C

PSA

AOtdpwiDpyeeptiSpetntRwhfacnrsw

ApBBmFstt(

Acg

A

©

RESEARCH

urrent Research

arental Pressure, Dietary Patterns, and Weighttatus among Girls Who Are “Picky Eaters”

MY T. GALLOWAY, PhD; LAURA FIORITO, RD; YOONNA LEE, PhD; LEANN L. BIRCH, PhD

CtibaTsmcJ

Pcbvahsa

fiptvpplHeBe

ttwAtiusvsehc

pe

BSTRACTbjective To determine whether mothers’ fruit and vege-able intake and mothers’ use of pressure in the feedingomain when their daughters were 7 years old predictedicky eating and dietary intake when their daughtersere 9 years old, and to examine diet and weight status

n picky and nonpicky eaters.esign/Subjects Participants were 173 9-year-old non-His-anic white girls and their mothers. A longitudinal anal-sis was used to assess maternal influences on pickyating and diet. A cross-sectional analysis was used toxamine diet and weight status between picky and non-icky eaters. Measures included maternal feeding prac-ices, daughters’ pickiness, mothers’ fruit and vegetablentake, daughters’ food intake, and weight status.tatistical Analyses t tests examined differences betweenicky and nonpicky eaters. Structural equation modelingxamined relationships among mothers’ fruit and vege-able intake; child feeding practices; daughters’ picki-ess; and fruit, vegetable, micronutrient, and fiber in-akes.esults Mothers consuming more fruits and vegetablesere less likely to pressure their daughters to eat andad daughters who were less picky and consumed moreruits and vegetables. Picky eaters consumed fewer fruitsnd vegetables, but also fewer fats and sweets. All girlsonsumed low amounts of vitamin E, calcium, and mag-esium, but more picky girls were at risk for not meetingecommendations for vitamins E and C and also con-umed significantly less fiber. In addition, picky eatersere less likely to be overweight.

. T. Galloway is an assistant professor with the De-artment of Psychology, Appalachian State University,oone, NC. L. Fiorito is a research assistant and L. L.irch is a distinguished professor of Human Develop-ent with the Department of Human Development andamily Studies, Pennsylvania State University, Univer-ity Park. Y. Lee is an invited senior researcher, Nutri-ion Research Team, Department of Health Care Indus-ry, Korean Health Industry Development InstituteKHIDI), Seoul, South Korea.

Address correspondence to: Amy T. Galloway, PhD,ssistant Professor, Department of Psychology, Appala-hian State University, Boone, NC 28608. E-mail:allowayat@appstate.eduCopyright © 2005 by the American Dietetic

ssociation.0002-8223/05/10504-0011$30.00/0

edoi: 10.1016/j.jada.2005.01.029

2005 by the American Dietetic Association

onclusions Mothers influenced daughters’ fruit and vege-able intake via their own patterns of fruit and vegetablentake and by influencing their daughters’ tendencies toe picky eaters. Both picky and nonpicky eaters hadspects of their diets that did not meet recommendations.aken together, these findings suggest that parentshould focus less on “picky eating” behavior and more onodeling fruit and vegetable consumption for their

hildren.Am Diet Assoc. 2005;105:541-548.

arents are concerned about children’s picky eating(1,2) and associated problems such as low fruit andvegetable intake. Although parental concerns about

hildren’s picky eating are common, picky eating has noteen systematically studied. Definitions of picky eatingary and have sometimes focused on the consumption oflimited variety of foods (3-5), whereas at other times

ave included things related to the amount of food con-umed (6). In this study we characterized the picky eaters a child who consumes an inadequate variety of food.Although research on picky eating is limited, recent

ndings indicate that prior food experience plays a role inicky eating. In a recent study, mothers who breastfedheir daughters at least 6 months and consumed moreegetables were less likely to have daughters who wereicky eaters (7). Fisher and colleagues (8) showed thatarents who ate fewer fruits and vegetables were moreikely to use pressure or coercion to increase food intake.owever, children who were pressured to “eat their veg-

tables” actually consumed fewer fruits and vegetables.oth the use of pressure as a feeding strategy and pickyating predict lower vegetable consumption (7,8).Parents also are more likely to pressure a child to eat if

hey perceive the child to be underweight (9). Accordingo Marchi and Cohen (10), picky eating in early childhoodas associated with anorexia nervosa in adolescent girls.lthough there are several reports that deal with rela-

ionships between pickiness and underweight or energyntake (6,11), the relationships among picky eating, these of pressure as a child-feeding strategy, and weighttatus are not well understood. Because low fruit andegetable consumption has been linked to increased con-umption of fats (12), it is possible that picky eaters, whoat fewer fruits and vegetables, are consuming dietsigher in fat and energy that may place them at risk forhildhood overweight (13).The aim of this research was to examine predictors of

icky eating (including mothers’ child-feeding practices,specially pressuring the child to eat), to evaluate picky

ating as a predictor of dietary intake patterns, and to

Journal of the AMERICAN DIETETIC ASSOCIATION 541

iw

MPPpmppaatcmcsalrcflt

san

RPp

DAflyp

MMt(twwmmt

DDyCswimfpe

T�

MM7dstst5ttmnnm

DFcoiiacsfwTnf2vlwtoptPtdifits2ntha

fitp(i

5

nvestigate possible links between picky eating and girls’eight status.

