;14:488-506.

32. Birch L, Fisher JO. Mothers’ feedingpractices influence daughters’ eating

and weight. Am J Clin Nutr.2000;71:1054-1061.

33. Pate R, Ross R, Dowda M, Trost S,Sirard J. Validation of a 3-day physicalactivity recall instrument in femaleyouth. Pediatr Exercise Sci. 2008;15:257-265.

34. Robinson JP, Godbey G.Time For Life:The Surprising Ways Americans UseTheir Time. 2nd ed. State College, PA:Pennsylvania University Press; 1999.

35. PettMA, LackeyNR, Sullivan JJ.Mak-ing Sense of Factor Analysis. ThousandOaks, CA: Sage; 2003.

36. Kim M-J, McIntosh WA, Anding J,Kubena KS,ReedDB,MoonG-S. Per-ceived parenting behaviours predictyoung adolescents’ nutritional intakeand body fatness. Matern Child Nutr.2008;4:286-303.

37. Miller DP, Han WJ. Maternal nonstan-dard work schedules and adolescentoverweight. Am J Public Health.2008;98:1495-1502.

38. Francis LA, Hofer SM, Birch LL. Pre-dictors of maternal child-feeding style:maternal and child characteristics. Ap-petite. 2001;34:231-243.

39. Romero ND, Epstein LH, Salvy SJ.Peer modeling influences girls’ snack in-take. J Am Diet Assoc. 2009;109:133-136.

40. Arcan C, Neumark-Sztainer D,Hannan P, van den Berg P, Story M,Larson N. Parental eating behaviours,home food environment and adolescentintakes of fruits, vegetables and dairyfoods: longitudinal findings from Pro-ject EAT. Public Health Nutr.2007;10:1257-1265.

Journal of Nutrition Education and Behavior � Volume 43, Number 3, 2011 McIntosh et al 149