ETHODSarticipantsarticipants were from central Pennsylvania and wereart of a longitudinal study of the health and develop-ent of young girls. There were 192 girls who partici-

ated in the study at 7 years of age and 183 girls whoarticipated again when they were age 9. Reasons forttrition included the following: family moved from therea, was no longer interested in participating, or failedo attend data collection session. Complete data wereollected from 173 non-Hispanic white girls and theirothers when girls were 7 and 9 years old. Eligibility

riteria for girls’ participation at the beginning of thetudy included living with both biological parents, thebsence of severe food allergies or chronic medical prob-ems affecting food intake, and the absence of dietaryestrictions involving animal products. Families were re-ruited for participation in the longitudinal study usingyers, newspaper advertisements, and letters of invita-ion.

Girls and their parents visited the laboratory in theummer. Trained staff interviewed the girls individually,nd mothers completed a series of self-report question-aires.The study protocol was approved by the Institutionaleview Board at The Pennsylvania State University.arents provided written consent for their daughter’sarticipation in the study.

esignlongitudinal analysis was used to assess maternal in-

uences on picky eating and diet. A cross-sectional anal-sis was used to examine diet and weight status betweenicky and nonpicky eaters.

EASURESaternal pressure to eat more and mothers’ dietary in-

ake were measured when girls were 7 years old7.3�0.3). Daughters’ picky eating, daughters’ dietary in-ake, weight status, and body composition were assessedhen they were 9 years old (9.3�0.3). Data obtainedhen the girls were 7 years old were predictors in theodel; data from the girls at age 9 were used for outcomeeasures in the model and for assessing differences be-

ween picky and nonpicky eaters.

aughters’ Picky Eatingaughters’ picky eating was assessed when they were 9ears old. Picky eating (three items) is a subscale of thehild Feeding Questionnaire (14). This subscale mea-ures mothers’ reports of how they perceive their child’sillingness to eat during mealtimes and included three

tems: (a) my child’s diet consists of only a few foods, (b)y child is unwilling to eat many of the foods that our

amily eats at mealtimes, and (c) my child is fussy oricky about what she eats. High mean scores on the picky

ating subscale represent higher levels of picky eating. m

42 April 2005 Volume 105 Number 4

he internal consistency of items on this subscale was�.85.

aternal Pressure to Eataternal pressure to eat was measured when girls wereyears old. Mothers’ use of pressure to encourage their

aughters to eat was measured using the pressure-to-eatubscale from the Child Feeding Questionnaire referredo earlier. The pressure-to-eat subscale (four items) mea-ures the extent to which mothers pressure their daugh-ers to consume foods. All items were rated using a-point Likert-type scale; responses ranged from disagreeo agree. Items were: (a) my child should always eat all ofhe food on her plate; (b) I have to be especially careful toake sure my child eats enough; (c) if my child says “I’m

ot hungry,” I try to get her to eat anyway; and (d) if I didot guide or regulate my child’s eating, she would eatuch less than she should. Cronbach ��.73.

ietary Intake of Mothers and Daughtersood Frequency Questionnaire (FFQ). This questionnaire wasompleted by mothers only when daughters were 7 yearsld. The FFQ was developed to measure the usual dietaryntake of free-living individuals (15,16). The Kristal FFQs quantitative and asks both frequency and usualmount consumed from an extensive list of foods. Wereated the within–food group frequency and varietycores for fruit and vegetable food groups. The within–ood group frequency score was based on the frequency ofithin–food group consumption during a 3-month period.he within–food group variety score was defined as theumber of different foods consumed within a particularood group over a 3-month period.4-Hour Dietary Recall. With the assistance of the Pennsyl-ania State University Diet Assessment Center, we col-ected dietary information from the mothers when girlsere 7 years old and from the girls at age 9 by adminis-

ering three 24-hour food recalls, two on weekdays andne on a weekend day, randomly selected within a 2-weekeriod. During each multiple-pass recall, mothers helpedheir daughters report food intake from the previous day.articipants were aware of the general time during whichhese calls would be made, but not of the exact time orate. For the mothers and daughters, 3-day food groupntake (including vegetables, fruit, dairy, meat, grain,ats, and sweets), as well as nutrient, fiber, and energyntakes were estimated using a computer-assisted Nutri-ion Data System for Research (Nutrient Database ver-ion 12A, Food Database version 4.02_31, release date001, Nutrition Coordinating Center, University of Min-esota, Minneapolis). Food portion posters (2D Food Por-ion Visual, Nutrition Consulting Enterprises, Framing-am, MA) were used to assist in the estimation of foodmounts.The total energy, macronutrient, micronutrient, and

ber values were assessed using dietary reference in-akes (DRIs) (17-21). Girls’ intakes at age 9 were com-ared with the Recommended Dietary Allowances (RDAs)18-21) for 9- to 13-year-old girls. For calcium, meanntakes were compared with Adequate Intake (AI) recom-

endations (21). The prevalence of inadequate nutrient

iRwbsDgosea

iDc(igbwssf

WBmiBhTfwwDactDius4mw

SWdadtsDascciinp

IratesriarittvttivhtawtcPd

RFdfnm(mm(cqtooGsbtfTogtl

mpcDeTwn

ntakes was determined using the Estimated Averageequirement (EAR) cutpoint method (22) for nutrientsith an established EAR. Nutrient intake estimates wereased on foods consumed and do not include intake fromupplements. The size of food servings was based on USepartment of Agriculture (USDA) Food Guide Pyramiduidelines for the appropriate age group (22). The meanf the following micronutrients were used to provide aingle micronutrient score that was used in the structuralquation model: calcium; iron; folate; vitamins A, C, E,nd B-6; magnesium; and zinc.The sum weight of all foods consumed was calculated,

ncluding those contained in mixed dishes. Nutritionata System summary file data containing each food

onsumed were edited to conduct food group analyses23). Editing involved summing ingredient gram weightsnto a single whole-food weight that was assigned to foodroups according to the Food Guide Pyramid. The num-er of servings was calculated from gram weights ofhole foods consumed and was based on closely matched

erving sizes used by the Food Guide Pyramid (22). Oneummary variable was created to represent combinedruit and vegetable intake for girls.

eight Status and Body Compositionody mass index. At age 9, girls’ heights and weights wereeasured in triplicate by a trained staff member follow-

ng procedures described by Gordon and colleagues (24).ody mass index (BMI) scores were generated based oneight and weight measurements (calculated as kg/m2).hree height and weight measurements were collected

or mothers and fathers when girls were 9 years old, andere used to calculate parents’ BMIs. Girls’ BMI valuesere converted to BMI percentiles using the Centers forisease Control and Prevention’s 2000 Growth Charts,nd girls were classified as overweight if their BMI per-entile was more than 85 and obese if their BMI percen-ile was more than 95 (25).ual-energy X-ray Absorptiometry. Girls’ body compositions,ncluding body fat percentage, were assessed at age 9sing dual-energy x-ray absorptiometry. Whole-bodycans were obtained using the Hologic QDR 4500W (S/N7261) (Hologic, Inc, Bedford, MA) in the array scanode. Scans were then analyzed using whole-body soft-are, QDR4500 Whole Body Analysis (Hologic, Inc).

tatistical Analysese used complete data from 173 cases and calculated

escriptive statistics for all variables used in the modelnd subsequent analyses. The maternal pressure-to-eatata were skewed, therefore we normalized using a logransformation. Due to skewed results, we used a medianplit of the pickiness data when t tests were calculated.ata from the FFQ and 24-hour recall were used to createsingle variable to represent frequency of maternal con-

umption of fruits and vegetables and variety using prin-ipal components analysis. Similarly, we used principalomponents analysis to create a single fruit and vegetablentake variable for girls. We used �2 to assess differencesn the proportion of picky and nonpicky eaters who didot meet recommendations for specified nutrients and the

roportion of girls who were at risk for overweight. s

Structural equation modeling software, Amos 4 (SPSS,nc, Chicago, IL) (26), assessed a model that representedelationships among various predictors of picky eatingnd fruit and vegetable intake in girls. Structural equa-ion modeling is a statistical procedure that tests thextent to which a theoretical model of proposed relation-hips among variables matches the observed pattern ofelationships that exists in a data set. More importantlyn our case, the structural equation modeling approach islso used as an alternative way of analyzing a series ofegressions for modeling associations among variables ofnterest. We used three common fit indexes to evaluatehe fit of the proposed models: the �2 test, the Compara-ive Fit Index, and the Normed Fit Index (27). A small �2

alue is indicative of good fit, but the �2 test is sensitiveo sample size where N approaches 200 and can makerivial differences seem significant. For this reason, otherndexes are typically used. The Normed Fit Index hasalues ranging from 0.00 to 1.00, and values nearing 0.95ave been proposed to indicate a good fit. The Compara-ive Fit Index compares the hypothesized model with thectual data and has values that range from 0.00 to 1.00ith values more than 0.95 indicating a good fit. Al-

hough we reported the traditional fit indexes, the pathoefficients were most salient for interpreting our data.ath coefficients in the model can be interpreted as stan-ardized regression weights.

ESULTSor the total sample, as shown in Table 1, girls’ meanaily energy intake was within the recommended rangeor 9-year-olds and they met recommendations for macro-utrient intake. Dietary fiber was less than recom-ended levels; none of the girls met recommendations

17). Girls’ mean intakes of most vitamins and mineralset recommendations except for vitamin E, calcium, andagnesium, which were less than recommended levels

see Table 2). Girls’ vitamin and mineral intakes wereompared EARs to determine the prevalence of inade-uate nutrient intakes. Many girls had inadequate in-akes of vitamin E and magnesium. As shown in Table 3,n average, girls consumed fewer than the seven servingsf fruits and vegetables per day suggested by the Dietaryuidelines for Americans (22), and fewer than the 3.5

ervings a day that Krebs-Smith and colleagues reportedeing consumed by this age group (28). When potatoes inhe form of french fries and potato chips were excludedrom the analyses, servings of vegetables were even less.able 3 also indicates that girls in the sample consumed,n average, fewer than the recommended servings ofrains and meat. Mothers’ total fruit and vegetable in-akes (including french fries and potato chips) were alsoess than the recommended seven servings per day (22).

One fourth of mothers gave picky eating scores of 3 orore to their daughters, in which 5 indicates maximum

ickiness (mean�1.9�1.1, range�0-5.0); however, weategorized picky eaters using a median split of the data.ifferences between picky and nonpicky eaters’ mean

nergy, macronutrient, and fiber intakes are presented inable 1. Picky eaters had lower energy intakes than girlsho were not picky eaters; however, this difference wasot significant between the two groups. Both groups con-

umed less fiber than recommended; however, picky girls’

April 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION 543

fifiiiieepFTnlvf

osBtpo(upg

evt

ky eate

5

ber intake was significantly less than the nonpicky girls’ber intake. Table 2 reveals that picky eaters had signif-

cantly lower intakes of vitamin E and folate, and signif-cantly more picky eaters were at risk for inadequatentakes of vitamins C and E. Table 3 indicates that pickyaters also had significantly lower intakes of fruits, veg-tables, fats, and sweets. Except for the dairy group, theercentage of girls meeting the recommendations of theood Guide Pyramid was less among the picky eaters.he percentage of girls consuming the recommendedumber of servings of fruit and vegetables at age 9 was

ow; none of the picky girls met the recommendations foregetables and only 2% met the recommendations forruit servings.

Table 1. Energy, macronutrient, and dietary fiber recommended and

Recommendationb Tota

4™™Energy (kcal) 1,500-2,400d 1,80Fat (% of energy) 32-35 3Carbohydrate (% of energy) 45-65 5Protein (% of energy) 10-30 1Dietary fiber (g) 26 1

aMean�standard deviation.bBased on Dietary Reference Intake (17) for girls age 9-13 years.cSD�standard deviation.dWe created an Energy Estimate Requirements range for girls 9 through 18 years, whereactive, active, and very active). The range of values in Energy Estimate Requirements ryzValues with different letters indicate significant differences between nonpicky and pic

Table 2. Dietary recommendations, mean intakes, and percent of 9-yeaters

Recommendationc

Nutr

Total sample(N�173)

N(n

4™™™™™™™™™™™™™Vitamin A (�g RAEe) 600 694.9�285.5 71Vitamin C (mg) 45 72.4�40.1 7Vitamin E (mg) 11 6.0�2.0Folate (�g) 300 315.9�88.7Vitamin B-6 (mg) 1.0 1.4�0.4Calcium (mg) 1,300 905.5�302.5Iron (mg) 8 12.4�3.8 1Zinc (mg) 8 8.8�2.6Magnesium (mg) 240 213.0�50.0 21

aMean�standard deviation.bAt risk cut-offs using Estimated Average Requirements cut-point method for nutrientscRecommendations based on the Dietary Reference Intakes for girls aged 9-13 years. Ade(18-21).dSD�standard deviation.eRAE�retinol activity equivalents.fNA�not available. Percent of a population with inadequate intakes cannot be assessedyzValues with different letters indicate significant difference within nutrient group compa*Significant difference within nutrient group comparisons at P�.05.

Information about weight status and body composition c

44 April 2005 Volume 105 Number 4

f the entire sample and each subgroup of girls is pre-ented in Table 4. Picky eaters had significantly lowerMIs, t�2.10 (P�.05), and significantly less body fat,

�1.97 (P� .05). In addition, Figure 1 shows that a higherroportion (43%) of nonpicky eaters are overweight andbese using the 85th and 95th percentile cutoffs for BMI25), �2�12.30, P�.01. Only one girl in the sample wasnderweight (�5th percentile cutoff), and she was not aicky eater. There were no differences between theroups in maternal or paternal BMI.Figure 2 presents the structural equation model that

xamined the relationship among predictors of fruit andegetable intake and the subsequent effect on micronu-rient and fiber intake in girls. The model’s path coeffi-

al intakes in 9-year-old girls classified as picky or nonpicky eaters

Nutrient Intakea

ple (N�173) Nonpicky (n�90) Picky (n�83)

™™™™™™™™™™™™™™™™™™ mean�SD c ™™™™™™™™™™™™™™™™™™™™™3330.5 1,837.6�374.5y 1,777.6�283.1y

4.6 31.9�4.6y 31.7�4.7y

5.9 57.1�5.6y 55.7�6.2y

2.5 13.7�2.2y 14.0�2.8y

3.5 12.7�3.6y 11.2�3.3z

ed girls’ mean age, weight and height, and physical activity coefficients (sedentary, lowpossible differences in the physical activity coefficients of participants (17).rs’ food group intake, P�.01.

ld girls at risk for inadequate intakes classified as picky and nonpicky

Intakea Girls at Riskb

ky)

Picky(n�83)

Total sample(n�173)

Nonpicky(n�90)

Picky(n�83)

�SD d ™™™™™™™™™™3 4™™™™™™™™™™™™™™ % ™™™™™™™™™™3288y 669.7�282y 15 12 1839y 67.0�39y 20 12 28*2.3y 5.6�1.5z 92 88 98*87y 303�92z 23 23 304.1y 1.5�.49y 5 6 5288y 911�320y NAf NA NA3.8y 12.0�3.9y �1 0 12.6y 8.7�2.6y 27 21 3052.7y 212.3�47.3y 43 41 46

take value listed for calcium; other nutrients listed as Recommended Dietary Allowances

trients with Adequate Intake values (50).at P�.05.

actu

l sam

™™™™4.6�1.8�5.9�3.9�1.9�

we useflects

ear-o

ient

onpic�90

mean8.1�8.0�6.8�330�1.4�905�2.9�9.0�3.7�

(50).quate In

for nurisons

ients showed that mothers with higher fruit and vege-

tavacncaamma(NvN

DReaefnT

mv

iErmqtcptmd

sgmaiwnfat

aters’

able intake had daughters who consumed more fruitsnd vegetables. Mothers who consumed fewer fruits andegetables were likely to pressure their daughters to eatnd to have daughters who were picky eaters. Girls whoonsumed more fruits and vegetables had higher micro-utrient and fiber intakes. Path coefficients in the modelan be interpreted as standardized regression weights,nd in structural equation modeling the path coefficientsre adjusted for the presence of other relationships in theodel. Several fit indexes show that the hypothesizedodel is an adequate representation of the relationships

mong variables. Although the �2 value was significant�2�63.59, df�8, P�.05), the Comparative Fit Index andormed Fit Index revealed that the model provided aery good fit of the data (Comparative Fit Index�0.96;ormed Fit Index�0.95) and the R2�20.

ISCUSSIONegardless of whether girls were classified as picky eat-rs, most girls consumed less-than-recommendedmounts of grains, fruits, vegetables, and meats; how-ver, picky girls consumed significantly fewer servings ofruits, vegetables, fats, and sweets than girls who wereot picky eaters. Fiber intake followed the same pattern.

Table 3. Recommended, actual food group intakes, and percentagenonpicky eaters

RecommendedNo. ofServingsb

No. of Se

Total sample(N�173)

Non(n�

4™™™™™™™™™™™™ mean�Grains 9 6.1�1.7 6.3�Vegetables 4 1.0�0.8 1.7�Fruit 3 1.3�1.0 1.5�Dairy 2-3 2.8�1.3 2.8�Meat 2 1.4�0.6 1.5�Fats and Sweets NAd 5.3�2.5 5.8�

aMean�standard deviation.bRecommendations are from the Food Guide Pyramid (22) for girls 6 years and older.cSD�standard deviation.dNA�not applicable.yzValues with different letters indicate significant difference between picky and nonpick*Significant difference within nutrient group comparisons between picky and nonpicky e

Table 4. Body mass index and body fat percentage in 9-year-old g

Total Sample

N Mean�SDa

BMIb 173 18.4�3.1Body fat %c 164 26.7�7.1

aSD�standard deviation.bBMI�body mass index; calculated as kg/m2.cBody fat percentage assessed using dual-energy x-ray absorptiometry. Not all participanyzValues with different letters are significantly different at P�.05.

hese results suggest that picky eaters did not consume

ore fats and sweets to compensate for lower fruit andegetable intake.The micronutrient intakes of both groups were similar

n that they consumed inadequate quantities of vitamin, calcium, and magnesium. Although most girls metecommendations for dairy intake, consuming the recom-ended number of dairy servings does not ensure ade-

uate calcium intake, and among these girls, consump-ion of calcium-rich vegetables was low and did notontribute substantially to calcium intake, which is im-ortant in the prevention of osteoporosis (29-31). Despitehe similarities in micronutrient intake, significantlyore picky eaters were at risk of not meeting recommen-

ations for vitamins C and E.The findings also show that picky eaters consumed

ignificantly less fiber than nonpicky eaters. However,irls in both groups consumed less fiber than recom-ended, possibly placing them at greater risk for long-

nd short-term health problems associated with low fiberntake, including constipation. Whereas nearly all girlsere obtaining the recommended intake for many micro-utrients, probably due to their consumption of fortifiedoods such as cereal (32), they did not obtain needed fibernd other potentially beneficial substances, such as phy-ochemicals.

-year-old girls meeting the recommendations classified as picky or

sa % Meeting Recommendations

Picky(n�83)

Total sample(N�173)

Nonpicky(n�90)

Picky(n�83)

™™™™™™™™™™™™3 4™™™™™™™™™™™™™™ % ™™™™™™™™™™™™™35.9�1.6y 6 8 41.3�.72z �1 1 01.0�0.94z 6 9 2*2.9�1.3y 36 36 37

y 1.4�0.59y 20 22 184.9�2.1z NA NA NA

s’ nutrient intake at P�.05, except for fruit at P�.01 and vegetables at P�.001.nutrient intake at P�.05.

assified as nonpicky and picky eatersyz

Nonpicky Picky

Mean�SD n Mean�SD

18.9�3.4y 83 17.9�2.7z

27.8�7.4y 82 25.6�6.6z

ided consent for dual-energy x-ray absorptiometry, hence the total sample was reduced.

of 9

rving

picky90)

SDc

1.8y

.89y

1.1y

1.2y

0.632.8y

y eater

irls cl

n

9082

ts prov

Results of the structural equation modeling revealed

April 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION 545

taaadntlocsieptw

wdathdwwlpairTtcrapmcw

bsinafppm(fica4tsa

Ftm*

Fpp

5

hat mothers directly influenced their daughters’ fruitnd vegetable intake by modeling and by making fruitsnd vegetables available; they indirectly affected fruitnd vegetable consumption by pressuring their pickyaughters to eat, which predicted higher levels of picki-ess and lower fruit and vegetable intake. Picky girlsended to consume less total energy, had significantlyower BMIs and body fat, and were less likely to beverweight. Although picky eating might compromisehildren’s nutrient intake in environments where food iscarce, these findings suggest that in today’s industrial-zed environments, where a wide variety of palatable,nergy-dense foods are readily and cheaply available,icky eating might be less problematic than previouslyhought and could help children maintain a healthfuleight.

igure 1. Structural equation model of influences on girls’ fruit and vehe model can be interpreted as standardized regression weights. Theodel. aIncludes calcium, iron, folate, vitamin A, vitamin C, vitamin E,

**P�.001.

igure 2. Proportion of 9-year-old girls classified as nonpicky andicky eaters with a body mass index (BMI) more than the 85th and 95thercentiles (25).

The finding that picky eaters are less likely to be over- r

46 April 2005 Volume 105 Number 4

eight diverges from previous studies, which reported noifferences in growth and weight status of picky toddlersnd preschool children (4,33). It is possible, however, thathe persistence of picky-eating behavior over time couldave long-term effects on weight status that are not evi-ent until children are older. In the present study, inhich the incidence of overweight is high, the picky girlsere not underweight, but they did have significantly

ower BMIs than nonpicky girls. This contradicts therevious finding of no relationship between picky eatingnd fat mass or fat-free mass in preteen children (34), butt is consistent with the findings of another study, whicheported that picky girls had a lower weight status (10).he difference in energy intake between picky eaters andheir nonpicky counterparts was not statistically signifi-ant. However, over time substantial weight gain canesult even when intake exceeds expenditure by smallmounts, such as the 50 kcal/day difference noted in theresent study, a value well within the margin of measure-ent error (35,36). This small difference between groups

ould, if it persisted over time, produce the differences ineight status we noted in the picky and nonpicky eaters.In this study, children who were pressured to eat more

y their parents consumed fewer fruits and vegetables,uggesting that mothers who modeled fruit and vegetablentake were less likely to use pressure. In contrast to theegative effects of pressure, making foods more available,ccessible, and increasing children’s familiarity withoods can increase acceptance and intake (37-39). Theresent findings contribute to the growing evidence thatarental modeling of fruit and vegetable intake (8,40) andaking fruits and vegetables available and accessible

41) promote fruit and vegetable consumption. Thesendings are also consistent with previous research indi-ating that pressuring children to eat a particular foodctually reduces their liking and intake of that food (42-7). Although there is some evidence that parents believehat pressuring children to eat a food is an effectivetrategy for promoting intake (48), recent work by Batsellnd colleagues (44) indicates that 69% of college students

le (F&V), micronutrient, and fiber intake (N�173). Path coefficients incoefficients are adjusted for the presence of other relationships in thein B-6, magnesium, and zinc. bSignificance values: *P�.05, **P�.01,

getabpathvitam

eported being pressured to eat foods as children, and

tTwsfa

fifcteimwutessci

CTesavwvuetdcqnti

WnbHbRsN

R

1

1

1

1

1

1

1

1

1

hat these incidents resulted in persistent food dislikes.he relationship between maternal use of pressure to eathen girls were 7 and girls’ picky eating at age 9 provides

tronger evidence that the use of parental pressure ineeding may promote both picky eating and lower fruitnd vegetable consumption in girls.Though both experimental research and retrospective

ndings suggest that pressure to eat is associated withood rejections and aspects of picky eating (44,47), theausal direction of these relationships has not been es-ablished. It seems likely that, as in many areas of par-nt-child interaction, influence between parent and childs bidirectional (49). Although parents’ use of pressure

ay promote picky eating in a child, the child’s initialeight status or eating style may also elicit the parents’se of pressure as a feeding strategy; children perceivedo be too thin or children who parents think are not eatingnough may elicit parental pressure to eat more (9). Iteems that pressure can exacerbate existing problems,uch as promoting dislikes for foods that parents wanthildren to eat, and perhaps reducing fruit and vegetablentake (8,44).

ONCLUSIONShese data indicate that the diets of picky and nonpickyaters were more similar than they were different anduggest that increasing fruit and vegetable intake amongll children should continue to be a major focus of inter-entions. Anticipatory guidance should provide parentsith strategies that promote children’s acceptance of aariety of foods, especially fruits and vegetables, partic-larly increasing availability and modeling fruit and veg-table intake, while discouraging parents from labelinghe child as a picky eater or pressuring children to eat. Toecrease parental use of pressure in child feeding, clini-ians should address parental anxieties about the ade-uacy of the child’s intake by emphasizing the child’sormal growth and pointing out that pressuring childreno eat is not an effective strategy for promoting children’sntake of healthful foods.

e thank Shelly Mannino for her assistance with theutrition analysis. This research was supported in party National Institutes of Health (NIH) Grant No. RO1D32973, The National Dairy Council, services providedy the General Clinical Research Center NIH Grant M01R10732, and the Diet Assessment Center of The Penn-ylvania State University. Dr Galloway was supported byIH NRSA Grant No. HD 08626.

eferences1. Tinsley BJ, Parke RD. The historical and contempo-

rary relationship between developmental psychologyof pediatrics: A review and empirical survey. In:Fitzgerald HE, Lester BM, Yogman MW, eds. Theoryand Research in Behavioral Pediatrics. New York,NY: Plenum Press; 1984:1-30.

2. Li Y, Shi A, Wan Y, Hotta M, Ushijima H. Childbehavior problems: Prevalence and correlates in ru-ral minority areas of China. Pediatrics Int. 2001;43:651-661.

3. Pliner P, Hobden K. Development of a scale to mea- 1

sure the trait of food neophobia in humans. Appetite.1992;19:105-120.

4. Carruth BR, Skinner JD, Houck K, Moran J, CollettaF, Ott D. The phenomenon of “picky eater”: A behav-ioral marker in eating patterns of toddlers. J Am CollNutr. 1998;17:180-186.

5. Potts HW, Wardle J. The list heuristic for studyingpersonality correlates of food choice behaviour: A re-view and results from two samples. Appetite. 1998;30:79-92.

6. Rydell AM, Dahl M, Sundelin C. Characteristics ofschool children who are choosy eaters. J Genet Psy-chol. 1995;156:217-229.

7. Galloway AT, Lee Y, Birch LL. Predictors and conse-quences of food neophobia and pickiness in children.J Am Diet Assoc. 2003;103:692-698.

8. Fisher JO, Mitchell DC, Smiciklas-Wright H, BirchLL. Parental influences on young girls’ fruit and veg-etable, micronutrient, and fat intakes. J Am DietAssoc. 2002;102:58-64.

9. Francis LA, Hofer SM, Birch LL. Predictors of mater-nal child-feeding style: Maternal and child character-istics. Appetite. 2001;37:231-243.

0. Marchi M, Cohen P. Early childhood eating behaviorsand adolescent eating disorders. J Am Acad ChildAdolesc Psychiatry. 1990;29:112-117.

1. Jacobi C, Agras WS, Bryson S, Hammer LD. Behav-ioral validation, precursors, and concomitants ofpicky eating in childhood. J Am Acad Child AdolescPsychiatry. 2003;42:76-84.

2. Dennison BA, Rockwell HL, Baker SL. Fruit andvegetable intake in young children. J Am Coll Nutr.1998;17:371-378.

3. Fisher JO, Birch LL. Fat preferences and fat con-sumption of 3-to 5-year-old children are related toparental adiposity. J Am Diet Assoc. 1995;95:759-764.

4. Birch LL, Fisher JO, Grimm-Thomas K, Markey CN,Sawyer R, Johnson SL. Confirmatory factor analysisof the Child Feeding Questionnaire: A measure ofparental attitudes, beliefs and practices about childfeeding and obesity proneness. Appetite. 2001;36:201-210.

5. Kristal AR, Patterson RE, Shattuck AS, Vizenor NC.Nutrient databases for food frequency question-naires. In: 23rd National Nutrient Databank Confer-ence; April 16, 1999; Washington, DC.

6. Patterson RE, Kristal AR, Tinker LF, Carter RA,Bolton MP, Agurs-Collins T. Measurement character-istics of the Women’s Health Initiative food frequencyquestionnaire. Ann Epidemiol. 1999;9:178-187.

7. Institute of Medicine, Food and Nutrition Board. Di-etary Reference Intakes for Energy, Carbohydrates,Fiber, Fat, Fatty Acids, Cholesterol, Protein, andAmino Acids. Washington, DC: National AcademyPress; 2002.

8. Institute of Medicine, Food and Nutrition Board. Di-etary Reference Intakes for Vitamin A, Vitamin K,Arsenic, Boron, Chromium, Copper, Iodine, Manga-nese, Molybdenum, Nickel, Silicon, Vanadium, andZinc. Washington, DC: National Academy Press;2001.

9. Institute of Medicine, Food and Nutrition Board. Di-

April 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION 547

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4

4

5

5

etary Reference Intakes for Vitamin C, Vitamin E,Selenium, and Carotenoids. Washington, DC:National Academy Press; 2000.

0. Institute of Medicine, Food and Nutrition Board. Di-etary Reference Intakes for Thiamin, Riboflavin, Ni-acin, Vitamin B6, Folate, Vitamin B12, PantothenicAcid, Biotin, and Choline. Washington, DC: NationalAcademy Press; 1998.

1. Institute of Medicine, Food and Nutrition Board. Di-etary Reference Intakes for Calcium, Phosphorus,Magnesium, Vitamin D, and Fluoride. Washington,DC: National Academy Press; 1997.

2. US Department of Agriculture, US Department ofHealth and Human Services. Nutrition and yourHealth: Dietary Guidelines for Americans. 5th ed.Washington, DC: US Government Printing Office;2000.

3. Mitchell DC. Database requirements to meet foodgrouping needs. J Food Composit Anal. 2001;14:279-285.

4. Gordon CC, Chumlea WC, Roche TG. Stature, recum-bent length, and weight. In: Lohman TG, Roche AF,Martorell R, eds. Anthropometric StandardizationReference Manual. Champaign, IL: Human KineticsBooks; 1988:3-8.

5. Kuczmarski RJ, Ogden CL, Grummer-Strawn LM,Flegal MK, Guo SS, Wei R, Mei Z, Curtin LR, RocheAF, Johnson CL. CDC growth charts: United States.Advance data from vital and health statistics.Hyattsville, MD: National Center for Health Statis-tics; 2000. Report No. 314.

6. Arbuckle JL, Wothke W. AMOS 4.0 user’s guide. Chi-cago, IL: Smallwaters; 1999.

7. Byrne BM. Structural equation modeling withAMOS: Basic concepts, applications, and program-ming. Mahwah, NJ: Lawrence Erlbaum Associates,Inc; 2001.

8. Krebs-Smith SM, Cook AD, Subar AF, Cleveland L,Friday J, Kahle LL. Fruit and vegetable intakes ofchildren and adolescents in the United States. ArchPediatr Adolesc Med. 1996;150:81-86.

9. Heaney RP, Matkovic V. Inadequate peak bone mass.In: Riggs BL, Melton LJI, eds. Osteoporosis: Etiology,Diagnosis, and Management. Philadelphia, PA: Lip-pincott-Raven Publishers; 1995:115-131.

0. Matkovic V, Ilich JZ. Calcium requirements forgrowth: Are current recommendations adequate?Nutr Rev. 1993;51:171-180.

1. Weaver CM. The growing years and prevention ofosteoporosis in later life. Proc Nutr Soc. 2000;59:303-306.

2. Berner LA, Clydesdale FM, Douglass JS. Fortifica-tion contributed greatly to vitamin and mineral in-takes in the United States, 1989-1991. J Nutr. 2001;131:2177-2183.

3. Carruth BR, Skinner JD. Revisiting the picky eater

48 April 2005 Volume 105 Number 4

phenomenon: neophobic behaviors of young children.J Am Coll Nutr. 2000;19:771-780.

4. Nicholls D, Wells JC, Singhal A, Stanhope R. Bodycomposition in early onset eating disorders. EurJ Clin Nutr. 2002;56:857-865.

5. Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity andthe environment: Where do we go from here. Science.2003;299:853-855.

6. Nestle M. Increasing portion sizes in American diets:More calories, more obesity. J Am Diet Assoc. 2003;103:39-40.

7. Birch LL, Marlin DW. I don’t like it; I never tried it:Effects of exposure on two-year-old children’s foodpreferences. Appetite. 1982;3:353-360.

8. Sullivan SA, Birch LL. Infant dietary experienceand acceptance of solid foods. Pediatrics. 1994;93:271-277.

9. Birch LL. Psychological influences on the childhooddiet. J Nutr. 1998;128(suppl 2):S407-S410.

0. Birch LL, Fisher JO. Development of eating behav-iors among children and adolescents. Pediatrics.1998;101:539-549.

1. Nicklas TA, Baranowski J, Cullen KW, RittenberryL, Olvera N. Family and child-care provider influ-ences on preschool children’s fruit, juice, and vegeta-ble consumption. Nutr Rev. 2001;59:224-235.

2. Birch LL, Birch D, Marlin D, Kramer L. Effects ofinstrumental eating on children’s food preferences.Appetite. 1982;3:125-134.

3. Newman J, Taylor A. Effect of a means-end contin-gency on young children’s food preferences. J ExpChild Psychol. 1992;64:200-216.

4. Batsell RW, Brown AS, Ansfield ME, Paschall GY.“You will eat all of that!”: A retrospective analysis offorced consumption episodes. Appetite. 2002;38:211-219.

5. Birch LL, Marlin D, Rotter J. Eating as the “means”activity in a contingency: Effects on young children’sfood preference. Child Dev. 1984;55:532-539.

6. Birch LL, McPhee L, Shoba BC, Steinberg L, Kreh-biel R. “Clean up your plate”: Effects of child feedingpractices on the conditioning of meal size. Learn Mo-tiv. 1987;18:301-317.

7. Fisher JO, Mitchell DC, Smiciklas-Wright H, BirchLL. Maternal milk consumption predicts the trade-offbetween milk and soft drinks in young girls’ diets. JNutr. 2001;131:246-250.

8. Casey R, Rozin P. Changing children’s food prefer-ences: Parent opinions. Appetite. 1989;12:171-182.

9. Holden GW, Miller PC. Enduring and different: Ameta-analysis of the similarity in parents’ child rear-ing. Psychol Bull. 1999;125:223-254.

0. Institute of Medicine, Food and Nutrition Board. Di-etary Reference Intakes Applications in Dietary As-sessment. Washington, DC: National Academy Press;

2000.